Concern for the Natural Environment and Its Effect on Pro-Environmental

Concern for the Natural Environment

and its Effect on Pro-environmental

Behaviour amongst the British Public

A thesis submitted to the University of Manchester for the degree of Doctor of

Philosophy in the Faculty of Humanities

2015

Rebecca Rhead

School of Social Sciences

Table of Contents

Chapter 1 Introduction ...... 17

1.1 Introduction ...... 18

1.2 The importance of anthropocentric climate change ...... 20

1.3 Environmental attitudes and behaviour ...... 22

1.4 Why this thesis? ...... 25

1.5 Research Aims ...... 27

1.6 Thesis structure ...... 30

Chapter 2 Theories of environmental attitudes and pro-environmental behaviour 33

2.1 Introduction ...... 34

2.2 Attitudes ...... 35

2.2.1 Conclusion ...... 44

2.3 Environmental concern ...... 45

2.3.1 Measures of environmental attitudes ...... 45

2.3.2 VBN and valuing the environment ...... 49

2.3.3 Social norms ...... 54

2.4 Social influences on environmental attitudes ...... 56

2.4.1 Age ...... 56

2.4.2 Income, education and lifestyle ...... 57

2.4.3 Political ideology ...... 59

2.4.4 Gender ...... 60

2.5 Environmental behaviour ...... 61

2.5.1 Defining pro-environmental behaviour ...... 61

2.5.2 Intent and impact ...... 63

2.5.3 Emotion and environmental behaviour ...... 65

2.5.4 Social context of behaviour change ...... 66

2.6 The relationship between environmental concern and behaviour ...... 69

2.6.1 The value-action gap ...... 71

2.6.2 Attitude-behaviour models ...... 73

2.6.3 Do attitudes only influence easy behaviours? ...... 78

2.6.4 Low priority ...... 82

2.7 Conclusion ...... 83

Chapter 3 Data sources and analytical methods ...... 89

3.1 Introduction ...... 90

3.2 Secondary Data Analysis ...... 91

3.2.1 Empirical reasoning ...... 93

3.2.2 Summary ...... 95

3.3 Data selection ...... 96

3.3.1 Shortlisted datasets ...... 98

3.4 Survey of Public Attitudes and Behaviours towards the Environment ...... 102

3.4.1 The distinction between reported and observed behaviour ...... 104

3.4.2 A critical reflection of the EAS ...... 106

3.5 Methods ...... 108

3.5.1 Latent variable analysis ...... 109

3.5.2 Examining direct and indirect relationships ...... 111

3.6 Methods of parameter estimation: Bayesian vs. frequentist ...... 113

3.7 Conclusion ...... 116

Chapter 4 Assessing the structure of British attitudes towards the natural environment ...... 118

4.1 Introduction ...... 119

4.2 The New Environmental Paradigm ...... 122

4.2.1 The VBN value frame ...... 123

4.2.2 Theory-driven questionnaires ...... 130

4.2.3 Aims ...... 132

4.3 Data ...... 133

4.3.1 Selecting variable for environmental concern ...... 133

4.4 Methods ...... 138

4.4.1 Exploratory Factor Analysis ...... 138

4.4.2 Confirmatory Factor Analysis ...... 142

4.4.3 Bayesian Structural Equation Modelling ...... 144

4.5 Analysis ...... 147

4.5.1 Part one – Exploratory Factor Analysis ...... 147

4.5.2 Part Two – Confirmatory Factor Analysis ...... 150

4.5.3 Modification indices ...... 152

4.5.4 Part three ...... 154

4.6 Discussion ...... 160

4.6.1 Similarities with the NEP ...... 160

4.6.2 Reflections on the data ...... 162

4.6.3 Reflection on the methods ...... 163

4.7 Conclusion ...... 166

Chapter 5 Producing a UK environmental attitude typology ...... 167

5.1 Introduction ...... 168

5.2 Previous research on population segmentation ...... 169

5.2.1 DEFRA segmentation ...... 170

5.2.2 Environmental attitude segmentation ...... 174

5.2.3 Theoretical frameworks used to study EC ...... 176

5.2.4 Summary ...... 180

5.3 Analysis ...... 182

5.4 Part one – Group classification ...... 184

5.4.1 Two-class model ...... 186

5.4.2 Four-class model ...... 189

5.5 Part two – Class profiles ...... 193

5.5.1 Within-class factor structures ...... 193

5.5.2 Gender ...... 194

5.5.3 Age ...... 196

5.5.4 Socio-economic status ...... 198

5.5.5 Climate change belief ...... 202

5.6 Part three – Regression analysis ...... 204

5.6.1 Missing data ...... 204

5.6.2 Regression ...... 206

5.7 Discussion ...... 208

5.8 Conclusion ...... 210

Chapter 6 The relationship between environmental attitudes and behaviour

212

6.1 Introduction ...... 213

6.2 Past studies of the environmental attitude behaviour relationship ...... 219

6.2.1 Low-cost behaviour ...... 221

6.2.2 Is the environment a low priority ...... 222

6.2.3 The effect of socio-economic status ...... 223

6.2.4 Summary ...... 224

6.3 Methods ...... 226

6.3.1 Measuring moderation ...... 227

6.3.2 Interaction effect ...... 228

6.3.3 Measuring mediation ...... 230

6.3.4 Indirect effects in cross-sectional data ...... 233

6.3.5 EAS measures of environmental behaviour ...... 233

6.3.6 Measure of social grade ...... 239

6.4 Direct effects of environmental class on behaviour ...... 240

6.4.1 The relationship between class membership and general level of pro-

environmental behaviour ...... 243

6.5 The moderated association between environmental concern and behaviour

247

6.6 The mediating effect of environmental class on the relationship between

social grade and behaviour ...... 250

6.7 Discussion ...... 253

6.8 Conclusion ...... 256

Chapter 7 Discussion ...... 257

7.1 Introduction ...... 258

7.2 Thesis summary ...... 258

7.3 Answers to research questions ...... 267

7.3.1 What are the components of environmental concern? ...... 267

7.3.2 How does environmental concern exist amongst the UK public? ...... 269

7.3.3 What social characteristics are associated with different forms of

environmental concern? ...... 271

7.3.4 How is concern for the environment associated with pro-environmental

behaviour, and what is the role of socio-economic status in this relationship? 273

7.4 Key messages ...... 275

7.4.1 Attitudes do influence behaviour ...... 275

7.4.2 The problem of apathy ...... 278

7.4.3 Moderation not mediation ...... 280

7.5 Further research ...... 284

7.5.1 Global environmental concern ...... 284

7.5.2 The three-step method ...... 285

7.5.3 Newly available data ...... 288

7.6 From a qualitative perspective ...... 290

7.6.1 Combining qualitative and quantitative research ...... 290

7.6.2 Further qualitative research ...... 292

7.7 Limitations of this research ...... 295

7.7.1 The advantages and disadvantages of cross-sectional research ...... 295

7.7.2 Subjective measures of behaviour ...... 297

Chapter 8 Conclusion ...... 298

8.1.1 Summary of findings ...... 301

8.1.2 How these findings relate to the broad debates in the environmental

social sciences ...... 303

8.1.3 Pro-environmental behaviour change as a policy tool ...... 305

8.1.4 Final message ...... 307

Bibliography ...... 309

Appendix a) ...... 354

Appendix b) ...... 393

Word Count: 62,388

List of Tables

Table 1: Original and Revised NEP items ...... 48

Table 2: Shortlist of datasets that capture environmental attitudes and

behaviours in the United Kingdom ...... 97

Table 3: Knott and Bartholomew (1999) classification of latent variable analysis

...... 110

Table 4: Analytical methods used to answer research questions ...... 116

Table 5: Environmental Concern Models Tested by Schultz (2000, 2001) .... 128

Table 6: Environmental Concern Models suggested by Snelgar (2006) ...... 129

Table 7: Excluded EAS measures of environmental attitudes ...... 134

Table 8: Indicator variables for subsequent latent variable analysis ...... 137

Table 9: Eigenvalues for original and parallel data ...... 149

Table 10: Variable loadings for EFA (rotated) ...... 150

Table 11: Standardised CFA results of EC model ...... 151

Table 12: Standardised BSEM coefficients for full sample ...... 155

Table 13: Standardised BSEM coefficients on 10% EAS sample ...... 158

Table 14: Maibach et al. (2011) environmental attitude segments ...... 174

Table 15 : A cultural theory-based interpretation of climate worldviews ...... 179

Table 16: Goodness of fit indices for LCA models containing two-six classes

...... 184

Table 17: Class sizes for 2-class model ...... 186

Table 18: Class probabilities for four-class model ...... 189

Table 19: Within-class household income ...... 200

Table 20: NRS social grade classification system ...... 201

Table 21: Climate change belief and most likely class membership ...... 203

Table 22: Multinomial regression of SES on most likely class membership .. 207

Table 23: Dichotomous recode of the DEFRA ‘Standard’ scale ...... 236

Table 24: Dichotomous recode of the ‘Regular Purchasing’ scale ...... 236

Table 25: Measures of specific pro-environmental ...... 238

Table 26: Logistic regression of environmental class on environmental

behaviour ...... 241

Table 27: Direct and indirect effect between environmental class, level of ... 251

Table 28: Environmental classes discovered in Chapter 5 ...... 264

Table 29: Environmental classes discovered in Chapter 5 ...... 270

List of figures

Figure 1: Attitudes formed of cognitive, affective and behavioural components

...... 40

Figure 2: Attitudes interacting with cognitive, behavioural and affective

components ...... 41

Figure 3: Attitudes forming relationships between cognitive, affective and

behavioural components ...... 42

Figure 4: Values Beliefs Norms theory of environmentalism (adapted from

Stern 2000) ...... 51

Figure 5: Information deficit mode for environmental behaviour ...... 73

Figure 6: The theory of reasoned action ...... 74

Figure 7: Theory of Planned Behaviour (Ajzen 1991) ...... 76

Figure 8: Low-cost high-cost model of pro-environmental behaviour

(Diekmann & Preisendöerfer, 1992) ...... 80

Figure 9: Proposed dual VBN value orientations ...... 127

Figure 10: Frequency distributions for indicator variables measured on a scale

of 1 (strongly disagree) to 5 (strongly agree) ...... 140

Figure 11:BSEM informative priors compared to ML – CFA priors in MPLUS

(Muthen 2012) ...... 145

Figure 12: Screeplot of eigenvalues for EFA model ...... 148

Figure 13: Model of environmental concern and goodness of fit indices ...... 153

Figure 14: DEFRA (2008) UK Environmental typology ...... 172

Figure 15: Cultural Theory (adapted from Schwarz & Thompson, 1990) ...... 178

Figure 16: Information criterion and log likelihood for one – six-class models

...... 185

Figure 17: Item probabilities for 2-class model ...... 187

Figure 18: Item probabilities for 4-class model ...... 190

Figure 19: Mean factor scores within the Pro-environment class ...... 214

Figure 20: Environmental classes in order of increasing environmental concern

...... 215

Figure 21: Moderating effect (adapted from1986) ...... 227

Figure 22: Moderation through interaction (adapted from Warner 2012) ...... 228

Figure 23: Mediated effect (adapted from Baron & Kenny, 1986) ...... 230

Figure 24: Response categories and distribution for single measure of

behaviour ...... 234

Figure 25: Response categories and distribution for participant assessment of

behaviour ...... 235

Figure 26: Within-class reported level of environmental behaviour ...... 243

Figure 27: Within-class satisfaction with environmental behaviour ...... 245

Figure 28: Pro-environment and Disengaged classes on level and satisfaction

with behaviour, moderated by social grade...... 248

Figure 29: New Environmental Paradigm (NEP) portion of the Theory of Values

Beliefs Norms ...... 259

Figure 30: Age distribution of Paradoxical and Disengaged class members 272

Figure 31: Bar chart depicting the relationship between Pro-environment class

membership probability and level of pro-environmental behaviour ...... 277

Figure 32: Within-class mean environmental attitude scores ...... 278

Figure 33: Pro-environment and Disengaged classes on general level of

behaviour, moderated by social grade...... 282

Figure 34: Spectrum of perspectives on combining qualitative and quantitative

research ...... 292

Abstract

Reports from the IPCC have been consistent in their findings: climate change is happening and human activity is the cause. The temperature of the earth’s climate has been steadily rising since the industrial revolution, with profoundly negative consequences for the natural environment. Britain is amongst the top 10 global contributors towards climate change, producing more CO2 per capita than China, and yet little is known about the relationship the British public have with the natural environment. Drawing upon DEFRA’s 2009 Survey of Public Attitudes and Behaviours Towards the Environment, a nationally representative sample of the UK, this study aims to (1) explore environmental attitudes in the DEFRA sample; (2) identify the types of environmental concern that exist in the UK and; (3) examine how environmental concern is associated with pro-environmental behaviours. The overall goal is to develop a better understanding this attitude-behaviour relationship. The thesis has 3 main findings.

First, environmental concern is formed of three environmental attitudes: (a) a cognitive appraisal of plant and animal welfare (ecocentric attitude); (b) welfare of the human race (human-centric attitude); and (c) a prioritisation of the self, alongside dismissal of environmental problems (denial).

Second, members of the British public can be assigned to one of four groups based on their environmental concern: Pro-environment, Neutral, Disengaged and Paradoxical (the latter 2 groups are apathetic towards environmental issues though in different ways).

Third, when examining behaviour variation across these environmental concern groups, it was found, unsurprisingly, that membership of the pro- environmental group is strongly predictive of pro-environmental behaviour. What was surprising was that pro-environmental concern predicts a variety of behaviours, both easy and challenging (i.e. easy behaviour such as recycling household waste as well more challenging behaviour such as an increase use of public transportation over driving), whereas previous studies have typically found such behaviours to be unaffected by attitudes. Membership of the Neutral group also predicts pro-environmental behaviours, although this relationship is weaker and exists for fewer measures of behaviour. Disengaged and Paradoxical forms of concern are not significant predictors of behaviour. Upon examining the effect of socio-economic status (SES) on group membership and this attitude-behaviour relationship, it was found that SES does not moderate the attitude-behaviour relationship, but it does influence group membership. Respondents with higher SES were more likely to belong to neutral or pro-environment groups.

After reviewing these findings, it is concluded that environmental attitudes do clearly predict behaviour, but a large portion of the UK population do not possess environmental attitudes strong enough to do so (the Disengaged and Paradoxical groups amount to 36% of the population). Future studies should focus on these apathetic groups in an attempt to understand them, determine effective methods of engagement and identify factors that increase the probability of members transitioning out of these groups.

Declaration

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

I. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes.

II. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made.

III. The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions.

IV. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available from the Head of School of Social Sciences.

Acknowledgements

I would like to express gratitude to my supervisors Mark Elliot and Paul Upham who have helped me enormously during my PhD and to whom I am deeply indebted.

It has been a privilege to study at the Cathie Marsh Institute for Social Research, and to work alongside its exceptional staff, especially my fellow PhD students Claire Shepherd, Jose Pina Sanchez, Pauline Turnball, Mollie Bourne and Ewan Carr.

I thank my wonderful and supportive family, without whom I could not have completed this thesis, and to my best friends Laura, Kristian and Steve, thank you for keeping me sane.

My ambition to do a PhD would have not been possible without the funding I received from the Economic and Social Research Council, which I have very much appreciated. It is through the generous ESRC Research and Training Support Grant that I have been able to present my research findings at a number of conferences and events.

Chapter 1 Introduction

1.1 Introduction

There is now evidence that climate change poses a real and serious threat to our natural environment. Natural scientists have described climate change as one of the most significant environmental risks confronting the world in the

21st century, yet we are far from doing everything we can to mitigate it.

Human civilisation is dependent on the natural environment. The ecosystems in which we live are not only essential to our health and quality of life, but also to our survival. Forests remove carbon dioxide and other pollutants from the air we breathe, they prevent soil erosion, landslides, and flooding. Wetlands store storm water and remove pollutants, recharging our aquifers with these filtered waters. Dune systems on our beaches provide natural barriers to storm waves and provide important habitat and travel ways for wildlife. Overall, the natural environment provides valuable raw materials for shelter, products and medicine. It allows us to pollinate crops without the need for chemicals or genetic engineering. But the natural environment is now under serious threat from climate change (IPCC 2007).

The UK is severely at risk from the effects of climate change, not only at home

(largely from flooding) but also from extreme weather elsewhere in the world that threatens to disrupt food imports, which as of 2008 accounted for 40% of all food sold in the UK. But the UK contribution to climate change is substantial. In 2009, the UK was the world's 10th largest producer of man-

made carbon emissions. The UK produces more CO2 emissions per capita than China. Reducing this detrimental environmental impact arguably requires

a better understanding of what the British public think about environmental problems, and how these thoughts correspond to pro-environmental behaviours.

Knowledge gained from understanding UK environmental concern, and disentangling the environmental attitude-behaviour relationship, can inform more effective public policy aimed at reducing UK carbon emissions. It is the goal of this research to gain such knowledge, achieved by drawing upon a nationally representative study of the British population: the Survey of Public

Attitudes and Behaviours towards the Environment. This study was commissioned by the UK Department for Environment Food and Rural Affairs

(DEFRA) and provides detailed information on attitudes towards the environment and the frequency of pro-environmental behaviours, as well as respondent characteristics. As this dataset has only recently been made available, this thesis is one of the first to examine this data.

This thesis investigates the structure and dimensions of environmental concern, its prevalence among the British population, and how environmental concern is associated with individual characteristics, such as age or income.

Furthermore, by examining the direct relationship and how this relationship is mediated and moderated by additional factors, this thesis also considers how concern for the environment translates into pro-environmental behaviour.

1.2 The importance of anthropocentric climate change

Climate change is the biggest modern day threat to the natural environment, to such an extent that the two topics cannot be discussed separately; a conversation on the natural environment cannot exist without also discussing the natural environment and vice versa.

Ultimately, tackling environmental welfare involves engaging with issues relating to climate change. This is not only due to the threat climate change poses to the natural environment, but also because the consequences of a wide variety of ‘environmentally detrimental’ not only damage the natural environment directly (e.g. causing harm to plant and animal species and causing cosmetic damage) also contribute to climate change. For example, not recycling, or simply buying products that cannot be recycling, increases the need for extracting virgin raw materials from the earth, through activities such as mining and deforestation. Mining and deforestation contribute substantially to pollution and the emission of greenhouse gases into the atmosphere. In addition, deforestation also reduces the amount of plants and trees available on earth to help reduce the accumulation of greenhouse gases

like CO2. Wasting food means the energy that has gone into the production, harvesting, transporting, and packaging of that wasted food, (an annual

average of 3.3 billion metric tons of CO2), is also wasted (Smith, R., 2015).

Travelling by car or plane rather than walking, cycling or using pubic transport

unnecessarily emits CO2 into the atmosphere. For example, according to The

Environmental Transport Association (2015) a one-way flight from London to

Manchester (185 miles) emits:

• Plane – 63.9kg per passenger if the plane is 70% full, and 44.7kg if the

plane is completely full.

• Car – based on the average car 19.8kg per person when carrying an

average 1.56 people and 7.7kg when carrying a family of four. A fuel-

efficient car with an emissions figure of 100g/km produces 11.8kg and

4.6kg respectively.

• Train – 5.2kg per passenger if the train is 70% full.

• Coach – 4.3kg per passenger if there are 40 people on the coach.

This production of greenhouse gases emitted by human activity is specifically referred to as anthropogenic climate change. This is an important problem to tackle, not only because the accumulation of environmentally detrimental

actions within a given society can add up to an avoidably large amount of CO2 emissions, but also because without intervention, such behaviours will continue and are likely be passed onto the next generation.

Understanding attitudes towards the natural environment in the UK is an important first step in tackling the problem of anthropocentric climate change.

From here, further understanding can be gained by examining how such attitudes influence behaviour, and how pro-environmental behaviour can be

encouraged through attitude change.

1.3 Environmental attitudes and behaviour

Attitudes are “unobservable cognitive constructs that are socially learned, socially changed and socially expressed” (Terry & Hogg, 2000, p. 1). During the 1920s attitudes were largely defined by behaviour, being conceptualised as behavioural tendencies (Allport 1954). Indeed, the field of social psychology was originally defined as the scientific study of attitudes (Thomas & Znaniecki,

1918; Watson 1925) because it was assumed that attitudes were the key to understanding human behaviour.

By the mid-1930s however, several studies had begun to identify a large discrepancy between attitudes and behaviour; challenging the synonymity of attitudes and behaviour. For example, LaPiere (1934) accompanied a young

Chinese couple as they travelled across America and recorded whether they received service in restaurants and hotels. LaPiere subsequently mailed a letter to each establishment they had visited, asking whether it would accept members of the Chinese race as guests. The Chinese couple were accepted and treated well in every establishment, but responses to the letter were almost entirely negative.

Corey (1937) assessed 67 university psychology students’ attitudes towards cheating at the beginning of their semester. Each student was given five objective tests at weekly intervals and was then asked to score their own

tests, giving them the opportunity to cheat. Corey found no correlation

between the students’ attitudes and their cheating behaviour. These, plus additional studies over the next 30 years (see Wicker 1969), suggested there was no relationship between attitude and behaviour. It was not until the 1970s, when researchers turned a critical eye on such evidence, that the study of the attitude–behaviour relation resuscitated. Publications from Kelman (1974) entitled ‘Attitudes Are Alive and Well and Gainfully Employed in the Sphere of

Action’ and Dillehay (1973) ‘On the Irrelevance of the Classical Negative

Evidence Concerning the Effect of Attitudes on Behaviour’, criticised previous attitude-behaviour research and argued for the predicative ability of attitudes.

This revitalisation of attitudinal research coincided with the revelation that environmental degradation is the consequence of ‘maladaptive human behaviour’ (Maloney & Ward, 1973, p. 583). This finding prompted social scientists to analyse individual motives underlying this behaviour. Such environmental studies have concentrated primarily on environmental concern or attitudes as predictors of environmental behaviour. Often the measured environmental attitudes have been broader in scope than the measured actions; for example, assessing how an individual cares about the environment and how this affects their recycling behaviour (e.g. Rajecki 1982).

Some such studies have concluded that there is an association between the two concepts, but that it is low to moderate (Eckes & Six, 1994; Fuhrer 1995;

Hines et al. 1987; Weigel 1983). Hines et al. (1987), for example, reported an average correlation of 0.35 in their meta-analysis of 128 studies. Eckes and

Six (1994) found only an average correlation of 0.26 in their meta-analysis (17 studies) and that environmental concern explains, at most, 10% variance of

specific environmental behaviours. In response to these findings, some have dismissed the claim that general environmental concern is a direct predictor of specific environmental behaviours, and have instead adopted the correspondence principle developed by Ajzen and Fishbein (1980). This posits that only when the attitudinal and behavioural measures correspond to each other concerning the relevant action, is there a substantial relationship.

Though this strategy could be seen as a temporary solution, and has allowed the study of environmental attitudes and behaviour to continue, it is not ideal.

Arguably, the goal of the correspondence principle is to find specific significant relationships, not to develop a fuller understanding of the relationship between broad-scope environmental concern and environmental behaviour. Past research suggests that caring about the environment does not make people more likely to engage in pro-environmental behaviour. Therefore there is a discrepancy between what people think about the environment and what they actually do. The mismatch between environmental attitudes and behaviours remains an unresolved research problem that poses a substantial challenge when it comes to designing climate-related policy. It is counterintuitive that on average the British population, for example, should be concerned about the environment, but that these concerns do not motivate more substantial behavioural and other change. To reduce further detrimental environmental impacts we need to unpick this attitude-behaviour disparity, determine if this occurs for all of the UK, across all environmental attitudes, and all measures of behaviour. This is what this thesis aims to do.

1.4 Why this thesis?

Despite the fact that scientific understanding of climate change is now sufficiently clear to justify taking prompt action, public consensus on climate change is out of alignment with scientific consensus. The opinion gap between scientists and the public in 2009 stood at 84% to 49% respectively, that global temperatures are increasing because of human activity. The 2011

Angus Reid Public Opinion poll found that 43% of the British public were likely to say that “climate change is a fact and is mostly caused by emissions from vehicles and industrial facilities”, compared to 49% of Americans and 52% of

Canadians. The same poll found that 20% of Britons think “global warming is a theory that has not yet been proven”. Individual-level environmental concern is only one component of the overall picture, but it is an important one that needs to be fully understood. The public will play a critical role, both in terms of their direct consumption of fossil fuels and resulting greenhouse gas emissions, and through their support for political leadership.

Unfortunately, relatively little is known about UK national public opinion or behaviour regarding climate change at a national level. This is the case because, as highlighted by Leiserowitz (2005), only a few national surveys have included even a single question on the issue. Public opinion is critical because it is a key component of the socio-political context within which policy makers operate. Public opinion can fundamentally compel or constrain political, economic and social action to address particular risks. Failure to account for public values and attitudes when taking decisions on climate risk

management will inevitably prove problematic. At a basic level, climate policies will require a certain level of support from those who will be affected if they are to be successfully implemented. Further, public support or opposition to climate policies (e.g. treaties, regulations, taxes, subsidies, etc.) will be greatly influenced by public perceptions of the risks and dangers of climate change. Leiserowitz (2005), points out that successfully mitigating / adapting to climate change will require changes in the behaviour of billions of human beings, who each day make individual choices that collectively have enormous impacts on the earth.

How the public perceive the natural environment and subsequently behave are thus vital matters. “Scientists need to know how the public is likely to respond to climate impacts or initiatives, because those responses can attenuate or amplify the impacts” (Bord, Fisher & O'Connor, 1998, p. 75). It is the responsibility of social scientists to examine this response. In order to do this, first an improved and nuanced understanding of public engagement with environmental problems must be obtained. Climate change science is well developed, relatively coherent in terms of theory and method and capable of measuring, analysing, and assessing what we do and do not know about the environmental consequences of climate change. Levin et al. (2012) argue that by comparison, social scientific research on climate change is more recent, far less coherent, and lacks consensus on either epistemological or substantive grounds. Now we need to understand environmental attitudes better, particular those held by the British public.

1.5 Research Aims

This thesis undertakes quantitative analysis of data obtained by the UK governmental Department for Environment, Food and Rural Affairs (DEFRA) in the 2009 survey ‘Public Attitudes and Behaviours towards the Environment’.

This survey captures reported attitudes and behaviours towards key environmental issues, such as energy use, climate change, water use, waste, transport and the natural environment. Data produced from this survey has been under-used; this research project is one of the first. As a result, this thesis provides a rare, national-scale examination of the British public’s reaction to and attitudes towards climate change. Significantly, this survey was designed without strict adherence to – or being constrained by – a particular theory of attitude-behaviour relationships. This in turn lends the resultant data amenable to inductive study, which again is relatively uncommon in environmental psychology. Drawing upon this data, this thesis aims to accomplish the following:

1) Explore environmental attitudes in the DEFRA sample

Environmental concern is here viewed as comprised of multiple

attitudes towards the environment. While past studies have sought to

map out the structure of environmental concern, studies testing such

models have yielded largely inconsistent results. Some studies have

sought to examine the structure of environmental attitudes, but have

done so using measures faithful to specific theories. Thus to an extent,

these studies have not set out to explore environmental attitudes but

to test a theory. Further to this, few studies use large, representative

samples and even fewer samples represent the British public; therefore

we know little about nationally representative UK attitudes towards the

environment. This thesis seeks to conduct a more inductive study of

environmental attitudes, by using a large representative UK dataset

designed without an a priori commitment to a specific theory. Results of

analysis will be interpreted by engaging with relevant theories, but

results will not be constrained by them.

2) Identify how environmental concern exists amongst the UK public

This thesis posits that environmental attitudes comprising

environmental concern co-exist in varying quantities, producing

potentially numerous ‘forms’ of environmental concern. This thesis

seeks to identify the most common forms of environmental concern that

exist amongst the British public. Grouping the British public by

environmental concern thus creates an environmental typology. Using

this typology, groups with particular socio-economic characteristics can

also be associated with particular environmental attitudes.

3) Learn how environmental concern is associated with pro-

environmental behaviours

There is a widespread assumption that concern for the environment is

positively associated with pro-environmental behaviours, yet empirical

findings remain less forthcoming.

Some studies have found that only particular types of pro- environmental behaviour are influenced by attitudes; others have found that only behaviour-specific attitudes significantly predict behaviour.

More still have found the relationship between environmental attitudes and behaviours to be weak. This study continues this line of enquiry and examines this relationship, but does so across numerous measures of behaviour, following from substantial analysis of environmental attitudes and concern. The indirect effects that socio-economic status has on this relationship are also studied.

1.6 Thesis structure

This thesis is comprised of eight chapters. Following this introduction, Chapter

2 provides a review of the relevant literature, conceptualising environmental attitudes and discussing their relationship with pro-environmental engagement.

Chapter 3 discusses the data sources and research methods used throughout the thesis. This chapter outlines the specific research questions of the thesis and describes in detail the data used, including available measures of environmental concern, pro-environmental behaviour and participant socio- demographic characteristics. Finally, the methods used in the analytical sections of the thesis are discussed in relation to their ability to provide answers to the specified research questions.

Three analytical chapters then follow. Though these analytical chapters are to some extent self-contained, each adopting a distinct methodological approach to address a specific research question, the chapters build on each other to support a central argument regarding the nature of environmental concern as exhibited in the UK population and its association with pro- environmental behaviour.

Chapter 4 is the first analytical chapter and examines the structure of environmental concern. Concern for the environment is a concept comprised of multiple attitudes, though there is a lack of consensus as to what these

attitudes are or how many there are. Observed measures of environmental

concern are selected from the dataset and are analysed using three factor- analysis methods (exploratory factor analysis, confirmatory factor analysis and

Bayesian structural equation modelling) to produce a robust, dimensional model of environmental concern.

Chapter 5 explores the possibility that individuals possess the attitudinal components of environmental concern but in varying quantities and combinations (an individual may possess strong attitudes towards the natural environmental but a weak attitude towards the earth’s resources, for example).

Various possible attitude combinations are theoretically and statistically possible. Therefore, a clustering method is adopted to identify homogeneity in participant responses to environmental attitude items, segmenting respondents according to homogeneous attitude clusters and producing a typology of environmental concern. Class members are profiled and the probability of possessing a common attitude cluster / form of environmental concern is investigated in terms of the relationship with various socio- demographic and economic factors. Chapter 6 examines the relationships between environmental attitudes and behaviour. The aims of this chapter are to determine: (i) in what ways environmental concern influences different environmental behaviours; (ii) whether different forms of environmental concern influence behaviour in the same way; and (iii) whether ‘anti- environmental’ attitudes predict a lack of environmental behaviour. Further to this, mediating and moderating factors of the relationship are also examined.

Chapter 7 revisits the aims and research questions of the thesis, providing in- depth answers and outlining the main findings of this body of work. The limitations of the data and the research are highlighted, feeding into potential questions for further research. Finally, Chapter 8 summarises the work conducted throughout this thesis, outlines what has been found and provides a closing statement.

Chapter 2 Theories of environmental attitudes and pro-environmental behaviour

2.1 Introduction

The overall purpose of this thesis is to examine British environmental attitudes and their association with pro-environmental behaviour. As outlined in Chapter

1, the specific aims of this research are to:

1) Explore the structure of environmental attitudes in the DEFRA

sample.

2) Identify the types of environmental concern that exist in the UK.

3) Learn how different forms of environmental concern are associated

with pro-environmental behaviours.

The purpose of this chapter is to define and discuss the key concepts mentioned in these aims: environmental attitudes, environmental concern, pro- environmental behaviour, and the attitude-behaviour relationship.

Sections 2 and 3 of this chapter explore and clearly define the two key operational concepts of this thesis: attitudes and environmental concern. In doing so these sections also critically examine the theories typically used to study human cognitive and behavioural responses to the environmental problems that were identified in Chapter 1. This examination identifies key weaknesses in the existing literature and thereby informs the overall research design of the thesis. Section 4 defines pro-environmental behaviour and section 5 explores evidence for the attitude-behaviour relationship from the field of environmental psychology.

2.2 Attitudes

The study of attitudes has been a prominent part of social science research for over a century. An attitude represents how people perceive important aspects of the world and guides their social lives. Its importance as a focal point of social study is enhanced by its versatility. The term ‘attitude’ is seemingly applicable across the full range of substantive social science research topics.

Much ontological and theoretical discussion of the formation, interpretation and definition of attitudes emerged from social psychology in the early twentieth century with the late 1910s and early 1920s being the formative period for attitude research and scholarship. It was at this time that the earlier introspective method of investigating attitudes was abandoned in favour of the concept of “social attitudes”, itself imported from sociology (Danziger 1997).

It is difficult to discuss attitudes without also discussing their impact on behaviour. Indeed, attitudes have been defined in relation to their role as indicators of behaviour since the origin of their study in the 1920s when they were conceptualised as behavioural tendencies, most notably by Allport

(1924). Further to this, the field of social psychology was originally defined by

Watson (1925) as the scientific study of attitudes, because it was assumed that attitude was the key to understanding human behaviour. Early attitudinal research gave no reason to doubt this assumption. Applying newly developed methods to assess attitudes1, divinity students were found to hold more

1 Often psychophysical research methods. 35

favourable attitudes toward the church than other college students (Thurstone

& Chave, 1929); military training groups, veterans, and conservative political groups had more favourable attitudes toward war than labour groups and professional men (Stagner 1942); and business men were found to be more opposed to prohibition of alcohol than Methodists (Smith 1932), and so forth

(for further examples see Bird 1940). In the mid-late 30s however, several studies demonstrating a large discrepancy between attitudes and behaviour challenged the notion that behaviour is almost entirely guided by attitudes, most famously studies by Lapiere (1934) and Corey (1937). Lapiere (1934) accompanied a young Chinese couple as they travelled across America and recorded whether they received service in restaurants and hotels. Lapiere subsequently mailed a letter to each establishment they had visited, asking whether it would accept members of the Chinese race as guests. There was no consistency between the attitudes expressed in responses to the letter and actual behaviour of the staff working at these establishments. The Chinese couple were accepted and treated well in every establishment, but responses to the letter were almost entirely negative. Corey (1937) assessed 67 university psychology students’ attitudes towards cheating at the beginning of their semester. Each student was given five objective tests at weekly intervals and were then asked to score their own tests, giving them the opportunity to cheat. Corey found no correlation between the students’ attitudes and their cheating behaviour, concluding that the “data presented in this study show overt behaviour, as measured by the amounts students will change their test papers when allowed to do their own grading, is not related to attitudinal scores derived from a highly reliable questionnaire measuring verbal opinions

toward cheating on examinations” (1937, p. 271). The Corey and Lapiere studies inspired further research into the attitude–behaviour discrepancy.

Research conducted in the 1950s and 1960s found attitudes to be very poor predictors of actual behaviour; several academics suggested abandoning the

‘attitude concept’ entirely, claiming the statistical relationship between attitude and action is so insignificant that further research was not warranted (e.g.

Blumer 1955; Campbell 1963; Deutscher 1965; Festinger 1964). After a thorough review of this literature, Wicker (1969) reached the following conclusion regarding the strength of the attitude–behaviour relation:

“Taken as a whole, these studies suggest that it is considerably more

likely that attitudes will be unrelated or only slightly related to overt

behaviours than that attitudes will be closely related to actions. Product-

moment correlation coefficients relating the two kinds of responses are

rarely above .30, and often are near zero.”

(Wicker, 1969, p. 65)

Thus, attitudes were deemed not only to be separate from behaviour, but had little to no influence upon it.

It was not until the 1970s, when researchers turned a critical eye on such evidence, that the study of the attitude-behaviour relation resuscitated.

Publications from Kelman (1974) entitled ‘Attitudes Are Alive and Well and

Gainfully Employed in the Sphere of Action’ and Dillehay (1973) ‘On the

Irrelevance of the Classical Negative Evidence Concerning the Effect of

Attitudes on Behavior’, criticised previous attitude-behaviour research and argued for the predictive ability of attitudes. Kelman (1974), for example, argued that Wicker (1969) only reviewed experimental studies, ignoring conclusions drawn from survey data that provided evidence for attitude– behaviour consistency. Dillehay (1973) pointed out that Lapiere’s research across America (1934) was a poor study of the attitude–behaviour relation, specifically as the person performing the behaviour may not have been the same person who responded to his letter. Such studies suggested that behaviours are guided by attitudes, but that the relationship between these two concepts is not what it was originally considered to be in the 20s and 30s.

Dillehay and Kelman point out that the original assumption, that attitudes are direct indicators of behaviour, is detrimental to attitudinal research, and that the “attitude is not an index of action, but a determinant component” (1974, p.

316). During the early 1970s, the attitude concept, and specifically its relationship with behaviour, was re-evaluated. In 1971, Triandis posited that

“an attitude is an idea charged with emotion which predisposes a class of actions to a particular class of social situations” (Triandis 1971, p. 21). Thus, attitudes were re-considered to be constructs that are somehow related to affective, cognitive, and behavioural components. The affective component is said to reflect the emotional underpinnings of an attitude (see Antonak &

Livneh, 1988), that is, the degree of positive or negative feelings toward a particular object. The cognitive component represents an individual’s ideas, thoughts, perceptions, beliefs, opinions, or mental conceptualisation (see

Olson & Zanna, 1993). The behavioural component includes behavioural

intentions and behavioural commitments, but not actual behaviour (Jorgensen

& Stedman, 2001). Some authors have however included behaviour itself in the notion of environmental concern (e.g. Dunlap & Jones, 2002). This multidimensional definition of attitudes is currently the most commonly used within social research, of which there are three interpretations.

1) Attitudes are formed of cognitive, affective and behavioural

components

The most common interpretation is that attitudes are composed of

cognitive, affective and behavioural components2, as shown in Figure 1.

Therefore, instead of a one-dimensional construct expressed

behaviourally, cognitively or affectively, an attitude is conceived as

being a multidimensional construct.

2 See Bagozzi (1978; 1979) and Breckler (1984)

Figure 1: Attitudes formed of cognitive, affective and behavioural components

Attitude

cognitive affective

behavioural

This interpretation proposes that behaviour is a component of attitudes.

The idea that attitudes and behaviours are closely related is something

that was criticised by Kelman (1974). This interpretation could be

problematic if the attitude in question is not directly related to a

particular behaviour. Not all attitude objects are behaviour specific, and

may indeed have little to do with subsequent behaviour (at least not

directly). For example, attitudes towards colours, patterns or music, are

more broad-scope attitudes, and are less behaviour specific than

attitudes towards driving or recycling. To assume that broad-scope

attitudes are defined partly by behaviour is arguably an inaccurate

conceptualisation.

2) Attitudes interact with cognitive, affective and behavioural

components.

Others posit a single dimension that is fundamentally evaluative (such

as Eagly & Chaiken, 1993). This evaluation then has cognitive,

behavioural and affective outlets (Fabrigar, MacDonald & Wegener,

2005), e.g. “affect, beliefs and behaviours are seen as interacting with

attitudes rather than as being their parts” (Albarracín et al., 2005, p. 5).

This is illustrated in Figure 2.

Figure 2: Attitudes interacting with cognitive, behavioural and affective components

cognitive behavioural affective

Attitude

3) Attitudes allow for relationships between cognitive affective and

behavioural components

Finally, a hierarchical interpretation proposes the existence of an

underlying general evaluative dimension responsible for the relationship

between cognitive, affective, and behavioural components (Fabrigar et

al., 2005).

Figure 3: Attitudes forming relationships between cognitive, affective and behavioural components

cognitive affective

Attitude

behavioural

Interpretations (2) and (3) suggest that attitudes are evaluations that interact with elements of affectation, behaviour, and cognition, as opposed to these three components forming a part of attitudes. This is a crucial difference.

When researching attitudes, consideration should be given as to the most appropriate interpretation, specifically if the attitude in question is related to behaviour enough to warrant the incorporation of behavioural intent within the conceptualisation of attitude. Ostrom (1969) and Kothandapandi (1971) both conceptualise attitudes as being formed of these three components, though

Kothandapandi studied the effect of behaviour-specific attitudes on behaviour, and Ostrom studied the effect of broad attitudes. Ostrom (1969) used this interpretation to examine attitudes towards the church. Measures of cognitive, affective and behavioural components were used to predict such religion- relevant behaviours as church attendance, monetary contributions to the church, or time spent in meditation. Correlations were generally low (median r

= .19), with particularly weak support for the superiority of the behavioural component measures. Kothandapani (1971) examined attitudes toward the use of birth control. There was some indication that behavioural measures were better predictors of behaviour than the cognitive and affective measures.

However, these findings had no bearing on the prediction of behaviour from attitudes because in this study, attitudes did predict behaviour: all cognitive, affective, and conative measures of attitude toward birth control correlated highly with contraceptive use (median r = .68). Ajzen and Fishbein (2005) reviewed these two studies and concluded that attitudes predicted behaviour better in the Kothandapani study than in the Ostrom study because

Kothandapani assessed attitudes toward the behaviour of interest, i.e., using birth control, whereas Ostrom assessed general attitudes toward the church to predict specific behaviours, such as donating money, attending church, and studying for the ministry.

When studying behaviour-specific attitudes, it is more appropriate to adhere to interpretation (1) of attitude structure. When studying broad attitudes, it is better to maintain an ontological distinction between attitude and behaviour.

This distinction is maintained throughout this thesis. This thesis interprets attitudes as being comprised of cognitive affective components that have the potential to interact with behaviour.

2.2.1 Conclusion

This section has discussed attitude conceptualisation, specifically with relation to behaviour, and asks the question, should attitudes be defined as having a behavioural component? The argument made here is that a behaviour-specific attitude is likely to have a behaviour component (i.e. behavioural intent) but broad-scope attitudes, for which the attitude object is removed from a specific action, will likely not have a behavioural component. The environmental attitudes being examined in this thesis are broad in scope, and as such are seen as concepts which interact with behaviour, but which are not formed of a behavioural component.

2.3 Environmental concern

This thesis uses the term environmental concern (EC) to refer to broad-scope attitudes towards the natural environment. Some studies have included a behavioural component in their definition of EC (Dunlap 2002). For these studies, EC not only consists of attitudes towards the environment and environmental problems, but also how people perceive their own level of environmental behaviour, and/or the behaviour of others. This behaviour- inclusive definition of EC will not be used in this body of work for the reasons discussed in the previous section. Behaviour should not be incorporated when conceptualising attitudes unless the attitude in question is behaviour specific.

Broad-scale attitudes, where the attitude object is more general and therefore more removed from specific behaviours, should not be considered as having a strong behavioural component.

2.3.1 Measures of environmental attitudes

Many EC measures have been produced since the study of environmental attitudes began in earnest in the 1970s. By 2002 there were at least 700, according to Dunlap and Jones (2002). Cooper and Pervin (Cooper & Pervin,

1998, p. 186) pointed out that, with regard to researchers, “most preferred to generate new scales rather than organise those already available”. Despite the development of several measures, few are widely used; for example, the

Environmental Concern Scale developed by Weigel and Weigel (1978) is a measure with documented validity and reliability that is rarely implemented.

There are two measures of EC that are widely and commonly used; the

Ecological Attitude Scale (EAS) and the New Environmental Paradigm (NEP)

Scale. Maloney and Ward (1973) developed the EAS, the original version of which contained 130 items. This was shortened two years later by Maloney and Ward to make it more convenient for researchers. The shorter version consists of 45 items and four scales:

Verbal Commitment (VC)

What the individual states they are willing to do to protect the

environment.

Actual Commitment (AC)

What an individual does to protect the environment.

Affect (A)

The degree of emotionality related to environmental issues.

Knowledge (K)

Specific factual knowledge related to environmental issues.

These scales adhere to the multi-component definition of attitudes discussed at the end of Section 2. The VC scale captures behavioural intention and the

AC scale captures actual behaviour, thus forming the behavioural component of attitudes. The A scale captures the affective attitude component and the K scale captures the cognitive component. The EAS captures behaviour, motion

and knowledge directly, capturing attitudes indirectly by using these

components as indicators of environmental attitudes. The EAS has a large behaviour component as it captures both behaviour intention (VC) and behaviour (AC). Given this, the EAS is likely best used when attempting to capture behaviour-specific attitudes where the attitude object is relevant to a particular behaviour. This is because behaviour and/or behavioural intention can be better indicators of the behaviour-specific attitudes compared to broad-scope attitudes.

The NEP has a comparatively smaller behavioural component. This scale focuses on beliefs about humanity’s ability to upset the balance of nature, the existence of limits to growth for human societies, and humanity’s right to rule over the rest of nature. Cluck (1998) commented that the NEP captures a general ‘environmental worldview’ while also examining components that contribute to this general form of concern, such as the ‘balance of nature’ or the effect of ‘humans over nature’. The NEP was developed by Dunlap and

Van Liere (1978) and at the time consisted of 12 items. It was later revised in

2000 (Dunlap, Van Liere, Mertig & Jones, 2000) to include additional items and to modernise the language (see Table 1).

Table 1: Original and Revised NEP items

Original NEP items (1978) Revised NEP items (2000)

We are approaching the limit of the number We are approaching the limit of the number of people the earth can support. of people the earth can support.

The balance of nature is very delicate and Humans have the right to modify the easily upset. natural environment to suit their needs.

Humans have the right to modify the When humans interfere with nature it often natural environment to suit their needs. produces disastrous consequences.

Mankind was created to rule the rest of Human ingenuity will ensure that we do nature. NOT make the earth unlivable.

When humans interfere with nature it often Humans are severely abusing the produces disastrous consequences. environment.

Plants and animals exist primarily to be Here has plenty of natural resources if we used by humans. just learn how to develop them.

To maintain a healthy economy we will Plants and animals have as much right as have to develop a “steady-state” economy humans to exist. where industrial growth is controlled.

Humans must live in harmony with nature The balance of nature is strong enough to in order to survive. cope with the impacts of modern industrial nations.

The earth is like spaceship with only limited Despite our special abilities humans is still room and resources. subject to the laws of nature.

Humans need to adapt to the natural The so-called “ecological crisis” facing environments because they can remake it humankind has been greatly exaggerated. to suit the needs.

There are limits to growth beyond which The earth is like a spaceship with very our industrialised society cannot expand. limited freedom and resources.

Mankind is severely abusing the Humans were meant to rule over the rest of environment. nature.

The balance of nature is very delicate and easily upset.

Humans will eventually learn enough about how nature works to be able to control it.

If things continue on their present course will soon experience a major ecological catastrophe.

A higher score on the NEP scale reflects a stronger pro-ecological worldview and stronger belief about human-nature interdependency. Many have demonstrated that this perspective reflects a higher concern about environmental issues and stronger pro-environmental attitudes (see Dalton,

Gontmacher, Lovrich & Pierce, 1999; Pierce, Dalton & Zaitsev, 1999). Dietz et al. (1998) have incorporated the NEP as a measure of beliefs in their widely used value–belief–norm (VBN) model of environmental concern and behaviour.

The VBN theory links value theory, norm-activation theory, and the NEP through a causal chain leading to behaviour: personal values, beliefs about the general environment, and personal norms contributing to pro-environmental action.

2.3.2 VBN and valuing the environment

Thomas and Znaniecki (1918) distinguished between social values and attitudes: the former are said to exist outside the individual in the broader social context; the latter are considered be individual and subjective. Since the

1960s, the cognitive hierarchy approach has been the most dominant social- psychological theory for understanding the relationship among values and attitudes. According to this model, values predispose attitudes, providing an important basis for understanding, maintaining, and/or influencing people’s attitudes toward relevant objects (Tesser & Shaffer, 1990). Researchers adopting this approach have produced results suggesting that values are significantly related to various forms of attitudes (Becker & Connor, 1981;

Donthu & Cherian, 1992; Prakash & Munson, 1985). The NEP scale is

commonly used to study environmental attitudes, conceptualised as reflections of values. Thus adhering to this value-based perspective (Hansla,

Gamble, Juliusson & Gärling, 2008; see, for example, Schultz 2001; Snelgar

2006). Indeed, the NEP, when interpreted as a component of the VBN (see

Figure 4), captures value-based environmental attitudes; that is, attitudes that derive from specific value orientations.

Figure 4: Values Beliefs Norms theory of environmentalism (adapted from Stern 2000) Activism behaviours behaviours Non-activist Non-activist organizations public-sphere public-sphere Behaviours in in Behaviours Behaviours Private-sphere Private-sphere action. Norms Personal Personal to take pro- environmental environmental Sense of obligation of obligation Sense (AR) ability to ability Perceived Perceived reduce threat reduce (AC) objects objects Beliefs Adverse Adverse for valued for valued consequences consequences (NEP) worldview worldview Ecological Ecological Values Egoistic Altruistic Biospheric

Stern and Dietz (1994) refer to value orientations as clusters of prioritised values and have proposed three forms of value orientation which effect worldview, attitude formation and environmental behaviour:

Anthropocentric

People with an anthropocentric value orientation prioritise values such as

‘a world at peace' and ‘equality', and judge phenomena on the basis of

costs and benefits to a human group or human beings in general.

Egoistic

An egoistic value orientation reflects a prioritisation of the self above the

wellbeing of the environment or others.

Biospheric

A biospheric value orientation takes into account costs and benefits to the

natural environment.

Therefore, the extent to which an individual values themselves (egoistic), other people (altruistic) or the environment (biospheric) shapes their attitudes towards the environment. Stern et al. (1993) found that these value orientations were associated with a willingness (or not) to engage in pro- environmental behaviour. Stern et al. (1995) later replicated this study on a larger sample and found the same results for egoistic and biospheric values; however in this case an anthropocentric value orientation was not significantly

related to willingness to take pro-environmental action.

Much empirical research has been conducted utilising the NEP portion of the

VBN model as a theoretical framework to clarify EC composition, though results have been inconsistent. For example, there is mixed empirical support for an independent biospheric value orientation, separate from egoistic and socio-altruistic orientations. A non-partisan biospheric component reflects the belief that nature has an intrinsic value, worth protecting for its own sake (see

Attfield 1981; Merchant 1992; Naess 1984). Steg et al. (2005) has reported direct evidence for a distinct biospheric value orientation, and has been distinguished from social-altruistic values in numerous studies (Stern et al.,

1993; Thompson & Barton, 1994).

However, in some factor analytic studies, social altruistic and biospheric value items have been found to load on the same factor (Schwartz 1992; Stern et al.,

1995; 1999). An amalgamation of biospheric concern with social-altruism suggests a desire to preserve the natural environment in order to reap its benefits (e.g. clean air, produce). Therefore the desire to protect the environment is driven by its instrumental value rather than its intrinsic value.

Alternatively Stern et al. (1995) has suggested that a combined biospheric/social-altruistic component of EC is representative of general altruism.

The research discussed above is based in the US, and the theory has not been tested on a large representative UK sample, which begs the question, are recognisable NEP/VBN components evident when using a nationally representative British sample, given the original US basis of the above? Using

the NEP component of the VBN as a theoretical framework in the analysis of this thesis can provide an answer to this question, and also aid the examination of the structure of environmental concern.

2.3.3 Social norms

Some studies, such as Newhouse (1990), suggest that environmental attitudes and values are bound by an individual’s social norm. A social norm is defined as an expectation held by an individual about how he or she ought to act in a particular social situation (Schwartz, 1977). These social norms are often enforced by the threat of punishment or promise of reward. For example, shaking hands when meeting someone new, and not cursing in polite conversation are example of social norms, adherence to which are often required in order to be accepted by those around you. However, in some circumstances, it may not be socially acceptable to behave in a way which is environmentally responsible. Newhouse (1990) for example, suggests that an individual may have negative attitude about wasting food, but is sometimes prevented from acting on this attitude by social norms, i.e. it is unacceptable to him to stand up in a banquet room and tell everyone to take only what they can eat. In addition, Oskamp et al. (1991) found that the degree of recycling by friends and/or neighbours predicted the degree of one's own recycling. In other contexts, Ajzen and Fisbein (1977) have demonstrated that attitudes predict behaviour better when no strong norms exist dictating how to behave.

Though a full exploration of the effect social norms have on the attitude

behaviour relationship is beyond the scope of this study, it is important to acknowledge the role of social norms, i.e. that the strength of values and attitudes, and the extent to which they can be acted upon, can be magnified or restricted by social norms.

2.4 Social influences on environmental attitudes

The question of what influences the conception of environmental concern has stimulated a considerable amount of research. A popular approach points to socio-demographic characteristics as potential determinants of environmental concern. This approach typically focuses on standard individual socio- economic-demographic characteristics such as gender, age, education and income, and their ability to affect levels of environmental concern.

2.4.1 Age

Van Liere and Dunlap (1980) suggested that younger members of society are comparatively more concerned about environmental deterioration than older individuals. This is because younger members of the public are less integrated into the social order, and are thus more open to change. Comparatively, older members of the public are more integrated and accustomed to certain ways of doing things, and as such, are more likely to feel threatened by change.

Younger individuals are thus more likely to adopt new practices and behaviours, and support actions against environmental deterioration. Many studies conducted in the 1980s and 1990s confirmed this, showing that young people were more concerned about environmental issues than older adults were. For example, Arcury and Christianson (1990) used a modified NEP scale to investigate the effects of experiencing drought on levels of environmental concern, the effects of which were more pronounced for younger participants.

Howell and Laska (1992) also found support for this hypothesis in a sequential,

cross-sectional study examining data from 1980, 1984, and 1988. Younger participants were found to express concern about environmental deterioration.

However, Howell and Laska (1992) also found that over these three time points the proportion of concerns over participants increased. They hypothesised that this was due to an increased media focus on environmental issues that occurred in the 1980s. As noted by Eagly and Kulesa (1997), in recycling campaigns persuasive appeals through the media have had an impact on people's attitudes. Nord et al. (1998) found a strong relationship between age and environmental concern, suggesting that differences between younger and older persons may have decreased but probably not necessarily disappeared.

More recent studies on age and environmental attitudes have continued to show that older age groups are less likely to show concern for the environment (Barkan 2004; Franzen & Meyer, 2010; Gelissen 2007). This relationship has also been demonstrated in studies of public opinion towards environmental issues in Britain, with younger people generally more likely to show environmental concern and be involved in ‘green’ behaviours (Christie &

Jarvis, 2002; Stradling, Anable, Anderson & Cronberg, 2008).

2.4.2 Income, education and lifestyle

Van Liere and Dunlap (1980) proposed that environmental concern is positively associated with increased levels of education and income. This hypothesis is derived from Maslow’s hierarchy of needs (Maslow et al., 1970), which stated

that the upper and middle classes have satisfied their basic needs (regarding physiology and safety), and are thus able to focus on satisfying other ‘higher' needs including environmental improvement/conservation. Another explanation indicates that those with higher income and higher education are able to assimilate environmental information more quickly. This increased the probability of these individuals possessing more knowledge about environmental problems. Furthermore, some studies suggest that high-income and well-educated people are more likely to have post-materialist views emphasising quality of life and environmental sustainability instead of economic growth and material possessions (see Inglehart 1995; Van Liere &

Dunlap, 1980). Dietz et al. (1998) and Kanagy et al. (1994) found income and occupation to be weak predictors of EC. However, level of education was found to be moderately associated with EC, with the well educated displaying more concern about environmental problems than their counterparts.

Alternatively, there is strong evidence that the middle/upper classes have a greater negative impact on the natural environment than lower classes. Both income and education have a strong influence on how individuals live their lives. Buying a large house, multiple cars and taking frequent holidays abroad are all examples of spending money in normative ways dictated by social class and lifestyle (Sanne, 2002). Lifestyle is relation to social class can therefore severely affect levels of pro-environmental behaviour. Though individual pro-environmental strategies may result in somewhat reduced environmental loads for some peoples, this reduction may not offset the total impact of prevailing trends in socio-economic determinants of lifestyle. High-

income consumers still tend to pollute more due to the normative level of consumer behaviour expected by their social class, and their ability to adhere to these norms due to their level of income (Laidley, 2013). For instance, recycling household waste and eating locally sourced food cannot offset environmentally detrimental behaviours expected by the middle classes e.g. frequent air travel abroad and the purchase of expensive cars. The situation becomes more complicated when middle/upper classes hold pro- environmental attitudes, and engage in some pro-environmental behaviours

(recycling), but despite their environmentally detrimental lifestyle, believe that they are helping the environment.

2.4.3 Political ideology

Greeley (1993) found that respondents who consider themselves left wing, liberal or moderate in their political views are more likely than conservatives to support stronger government regulations and/or more public spending to protect the environment. Dunlap (1975) suggested three possible reasons for this which, though proposed in the 1970s, are still applicable:

1. Business and industry that often supports conservative political parties

are opposed to environmental reforms.

2. Environmental reforms need additional government activities and

regulations to which conservatives are opposed.

3. Environmental reforms often require innovative action that is opposed

by conservatives.

2.4.4 Gender

Van Liere and Dunlap (1980) reported that initial studies of environmental concern in the 1960s and 1970s typically found no gender differences in environmental attitudes and support for environmental policies. However, studies since then report that women typically express more concern than men about local environmental problems, especially those posing health and safety risks to community members, such as nuclear waste (e.g. Davidson &

Freudenburg, 1996). Gender differences persist, though the extent seems lesser when the focus is general environmental concern and non-local problems with no identifiable health and safety risks. Bord and O’Connor

(1997) found that when measures of general environmental concern explicitly focus on risk perceptions, women consistently express comparably higher levels of concern. It was concluded that both men and women express similar levels of fondness for the natural environment, but women perceive the threat of environmental problems more acutely, and are thus better able to recognise risks involved in allowing environmental problems to continue. Overall, Bord and O’Connor (1997), as well as Blocker and Eckberg (1997) and Davidson and Freudenburg (1996), concluded that women are moderately more concerned about general environmental issues than men.

2.5 Environmental behaviour

There has been increasing interest in the behavioural component of environmental problems in the past 20 years, specifically; researchers have attempted to understand the cause of environmentally detrimental behaviour, in order to find an effective way of encouraging pro-environmental behaviour change.

2.5.1 Defining pro-environmental behaviour

Attitudes have been discussed and conceptualised in the previous sections largely regarding their relationship with behaviour, specifically the ability of attitudes to define and predict behavioural outcomes. Like attitudes, the conceptualisation of behaviour has varied over the course of their study. Early research on pro-environmental behaviour presumed it to be a unitary, undifferentiated concept. But more recent research has found pro- environmental behaviour is not clear cut. Such behaviour is most commonly defined as ‘intentionally reducing the negative impact that an action can have on the environment’ (Kollmuss & Agyeman, 2002), and has been operationalised as ‘everyday environmental behaviour’ (Tindall, Davies &

Mauboules, 2003), ‘conservation behaviour’ (Monroe 2003), ‘recycling’

(Schultz, Oskamp & Mainieri, 1995; Vining & Ebreo, 1990), ‘transport use’

(Joireman, Van Lange & Van Vugt, 2004), ‘household consumption’

(Gatersleben, Steg & Vlek, 2002), and ‘household energy use’ (Poortinga, Steg

& Vlek, 2004). Stern (2000) argued that environmentally significant behaviour is

complex, due to its variety and the number of potential causal influences on it.

In response to what he refers to as a lack of adequate conceptualisation, Stern

(2000) produced a conceptual framework of environmental behaviour, consisting of three categories:

Environmental activism

This is the active participation in pro-environmental activism, including

involvement with environmental organisations and engaging in public

demonstrations and protests.

Non-activist behaviours in the public sphere

This includes both active kinds of environmental citizenship (e.g.

petitioning on environmental issues, joining and contributing to

environmental organisations) and support or acceptance of public policies

and environmental regulations.

Private-sphere environmentalism

These are pro-environmental behaviours that are exhibited in the home,

such as reduction of household energy usage and the disposal of

household waste. Consideration for the environment when purchasing

major household goods and services are also categorised as a form of

private-sphere environmentalism; this includes the purchase of cars, white

goods, recreational travel and everyday ‘green’ consumerism. Green

consumerism is more intent focused than impact, and refers to

consumerism where consideration is given for the environmental impact of

the production processes, such as refusing to purchase products unless

their packaging is recyclable and purchasing local / seasonal produce.

Other environmental behaviours

Stern uses this category to refer to more systemic influences that exist

through organisations.

As noted by Perrin and Benassi (2009), researchers often obscure different forms of environmental behaviour into a single measure. However, Stern

(2000) argues that it may be more appropriate to consider each form of behaviour separately, as causal influences vary across the three behaviour categories.

Stern (2000) further distinguishes between different forms of pro- environmental behaviour, defining such behaviours by the intention that motivates them, and by the positive environmental impact achieved once the behaviour has been actioned.

2.5.2 Intent and impact

Environmental behaviour is often conceptualised in terms of its impact, and is therefore defined by Stern (2000) as “the extent to which it changes the availability of material or energy from the environment or alters the structure and dynamics of ecosystems or the biosphere itself” (Stern 2000, p. 408).

However, the measurement of these potential changes and/or alterations presents a challenge. The level of positive environmental impact produced by

a specific form of pro-environmental behaviour is relative; for example, cycling has a positive environmental impact compared to driving, and using energy efficient light bulbs has a positive environmental impact compared to the use of standard light bulbs. It is difficult to define environmental impact solely on its own merit, and due to its conditional nature, the level of impact is dependent on the comparative activity.

Only in the past 30 years has environmental protection become an important consideration in mainstream decision making. This development has given pro-environmental behaviour a second meaning. It can now be defined from the individual’s standpoint as “behaviour that is undertaken with the intention to change (normally to benefit) the environment” (Stern 2000, p. 408).

However, intention to create a positive impact on the natural environment does not guarantee its actualisation. An individual’s understanding of what damages or helps the environment may be flawed, and as a result, their intent may cause actions that do not have a positive environmental impact. Despite this potential problem intent is important to understanding environmental behaviour. Stern and Gardner (1981) argue that it is necessary to adopt an impact-oriented definition to identify behaviours that can achieve a positive difference to the environment. However, it can also be necessary to adopt an intent-orientated approach that focuses on motivation. This is best suited to studies that seek to understand the mechanisms of environmental behaviour and account for more complex and often sustained forms of action, such as voting behaviour and environmental vegetarianism.

2.5.3 Emotion and environmental behaviour

Since the early 1990s there has been a growing body of research into pro- environmental behaviour as an emotional response. Performing pro- environmental behaviours produce neither tangible external benefits, nor are people forced to perform them. Instead, Thøgersen (1996) identifies that those who do engage in such behaviours are compelled to by moral and emotional factors. In support of this, Kaiser et al. (1999) has found that the majority of people, at least those in affluent industrial societies (Thøgersen 1996), view ecological or environmentally friendly behaviour as part of the ‘moral domain’.

Kals et al. (1999) focus on the emotional motivations for nature-protective behaviour sighting that rational and cognitive approaches alone are insufficient to explain the motivations behind such behaviours, arguing that emotional affinity towards nature is the primary cause for this engagement. They further claim that the greater the emotional affinity the more frequent the behaviour.

Kals et al. (1999) distinguish between a ‘love of nature’ and an ‘interest in nature’, the purpose of the latter being to satisfy a scientific curiosity regarding the function of flora and fauna. In doing this, Kals et al. make a point that nature is often admired either for its intrinsic value or for its ability to provide valuable resources.

There has been further research into fear of small scale environmental hazards and how this fear affects environmental behaviour. Both Johnson and Luken

(1987) and Fisher et al. (1991) investigated the perceived environmental threat

of domestic level radon exposure and how this was related to subsequent behaviour. Though Johnson and Luken (1987) suggested that the link between perception of risk and behaviour is weaker than initially thought, this was later contradicted by Fisher et al. (1991) who found that perceptions of higher risk are linked to risk-reducing behaviour for radon in homes. Abdalla, Roach and

Epp (1992) examined the perceived risks of drinking contaminated water on behaviour outcomes and found a significant correlation between perceptions of cancer risk from a contaminated community drinking water supply, and household decisions designed to avoid using this water.

There are many things to be considered when conceptualising pro- environmental behaviour. It should be considered whether such behaviour is an emotional response (fear or love of nature), if it is defined by intention or impact, and which of the four categories outlined by Stern (2000) the behaviour being examined belongs to.

2.5.4 Social context of behaviour change

Within environmental psychology, the dominant aim has been to understand the relationships between environmental attitudes / values, and everyday behaviour. In attempting to overcome a persistent “value action gap” (Blake

1999), researchers in this field have proposed countless models and frameworks, each identifying a different array of intermediary variables between attitudes and behaviour (see Jackson 2005 for a summary). In developing these models, researchers have recognised that individuals

behave with bounded rationality, and that they are influenced by their surrounding contexts in various ways. Different researchers have understood this context in different ways. Barr (2003) for example, refers to the impact of contextualised social norms, values and attitudes on individuals’ behaviour.

Martin et al. (2006) refer to various practical “situational variables” such as lack of a doorstep-recycling scheme or space for a recycling bin as inhibiting pro- environmental decisions. Olli et al. (2001), highlight an individual’s social network as vital in shaping his/her pro-environmental attitudes and actions.

Finally, within more recent work developed by behavioural economists, this surrounding context is understood as a “choice architecture” that influences particular decisions in various. Often these ways are obvious and usually financial in nature, for example the use of a carbon tax. The primary purpose of such as tax it to financial penalise people who engage in environmentally detrimental behaviour (such as car and air travel) in an effort to dissuade them frequently engaging in such behaviour. However, sometimes more subtle approaches are employed which have the potential to influence people actions in unconscious ways. Recycling rates can be increased by matching the shape of lid of the recycling bin to the shape of the object to be recycled: small circles for cans and bottles, slits for paper. Thaler and Sunstein (2008) insist that this works by reducing the cognitive demand required to complete the task, thus increasing compliance. Hotels have been able to encourage customers to reuse towel (decreasing environmental impact without decreasing the quality of service) by putting cards in bathrooms which state

“Help us preserve natural resources by reusing your towel.” Those messages, though aimed at rational and moral processes, are reported to be less

effective than social norm based messages (Thaler and Sunstein 2008; Dolan et al. 2010).

The research in this thesis adheres more to Barr’s interpretation of context.

This research is based on the understanding that contextualised social values and attitudes can encourage individuals to choose incremental shifts to their everyday lifestyles that will, in aggregate, can add up to a more sustainable lifestyle. This thesis also acknowledges that choice architecture is effective, and that more radical call for a shift in social systems and practices is valuable

(and likely necessary) but that without first achieving fundamental shifts in social norms, values and attitudes, such radical shifts may be rejected and ultimately unsuccessful.

2.6 The relationship between environmental concern and behaviour

As discussed in the previous sections, the term environmental concern (EC) refers to broad-scope attitudes towards the natural environment, and pro- environmental behaviour is often defined as behaviour that minimises an individual’s negative impact on the natural world (e.g. reducing energy consumption/waste production). While it has been found that the latter can be influenced by economic incentives (see Stern 1999), environmental concern has had less consistent effects. Research on the relation of environmental concern and ecological behaviour has shown that, although positive, the strength of correlation varies, and on average, there is a moderate correlation between environmental attitudes and behaviour (Hines et al., 1987; for reviews see Weigel 1983). This section discusses the complicated relationship between these two concepts.

Early social science research worked on the assumption that attitudinal constructs were direct indicators of related behaviour. Though as discussed in

Section 2.2, this was soon found to be a faulty assumption. Indeed, much research since the 30s has found a discrepancy between attitudes and behaviour; the attitudes held by an individual are not always perfectly aligned.

Researchers often consider how behaviour can be brought into alignment with attitudes, under the assumption that attitudes dictate behaviour (to an extent).

This is a basic assumption underpinning the majority of attitude – behaviour models, as well as the research in this thesis. The widespread adherence to

this assumption is not only because this order of influence seems the most logical, i.e. we develop an assessment of the world we live in which influence behaviour, but also because it has been reinforced by almost a decade of social and psychological inquiry as the dominant order of causality between these two concepts.

Despite the fact that it will not be pursued in this thesis, it should be acknowledged that in some circumstances, attitudes might be altered so that they exist in alignment with current behaviours. For example, if an individual recycles because they are told to, this individual may develop positive attitudes towards recycling and or the environment. This has previously been utilised at a method on encouraging behaviour change. In 1955, clinical psychologist George Kelly developed the concept of personal constructs.

Kelly's (1955) constructs were based on the idea that each individual looks at the world through his or her own unique set of preconceived notions about it

(i.e., constructs). These constructs change and adapt as the individual is exposed to new and different situations. At the heart of Kelly's theory is the idea that individuals can seek new experiences, and practice and adapt new behaviours in order to change their attitudes (or constructs) towards the world.

He recommended that therapists encourage their patients to try out new behaviours and coping strategies; he and others often found that patients would adapt to the new behaviour patterns and subsequently change their attitudes. Though attitude change only occurred after sustained and frequent engagement with such behaviour. It has been speculated that such examples of attitude change occur due to cognitive dissonance. Cognitive dissonance is

a phenomenon in which a person experiences psychological distress due to conflicting thoughts, behaviours or beliefs. In order to reduce this tension, people may change their attitudes to reflect their other beliefs or actual behaviours (Smith and Mackie, 2007).

2.6.1 The value-action gap

The value-action gap, sometimes referred to as the attitude-behaviour gap

(Blake 1999), is the gap that can occur when the values or attitudes of an individual do not correlate to his or her actions. This disparity between attitudes and behaviour exists in varying degrees across all substantive topics, but is particularly prominent regarding environmental attitudes. The outcome is that there is a divergence between the value people place on the natural environment and the relatively low level of action taken by individuals to counter environmental problems. Research into the environmental value- action gap often focuses on cognitive theories of attitude formation and how this affects individuals’ behaviour, endeavouring to explain why high regard for environmental issues does not translate into behaviours to solve environmental issues. Results have thus far suggested there are any number of things that dictate, at any point of time, whether an individual will do X rather than Y – environmental attitudes are just one of those.

Most commentators agree there is no simple correspondence between attitudes and behaviour, but many suggest that attitudes are likely to be better predictors of behaviour if the attitudes in question are particularly strong, and

based on direct experience. It has also been suggested that situational constraints influence this relationship. Behaviour should ideally be in line with the individual's social norms, which in turn are influenced by different social, economic, demographic and political contexts. Socio-economic and demographic factors can thus potentially influence behaviour as well as attitudes towards the environment (as discussed in Section 3.2). This also suggests that attitudes are more likely to influence behaviours that are (or are perceived to be) easy to engage in, which will be discussed in more detail throughout this section.

In an attempt to understand both the relationship and the discrepancy between these concepts, much social research over the past 40 years has explored the relationship between concern for the environment and pro- environmental behaviour, developing theoretical models to explain this association.

2.6.2 Attitude-behaviour models

The oldest and simplest models of pro-environmental behaviour were based on a linear progression of environmental knowledge leading to environmental awareness and concern (environmental attitudes), which in turn was thought to lead to pro-environmental behaviour. These models assumed that educating people about environmental issues automatically resulted in pro- environmental behaviour. These have been termed (information) ‘deficit’ models of public understanding and action by Burgess et al. (Burgess,

Harrison & Filius, 1998, p. 1447).

Figure 5: Information deficit mode for environmental behaviour

Environmental Environmental Pro-environmental Knowledge Attitude Behaviour

This model (Figure 5) does not address the weak relationship between attitudes and behaviour, but instead proposes an explanation for low levels of pro-environmental behaviour. It suggests that an individual cannot have pro- environmental attitudes without knowledge; therefore knowledge is the key to increasing levels of behaviour. These models from the early 1970s were soon proven to be wrong. Research showed that in most cases, increases in knowledge and awareness did not lead to pro-environmental behaviour (see

Watson, Murphy, Kilfoyle & Moore, 1999). Later models placed the emphasis

on the direct relationship between attitude and behaviours, such as Ajzen and

Fishbein’s Theories of Reasoned Action (TRA) and Planned Behaviour (TPB).

These theories propose that, for a particular behaviour to occur, an individual’s attitudes must include an intention to carry out a specific action that reflects a reasoned evaluation of the likely consequences of that action

(Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975).

The TRA dictates that belief about behavioural outcomes, combined with the evaluation of those outcomes, determines attitudes towards that behaviour; bridging the gap between attitudes and behavioural outcomes by inserting the construct of ‘intentions’. These intentions then directly lead to behaviour (as shown in Figure 6).

Figure 6: The theory of reasoned action

Attitude toward act or behaviour

Behavioural Behaviour intention

Subjective Norm

This model assumes that rational considerations govern behaviours (Ajzen

1991; Ajzen & Fishbein, 2005). Specifically, according to the TRA, behaviour is determined by the intention (the explicit plans or motivations to commit a specific act). For example, intention to quit smoking refers to an explicit commitment to this abstinence, reflecting both a cognitive appraisal of the act

(smoking is harmful) as well as an affective evaluation (smoking is unseemly).

These intentions partly reflect the personal attitudes of individuals and thus the extent to which acts are perceived as desirable or favourable.

Applying this theory to the act of recycling, for it to occur, recycling must be seen as a positive and beneficial action (i.e. attitude towards act or behaviour).

Though arguably for a positive attitude towards recycling to be formed, the individual must first believe in the reality of environmental problems, the effects of which they are attempting to mitigate through recycling. If you do not believe there is a problem, why try to address it through positive action?

The individual may also have a strong internal locus of control. “A locus of control orientation is a belief about whether the outcomes of our actions are contingent on what we do (internal control orientation) or on events outside our personal control (external control orientation)” (Zimbardo 1985, p. 275).

Therefore, an individual with an internal control orientation believes that they themselves are able to achieve a positive impact.

Alongside the positive attitude towards recycling, it must also be an act that is considered to be good or acceptable within the individual’s social environment. This is because the degree to which significant individuals – such as relatives, friends, or colleagues – condone or criticise a particular behaviour also affects intentions. This is referred to in the TRA as the

‘subjective norm’ (Ajzen & Fishbein, 2005). The perceived importance or relevance of these significant individuals affects the extent to which their approval will shape intentions, though these vary across contexts. For

example, the beliefs of relatives are likely to influence behavioural intentions related to family life. In contrast, the beliefs of managers might be more likely to shape behavioural intention in the workplace. Therefore, if an individual’s friends and family do not recycle, and/or view this action unfavourably, then the probability of the individual recycling is diminished. Overall, an individual’s attitude towards recycling, combined with how well recycling fits with their social norms, influences their intention to actually recycle.

As noted by Werner (2004), though the TRA contributed to the study of attitudes on behaviour, it received a large amount of criticism for its inability to take into consideration social and environmental constraints, thought to be significant determinants for individual behaviour. Addressing this weakness,

Ajzen (1991) produced his Theory of Planned Behaviour (TPB) shown in Figure

Figure 7: Theory of Planned Behaviour (Ajzen 1991)

The TPB introduced perceived behavioural control as an additional component. Perceived behavioural control is the subjective perception of how easily a specific behaviour can be performed (Ajzen 1991). This is based on levels of self-efficacy and the extent to which individuals feel other factors could inhibit or facilitate the behaviour (often referred to as external control – see Kraft, Rise, Sutton & Røysamb, 2005). Perceived behavioural control is partly, but not absolutely, related to actual behavioural control (Armitage &

Conner, 2001), which in turn affects the extent to which intentions are associated with the corresponding behaviours. Perceived and actual behavioural control can sometimes diverge, such as when individuals are oblivious to factors that obstruct or facilitate the intended behaviour. The most common criticism of the TPB is the length of time it takes for the behavioural intention to manifest as actual behaviour. The model assumes that once intention has formed it will remain unchanged. Though, as pointed out by

Werner (2004), the greater the length of time between intention and behaviour, the higher the probability that intention may alter; as such, the TPB only exists within a small timeframe. This makes testing the relationship between behavioural intention and behaviour challenging.

Further to this though, the model assumes that behavioural intention directly leads to behaviour. Fishbein and Ajzen suggested that attitudes and subjective norms affect behaviour by promoting the formation of a decision or intention to act. This makes behavioural intention the direct determinant of behaviour and mediates the influence of both attitude and subjective norms as well as perceived behavioural controls. Thus, according to the TPB, intention is the

most immediate and important predictor of behaviour.

Ajzen and Fishbein (1977) emphasised the importance of using these models to measure attitudes and behaviour of equivalent levels of specificity. In doing this, the researcher is able to address the value-action group, and increase the chance of finding that attitudes are significant predictors of behaviour. For example, Kelley and Mirer (1974) found that voting behaviour was predicted by voting attitudes for 85% of their sample. Goodmonson and Glaudin (1971) also found that attitudes towards organ transplant correlated highly with participants who signed legal documents donating their own organs after death. Regarding pro-environmental behaviour, Hines et al. (1987) found a stronger relationship between pro-environmental behaviour and behaviour- specific attitudes, compared to more general attitudes towards environmental protection. Though this line of enquiry is useful, it detracts from the study of perceptions towards the natural environment and the ability of such perceptions to influence levels of pro-environmental activity.

2.6.3 Do attitudes only influence easy behaviours?

According to Rational Choice Theory (RCT) an individual is presumed to weigh the potential benefits against the potential disadvantages of a particular action. He or she will then conform to the behaviour that optimally satisfies their preferences resulting in maximum utility (Fishbein & Ajzen, 1981). For example, in the application of RCT to environmental behaviour, an individual would only separate household waste into recyclables and non-recyclables if

the advantages of doing so exceeded that of simply depositing of all waste together. From an economic perspective, both recycling and not recycling produce the same desired outcome (the removal of waste from the home).

Whereas the behavioural cost may differ in terms of effort and time required.

In their research on energy consumption, Stern and colleagues (Black, Stern &

Elworth, 1985; Gardner & Stern, 1996; Stern 1992; Stern & Aronson, 1984) posit that “a personal norm supporting energy conservation is most likely to be converted into action if the action involves little cost in time and money”

(Stern & Aronson, 1984, p. 73). Furthermore they state that “psychological variables such as attitudes and personal norms appear to have more effect on relatively inexpensive, easy-to-perform energy-saving actions” (Stern 1992, p.

285). Best et al. (2011) also suggests that a reduction in behaviour cost should lead to higher rates of recycling. A transition from a drop-off scheme to a curb-side scheme would save time and effort thus reducing behaviour cost.

By the same argument, rates of recycling would be expected to decline with increasing distance to the next container if a drop-off scheme is implemented3.

These propositions are in support of Diekmann and Preisendöerfer’s (1992) low-cost theory of environmental attitude-behaviour relation. This theory states that environmental concern influences ecological behaviour primarily in

3 Such effect of utility considerations and opportunity structures on recycling behaviour is well documented by Barr and Gilg (2005) and Do Valle et al. (2005).

situations and under conditions connected with low costs and little inconvenience for individual actors. ‘Cost’ is not only defined in an economic sense but also in a broader psychological sense that includes, among other factors, the time and effort needed to undertake a particular behaviour. The lower the pressure of costs in a situation, the easier it is for actors to transform their attitudes into corresponding behaviour. If costs are high, environmental concern does not help overcome one's reservations, and they will have little- no effect on behaviour (as illustrated in Figure 8).

Figure 8: Low-cost high-cost model of pro-environmental behaviour (Diekmann & Preisendöerfer, 1992)

Therefore, higher correlations can be expected between environmental concern and behaviour in circumstances characterised by low cost. Diekmann and Preisendöerfer (1992) conclude that positive environmental attitudes can directly influence easy, low-cost pro-environmental behaviour such as recycling, but that people with high levels of environmental awareness might

not be willing to make bigger lifestyle sacrifices (such as a significant change in travel behaviour).

Further to the low cost theory of behaviour, Tyler et al. (1982) proposed a

‘defensive denial hypothesis', which states that under high-cost conditions, those who are highly concerned about energy define and interpret the situation in a way that makes activation of environmental concern seem inappropriate. To avoid cognitive dissonance and to maintain positive self- esteem, individuals downgrade or eliminate environmental aspects in high- cost situations. Thus, this psychological mechanism makes attitudes less important in high-cost situations, compared to low-cost situations. This mechanism may explain the result of many studies (e.g. Littig 2000) that show that recycling is dominated by environmental considerations, whereas most people do not apply such considerations to their travel behaviour. Whitmarsh

(2012) supports this, finding recycling to be the most common mitigating response to environmental problems, while also finding strong resistance to changing travel habits. When provided with a list of alternative mitigation strategies, most British people claim they would recycle household waste and improve home energy efficiency, while fewer would change their transport habits or pay more to travel. US researchers have found a similar resistance to changing driving habits, while there is a greater willingness to adopt domestic energy conservation practices (Bord, O'Connor & Fisher, 2000; Fortner et al.,

2000).

2.6.4 Low priority

Another common explanation for environmental attitude-behaviour discrepancy is that the environment is a low priority compared to other social issues. Health, security, and finances are found to be more important than environmental issues for the public (Bord et al., 2000; e.g. Norton & Leaman,

2004; Poortinga & Pidgeon, 2003). So, although attitudes are likely to influence behaviour, they lose their predictive power, as pro-environmental behaviours are de-prioritised in favour of actions that directly increase income, happiness and wellbeing. This low ranking of the environment reflects a widespread perception that environmental problems are removed from space and time – that the effects are neither severe nor imminent. For example, whilst considered socially relevant, most individuals do not feel that environmental problems pose a prominent personal threat (Lorenzoni & Pidgeon, 2006). In the UK, 52% of people believe that climate change will have ‘little’ or ‘no effect’ on them personally (BBC 2004; see also Hillman 1998). The Energy

Savings Trust (2004) found that 85% of UK residents believe the effects of climate change will not be seen for decades. This lack of urgency and removal of personal threat allows other social issues to be prioritised. If conditions are met for these social issues however, it can potentially allow for an increased prioritisation of the environment; as a result those in good health, with wealth and security may have a higher probability of engaging in pro-environmental behaviour.

2.7 Conclusion

This chapter has presented a brief history of attitudinal research, to demonstrate a) how attitudes have long been defined by their influence on behaviour, b) how researchers should use caution when defining attitudes in relation to behaviour, and c) why this research maintains an ontological distinction between attitude and behaviour. Following on from this, attitudes towards the natural environment were discussed and the term environmental concern was introduced. This term is used throughout this thesis to denote broad-scope, general attitudes towards the natural environment. One of the most common conceptualisations of environmental concern is the Values

Beliefs Norms (VBN) theory.

This theory argues that environmental concern is comprised of value-based attitudes. Value is placed on the self, other people or nature, these values then influence and shape environmental attitudes. Placing high value on society for example (referred to in the VBN as a social-altruistic value orientation) forms attitudes regarding the earth’s ability to sustain the human population. Highly valuing nature (bioshpheric value orientation) forms attitudes towards the natural environment and how it is at risk from environmental problems such as climate change. Finally, valuing the self (egoistic value orientation) produces attitudes focusing on how the individual should interact with, and will be personally affected by, the environment. This theory however has not been tested on a large, representative UK sample, so it is uncertain whether these components of environmental concern exist amongst the UK population.

Further to this, little-no studies have considered that these components can exist in combination with each other in varying quantities, potentially producing multiple forms of environmental concern.

Studies since the 1970s have found that concern for the environment varies according to socio-demographic and economic factors. Though age and gender are, overall, found to be inconsistent in their predictive abilities, socio- economic variables such as education and income do consistently correlate with high levels of concern. Finally, the complex relationship between environmental concern and environmental behaviour was reviewed. There is evidence to suggest that environmental concern is affected by socio- economic status. There is also evidence to suggest that behaviour too is affected. Research generally suggests that members of higher social grades

(those who are well educated with a high managerial or professional job and high household income) are more likely to be concerned with environmental issues, and are better able to engage in actions that mitigate the effects of such issues. Very few studies have examined the role socio-economic status plays in the relationship between environmental attitudes and behaviour. For example, it is uncertain whether socio-economic status (SES) increases the probability of possessing strong environmental concern, which then translates into increased rates of behaviour, or whether high SES enables concern to lead to pro-environmental action.

From this review of the literature, and considered in context with the research aims of this thesis, the following four research questions have been

formulated:

• What are the components of environmental concern?

• What ‘forms’ of environmental concern exist amongst the UK public?

• What social characteristics are associated with different forms of

environmental concern?

• How is environmental concern associated with pro-environmental

behaviours, and what is the role of socio-economic status in this

relationship?

These are considered in more detail below.

What are the components of environmental concern?

Environmental concern is a term used to refer to broad-scope attitudes

towards the natural environment. Value-based perspective proposes that

environmental attitudes are derived from the level of value placed on the

environment. One of the most dominant theories dictating EC structure is

the NEP portion of the VBN. This states that EC is dictated by

anthropocentric, egocentric or bio-centric value orientations. Attitudes are

derived from these values. This theory of environmental concern has not

been tested on a large representative UK sample so it is uncertain whether

these components of environmental concern are present amongst the

British population. This thesis will attempt to determine whether these

value orientations exist amongst the British public.

What forms of environmental concern exist amongst the UK public?

It is hypothesised that the attitudes comprising environmental concern

exist in combination with each other, but that the strength of each attitude

may vary. If this hypothesis is correct, different forms of environmental

concern exist within a given population. For example a strong positive

cognitive appraisal of plant and animal species in combination with a weak

cognitive appraisal of natural resources would result in a form of

environmental concern that centres on the welfare of the biosphere and

natural ecosystems over energy production derived from natural

resources. This is one example of a form of environmental concern,

though there could potentially be several more. Accordingly, this thesis

examines some of the forms of environmental concern that exist amongst

the UK population.

What social characteristics are associated with environmental concern?

Van Liere and Dunlap (1980) proposed five hypotheses on how

environmental concern is associated with age, gender, social class,

residence and political ideology. Studies since then have produced mixed

results regarding age and gender, but have supported the argument that

increased environmental concern is positively associated with socio-

economic status.

Data allowing, these hypothesis will be tested to determine if, in the UK,

demographic and economic characteristics impede or facilitate the

manifestation of pro-environmental attitudes. Or alternatively whether the

cognitive appraisal of the natural environment is not influenced by such social characteristics.

How is environmental concern associated with pro-environmental behaviours, and what is the role of socio-economic status in this relationship?

Much debate has occurred over the level of influence attitudes have over behavioural outcomes, given that both have the potential to be affected by other factors. Previous investigations into the influence of attitudes on environmental behaviour have yielded largely inconsistent results. Some studies have gone so far as to suggest that attitudes are not an influential factor in the performance of pro-environmental behaviour.

Stern (2000) posits that there are many forms of environmental behaviour and that mechanisms responsible for these behaviours vary. Further to this, the low-cost high-cost model dictates that as the cost of behaviour increases, the influence of attitudes decreases. Consequently the degree of impact that attitudes have will vary across different pro-environmental behaviours.

Socio-economic status has been found to influence both frequency of environmental behaviour and strength of environmental concern. But few studies have examined the indirect role of socio-economic factors on the environmental concern-behaviour relationship. Does SES account for the relationship between these two concepts? Does it alter the strength of the relationship? Or does it have no indirect effects? This thesis will not only

examine the direct effects of environmental concerns and attitudes

towards various measure of environmentally significant behaviours, but

also how socio-economic factors influence these relationships.

The following chapter will discuss the data and methods this thesis will use to provide answers to these research questions.

Chapter 3 Data sources and analytical methods

3.1 Introduction

The purpose of this chapter is to outline the research objectives of this thesis and provide details of how these objectives will be achieved, specifically regarding data and analytical methods. This chapter is structured in the following way. In the following section, the advantages and disadvantages of secondary data analysis are discussed, in acknowledgment that this methodology will have significant implications for the results and their interpretation. Section 3 discusses the data selection process for this thesis.

The potential datasets are outlined, and justification for the selection of

DEFRA’s ‘Survey of Public Attitudes and Behaviours towards the Environment’ is provided along with details of how the data were collected. Descriptive statistics of respondent characteristics, and knowledge of environmental terms used throughout the survey will be briefly outlined. Section 5 broadly discusses available methods that can be used throughout the thesis to supply answers to the research questions posited in Chapter 2. Finally, this chapter concludes by outlining the structure of the forthcoming analytical chapters, each of which is tasked with answering a specific research question using one of the methods discussed in Section 5.

3.2 Secondary Data Analysis

One of the most important decisions when conducting social science research is choosing an appropriate source of data with which to effectively answer designated research questions. As the current thesis reports a quantitative

(rather than qualitative) body of research, choice of appropriate data is narrowed considerably. Though what narrows the selection further is the requirement of a large, representative UK sample.

Gathering a large UK sample constructed of primary data is expensive and time consuming. Though the problem of having limited time and monetary resources could be resolved through the use of online data gathering, such methods can be highly problematic. Online data collection restricts the sample to those who have the skill and facilities necessary to access and complete an online research questionnaire. And with no interviewer present, any difficulties or misunderstandings by the participant cannot be addressed (possibly resulting in item non-response or inaccurate data). Furthermore, though the use of an online questionnaire reduces time spent collecting data compared to the physical alternative, it is still time consuming and challenging to both design an appropriate questionnaire and arguably even more work to obtain a representative and sufficient sample. The use of Secondary Data Analysis

(SDA) is the most practical approach to examining the UK population. This form of analysis allows for extensive research to be conducted without the need for robust financial resources, manpower or large quantities of time.

SDA is the further analysis of an existing data gathered by other researchers, institutions or non-government organisations (Cnossen 1997). In doing,

“interpretations, conclusions or knowledge additional to, or different from, those presented in the first report on the inquiry as a whole and its main results” are produced (Hakim 1998, p. 1). SDA is as such the reshaping of data that has already undergone analysis (Arber, Dale & Proctor, 1988): an adjunct to new research through the exploration of a dataset in greater detail (Hakim

1998). However, it is important to acknowledge the methodological implications for engaging in this form of secondary research. Secondary data analysis has the potential to both support and hinder the research in this body of work.

SDA can potentially improve the analytical process as it redirects the researcher’s focus away from the practical issues of data collections and instead towards considering the theoretical aims and substantive issues of the study (1998). However, a period of familiarisation is necessary. Understanding the dataset’s variables, what they measure and how they are coded can be particularly time consuming when using large datasets. An advantage of SDA is that, as Bryman and Hardy (2004) point out, larger datasets are gathered by experienced research organisations that employ rigorous sampling procedures. Furthermore, due to the relative ease in accessing secondary datasets, researchers with different perspectives and from different disciplines are able to examine the same data and therefore analyses can be replicated and cross-examined. This helps to improve the validity and quality of the research and widens accessibility and usage. A secondary researcher is able

to ‘bring a fresh perspective to the strengths and limitations of any dataset, and be more innovative in exploiting it’ (Hakim 2000, p. 31). SDA has some disadvantages as a result of the necessary compromise made between available data and desired analysis. One major disadvantage is that secondary data is gathered to meet a set of requirements dictated by the research body commissioning the survey; the data has not been collected with the secondary analyst’s research questions in mind, leading to the possibility that the information required to adequately answer specific research questions is not available. Thus secondary analysis is limited by the measurements and conceptualisations already built into the dataset.

3.2.1 Empirical reasoning

Abduction, deduction and induction describe forms of reasoning.

“Abduction and deduction are the conceptual understanding of

phenomena, and induction is the verification. At the stage of abduction,

the goal is to explore the data, find out a pattern, and suggest a plausible

hypothesis with the use of proper categories; deduction is to build a

logical and testable hypothesis based upon other plausible premises; and

induction is the approximation towards the truth in order to fix our beliefs

for further inquiry. In short, abduction creates, deduction explicates, and

induction verifies.”

(Yu 1994, p. 23)

Fisher (1935) considered significance testing as “inductive inference” and argued that this approach is the source of all knowledge. Neyman and

Pearson (1928) maintain that only deductive inference was appropriate in statistics as shown in his school of hypothesis testing tradition. However, both deductive and inductive methods have been criticised for various limitations such as their tendency to explain away details that should be better understood and their incapability of generating new knowledge (Hempel 1965;

Thagard & Shelley, 1997). Pierce (1999) dictates that the logic of abduction and deduction contribute to our conceptual understanding of phenomena, while the logic of induction provides empirical support to conceptual knowledge. Therefore, abduction, deduction, and induction work together to explore, refine and substantiate research questions (Hausman 1997).

As the forms of secondary data are rarely designed to test the specific hypotheses for which they are subsequently used, research conducted using

SDA is likely to involve abduction. This arises from the inherently exploratory nature of the process of analysing data obtained for other purposes. This is not to say that induction and deduction are impossible with SDA, rather that strictly adhering entirely to either induction or deduction is more challenging when working with secondary, rather than primary, data. The implementation of inductive and/or deductive reasoning is heavily dependent on secondary data being capable of supporting the test of hypotheses specified after the collection of that data.

This thesis primarily follows a strategy of abduction, followed by induction. As

stated by Yu (1994, p. 1) in his description of abduction, part of this thesis will

“explore the data, find out a pattern and suggest a plausible hypothesis”.

Further to this, analysis will also exhibit elements of induction by drawing comparisons with dominant theoretical models. Crucially however, the thesis does so without limiting the terms of the analysis to those within the remit of such theories.

3.2.2 Summary

In sum, secondary analysis provides a viable method for undertaking a representative study of environmental concern and subsequent pro- environmental behaviour. Like all research methods, secondary analysis brings both advantages and limitations. The use of secondary data is a trade off between practicality and suitability, though if the dataset meets certain criteria, the penalties from this trade off will be minimal.

3.3 Data selection

In order for this thesis to provide answers to the prescribed research questions, the data analysed must adhere to the following criteria. First, the sample should be representative of the UK population, so as to accurately generate inferences regarding this population based on the analyses in this thesis. Further to this, both environmental attitudes and behaviours should be captured using multiple measures. As discussed in Chapter 2, environmental concern and behaviour are both complex concepts and are thus unlikely to be adequately captured with a single measure. Stern (2000) also stated that if possible, environmental behaviour should be captured with multiple measures that enquire about different kinds of behaviour. This is because previous studies have found the strength of attitudinal and socio-economic influence varies across different forms of behaviour. The sufficiency of available measure to investigate both a wide variety of environmental behaviours as well as components of environmental concern will be discussed in the reminder of the section. Finally, the dataset should be accompanied with details of survey design, as these will be necessary to establish whether the dataset is representative, but such details will also be used to account for the complex nature of the survey design when analysing it. Not doing so when analysing the data will likely lead to inaccurate standard errors. Datasets that meet these criteria are shortlisted and displayed in Table 1.

Table 2: Shortlist of datasets that capture environmental attitudes and behaviours in the United Kingdom

Sample Survey Country Organisation Year Size

Public Perceptions of University of East Nuclear Power, Climate England Anglia, Tyndall 2005 1,491 Change and Energy Centre Options in Britain

Survey of Public Attitudes and Behaviours Towards UK DEFRA 2009 2,929 the Environment

2008 26,730 Attitudes of European

Citizens Towards the Europe Euro barometer 2011 26,825

Environment 20144 27,998

International 1993 31,320

ISSP Special Environment Social Survey Various 2000 32,547 Module Programme

(ISSP) 2010 45,199

4 The 2014 wave of the Eurobarometer survey was not available while research within this thesis was being conducted.

The British Social Attitudes Survey, Understanding Society and the European

Values Survey were also considered for the shortlist. However, though these surveys cover a very broad range of topics, at the time this thesis was conducted, none of these surveys contained a module specifically designed to capture environmental attitudes or behaviours. As such, these surveys contained only a few questions pertaining to environmental issues providing insufficient data to answer this study’s research questions. More recently however, the British Social Attitudes Survey has begun to include measures of attitudes towards climate change, and Understanding Society now includes a module on environmental attitudes and behaviours. Many of the measures included in this module are taken from the DEFRA EAS, and as Understanding

Society is a longitudinal survey, results will eventually provide an understanding of environmental attitudes and behaviours change over time.

3.3.1 Shortlisted datasets

The survey of ‘Public Perceptions of Nuclear Power, Climate Change and

Energy Options in Britain’ was conducted in 2005 and has the lowest number of participants from the shortlist with only 1,491. Data from this survey were gathered via a nationally representative quota sample consisting of face-to- face interviews. The survey focus is public perceptions of climate change, specifically the contribution of large-scale energy production towards global warming. Participants were asked to convey their attitudes towards climate change, current methods of energy production and of nuclear energy as an alternative energy source. Although this dataset does provide information on

both behaviour and attitudes, it is extremely limited. Because the focus of the study is firmly on public perception of energy production in relation to climate change, many questions aim to capture attitudes towards energy suppliers and relevant government bodies. There are an insufficient number of items to examine the relationship between cognitive appraisal of the natural environment and pro-environmental participants, and therefore this dataset is not suitable for this piece of research.

There are two repeated cross-national datasets shortlisted in Table 1 from the

Eurobarometer and the ISSP. The ‘Attitudes of European Citizens towards the

Environment’ survey conducted by Eurobarometer in 2007 is the largest of the shortlist with a collective sample size of 26,730 taken from the 27 member states of the European Union via face-to-face interviews. In 2008, 2011 and recently in 2014, a special Eurobarometer survey was conducted to capture the attitudes of European citizens towards the environment, with a specific focus on the perceived role of the EU in tackling environmental problems. At the time of this research, the environmental attitudes Eurobarometer survey had data from two time points, as the 2014 data were unavailable. The survey is conducted across 28 European Union Member States. The sample for each country is representative of the national population 15+. The study was commissioned by the Environment Directorate-General to measure the opinions, attitudes and behaviour of Europeans towards the environment.

Participants were asked questions regarding their domestic behaviour and their attitudes towards both national and global environmental issues as well

as the role of the EU in tackling such issues.

The International Social Survey Program (ISSP) is an annual cross-national survey of public attitudes towards a range of social and moral issues. The

ISSP survey is a global survey conducted across 32 countries that contains a rotating module on environmental concern. The ISSP Environment module consists of three surveys from 1993, 2000 and 2010. The 2010 survey is a partial replication of the 2000 study and the 2000 survey is a partial replication of the 1993 study. The ISSP Environment module focuses on attitudes to the environment, environmental protection, respondents' behaviour and respondents' preferences regarding governmental measures on environmental protection. Specifically, analysis was undertaken of data collected in 1993, 2000 and 2010, which focused on measuring attitudes, perceptions, knowledge and behaviour in relation to the environment. At each point in time, samples of randomly selected individuals in the countries participating in the Programme completed a centrally developed questionnaire instrument, measuring their attitudes, perceptions and behaviour in relation to the environment as well as a range of generic and country-specific socio- demographic measures.

These two datasets would allow for a multilevel study of environmental attitudes and behaviours both within the UK and between European countries.

Data obtained at a global level over multiple time points should be analysed using multilevel or multiple-group repeated cross-sectional analysis. Results would demonstrate how environmental concern has evolved over time across

100

multiple countries, and how other countries compare to the UK. However, these surveys are not optimal for this study. The 26 countries that participated in the ISSP are not a random sample of all nations. Specifically, two-thirds of them are members of the Organisation for Economic Co-operation and

Development (OECD) and only one-third are developing nations. This selectivity could be problematic when attempting to generalise the macro-level findings. In addition, both the ISSP and Eurobarometer contain few measures of pro-environmental attitudes and behaviours – less than ten measures per survey, compared to DEFRA which has 50+. With few available measures, using the ISSP or Eurobarometer surveys could make it challenging to gain a detailed understanding of these concepts. Further to this, the focus of the

Eurobarometer survey is to capture the perceived role of the EU in combating global warming and improving environmental sustainability. The lack of data pertaining to participant attitudes and behaviours leaves an inadequate amount of data to address this study’s research questions. This is unfortunate given the research potential of this dataset with its inclusion of other European countries. Despite this potential, an examination of European countries and the perceived role of the EU are beyond the scope of this study. Neither dataset was selected for use in this thesis, but would be ideal for further research.

101

3.4 Survey of Public Attitudes and Behaviours towards the Environment

The analysis in this thesis is conducted using data from DEFRA’s Survey of

Public Attitudes and Behaviour towards the Environment (hereafter referred to at the EAS – Environmental Attitudes Survey). DEFRA is the UK government department responsible for policy and regulations on environmental, food and rural issues. The 2009 wave of EAS is used, with a sample of 2929 British participants.

Data was gathered using quota sampling via face-to-face interviews and a two-stage stratified sample design. Interviews were carried out using census output areas as sampling units. Census output areas are small, homogeneous areas, comprising about 125 – 150 households (see Vickers &

Rees, 2007 for a description of the creation of the Office for National Statistics output area classification). Output areas were also stratified by socio- economic variables within region, to ensure a representative sample of all areas. Finally, quotas were applied in each output area to control for the likelihood of selected respondents being at home. These quotas were set on sex, working status and presence of children in the household.

Using demographic quotas effectively form a second level of stratification and avoid over-representation of those groups who are more likely to be at home when interviewers call. Interviewers worked between 2pm and 8pm on weekdays and at weekends to further minimise the response bias that is introduced by only working during standard working hours. Residual non-

102

representativeness is dealt with through the use of population and design weights.

The objectives of the survey were to:

• Identify people's beliefs, attitudes, and values in relation to the

environment, climate change and pro-environmental behaviour.

• Determine reported behavioural motivations and barriers.

• Monitor self-reported behaviours across a range of behavioural areas.

• Enable DEFRA and the Energy Saving Trust to assess and baseline

attitudes and behaviours in new areas.

The questionnaire was developed jointly by Taylor Nelson Sofres (TNS),

DEFRA and the Energy Saving Trust5. Cognitive testing was used to pilot the final questionnaire, to ensure that questions were correctly interpreted and that the interview flowed properly. Twenty interviews were carried out among a cross section of respondents, with each interview lasting 30 minutes.

Cognitive interviewing is a form of in-depth interviewing which pays explicit attention to the mental processes respondents use to answer survey questions. After the cognitive interviewing, further revisions were made to the questionnaire.

5 See Appendix a) for the full questionnaire.

103

This 2009 survey is preceded by several waves conducted in 1986, 1989,

1993, 1996-7, 2001 and 2007. Data from the 2007 and 2009 waves are available from the UK Data Service. DEFRA was contacted in an attempt to obtain data from previous waves but they were unable to provide it. No reason was given. A decision was made to examine the 2009 data only. Examining data from 2 time points, spanning only 2 years apart is unlikely to produce useful results due to the small window of time being examined. So this thesis is a cross-sectional analysis of the 2009 wave of the EAS. This dataset is explicitly divided into three sections: Household and Respondent

Characteristics, Environmental Behaviours, and Environmental Attitudes.

Relevant variables are selected and discussed in the following analytical chapters.

3.4.1 The distinction between reported and observed

behaviour

The EAS, like most surveys, contains self-reported measures of environmental behaviour. Self-reports are among the most common methods used in assessing environmental conservation behaviour. The main advantage of this method being that by simply asking the respondent to indicate their level of behaviour, the researcher is able to quickly, cheaply and easily collect a wide variety of information related to varied and diverse aspects of behaviour.

However, the primary concern over this method is its potential lack of accuracy. Some researchers have suggested that the pressures to appear socially responsible may lead individuals to overstate their conservation

104

behaviour and understate the amount of resources spent (Geller, 1981;

Luyben, 1982; Warriner et al., 1984).

To test this assumption, Corral-Verdugo, V. (1997) conducted a study to compare reported and observed levels of pro-environmental behaviour. For this study, 100 housewives from Mexican families were randomly selected, their self-reported re-use/recycling behaviour, along with direct observations were analysed. Here, observational data was coded according to the frequency of re-used and recycle products found throughout the household by the researcher (an undergraduate psychology student). The frequency of observed items were then weighted by the number of family members to avoid quantification bias due to family size.

A comparison between reported and observed measures revealed low correlations of re-use/recycling. Though it should be noted that this method of quantifying reuse and recycling behaviour was not necessarily a reliable one.

The number of recycling or re-used items found during one brief single examination of a stranger’s house by a young inexperienced researcher could conceivable vary for many reasons. Furthermore this study looked only at recycling and re-using behaviour, and as such, its findings may not be true for other measures of environmental behaviour. Despite these limitations, and the fact that Corral-Verdugo’s (1997) findings are outdated and not representative of UK households, it is still important to acknowledge the distinction between observed and reported behaviour.

105

Though it is likely that reported behaviour is (to an extent) indicative of actual behaviour, until effective measures of objective pro-environmental behaviour are routinely incorporated into UK surveys, the size of this distinction is uncertain. However, it is acknowledged that the results obtained in this thesis from analysing self-reported measures of environmental behaviour are potentially subject to respondent bias.

3.4.2 A critical reflection of the EAS

Though the EAS is considered to most suitable dataset available, it is not without its limitations. Though the EAS has a wide variety of measures that capture environmental behaviour, they lack measures on voting behaviour and eating habits, two forms of behaviour influenced by environmental attitudes.

Environmental vegetarianism and veganism are two rapidly growing practices in the western world which have emerged in response to the increasingly intensive production of animal products and the detrimental effect this production has on the quality of the environment (Fox and Ward, 2008).

Furthermore, environmental attitudes and behaviours are potentially linked with political inclination and voting behaviour (Leiserowitz, 2006), but no such indicators exist in the EAS.

Aside from the content of this dataset, fully understanding the measures developed by DEFRA and their method of data gathering may be difficult as there is limited documentation available and few people left at DEFRA who know details of the EAS. This is largely due to massive budget cuts recently

106

experienced by DEFRA, as a result, many of its employees involved in producing the EAS have left. Also, the DEFRA website which previously contained information on all their surveys has also been shut down. These are inconvenient but not insurmountable problems and do not prohibit the usage of the EAS.

107

3.5 Methods

Observed and unobserved constructs are studied in this research, requiring appropriate methods. To reiterate, the first four research questions of this thesis are:

• What are the components of environmental concern?

• What ‘forms’ of environmental concern exist amongst the UK public?

• What social characteristics are associated with different forms of

environmental concern?

• How is environmental concern associated with pro-environmental

behaviours, and what is the role of socio-economic status in this

relationship?

The first two research questions look at the structure of environmental concern, and how environmental concern exists amongst the public. Latent variable methods of analysis are required to answer these research questions.

Answering the second two research questions requires methods that capture associations between measures. Consequently for these two questions path and regression analyses are used.

108

3.5.1 Latent variable analysis

Environmental concern is a latent variable (the word latent is derived from the

Latin, meaning to ‘lie hidden’). That is to say that environmental concern cannot be observed directly and as such cannot be measured by means of observable indicators. Instead, concern is measured with the use of indirect indicators reflective of the concept (Vermunt & Magidson, 2004). Historically, the concept of latent variables arose from the field of psychometrics, beginning with the general intelligence factor developed by Spearman (1904) and continuing with other psychologists such as Thomson (1951), Thurstone

(1947), and Burt (1940), who investigated the mental ability of children as suggested by the correlation and covariance matrices from cognitive test variables. Given that environmental concern is latent in nature, and in order to answer research question 1 – what are the components of environmental concern? – methods of latent variable analysis are required.

Latent variable analysis is a general term within which broadly exist four sub- types of method. Knott and Bartholomew (1999) outline these methods which are differentiated by the nature of the indicator and outcome variables. There classification is replicated in Table 3.

109

Table 3: Knott and Bartholomew (1999) classification of latent variable analysis

Indicator Variables Latent outcome Continuous Categorical

Continuous Factor Analysis Latent Trait Models

Categorical Latent Profile Analysis Latent Class Models

Heinen (1996) argues that the distribution of a continuous variable can be approximated by a categorical distribution, suggesting that the distinction between continuous and categorical indicators is not fundamental. Rather the emphasis should be on the latent outcome. Treating the latent variable as categorical groups participants according to homogeneity in participant response probabilities. Alternatively, treating the variable as continuous does not attempt to classify participants, and instead seeks out homogeneity amongst correlated coefficients. Therefore, whether environmental concern is treated as continuous or categorical dictates the method used in its analysis.

In order to answer research question 2 (i.e. What are the ‘forms’ of environmental concern that exist amongst the UK public?) this thesis will treat environmental concern as categorical in its examination of how concern exists amongst the UK population. This will group the population according to their form and/or level of environmental concern. Once groups are formed, socio- economic and demographic characteristics of group members can be examined, providing answers to research question 3 (What social

110

characteristics are associated with different forms of environmental concern?)

To answer research question 1, environmental concern will be treated as a continuous variable; using indicators of concern to reveal its dimensional components.

3.5.2 Examining direct and indirect relationships

Research question 5 (How is environmental concern associated with pro- environmental behaviours, and what is the role of socio-economic status in this relationship?) requires both direct and indirect relationships to be examined. The direct effect of environmental attitudes on the frequency of pro-environmental behaviour will be examined. How socio-economic status indirectly affects this relationship is also examined.

Fabrigar and Wegener (2011) state that there are three forms of relationship between variables: direct effect, indirect mediated effect, and indirect moderated effect. Regression is commonly used to assess direct effects between variables, and will be used to assess the relationship between concern and behaviour. To examine the indirect effect of SES on this relationship requires the use of additional methods. Path analysis (an extension of regression analysis) is a statistical technique used to examine the comparative strength of indirect effects among variables (Lleras 2005). This method provides estimates of the magnitude and significance of hypothesised causal connections between sets of variables. Variables responsible for indirect effects are classed as either moderators or mediators. Moderator

111

variables “affect the direction and/or strength of the relation between an independent or predictor variable and a dependent or criterion variable” (Baron & Kenny, 1986, p. 1174), while mediator variables “account for the relation between the predictor and the criterion” (1986, p. 1174). In short, a moderator influences the strength of a relationship between two variables, and a mediator explains the relationship between the two variables.

Testing for mediation involves examining the relation between the predictor and the outcome variables, the relation between the predictor and the mediator variables, and the relation between the mediator and outcome variables. This can be tested using path analysis; a statistical technique used to examine the comparative strength of direct and indirect effects among variables (Lleras 2005). Additionally, an effective method of testing for moderation is by examining the interaction between concern and SES. This interaction represents the combined effects of both variables on the dependent measure; therefore the impact of concern is conditional/depends on SES. This interaction will be included as an additional explanatory variable in a regression model that assesses the effect of SES and concern on levels of behaviour. Assessing the association between the interaction, and the outcome variable will show how concern, conditional on SES, affects behaviour. Both the use of interaction effects and path analysis will be discussed further in Chapter 6.

112

3.6 Methods of parameter estimation: Bayesian vs. frequentist

The methods discussed in Sections 3.5.1 and 3.5.2 are primarily ‘frequentist’ in their parameter estimation. Frequentist inference is one of a number of possible techniques of formulating generally applicable schemes used to make statistical inference. In other words, frequentist inference is drawing conclusions from sample data by the emphasis on the frequency or proportion of the data. This is the inference framework in which the well-established methodologies of statistical hypothesis testing and confidence intervals are based. But, there are two main approaches to statistics. This frequentist approach to statistics is the most widely used, and hence is sometimes called the orthodox approach or classical approach. Bayesian estimation is an alternative and considerably more modern method of estimation.

Parameters are interpreted differently in both perspectives. From a frequentist viewpoint these are fixed but unknown quantities. From a Bayesian viewpoint these are random variables. A frequentist may evaluate the parameters of a particular estimator by considering the distribution underlying the data generating process. For example, Fisher (1922) posits that the distribution of the maximum likelihood estimation is approximately normal. However,

Bayesian inference adheres to the likelihood principal conditional on the data observed. This is to say that inferences from data should depend on the likelihood of actual data under competing hypotheses, not on how data could potentially have been under a single “null” hypothesis (O'Hagan, Forster &

113

Kendall, 2004). As such, from this perspective, what is important is the observed information and the log-likelihood in one's experiment, not an average over all experiments. Bayesian inference as such, depends only on the probabilities assigned due to the observed data, not due to other data that might have been observed. Bayesian analysis typically begins with a ‘prior distribution’; a probability distribution reflecting the state of knowledge about X before data is collected. This prior distribution is updated after reviewing the new data, producing a new probability distribution for X known as the posterior distribution. From a frequentist perspective, X is an unknown constant that either exists within estimated parameters or does not.

Probability statements are only made regarding sampling, where X is a random drawn from a specified population. Frequentist techniques, such as confidence intervals and hypothesis tests, provide ways to make statements that resemble Bayesian probability statements but which only use probability in the frequentist way (Gelman & Meng, 1996).

Bayesian estimation is often used for hierarchical analysis. Though recently, new Bayesian techniques have been developed for use in latent variable analysis (more typically considered a frequentist group of methods).

Asparouhov and Muthén (Muthén 2010; Muthén & Asparouhov, 2011, 2012a) have developed multiple new Bayesian techniques for latent variable analysis.

Notable amongst them is Bayesian structural equation modelling (BSEM).

BSEM is essentially confirmatory factor analysis performed from a Bayesian perspective. It is a useful tool for analysing continuous latent variables as it is able to address several weaknesses of its frequentist equivalent, confirmatory

114

factor analysis (CFA). CFA applies unnecessarily strict models to represent hypotheses derived from substantive theory. This is largely due to the fact that

CFA automatically fixes cross loadings at exact 0 (mean = 0, variance = 0).

This strictness introduces the possibility of unnecessary model rejection (see

Marsh et al., 2009) and, furthermore, goodness of fit statistics that may not be an accurate reflection of the specified model, instead capitalising on chance

(see MacCallum, Roznowski & Necowitz, 1992). (Muthén & Asparouhov,

2012a) suggests BSEM produces an analysis that better reflects substantive theories.

BSEM is a more sophisticated approach to factor analysis and as such could benefit the study of environment concern. But Bayesian estimation has only recently been used with latent variable analysis, and little-no research has used BSEM to study environmental attitudes. As a result, it has not yet been determined how the use of BSEM can contribute to the study of environmental concern. The contribution this method can make to the study of environmental concern will be assessed in this thesis.

115

3.7 Conclusion

This chapter has discussed the use of secondary data analysis and how doing so affects the form of empirical reasoning used throughout this body of work.

Suitable datasets were then outlined as potential sources of data for this analysis. The pros and cons of these data were discussed leading to the justification for the usage of DEFRA’s EAS data in this thesis. Finally, the best methods with which to answer the outlined research questions have been discussed. As shown in Table 4, specific methods have been selected to answer this thesis’ research questions throughout the following chapters.

Table 4: Analytical methods used to answer research questions

Chapter Research Question Methods

RQ1: What are the components of Factor analysis / 4 environmental concern? BSEM

RQ2: What ‘forms’ of environmental

concern exist amongst the UK public? Latent Class

5 RQ3: What social characteristics are Analysis / associated with different forms of Regression environmental concern?

RQ4: How is environmental concern associated with pro-environmental Path Analysis / 6 behaviours, and what is the role of socio- Regression economic status in this relationship?

116

Factor analysis will be used to examine the components of the latent concept of environmental concern which will, through the analysis of relevant observable indicators, reveal its dimensional components. Latent class analysis will then be used in the following analytical chapter to group the UK public according to homogenous item response probabilities. Treating the latent outcome as categorical in this way will uncover the forms of environmental concern that exist amongst the UK population. Finally, logistic regression will be used to assess the direct effect between environmental concern and pro-environmental behaviour. How SES indirectly affects this relationship will be assessed through the use of regression analysis with an added interaction effect and by using path analysis.

117

Chapter 4 Assessing the structure of British attitudes towards the natural

environment6

6 A modified version of this chapter has been accepted for publication in the Journal of Environmental Psychology. A copy of this paper is included in Appendix b).

118

4.1 Introduction

As a psychological phenomenon, concern for the environment has been under investigation for four decades. Its study has provided a greater understanding of how individuals relate to the natural environment. When research began in earnest in the 1970s, the emphasis was on how environmental concern (EC) could spread from small groups of environmental activists to the wider population. Such research was conducted by Buttel and Flinn (1974; 1978),

Dunlap (1976) and Dunlap and Gale (1972). After the political intensity of that decade began to fade, research on environmentalism turned inward, focusing on the examination and conceptualisation of environmental attitudes, driven by the need to understand attitudinal and behavioural engagement with environmental issues. In terms of practical application, understanding public engagement with ‘the environment’ might lead to more targeted and nuanced environmental messaging, encouraging greater concern and increasing the frequency of pro-environmental behaviour. Yet achieving such benefits requires an understanding of what environmental concern is – and therein lies the focus of this chapter.

In the literature, EC is taken to refer to various combinations of: the degree to which people are aware of environmental problems, their support of efforts to solve such problems and a willingness to contribute personally to their solution (Dunlap & Jones, 2002, p. 485). This definition rightly indicates that

EC is a very broad concept covering a wide range of phenomena with multiple aspects and dimensions (Alibeli & White, 2011; see also Xiao & Dunlap, 2007).

119

Both (Dunlap & Jones, 2002) and Klineberg et al. (1998) emphasise that this broadness implicitly requires researchers to “think clearly at the outset about what aspects or facets of environmental concern they want to measure, and then carefully conceptualize them prior to attempting to measure them”

(Dunlap & Jones, 2002, p. 515), thus avoiding further ambiguity in concept definition and variations or errors in variable measurement.

EC is generally considered to be attitudinal in nature. Minton and Rose (1997) conceptualise EC as constructed from a broad range of environmental attitudes. Similarly, Vining (1992) treats EC and environmental attitudes as synonymous, defining EC as the development of an array of attitudes toward the environment. However, there is only weak consensus on the specific structure of these attitudes and as such the composition of EC varies across studies. Furthermore, on-going EC research is required due to the continuously changing nature of both environmental problems and the relationship of the human population to these. As the effects of climate change are experienced and perceived in different ways by different people in different countries and mediated by a host of differing factors, attitudes are likely to change in unpredictable ways. To reiterate Stern et al. (1995), “Although it is safe to expect many newly described environmental conditions to take form as social attitude objects, it is not easy to predict what form they will take, what attitude will form about them, or whether public opinion will be of one mind or be fragmented” (Stern et al., 1995, p. 1612). Without greater clarification of the structure and composition of EC in any given study, a clear understanding of attitudinal and behavioural engagement with current environmental issues is

120

unlikely to emerge.

Many researchers such as Yin (1999), Cottrell (2003), Schultz et al. (2004) and

Milfont and Duckitt (2010) adhere to the orthodox three-component attitude model as an approach for specifying the broad structure of environmental attitudes. However, some contemporary attitude theorists hold that cognition, affect and behaviour form the basis of evaluations of particular psychological objects. For example, Albarracín et al. (2005, p. 5) state “affect, beliefs and behaviours are seen as interacting with attitudes rather than as being their parts”. This contemporary approach suggests that attitudes should be conceptualised as evaluative tendencies that can both be inferred from and have an influence on beliefs, affect and behaviour.

A combination of these two theoretical perspectives is used throughout this thesis. Here, EC is considered to be a concept that consists of cognitive and affective components, which behaviour interacts with but does not form a part of. As with many other attitudinal constructs, EC has many mediating and moderating influences between the internal, latent concern and the outward environmental behaviour, therefore it seems most appropriate to treat behaviour and attitudes as ontologically distinct. This thesis maintains a strong theoretical statement about what EC is throughout.

121

4.2 The New Environmental Paradigm

There are an extraordinary number of measures of environmental attitudes, a fact which led Stern (1992, p. 279) to describe the situation as an “anarchy of measurement”; however, only three have been widely used and had their validity and reliability assessed. These are the Ecology Scale (Maloney &

Ward, 1973; 1975), the Environmental Concern Scale (Weigel & Weigel, 1978), and the New Environmental Paradigm (NEP) Scale (Dunlap & Van Liere, 1978;

2000). These three scales examine multiple phenomena or expressions of concern, such as beliefs, attitudes, intentions and behaviours. They examine concerns about various environmental topics, such as pollution and natural resources. Hence, according to Dunlap and Jones’ (2002) typology these measures are all multiple-topic/multiple-expression assessment techniques.

Although widely used, both the ecology scale and the environmental concern scale include items tapping specific environmental topics that have become dated as new issues emerge (Dunlap & Jones, 2002; 2003). The NEP Scale avoids this issue by using only general environmental topics that do not become dated, therefore better measuring the overall ongoing relationship between humans and their environment. The NEP Scale deliberately measures an ecocentric system of beliefs (i.e. humans as just one component of nature) as opposed to an anthropocentric system of beliefs (i.e. humans as independent from, and possibly superior to, other organisms in nature)

(Bechtel, Corral-Verdugo, Asai & Riesle, 2006).

122

The original NEP Scale was published in 1978 by Dunlap and Van Liere

(Dunlap & Van Liere, 1978) and consists of 12 items (8 pro-trait and 4 con-trait) responded to on a 4-point Likert scale (anchored by strongly agree to strongly disagree). This was later updated in 2000 by Dunlap et al. (Dunlap et al., 2000) who included additional items to make the scale more psychometrically sound, and updated the terminology used. The items in both versions were developed based on both the literature on the emergence of the NEP worldview and consultation with environmental experts. The items are intended to tap three main facets of EA: a belief in (1) humans’ ability to upset the balance of nature, (2) the existence of limits to growth, and (3) humans’ right to rule over the rest of nature.

The NEP Scale measures the overall relationship between humans and the environment; higher NEP scores indicate an ecocentric orientation reflecting commitment to the preservation of natural resources, and lower NEP scores indicate an anthropocentric orientation reflecting commitment to exploitation of natural resources.

4.2.1 The VBN value frame

Since the late 1990s, a second wave of EC study has, by asking fundamentally different questions, opened promising new lines of inquiry. Rather than investigating general attitudes about environmental issues, this research seeks to identify underlying values that provide the basis for environmental attitudes

(e.g. Schultz & Zelezny, 1999), thus moving towards a more differentiated

123

conceptualisation of environmental attitude formation. Values are usually theorised as being relatively stable over the life course and allow individuals to subjectively judge what is important (Slimak & Dietz, 2006). Attitudes, by contrast, are mutable and can appear, disappear and change over time (Stern et al., 2000). Understanding the link between values and attitudes is important.

One approach is to view relatively enduring value orientations interacting with more fluid contextual (and life course) factors to produce attitudes. A key theory that embodies this approach is the value-belief-norm theory described by Stern et al. (2000; 1995; 1999).

The VBN links three theoretical models – norm-activation theory, the theory of personal values, and the NEP – into a unified explanation for environmentalism. It postulates that the consequences that matter in activating personal norms are those that are perceived as adverse to whatever the individual values. Thus, people who value other species highly will be concerned about environmental conditions that threaten those valued objects, just as those who care about other people will be concerned about environmental conditions that threaten the other people’s health or wellbeing and so on. Values are, here, considered as an orientation.

“Values are defined as prescriptive beliefs about end states of existence

(e.g. peace) and modes of conduct (e.g. justice) that transcend specific

objects and situations and that are held to be personally and socially

preferable to opposite end states of existence (e.g. war) and modes of

conduct (e.g. injustice) (Rokeach 1973). Rokeach’s definition captures the

124

traditional view of values as principles or ideals about what ought to

happen in a society (Kluckhohn 1951; Smith 1963; Williams Jr 1968),

regardless of context and situation (Smith 1965). Values are widely

regarded as principles that guide the formation of attitudes and actions

(Ball-Rokeach & Loges, 1994; Kluckhohn 1951; Rokeach 1973; Smith

1963; Williams Jr 1968). Compared with attitudes, values are regarded as

more central, deeply considered, strongly held, stable, limited in number

and connected with many other beliefs (Ball-Rokeach & Loges, 1994;

Inglehart 1977; Rokeach 1973).”

(Braithwaite, 1998, p. 224–225).

While the VBN theory is primarily intended to explain behaviour, embedded within it is a theory of environmental concern, specifically the NEP portion

(highlighted in Figure 4). The theory postulates that values are at the core of environmental concern (Slimak & Dietz, 2006) and that an individual’s value orientation is focused on the self, other people or nature, and from these value orientations, corresponding attitudes of EC are formed. Of these components, egoistic concerns are based on a person valuing himself or herself above other people and other living things. As Stern and Dietz (1994, p. 70) observe,

“Egoistic values predispose people to protect aspects of the environment that affect them personally, or to oppose protection of the environment if the personal costs are perceived as high”. Social-altruistic values lead to concern for environmental issues when a person judges environmental issues on the basis of costs to or benefits for people in general. Biospheric EC is based on a value for all living things, regardless of any social or personal benefits the

125

natural environment may yield.

Much empirical research has been conducted utilising the NEP portion of the

VBN model as a theoretical framework in attempts to clarify the composition of EC, though results have been inconsistent. For example, there is mixed empirical support for an independent biospheric-value orientation, separate from egoistic and socio-altruistic orientations. A non-partisan biospheric component reflects the belief that nature has an intrinsic value, worth protecting for its own sake (see Attfield 1981; Merchant 1992; Naess 1984).

Steg et al. (2005) have reported direct evidence for a distinct biospheric-value orientation, which has been distinguished from social-altruistic values in numerous studies (Gagnon, Thompson & Barton, 1994; Stern et al., 1993).

However, in some factor analytic studies, social-altruistic and biospheric-value items have been found to load on the same factor (Schwartz 1992; Stern et al.,

1995; Stern et al., 1999). Stern et al. (1995) has suggested that a combined biospheric/social-altruistic component of EC is representative of general altruism.

In another permutation, Schultz (2000; 2001) found a distinct biospheric concern, with egoistic and social-altruistic concerns combining into a single component. This result is in line with Thompson and Barton’s (1994) proposition that environmental attitudes originate from either an anthropocentric or ecocentric-value focus. These varied findings challenge the

VBN model, in that they do not conform to the notion of three separate and distinct value orientations. Instead attitudes of EC seem to be derived from

126

two possible dichotomised values sets, as shown below in Figure 9. Both of these dichotomous value orientations represent the appreciation of nature for either its intrinsic or instrumental value.

Figure 9: Proposed dual VBN value orientations

Thompson and Barton (1994) Stern (1995) Ecocentric Anthropocentric General Altruistic Individualistic

egoistic biospheric biospheric egoistic social- social- altruistic altruistic

These contrasting findings and reflections raise the question of the veridical value/attitude structure for EC. In response to such inconsistencies, both

Schultz (2000; 2001) and Snelgar (2006) have tested several different factor structures for EC. As shown in Table 2, Schultz (2000; 2001) tested one-, two- and three-factor measurement models for EC. The three-factor model

(highlighted in Table 5) was found to be both theoretically and statistically optimal, adhering to the VBN model and satisfying both the K1 and scree plot tests.

127

Table 5: Environmental Concern Models Tested by Schultz (2000, 2001)

Model 1 One-factor model: Uni-dimensional EC

Two-factor model: Biospheric items loading onto one factor with both egoistic and altruistic items loading on another factor. This is Model 2 consistent with Thompson and Barton's (1994) classification of environmental attitudes.

Three-factor model: Egoistic, altruistic, and biospheric concerns Model 3 fitted the data well, providing support for the notion that three value-orientations underlie EC.

Snelgar's (2006) later study tested a total of five models, including both dichotomous value orientations shown in Table 6, and found that a two-factor model with a distinct biospheric component had the best fit to the data.

Overall however, the best model was a four-factor structure, in which the biospheric attitude split into two separate biospheric concerns for plant and animal life.

128

Table 6: Environmental Concern Models suggested by Snelgar (2006)

Model 1 One-factor model: Uni-dimensional EC.

Two-factor model: Egoistic items load onto one factor, both altruistic and biospheric items load onto a second. This is based on Model 2 Stern et al.'s (1995) suggested that biospheric value may be part of a general-altruistic cluster.

Two-factor model: Egoistic and altruistic items load onto one factor, biospheric load onto a second. This provided a better fit of Model 3 the data than Model 2, supporting Thompson and Barton's (1994) dichotomous value orientation.

Three-factor model: Separate biospheric, egoistic and social- Model 4 altruistic components, as suggested by the VBN model.

Four-factor model: Distinct egoistic and social-altruistic Model 5 components, as well as two separate biospheric components for plant and animal life. This model provides the best fit to the data.

Overall therefore, studies suggest that the biosphere is something that is intrinsically valued by some people. However, Snelgar’s (2006) study suggests that there is a distinction between concern for the welfare of animal species and the preservation of plant life. This presents the additional possibility that biospheric concern is either formed of two components, or forms the basis for two separate biospheric attitudes.

129

4.2.2 Theory-driven questionnaires

Studies that aim to examine EC from the VBN perspective naturally use question scales designed to reflect components of the VBN. Hence the

Environmental Concern (EC) and Adverse Consequences (AC) scales were used to generate data for both the Snelgar and Schultz studies. The EC scale constructed by Schultz (2000) employs the statement: I am concerned about environmental problems because of consequences for ‘______’. Respondents are then asked to rate nouns such as: me, my health, people in the community, future generations, plants, trees, whales, etc. The AC scale has been described as a measure of general beliefs about environmental consequences (Stern et al., 1995). A set of items on a Likert scale measures awareness of consequences relating to each of the egoistic, social and biospheric value orientations.

EC studies that have implemented these scales have reported exploratory and confirmatory analyses verifying Stern’s VBN structure (Hansla et al., 2008;

Milfont & Gouveia, 2006). This, though, is in one sense unsurprising, given that these scales only measure components of the VBN structure. Granted the studies could have found no evidence in support of the VBN framework, but if other possibilities were not sought, it is debatable whether they would be found, a view consistent with Duhem’s view (1906, p. 1954) that theories give meaning to ‘facts’ (Oberheim & Hoyningen-Huene, 2009); work conducted in a deductive mode may or may not be open to finding other patterns in the data, depending on the researcher’s objective. Such scales also oblige respondents

130

to answer according to the structure of the questions, thus increasing the likelihood of them confirming the VBN structure in their responses.

One particular problem with theory-driven survey design is that the instrument is not an independent tool for testing the theory. The survey instrument used in many of the above studies is precisely designed to tease out the structure of EC, the likelihood of finding the NEP structure and no other is therefore greatly enhanced. Inductive, secondary data analysis of representative survey data provides at least a partial solution to this problem of circularity. Data generated without an a priori commitment to a specific theoretical framework places fewer limitations on participant responses, potentially reducing bias and more fully allows for results that are outside of the model. If when using secondary data that, while palpably about environmental concern is not theory-specific, and the same structure emerges, then the evidence for theory is much stronger. If it does not, then modifications to the theory should be considered. This approach shifts the emphasis of the research away from measuring the extent to which people subscribe to NEP components, and towards determining whether a population exhibits NEP components at all.

Thus, this approach has the potential to not only independently test theories of

EC, but also to reveal alternative EC attitudes.

A secondary problem of theory-driven scale implementation is the burden placed on researchers to gather a suitable sample, ideally a representative one. Given the high demand on time and resources required to gather primary data, such a sample often cannot be obtained. For example, conclusions

131

drawn by Schultz (2000; 2001) cannot be generalised to their respective populations given their use of small and unrepresentative samples; both studies consisted of psychology undergraduate students from the United

States (samples of 400 and 1010 respectively). Stern seems to have specialised in idiosyncratic sampling. For example, Dietz, Stern et al. (1998) actively dropped 10% of his sample who are in “other” or “Jewish” categories.

Stern et al. (1995) used random digit dialing to select 199 Virginia households.

Snelgar (2006) obtained a convenience sample of 368 participants. Of these participants, 296 were undergraduate students taking psychology modules at the University of Westminster. The remaining 72 participants were recruited with the use of snowball sampling. Snelgar acknowledges that due to these sampling methods, conclusions about larger populations cannot be drawn.

Results that cannot be generalised to the wider population are diminished in value: it is uncertain whether the findings relate to the wider population or if they are simply characteristics of the sample acquired.

4.2.3 Aims

This chapter aims to answer three main questions. First, can and do theoretically familiar EC constructs emerge from large-scale environmental attitude and behaviour survey data without the use of strict EC scales?

Second, if so, are recognisable NEP / VBN components evident when using a nationally representative British sample? Third, what is the value of an ontological distinction between attitudes and behaviours in this context?

132

4.3 Data

Analysis is performed on DEFRA’s 2009 wave of the EAS. This dataset is explicitly divided into three sections: Household and Respondent

Characteristics, Environmental Behaviours, and Environmental Attitudes. The latter portion of the EAS dataset will be used for this analysis7. From this section, variables that in some way express belief or affect in respect of the environment were selected.

4.3.1 Selecting variable for environmental concern

Variables for this analysis were selected from those explicitly defined by the dataset as reflections of environmental attitudes. These items were developed to measure British public attitudes towards the environment, without commitment to one specific theoretical framework. Selection was based on a theoretical assumption that EC is primarily a cognitive and affective state.

Variables that were not compatible with this assumption were excluded from analysis and are shown in Table 7.

7 Please see Appendix a) for information on all variables from this portion of the dataset

133

Table 7: Excluded EAS measures of environmental attitudes

I don't really give much thought to saving energy in my home

People have a duty to recycle

I don't pay much attention to the amount of water I use at home

I find it hard to change my habits to be more environmentally- friendly

I would only travel by bus if I had no other choice

Any changes I make to help the environment need to fit in with my lifestyle These measures capture attitudes I often talk to friends and family about the things they can do to help the environment towards environmentally I try to persuade people I know to be more environmentally friendly friendly behaviour I’ve suggested improvements at my workplace/the place where I study to make it more environmentally friendly

I don’t believe my everyday behaviour and lifestyle contribute to climate change

We should all try and save water regardless of whether it rains or is sunny

I sometimes feel guilty about doing things that harm the environment

I make an effort to buy things from local retailers and suppliers

If government did more to tackle climate change, I’d do more These measures too

capture self If business did more to tackle climate change, I would too

efficacy – and It's not worth Britain trying to combat climate change, because attitudes in other countries will just cancel out what we do

relation to others It's not worth me doing things to help the environment if others don't do the same

134

It's only worth doing environmentally-friendly things if they save you money These measures For the sake of the environment, car users should pay higher capture financial taxes status and People who fly should bear the cost of the environmental concern over damage that air travel causes income I would be prepared to pay more for environmentally-friendly products

It’s important to me that I can be proud of my local environment

There are many natural places that I may never visit, but I’m glad they exist

It really disappoints me when I see big offices and public buildings with their lights on when the building is empty These measures indirectly capture It really bothers me when I see people wasting energy or food general It would embarrass me if my friends thought my lifestyle was purposefully environmentally friendly environmental concern Being green is an alternative lifestyle it's not for the majority

I need more information on what I could do to be more environmentally friendly

‘Waste not want not' sums up my general approach to life

The Government is doing a lot to tackle climate change

135

Environmental attitude statements that were in part behavioural – that is, statements that commented on the execution, frequency or opinion of environmental behaviour – were excluded, in order to maintain an ontological divide between attitude and behaviour. Statements that remarked on the willingness of participants to incur a financial penalty for engaging in environmentally detrimental activities, or pay an increased price for comparatively environmentally friendly products were also excluded.

Responses to such statements are indicative of participant willingness to dispense with monetary resources to achieve a positive effect (or avoid a negative effect) on the environment. Consequently, responses are potentially influenced by participant income or wealth (and indeed their attitudes to the same). To include such variables would be to introduce additional variance into the analysis – constraining EC and potentially producing results relating to income or wealth. Undoubtedly such variables do have a relationship with environmental concern but they are almost certainly confounding. What remains are raw belief statements unmoderated by extraneous influences.

These items were derived from responses to the statements shown in Table 8 with which participants indicated levels of agreement on a five-point Likert scale.

136

Table 8: Indicator variables for subsequent latent variable analysis

Variable Name Statement

If things continue on their current course, we will Major Disaster soon experience a major environmental disaster.

Limited Resources The Earth has very limited room and resources.

The so-called ‘environmental crisis’ facing Crisis Exaggerated humanity has been greatly exaggerated.

The effects of climate change are too far in the Too Far in Future future to really worry me.

We are close to the limit of the number of people Over Populated the earth can support.

I do worry about the changes to the countryside in Changes to Countryside the UK and the loss of native animal and plants.

I do worry about the loss of animal species and Loss of Animal Species plants in the world.

Climate change is beyond control – it’s too late to Beyond Control do anything about it.

The environment is a low priority compared to Low Priority other things in my life.

137

4.4 Methods

For this analysis, a combination of Exploratory Factor Analysis (EFA) and

Confirmatory Factor Analysis (CFA) methods are used, as well as an additional new method, Bayesian Structural Equation Modelling (BSEM).

4.4.1 Exploratory Factor Analysis

Deciding upon the optimal number of factors to be retained from EFA is crucial. It is important to distinguish between major and minor factors; specifying too few or too many can distort results. There is no clear consensus for factor retention criteria. The most commonly used method is known as the

K1 rule, which retains factors with eigenvalues greater than 1 (see Kaiser

1960). Another method for retaining factors is through the examination of

Cattells (1966) scree plot for breaks and discontinuities, only retaining factors above a significant inflection. This method suffers from subjectivity and ambiguity, particularly if there is no clear inflection. A third method is Parallel

Analysis (PA), a Monte Carlo simulation approach which uses random data with the same number of observations and variables as the original data (see

Fabrigar, Wegener, MacCallum & Strahan, 1999; Hayton, Allen & Scarpello,

2004). The correlation matrix of random data is used to compute eigenvalues; these eigenvalues are then compared to the eigenvalues of the original data.

The optimum number of factors is the number of the original data eigenvalues that are larger than the random data eigenvalues. This method adjusts for sampling error and is a sample-based alternative to the K1 rule and scree plot

138

examination. Ironically, PA has enjoyed both a substantial affirmation in the methods literature for its performance relative to other retention criteria, while at the same time being one of the least often used methods in actual empirical research (cf. Hayton et al., 2004; Patil, Singh, Mishra & Donavan, 2008;

Velicer, Eaton & Fava, 2000). Methods papers making comparisons between retention decisions in PCA and FA have tended to ratify the idea that PA outperforms all other commonly published component retention methods, particularly the commonly reported Kaiser rule and scree test (Cattell 1966) methods. In most studies, one or two of these methods are used, however in this analysis all three are used to ensure the best possible model fit and accurate interpretation of retained factors. The production of factors through the use of EFA is generally followed by their rotation so as to improve their interpretability and to simplify the factor structure (Thurstone 1935; 1947).

Oblique rotation is used here as it allows factors to correlate and given that factors within this model form the EC attitude object, it is highly likely that they will correlate. The Maximum Likelihood (ML) EFA fitting procedure is used for this analysis. Though most research typically uses Principal Components

Analysis (PCA) or Primary Axis Factoring (PAF) methods of EFA, maximum likelihood allows researchers to test for the statistical significance of and correlations between factors, as well as generating goodness of fit statistics.

Gorsuch (1990) has shown important differences between PCA and common factor solutions such as principal axis and maximum likelihood factoring. In such cases, the evidence favours the common factor model as the more accurate. Conway and Huffcutt (2003) therefore urge researchers to make greater use of common factor model approaches (maximum likelihood in

139

particular due to the fit indices that can be used to help determine the number of factors).

Variable distributions

The ML method of EFA is best suited to indicator variables that are normally distributed. The frequency distributions of the nine indicator variables are displayed in Figure 10. All measures exhibit a skew, and the distribution of the species variable is severe, demonstrating a floor/ceiling effect.

Figure 10: Frequency distributions for indicator variables measured on a scale of 1 (strongly disagree) to 5 (strongly agree)

crisis people resource

future countryside disaster

control priority species

140

A potential strategy for remedying non-normality is to transform raw variable scores so that they more closely reflect a normal distribution. However this approach has a few possible disadvantages. First, transformation is not always successful at reducing the skewness or kurtosis of a variable.

Transformed variables must be reassessed in order to verify the success of the transformation in approximating normality. Though this may not be strictly speaking a disadvantage, it can potentially exacerbate the skewness. Second, in addition to altering the distribution of variables, nonlinear transformation often changes relationships between the transformed variable and other variables in the analysis. Therefore, fit statistics, parameter estimates and standard errors derived from an EFA model using transformed indicator variables may differ dramatically from such analysis that uses original variables

(West, Finch & Curran, 1995). This not only makes interpreting parameter estimates difficult, but it also strains reality if the true population distribution is not normally distributed. An alternative to transformation is the use of robust methods of estimation. Generally, robust estimation seeks to provide methods that emulate popular statistical methods, but in a way that is not overly affected by outliers or other small departures from model assumptions.

Classic estimation methods rely heavily on assumptions that are sometimes, as here, not met in practice. In particular, it is often assumed that the errors are normally distributed, at least approximately, or that the central limit theorem can be relied on to produce normally distributed estimates. Robust estimation takes into account non-normality of indicator variables, allowing for good performance despite a departure from parametric distributions. MPLUS provides maximum likelihood robust (MLR) estimation of standard errors and

141

robust chi-square tests of model fit, and so it will be used for all frequentist models in this analysis.

4.4.2 Confirmatory Factor Analysis

Once the optimal number of factors is established and a factor model is generated, its structure is specified and tested through CFA. Goodness of fit indices are used. There are many such indices; reporting them all would be a hindrance to the interpretation of model validity. The most common index is the chi-square, which should always be reported as it shows the difference between expected and observed covariance matrices (Hu & Bentler, 1999).

According to various studies (Hu & Bentler, 1999; MacCallum, Browne &

Sugawara, 1996; Yu 2002) the TLI, CFI, and RMSEA indices should also be reported alongside the chi-square statistic.

Root Mean Square Error of Approximation (RMSEA)

This evaluates how close the model fits the data and takes a value

between 0 and 1 with a smaller value indicating better model fit. According

to Chau and Hocevar (1995), acceptable model fit is indicated by a value

of 0.06 or less.

Comparative Fit Index (CFI)

CFI compares the existing model fit with a null or independent model

where the latent variables are assumed to be uncorrelated with the

observed ones. The CFI ranges from 0 to 1, where 1 is the best possible

142

fit. Garson (1998) suggests that a CFI of 0.9 or greater is indicative of a

successful model.

Tucker-Lewis Index (TLI)

Like the CFI, the TLI reflects the proportion by which the specified model

fits the data compared to the null model. Thus, a value of .50 indicates an

improved fit of 50% compared to the null model. Generally a value of less

than 0.9 is taken to indicate that the model may need to be re-specified.

The standard CFA model imposes the assumption that all cross-loadings are fixed at zero. To relax this assumption, modification indices are reviewed to identify any additional cross-loadings that would improve the overall model fit.

The modification indices present the reduction in the overall chi-square value that is produced by allowing a given cross-loading to be freely estimated.

However, caution should be exercised when considering the incorporation of any additional crossloadings into the model (MacCallum et al., 1992); allowing modification indices to dictate the substantive content of the model can compromise its interpretation and lead to overfitting. Any modifications made to the model should make theoretical sense, and not simply because the analysis indicates the addition or subtraction of a parameter (Schreiber, Nora,

Stage, Barlow & King, 2010).

143

4.4.3 Bayesian Structural Equation Modelling

BSEM is a newly developed form of CFA that incorporates the use of Bayesian estimation and priors to relax the (potentially unnecessary) strict CFA assumptions regarding crossloadings.

It is argued that current analyses using ML apply unnecessarily strict models to represent hypotheses derived from substantive theory. This is largely due to the fact that CFA automatically fixes cross loadings at exact 0 (mean = 0, variance = 0). This strictness introduces the possibility of unnecessary model rejection (see Marsh et al., 2009) and furthermore, goodness of fit statistics that may not be an accurate reflection of the specified model, instead capitalising on chance (see MacCallum et al., 1992). Muthén & Asparouhov

(2012a) suggest BSEM as an alternative approach, intended to produce an analysis that better reflects substantive theories.

Unlike the frequentist approach, in Bayesian estimation parameters are viewed as variables as opposed to being a constant (which is estimated by maximising a likelihood computed for the data). Bayesian estimation combines prior probability distribution (the prior) to an uncertain parameter θ with the data likelihood to form posterior distributions for the parameter estimates. The posterior provides an estimate in the form of a mean, median, or mode of the posterior distribution (Muthén 2010).

When prior information is available about θ, it is included in the prior

distribution of θ (informative prior). For this analysis, prior information of θ

144

is obtained through the use of EFA. Therefore the posterior distribution of θ from the EFA model is used as the prior distribution of θ for the BSEM model.

Figure 11:BSEM informative priors compared to ML – CFA priors in MPLUS (Muthen 2012)

BSEM uses Bayesian analysis to specify informative priors for such parameters, replacing parameter specification of exact zeros (0 mean, 0 variance) used in standard CFA with approximate zeros (0 mean, small variance), as shown in Figure 11. The model is therefore identified through substantively-driven (informative) small-variance priors, which are based on previous EFA analysis.

This method could potentially be valuable for the study of environmental concern. Environmental concern, though a singular concept, is formed of numerous components that are undoubtedly related to each other in some form or another. Imposing the assumption that cross loading should be restricted to exact zero when defining a model which dictates the composition of environmental concern, is unlikely to be necessary.

145

A disadvantage of BSEM is that unlike CFA there are very few indices for determining model fit. The posterior predictive p-value (ppp) (Gelman, Meng &

Stern, 1996; Guttman 1967; Meng 1994; Rubin 1984) is the primary method for carrying out model evaluations and comparisons that has become common in

Bayesian model checking, partly in consequence of its easy implementation by MCMC methods. The intention is to quantify the degree of surprise by observing what we actually have observed, in view of the prior and model.

BSEM is a more sophisticated approach to factor analysis and as such could benefit the study of environment concern. But, Bayesian estimation has only recently been used with latent variable analysis, and little-no research has used BSEM to study environmental attitudes. As a result, it has not yet been determined how the use of BSEM can contribute to the study of environmental concern. As such, this chapter will use both BSEM and CFA following EFA analysis to study the composition of environmental concern.

146

4.5 Analysis

This analysis is divided into three parts. First, EFA is used to determine the basic structure of the model. This is done through the use of PA and K1 factor retention criteria, as well as consideration of Cattell’s (1966) scree plot. Parts two and three assess the fit of this model through the use of both CFA and

BSEM. As well as measuring the model structure and fit, both methods will be evaluated relative to their ability to assess the structure of environmental concern.

4.5.1 Part one – Exploratory Factor Analysis

EFA is performed on the nine indicators variables shown in Figure 6. Three factor retention criteria are implemented to determine optimal number of factors. The scree plot displayed in shows no single point of inflection and appears to suggest the retention of between two and four factors.

147

Figure 12: Screeplot of eigenvalues for EFA model

2.5

1.5

Eigenvalue 1

0.5

0 1 2 3 4 5 6 7 8 9 Factors

According to K1 factor retention criteria, factors generated with an eigenvalue

>1.0 are to be retained. Parallel analysis produces eigenvalues from randomly generated parallel data. If eigenvalues from this parallel data are smaller than those from the original data, then this is indicative of an optimal model. As shown Table 9, both the K1 and parallel analysis methods emphasise a three- factor model.

148

Table 9: Eigenvalues for original and parallel data

Eigenvalues

Factors Original Parallel Data Data

1 2.76 1.11

2 1.48 1.08

3 1.12 1.05

4 0.77 1.03

5 0.67 1.02

6 0.62 1.00

7 0.57 0.98

8 0.53 0.96

9 0.49 0.93

The rotated factor loadings of the three-factor model are displayed in Table

10. Variables with a coefficient above the minimum criteria of .3 are highlighted to indicate their contribution to that factor.

149

Table 10: Variable loadings for EFA (rotated)

Factor Variable 1 2 3

Exaggerated Crisis 0.61 0.19 -0.05

Over Populated -0.10 0.66 0.00

Limited Resources -0.02 0.60 0.02

Too Far in Future 0.74 -0.01 0.00

Major Disaster 0.22 0.42 0.04

Changes to Countryside 0.00 0.04 0.64

Beyond Control -0.52 0.18 -0.05

Low Priority 0.49 0.00 0.16

Loss of Animal Species 0.01 -0.01 0.73

F1 1

F2 -0.26* 1

F3 -0.36* 0.43* 1

p<.05

4.5.2 Part Two – Confirmatory Factor Analysis

Table 11 shows loadings for the CFA model as well as correlations between factors. The structure of this model is dictated by the results of the EFA. Items in Table 10 that have a coefficient >0.3 are allowed to load freely onto their respective factors; all other loadings are restricted to 0. There is a clean cut between the loadings above and below the threshold. The lowest that is

150

included is 0.42 and the highest that is not included is 0.22, suggesting the emergence of a clear model.

Table 11: Standardised CFA results of EC model

Variable Estimate S.E.

Exaggerated Crisis 0.63* 0.02

Too Far in Future 0.74* 0.02 F1 Beyond Control -0.46* 0.03

Low Priority 0.57* 0.02

Major Disaster 0.53* 0.03

F2 Limited Resources 0.62* 0.03

Over Populated 0.57* 0.03

Changes to Countryside 0.67* 0.03 F3 Loss of Animal Species 0.71* 0.03

F1 F2 -0.38* 0.04

F2 F3 0.49* 0.04

F3 F1 -0.42* 0.03

*p<.05

Factor 1

This factor is captures Denial and is formed of the following key

components:

• Scepticism (positive loading of the exaggerated crisis variable)

151

• A desire not to act in response to environmental problems at present

(positive loading of both the low priority and too far in the future

variables)

• The belief that it is not too late to do something about the environment,

that problems can be controlled as necessary (negative loading of the

control variable)

Factor 2

This factor captures Human-Centric concern. The Over Populated and

Limited Resources variables together indicate an EC with respect to the

human population, specifically their impact on the planet and its ability to

sustain them.

Factor 3

This factor captures Ecocentric concern, demonstrating a distinct

ecocentric component, encapsulating concern for both animal species

and countryside.

4.5.3 Modification indices

Modification indices (MI) indicate some possible improvements to the model.

However, of the potential cross loadings that have a high MI, most contradict the interpretation of the factors and as such are not incorporated into the model. Thus heeding warnings from many academics who emphasise that MI

152

should not drive the substantive content of the model. Potential crossloadings that support current interpretations only marginally improve model fit when incorporated into the model. Thus for the model to remain parsimonious, none of the crossloadings suggested were included.

Figure 13: Model of environmental concern and goodness of fit indices

e1 e2 e3 e4 e5 e6 e7

Too Far in Beyond Exageratted Major Limited Over Low Priority the Future Control Crisis Disaster Resources Populated

Human-Centric Denial Concern

e8 e9

RMSEA 0.037 Changes to Loss of Animal CFI 0.982 Countryside Species TLI 0.947

SRMR 0.015 Ecocentric Chi-2 61.041 Concern df 12

Figure 13 shows the composition of this model and its goodness of fit statistics. CFI, TFL and SRMR statistics all indicate good fit.

153

4.5.4 Part three

Model estimation was performed with 50,000 iterations using the Markov chain Monte Carlo algorithm and the Gibbs sampler. Two chains are used for this analysis. The details of the technical implementation of BSEM are described in Asparouhov and Muthén (2012b). Model convergence was assessed with the potential scale reduction factor (PSRF) diagnostic, with a

PSRF value of 1.1 or smaller regarded as evidence of convergence. BSEM model fit was assessed with ppp, a well-fitting model is expected to show a posterior predictive p value around .5 (Fong & Ho, 2013).

In this model, the Disaster variable loads negatively onto factor 1, unlike its behaviour in both EFA and CFA models where its coefficients were positive.

This negative loading is in line with the interpretation of this factor (that it captures denial towards environmental problems). However, the control variable now has a positive loading. Variable loadings for factors 2 and 3 are similar to those from the CFA model.

154

Table 12: Standardised BSEM coefficients for full sample

Factor Variable Estimate S.D. Lower Upper

Crisis 0.84* 0.04 0.76 0.93

Future 0.66* 0.05 0.58 0.76

Control 0.58* 0.04 0.50 0.67

Priority 0.54* 0.04 0.45 0.62

F1 Disaster -0.24* 0.05 -0.35 -0.14

Resource -0.03 0.07 -0.17 0.12

People 0.15* 0.07 0.00 0.30

Countryside 0.01 0.11 -0.18 0.24

Species 0.13 0.18 -0.17 0.56

Disaster 0.45* 0.04 0.36 0.53

Resource 0.68* 0.05 0.60 0.79

People 0.73* 0.05 0.64 0.84

Future 0.01 0.06 -0.11 0.13

F2 Control 0.20* 0.05 0.10 0.30

Priority 0.04 0.05 -0.05 0.14

Countryside 0.08 0.05 -0.03 0.17

Crisis -0.19* 0.05 -0.29 -0.08

Species -0.01 0.06 -0.13 0.11

F3 Countryside 0.65* 0.07 0.53 0.81

155

Species 0.98* 0.12 0.78 1.27

Crisis 0.07 0.06 -0.04 0.19

Future 0.03 0.06 -0.09 0.16

Control -0.06 0.05 -0.16 0.05

Disaster 0.05 0.05 -0.06 0.15

Resource 0.01 0.07 -0.13 0.14

Priority -0.16* 0.05 -0.25 -0.06

People 0.01 0.07 -0.14 0.14

F2 F1 -0.36* 0.12 -0.58 -0.12

F1 -0.51* 0.14 -0.76 -0.21 F3 F2 0.44* 0.11 0.22 0.63

PPP = .012

*p<.05

The ppp of this model suggests extremely poor goodness of fit, contradicting the goodness of fit indices of the CFA model. However, this may be because the ppp is a poor indicator of model fit. The ppp is a chi-square based statistic. Because the chi-square statistic is a statistical significance test, it is sensitive to sample size which means that the models with large sample sizes are often rejected (Bentler, 1980; Jöreskog and Sörbom, 1993).

To determine if this model rejection is a type 2 error, a small subsample is

randomly generated from the EAS sample, approximately 10% in size

156

(n=287), and analysed (see Table 13). The model produced using this subsample produces a ppp = .501, indicating that the model fits the data well.

This drastic change in ppp suggests that it is overly sensitive to sample size.

157

Table 13: Standardised BSEM coefficients on 10% EAS sample

Estimate S.D. Lower Upper

Crisis 0.55 0.42 -0.78 0.89

Future 0.59 0.42 -0.74 0.90

Control -0.31 0.24 -0.56 0.41

Priority 0.39 0.27 -0.45 0.64

F1 Disaster 0.20 0.13 -0.16 0.38

Resource 0.13 0.11 -0.15 0.31

People -0.05 0.10 -0.23 0.20

Countryside 0.09 0.10 -0.11 0.27

Species 0.07 0.27 -0.27 0.55

Disaster 0.45* 0.10 0.28 0.68

Resource 0.57* 0.11 0.34 0.78

People 0.90* 0.15 0.37 1.00

Future -0.10 0.09 -0.27 0.11

F2 Control 0.10 0.08 -0.08 0.26

Priority 0.07 0.08 -0.08 0.22

Countryside -0.03 0.07 -0.16 0.11

Crisis 0.09 0.08 -0.07 0.26

Species 0.00 0.02 -0.04 0.04

F3 Countryside 0.59* 0.09 0.41 0.76

158

Species 1.02* 0.19 0.85 1.37

Crisis -0.03 0.12 -0.21 0.29

Future -0.05 0.12 -0.25 0.26

Control -0.08 0.09 -0.27 0.11

Disaster 0.13 0.09 -0.10 0.29

Resource -0.06 0.09 -0.23 0.11

Priority 0.15 0.10 -0.05 0.34

People -0.02 0.08 -0.22 0.11

F2 F1 0.14 0.31 -0.85 0.57

F1 0.30 0.43 -0.98 0.73 F3 F2 0.37* 0.19 0.12 0.99

PPP = .501 *p<.05

159

4.6 Discussion

The purpose of this chapter has been to uncover latent components of EC from the EAS dataset; a large, nationally representative British dataset complied from a survey without an explicit, particular theoretical basis. Items corresponding to a specified theoretical understanding of EC were selected from the environmental attitudes section of the dataset. Exploratory and confirmatory factor analyses, as well as Bayesian structural equation modelling, were performed and a three-factor model of EC was produced.

4.6.1 Similarities with the NEP

This model has some similarities with those produced by Stern and Dietz

(Stern & Dietz, 1994). Though the third factor (denial) may not routinely map onto the third factor of the VBN (egocentric concern), some previous research suggests that denial is related to an egoistic/self-enhancement value orientation (Hansla et al., 2008).

It could be that the drivers of denial and those of egocentric concern co-occur and it would certainly be an interesting study to establish if this was so. It is suggested that those who score highly on this factor may be exhibiting a form of denial or resignation, where an expressed lack of EC is used as a coping mechanism in the face of numerous environmental problems. Thus, environmental problems may be recognised but not fully accepted. Indeed,

Oskamp (2000) argues that because environmental problems are so large,

efficacy amongst the public is likely to be low. Low efficacy in the face of

160

an alarming issue such as climate change can produce defensive responses including message avoidance, issue derogation, or denial. This supposition is supported by Stoll-Kleeman et al.’s (2001) study. Conducted in Switzerland using focus groups, Stoll-Kleeman et al. (2001) identified widespread denial in the face of climate change, arising in part from a sense of helplessness and despair at the magnitude of the problem and the smallness of individual contributions to its solution. Efficacy would, therefore, appear to be an essential construct in understanding the public’s denial to climate change.

Although simply unwillingness may be a factor, low efficacy and denial are likely to be important inhibitor of action.

This interpretation would explain why the MI output indicated an improved model fit with an additional cross loading for the major disaster variable, as well as this item’s modest loading for both BSEM and EFA models; i.e. it is to some degree recognised that there will be an environmental disaster if things continue, but individuals may either be unwilling to change or unable to cope with this prospect. While a psychological relationship between egocentrism and denial is intriguing, it is not one that we are able to explore directly here, but which merits follow-up work. Regarding these differences, the denial factor could be interpreted as mapping onto the egoistic component of the

VBN.

Factor two closely corresponds to the social altruistic component of the VBN model. The variable loadings of this factor suggest recognition of society’s environmental impact, though the focus is on the earth’s ability to continue

161

meeting growing needs of this population. Due to the limitations of the data, this factor not altruistic in the sense intended by Stern (1994): the variables that have loaded onto this factor appear to indicate a concern for the earth’s ability to continue meeting the needs of human society rather than a concern for the welfare of society. Therefore due to the lack of solely altruistic variables in the EAS data, this factor has been labeled here as Human Centric, though it is plausible to theorise that these two things are conceptually isomorphic.

The final factor reflected an ecocentric concern in that it concerns the impacts on the non-human parts of the biosphere. Overall therefore, the configuration of the factors extracted does align with the VBN model, though the meaning of the denial factor does need to be considered in more detail.

4.6.2 Reflections on the data

The EAS is intended to measure environmental attitudes, norms, values and behaviours, including barriers to pro-environmental behaviour. The survey is not intended to embody a particular theoretical commitment. The results produced from the analysis of the EAS provide broad support for the VBN, in that the factors found could conceivably be attitudes of EC derived from the three value orientations outlined by the VBN.

This analysis suggests that there is value to this dataset in terms of its ability to characterise EC in the UK. However, whilst the 2009 EAS is part of a series of public attitude surveys run by DEFRA, data from the majority of previous

162

waves can no longer be obtained from the commissioning government department. In light of this and the nuances in the results presented here and meriting follow-up work, it is recommended that serious consideration be given to longitudinal maintenance of the EAS. Longitudinal methods of data analysis are particularly appropriate, given that attitudes are subject to change, particularly environmental attitudes, as previously noted by Stern

(2000) and more recently Melis et al. (2014).

4.6.3 Reflection on the methods

Measurement artifacts

In EFA, multiple factors emerge for several reasons. One possibility is that

there are different conceptual constructs being measured by the

questionnaire. Another possibility is that the wording of the questionnaire

is such that factors emerge due to ‘measurement artifacts’. Measurement

artifacts are findings that are not the result of what has been captured

within the data, but are instead the result of confusing questionnaire

design; specifically, the negative wording of questions. Negatively worded

and correspondingly reverse-scored items have been (and are still) used

extensively in survey constitution to guard against acquiescent behaviours

(Cronbach 1950) or the tendency for respondents to generally agree with

survey statements more than disagree. In addition, such items are used to

guard against participants developing a response set in which they pay

less attention to the content of the item and provide responses that relate

more to their general feelings or the subject than the specific content of

163

the item. However, a respondent may misunderstand a negatively worded

question and provide an incorrect answer. If this is response pattern is a

common occurrence amoungst the population, negatively worded

measures, when incorporated into an EFA model, can potentially load onto

a single factor. As factor one of the model generated in this analysis

(denial) is mostly comprised of negative measures, it is possible that it is

the result of a methodological artifact, rather than an actual latent

dimensional construct.

In defence of this analysis, the three negative items which load onto factor

one are not strictly speaking negatively worded, but are instead negative

statements (i.e. relating to a negative opinion towards an event, person,

place or object). For example, on a multiple-choice attitude measure in

which most items do not contain any negative words (e g. not), an item

containing such a word would be deemed negatively worded. This is not

the case for the three items in questions, which consist of short and

clearly written statements. Furthermore, given that data for the EAS was

gathered via Computer Assisted Personal Interviewing (CAPI) (and as such

an interviewer was present while the respondent was completing the

questionnaire to help / guide them) carelessness of the respondent is

unlikely to have greatly interfered with the data to the extent that its

analysis would produce a methodological artifact.

164

Bayesian Structural Equation Modelling

BSEM is a new method with which the mean and variance of cross loadings can be specified through the use of priors, providing a more sophisticated method for establishing model fit for latent variables models.

However, an appealing aspect of traditional CFA is its production of goodness of fit indices (though the shear abundance of them can itself be problematic, model fit is established by examining an appropriate combination of these indices). BSEM therefore is at a disadvantage; model fit is currently measured solely by the posterior predictive p-value. This is a chi-square based statistic, and as such is particularly sensitive to sample size. Given the large sample size of the EAS, the ppp is not a reliable measure of model fit as demonstrated by the drastic increase in its value when the model was restricted to a small sub-section of the sample.

Indeed, studies such as Fong and Ho (2013) and Golay et al. (2013), which have reported the benefits of BSEM over ML-CFA (often finding model fit when CFA wouldn’t), have used samples of N=312 and N=249 respectively. Overall, BSEM theoretically is more sophisticated and nuanced compared to CFA as it does not apply unnecessary restrictions to variable loadings, making it particularly useful if all factors within the model are theoretically linked (as they often are). However, its heavy reliance on its one, problematic chi-square based model fit statistic (ppp) currently renders it unsuitable for studying environmental concern within this thesis. Though, with further development, this method is likely to have great value within this field of study.

165

4.7 Conclusion

In this chapter the concept of EC and its structure has been examined. EC is defined as based on a two-component attitudinal model based upon relevant affect, keeping behaviour ontologically distinct. Through a series of analyses of a representative sample of UK residents focusing particularly on those questions that express environmental concern defined as above, a three- factor solution emerges. That structure largely overlaps with the VBN model of environmental concern. To the extent that these emerge from a different set of items to those contained in the NEP questionnaire this can be interpreted as an affirmation of that component the VBN model. The denial component of the model is interpreted as an expression of egoistic value orientation, though further research should be conducted to investigate this possibility.

166

Chapter 5 Producing a UK environmental attitude typology

167

5.1 Introduction

The previous chapter examined the structure and dimensions of environmental concern, finding three components: denial, ecocentric and human-centric. It is important to understand the components of environmental concern; to do so is to examine environmental concern at a conceptual level. But how does environmental concern exist amongst the UK population? This chapter answers this question by adopting an alternative approach to the study of EC, using latent class analysis (LCA). Participants of the EAS dataset, based on their responses to questions measuring EC, are assigned to four latent classes. This creates a typology, where the British public are grouped into classes based on their concern for the natural environment.

168

5.2 Previous research on population segmentation

Population segmentation is common in the social sciences, but comparatively rare in environmental psychology. Segmenting a given population according to a set of designated criteria is often used in disciplines with an emphasis on clearly communicating research findings and in order to improve or inspire behavioural change/engagement such as epidemiology, political science and market research (Maibach, Leiserowitz, Roser-Renouf & Mertz, 2011). For example, segmenting the population into groups of people with sufficiently homogenous health prospects and priorities can subsequently facilitate and improve commonly needed support services (Lynn, Straube, Bell, Jencks &

Kambic, 2007). Researchers in social science have segmented samples with attitudinal data. For example, Genge (2014) segmented the Polish population by public attitudes towards the countries adoption of the euro. Sandell et al.

(2006) clustered therapists’ values and attitudes towards therapeutic matters to determine if this accounted for the large difference between treatment results.

Of the few studies that have segmented the population according to environmental attitudes, the most relevant will be discussed before this chapter analyses the EAS data to create a typology of UK environmental concern.

169

5.2.1 DEFRA segmentation

DEFRA funded an environmental attitude/behaviour segmentation study in

2006 (Barr, Gilg & Shaw, 2006), and conducted a similar study in 2008 (DEFRA

2008). Barr et al. (2006) examined how pro-environmental behaviours were practiced on a daily basis and how such practices varied according to lifestyle.

Data were collected from 1265 participants residing in the Devonshire, capturing environmental knowledge, attitudes and behaviour. Individuals were grouped into distinctive segments according to their level of behavioural commitment using factor and cluster analysis. The study identified four distinct groups of people defined by their behaviour patterns: (a) committed environmentalists, (b) mainstream environmentalists, (c) occasional environmentalists, and (d) non-environmentalists. Additional qualitative data were then collected from eight focus group discussions based on the quantitative findings. Thus a mixture of quantitative and qualitative analysis is used to segment the population and gain meaningful incite into the attitudes of group members.

Barr et al. (2006) differs from the analysis in this thesis in that the emphasis was on levels of pro-environmental behaviour, as well as attitudes towards such behaviour. This thesis by contrast is focused on assessing attitudes toward the environment/environmental problems, and how such attitudes are associated with pro-environmental acts; deliberately neglecting to assess attitudes towards such behaviour. Unfortunately, Barr et al. (2006) has a limited sample, gathered from one small part of the UK. As such, any

170

conclusions derived from the analysis are pertinent to that area only and cannot confidently be generalised to the rest of the country.

Subsequently, DEFRA (2008) produced a UK-wide environmental typology based on self-reported pro-environmental behaviours using data taken from the 2007 Environmental Attitudes and Behaviours Survey. This segmentation was intended to be used to understand and promote ‘green’ or pro- environmental behaviours and was widely reported in the media.

Unfortunately, little information is given on the methods used in the study. The segmentation model is described as “the outcome of an extensive three stage research process: (1) desk research, (2) qualitative research, and (3) quantitative research” (p. 43). This study produced a model consisting of seven clusters: (1) positive green; (2) waste watches; (3) concerned consumers; (4) side-line supporters; (5) cautious participants; (6) stalled starters; and (7) honestly disengaged (shown in Figure 14).

171

Figure 14: DEFRA (2008) UK Environmental typology

172

DEFRA claim that the model displayed in Figure 14 contributes to an understanding of how environmental attitudes, values, current behaviours and motivations and barriers are packed together for defined segments of the population. It is likely that the methods used are similar to those in the 2006 study (i.e. a combination of factor and cluster analysis as well as qualitative research to contextualise the results), but the paucity of methodological detail is concerning.

DEFRA (2008) combine measures of both broad-level environmental concern and behavioural in the same model, conflating the concepts empirically. Given their conceptual differences and empirical distance this is undesirable.

Merging these concepts within a singular model without justification introduces arbitrary variance and complicates the interpretation of results.

Furthermore, it can also be argued that treating behaviour as a latent variable is problematic, as strictly speaking behaviour is not a latent concept as it consists of observable actions and a latent concept by definition is unobservable, but attitudes towards behaviour and behavioural intent are.

Overall, it is not clear which aspects of behaviour or attitude these studies are capturing, or indeed, how they have been captured.

173

5.2.2 Environmental attitude segmentation

Maibach et al. (2011) produced an environmental attitude typology with the goal of improving the efficacy of public engagement campaigns. The study assessed belief in climate change and support for environmental policies among a nationally representative survey of American adults (N = 2164). The sample was segmented according to homogenous item response patterns using LCA, producing six segments shown in Table 14.

Table 14: Maibach et al. (2011) environmental attitude segments

Class Description

These are individuals who are the most engaged in the issue of Alarmed climate change. They are convinced it is happening, human- caused and a serious and urgent threat.

The concerned group is also convinced that global warming is a serious problem, but while they support a vigorous national Concerned response, they are distinctly less involved in the issue, and less likely than the ‘alarmed’ to be taking personal action.

These people believe that global warming is a problem, although they are less certain that it is happening than the Cautious ‘alarmed’ or the ‘concerned’. They don’t view it as a personal threat, and don’t feel a sense of urgency to deal with it through personal or societal actions.

The disengaged haven’t thought much about the issue. They are most likely to say that they could easily change their minds Disengaged about global warming and are the most likely to select the “don’t know” option in response to survey questions on global warming.

174

This group is evenly split among those who think global warming is happening, those who think it isn’t and those who don’t know. Many within this group believe that if global Doubtful warming is happening, it is due to natural changes in the environment, that it won’t harm people for many decades into the future (if at all) and that America is already doing enough to respond to the threat.

Like the ‘alarmed’ these people are actively engaged in the issue, but at the opposite end of the spectrum. The large majority of the people in this segment believe that global Dismissive warming is not happening, is not a threat to either people or non-human nature, and is not a problem that warrants a personal or societal response.

Unfortunately, class members from this study were not profiled in anyway, so nothing is known of any potential associations between class membership and socio-demographic characteristics such as age, gender or social class.

These six classes can broadly be divided into three groups of pro-, neutral- and anti-environmental perceptions. It is notable, then, that six classes were produced and not three. This suggests a higher level of complexity that perceptions towards the environment will differ not only between positive and negative views, but are likely to further differentiate within these segments.

Though arguable, these segments still appear to be an ordinal variable.

The primary distinction between the two negative environmental classes – the

Doubtful and the Dismissive classes – is the belief in the human contribution to climate change. The Dismissive class denies the existence of climate entirely,

175

where as most the Doubtful class believes that climate change is caused by natural changes.

Sibley and Kurz (2013) examined this distinction among climate change sceptics. Based on data from New Zealand, the authors used LCA to cluster people according to their views on climate change. They hypothesised that a distinction exists between those who are sceptical of climate change and those who are sceptical of human involvement, and that these attitudes had differing associations with pro-environmental behaviour. It is necessary, therefore, to understand this distinction in order to facilitate effective climate change policy. Sibley and Kurz (2013) produced a four-class model consisting of (a) Climate Believers, (b) Undecided/Neutral, (c) Climate Skeptics, and (d)

Anthropogenic Climate Skeptics, supporting their hypothesis of two distinct forms of climate change scepticism. It was also found that belief in the reality of climate change was significantly more predictive of pro-environmental behaviour and policy support than belief in the human involvement in climate change. From this, the authors concluded that it more important to convince people of the existence of climate change rather than its causes or the level of human involvement.

5.2.3 Theoretical frameworks used to study EC

The studies discussed thus far have used exploratory clustering techniques; classes are constructed and respondents are assigned to classes inductively, rather than on the basis of pre-conceived theoretical assumptions. Other

176

studies have approached the problem theoretically. Pendergraft (1998), for example, developed a typology of EC that hypothesised that climate change attitudes were derived from cultural belief systems. In developing this typology

Pendergraft drew on pre-existing taxonomies, most of which were dichotomous, such as Inglehart’s (1990) comparison of materialism and post- materialism or the collectivism/individualism paradigm (Triandis, Brislin & Hui,

1988). However, Pendergraft highlights that dichotomous classification sometimes overly simplistic, obscuring important differences within the two groups. Dividing a population into two or three clusters will identify intra-class differences, but at the cost of ignoring potentially important inter-class variation. The political system in the United States, for example, can be typified as a dichotomy between liberal and conservative. However, within the liberal group exist Marxists, socialists and postmodernists, whilst conservatives are comprised of republicans, tea party etc. These groups may hold similar values but they are certainly not the same.

At the opposite extreme, a typology with many classes can be complex and highly detailed, thus making it highly likely to only be relevant to a specific time and place. Consequently, such a model will have few applications elsewhere, diminishing its value. Pendergraft found a compromise with Cultural Theory

(Thompson, Ellis & Wildavsky, 1990) that suggests a typology of four classes.

All societies and their underlying worldviews, irrespective of time or place, must be hierarchic, individualistic, egalitarian or fatalistic as shown in Figure

15.

177

Figure 15: Cultural Theory (adapted from Schwarz & Thompson, 1990)

External locus of control

The Fatalist The Hierarchist

Indivisualised Collectivised The Individual The Egalitarian

Internal locus of control

This model is credited as being “parsimonious as it is comprehensive”;

(Thompson et al., 1990, p. 83). The matrix in Figure 15 plots acceptance of social controls (on the vertical axis) against levels of social commitment (on the horizontal axis). Pendergraft (1998) claims that the potential for cultural theory to “cross temporal and spatial comparisons makes it a particularly attractive instruments for study of human dimensions of global climate change” (1998, p. 8). Table 15 shows a cultural theory-oriented interpretation of environmental concern.

178

Table 15!!: A cultural theory-based interpretation of climate worldviews

Class Description

International protocols and national commitments are needed to Hierarchist address the tragedy of the atmospheric commons and reduce greenhouse gas emissions.

The underlying problem is consumption (resource throughput). Egalitarian Precaution, lifestyle simplicity and grass roots action are the most effective responses.

To address climate change, rely on laissez-faire markets to spur Individualist competition and innovation. The benefits of climate change may even balance out the costs.

Natural forces are beyond human understanding, much less Fatalist human influence

Responses were collected from 441 participants between August 1994 and

January 1997. Respondents assessed statements relating to climate change, and based on their responses were categorised as belonging to one of the worldviews shown in Table 15. This research was based around the assumption that (a) these worldviews are mutually exclusive and (b) an individual can subscribe to only one. However, these assumptions were not upheld by the empirical findings. Some respondents held the worldviews outlined by cultural theory, but in many cases these worldviews overlapped, with respondents reflecting multiple positions.

179

5.2.4 Summary

DEFRA (2008) conducted a UK environmental segmentation study using similar data to that used in this thesis. However, DEFRA withheld details of their methods and produced uninformative segmentation models.

Furthermore, variables relating to engagement with the natural environment appear to have been included in the model with no real consideration for what they represent – whether attitude, knowledge, behaviour or intention. It is unclear, therefore, what has been captured by this Study.

Using theory to dictate class membership (rather than allowing it to be determined empirically) is problematic. Strict application of a theoretical framework forces respondents to fit inside a pre-conceived notion of how they should be categorised rather than allowing the results to reflect the data. This echoes the problems highlighted in the previous chapter regarding the use of strict scale implementation. A more moderate methodological position is that theoretical frameworks should inform the interpretation, but not to the point where they impede alternative substantive conclusions.

Of these studies only the 2008 DEFRA study examined social characteristics of class membership, and none have investigated the association between social characteristics and class membership, despite the wealth of evidence suggesting that environmental attitudes are heavily influenced by social, economic and demographic factors (Black, Collins & Snell, 2001; Exley &

Christie, 2002; Steg, Geurs & Ras, 2001). People are not defined by broad

180

cognitive appraisals of the environment, but, as these studies show, many smaller and more specific classes of pro, neutral and negative environmental attitude classes.

181

5.3 Analysis

This section uses latent class analysis to cluster participants based on the nine measures of EC, chosen in the previous chapter. There are three parts.

First, the optimal number of classes is determined based on established goodness-of-fit criteria. Second, the chosen latent class model will be interpreted by drawing on the findings of the previous chapter. Third, how class membership is associated with individual social and economic characteristics will be examined.

LCA is a statistical method for identifying latent classes based on a set of observed response items. Latent classes are unobservable subgroups or segments containing individuals who are homogeneous on certain criteria.

Formally, latent classes are represented by categories of a nominal latent variable. LCA estimates conditional class membership probability (while also assigning individuals to their most likely class based on this conditional probability) and item response probability. The conditional membership probability represents the probability that an individual belongs to a latent class conditional on their answers to the EC indicators. Item response probability is the probability that class members will give response x to EC indicator y. These item response probabilities are the foundation of the latent class model.

The analysis in this chapter does not use an a priori hypothesis to dictate the number or nature of latent classes for EC, and so LCA is performed in an

exploratory manner with these data. Several proposed models will be

182

tested, differentiated by the number of latent classes. The resulting fit indexes will be compared to determine which best corresponds to the observed data

(as outlined by Finch & Bronk, 2011).

This decision to use LCA over cluster analysis was made based on the superiority of LCA as a clustering method. LCA is model based, and as such, conclusions from the chosen model can be generalised to the population from which the sample was drawn (in this case, the UK population). LCA also imposes the assumption that data are generated by a mixture of underlying probability distributions (Vermunt & Magidson, 2004). Therefore, based on the statistical concept of likelihood, participants are not only assigned to classes, but also have a probability class membership for all classes. Another advantage of LCA is that it does not require decisions to be made about the scaling or transformations of the observed variables. For example, when working with normal distributions with unknown variances, results will be the same irrespective of whether the variables are normalised. LCA is unaltered by linear transformations on variables, so standardisation is not necessary

(Francis 2006). This is very different to standard non-hierarchical cluster methods like K-means, where scaling is often an issue. LCA also provides diagnostics, such as the Bayesian information criterion (BIC), to determine the number of classes. Determining the optimal number of clusters for a cluster analysis model is less sophisticated and often relies on the (subjective) interpretation of a dendogram or somewhat arbitrary statistics such as the

VRC criterion.

183

5.4 Part one – Group classification

To determine the optimal number of classes, goodness of fit statistics are examined for two- to six-class models (Table 16). The results suggest both a two-class and a four-class model fit the data, as indicated by the adjusted Lo-

Mendell-Ruben (ALMR) p-value8 and high entropy. Furthermore, inflections can be observed in both information criterion and log likelihood for the two and four-class models as shown in Figure 16.

Table 16: Goodness of fit indices for LCA models containing two-six classes

Model ALMR p-value Entropy

Two-class 0.00 0.786

Three-class 0.09 0.733

Four-class 0.01 0.754

Five-class 0.49 0.771

Six-Class 0.76 0.745

8 Indicating that the mixture model with k classes fits the data better than the simpler k-1 class model.

184

Figure 16: Information criterion and log likelihood for one – six-class models

72000 Information Criterion

70000

68000 AIC ABIC 66000

64000

62000 1 2 3 4 5 6

-31000

-32000

-33000

-34000

-35000

Log Likelihood -36000

185

It is hypothesised that the significance of the two- and four-class models is indicative of there being two broad groups of participants (two-class model), within each of which are two subgroups (four-class model). The remainder of part one will determine which model has the greatest substantive value. If a four-class model does not offer further insight, it will be more constructive to only retain the more parsimonious two-class model.

5.4.1 Two-class model

There are two forms of model parameters for latent class models with categorical indicators: conditional item probabilities and class probabilities.

The conditional item probabilities are specific to class x and provide information about the probability that an individual belonging to class x will endorse a specific item. The class probabilities specify the relative size of class x, or the proportion of the population that is in this class. Table 17 shows the size of the classes for this model.

Table 17: Class sizes for 2-class model

N %

Class 1 1151 39.32%

Class 2 1777 60.68%

186

Figure 17: Item probabilities for 2-class model Special Loss of Animal Class 1 Class 2 Changes to Countryside Major Disaster Limited Resources Over Populated Beyond Control Exaggerated Crisis Too Far in the Future Low Priority 0 0 0 0 0 0 0 0 0 0 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

187

Figure 17 shows the probability that class members agree9 with the nine items used throughout this analysis as indicators of EC. Class 1 members have a high probability of agreeing with pro-environmental statements (particularly those relating to the countryside and animal species), while having a low probability of agreeing with the three negative statements. Class 2 members demonstrate a comparatively lower level of EC, though the same pattern can be observed (high probability of agreeing with positive statements and a low probability of agreeing with negative statements).

The biggest difference in probability between the two classes is regarding the major disaster item (with a difference of 0.34). Though both classes demonstrate a concern for the environment the EC expressed by Class 1 appears to be stronger, possibly because they have a higher probability of agreeing that a major disaster is imminent, and as such consider environmental disaster a real and approaching threat. Interestingly, both classes have the same probability of agreeing that climate change is beyond control (0.14 and 0.15 respectively).

This two-class model appears to be overly simplistic, only demonstrating that some individuals express a high level of EC, while others slightly less so. It is likely that this model has grouped smaller clusters, obscuring potentially valuable subgroups. This is common in binary class models as highlighted by

Pendergraft (1998).

9 If either the ‘Strongly Agree’ or ‘Agree’ categories were selected.

188

5.4.2 Four-class model

Item probabilities from the four-class model are displayed in Figure 18, revealing distinct and interesting subgroups. Class probabilities for this model are shown in Table 18.

Table 18: Class probabilities for four-class model

N Class %

Class 1: Pro-environmental 837 28.59%

Class 2: Neutral Majority 1036 35.38%

Class 3: Paradoxical 513 17.53%

Class 4: Disengaged 542 18.50%

189

Figure 18: Item probabilities for 4-class model

l a m s i e n i A c

f e o p

S s s o L

e o t

d i s s e y r g t n n a h ou C C r e t s sa i D

r o j a M s e c r u o s Re

d e t i m i L d e t a l opu P

r ve O l o r t n o C

ond y Be s i s i r C

d e t a er g a x E e r u t u t n e in F r

m a l d n F a e o c r i i oo x v T o n e - ad engag r o s i r a

Neutral Majority P P D

: : : 1

2 3 4 y

t i s

s s s r s

s s s o i a a

a a r P

Cl Cl

Cl Cl Low 0 0 0 0 0 0 0 0 0 0 0 9 8 7 6 5 4 3 2 1 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. Class 1 members have the overall highest probability of agreeing with positive

190

statements and lowest probability of agreeing with negative statements. This class has been accordingly labelled as “Pro-environment”. Class 2 members show a similar pattern, but with less extreme item probabilities for positive and negative statements. This class’ item probabilities in conjunction with it having the highest proportion of participants, has been labelled as “Neutral Majority”.

Class 3 members exhibit a paradoxical combination of item probabilities, with similar scores for both positive and negative statements. Item probabilities do not decrease for negative statements, nor increase for positive statements as is shown in classes 1 and 2. Instead, probabilities remain between 0.4 and 0.7 for all nine items. Members of this class therefore are reasonably likely to agree that an environmental crisis has been exaggerated, that it is a low priority and too far in the future to be of concern. However, class members are also reasonably likely to be concerned about the countryside and animal species, the planets ability to sustain an ever-growing human population, and acknowledge that if things continue there will be a major environmental disaster. This combination of views is not only complex but also contradictory, and is reminiscent of the “Denial” EC component uncovered in the previous chapter. This attitude cluster could reflect a form of denial, where environmental problems are recognised but then dismissed and trivialised as a coping mechanism. Given these contradictory item probabilities, this class has been labelled as “Paradoxical”.

Class 4 members have the lowest probability for positive items. Probability for negative items is also low, although not as low as Neutral and Pro-

191

environment classes. While the low probability for positive items is indicative of scepticism or denial, item probability for negative items is too low to support this interpretation. Given the low item probability for all items, it is likely that class members are disengaged or apathetic towards environmental issues. This class has been labelled “Disengaged”.

What is noteworthy are the near identical ‘beyond control’ item probabilities for three of the four classes (Pro-environmental: 0.11, Neutral Majority: 0.08,

Disengaged: 0.09). That three, seemingly different classes have similar probabilities of agreeing with this item, suggests that participant interpretation of this item is highly variable.

192

5.5 Part two – Class profiles

5.5.1 Within-class factor structures

The model of EC developed in the previous chapter will be used here to aid interpretation of the attitude clusters captured by LCA. Mean factor scores for each factor of the EC model within each class are shown in Figure 1.

Figure 1: Within-class mean factor score

Eco-centric Human-centric Denial

-.5

Neutral Pro-environment Paradoxical Disengaged Majority

Unsurprisingly, the Pro-environment class score extremely low on the denial factor (suggesting that they have fully accepted that there are current environmental problems) and also have the highest mean score for eco- and human-centric factors.

All mean factor scores for the Neutral Majority class are low. Members are marginally concerned about the ecosystem and accept that there are

193

environmental problems (given the negative denial score), but overall, there does not appear to be a strong response (positive or negative) towards environmental issues.

The low probability of expressing eco- and human-centric concern, combined with the extremely high denial score for the Paradoxical class provides further evidence for the proposition that these class members are in denial regarding the severity of current environmental problems. The Disengaged class has the lowest average scores for eco- and human-centric concern, supporting the interpretation of these class members as disengaged or apathetic.

5.5.2 Gender

Although the relationship between EC and gender has received much considerable attention the findings have often been inconsistent. For example,

Blocker and Eckberg (1997) and Arcury and Christianson (1990) report that men are more active, knowledgeable, and concerned about the environment than women. Conversely, Zelezny et al. (2000), Olofsson and Öhman (2006) and Uyeki and Holland (2000) found that women tend to be more concerned about the environment than men. The latter study emphasised that women are significantly more concerned about the environment, nature, and animals. To complicate the matter further, other studies have reported that gender does not influence EC and women “are not more concerned about the environment than men” (Hayes 2001, p. 657). Arcury and Johnson (1987) also down play the influence of gender on EC noting that the gender effect is weak; as such

194

no definite conclusion could be drawn. Bord and O’Conner (1997) suggest that if gender differences are found, they are likely to be related to the divergences in perceptions of the harmful consequences of environmental problems rather than to gender per se.

Figure 2: Proportion of class members by gender

.30

.20

oportio r

P .10

.00 Male Female

Pro-environment Neutral Majority Disengaged Paradoxical

Figure 2 shows little difference between men and women regarding the proportion of class members. Amongst women, there is a slightly higher proportion of Pro-environment, Neutral and Paradoxical class members, while male respondents have a higher proportion of Disengaged class members.

Overall there is a significant but extremely weak relationship between gender and class membership (Cramer’s V = .07), supporting the supposition that

195

gender has little-to-no effect on levels of EC.

5.5.3 Age

Many past studies have found EC to be higher among young people (Arcury &

Christianson, 1993; Schultz 2001). This difference is frequently attributed to younger people being less invested in social roles that emphasise industriousness and self-interest, and instead have been found more willing to consider new and different ways to approach problems (Van Liere & Dunlap,

1980). However, some recent studies have failed to replicate these findings, which may be due to conflating of age and cohort effects (Jackson, Ones &

Dilchert, 2012). This implies that age differentials in EC identified in recent studies could be the result of a generation effect whereby the cohort of young individuals growing up in the 1970s became more concerned about the environment (as a result of the growing pro-environmental movement of that time). As these concerned individuals have grown older, they have passed on their belief to their offspring, reducing any age difference in environmental attitudes.

196

Figure 3: Within-class age distributions and class member proportions by age group

Pro-environment Neutral Majority .30

.20

.10 n .00 Paradoxical Disengaged

oportio .30

r P

.20

.10

.00 20 40 60 80 100 20 40 60 80 100 Age

1.00

.80

n .40 oportio

r .20 P

.60

.00 16 - 24 25 - 34 35 - 44 45 - 54 55 - 59 60+ Age Group

Pro-environment Neutral Majority Disengaged Paradoxical

197

As shown in Figure 3, the age distributions for members of the Pro- environment and Neutral Majority classes are relatively normal (although the

Neutral Majority class does demonstrate a slight negative skew). Paradoxical age distribution is particularly flat and has the highest proportion of participants aged 70+. The Disengaged class appears to be negatively distributed and has the highest within-class proportion of young participants.

This goes against previous research that suggests that younger people tend to have greater pro-environmental attitude. Indeed, the majority of the Pro- environment class appears to be middle-aged.

5.5.4 Socio-economic status

A large body of literature has examined the question of whether the socio- economic status (SES) of individuals increases levels of EC. Marquart-Pyatt

(2008), using data for 19 European countries, consistently found education and income to be positively associated with pro-environmental attitudes and behaviour. Franzen and Meyer (2010) and Givens and Jorgenson (2011) also document positive individual-level associations between EC and education, income and social class. Other studies have found the relationships between these variables and EC to be more variable (Pampel 2014). As such, measures of highest qualification, social grade and household income from the EAS dataset are used in this research as indicators of SES.

198

Figure 4: Proportion of class members by highest qualification

.40

.30 n

.20

oportio

r P .10

.00 No formal A levels /O levels Degree level qualiﬁcations

Pro-environment Neutral Majority Disengaged Paradoxical

As shown in Figure 4, amongst those who are educated to degree level, a high proportion are members of the Pro-Environment and Neutral class, with very few belonging to the Disengaged or Paradoxical group. Comparatively, those with no formal qualifications and those educated to A / O level have low proportions of Pro-environment and Neutral class members. This suggests that perhaps an increased level of formal education also increases the probability of possessing pro-environmental attitudes, while a lack of education has the opposite effect.

A cross-tabulation of household income and class membership is given in

Table 19. The Paradoxical class has the largest proportion of low-income households, while Pro-environment and Neutral classes have the highest.

199

Table 19: Within-class household income

£10k - £19k £20k - £39k £40k - £50k

Pro -environment 253 (42.57%) 174 (31.52%) 140 (25.91%)

Neutral Majority 301 (38.43%) 250 (36.38%) 172 (25.19%)

Paradoxical 200 (57.96%) 86 (26.57%) 51 (15.47%)

Disengaged 169 (49.70%) 109 (33.00%) 56 (17.30%)

Total 44.90% 32.74% 22.36%

Chi2 p =.00 Cramer’s V = .11

The EAS uses the NRS (National Readership Survey) social grades system of socio-economic classification. This is a classification system based on occupation and it enables a household and all its members to be classified according to the occupation of the Chief Income Earner (CIE). In addition, if the respondent is not the CIE and is working, then the social grade of that individual is also recorded. A number of questions need to be asked in the interview in order to assign social grade accurately. The interviewer probes the respondent for information about the occupation of the CIE, the type of organisation he or she works for, duties and responsibilities, job title/rank/grade, and whether the CIE is self-employed. Also relevant are details of the number of people working at the place of employment and whether the CIE is responsible for anyone, together with confirmation of qualifications. Income is not part of the social grade classification. Categories of social grade are displayed in Table 20. Poortinga et al. (2011) also used this system and found social grade to be a significant predictor of public

scepticism about anthropogenic climate change.

200

Table 20: NRS social grade classification system

Grade Social class Chief income earner's occupation

A Upper-Middle Higher managerial, administrative or professional

Intermediate managerial, administrative or B Middle professional

Supervisory or clerical and junior managerial, C1 Lower-Middle administrative or professional

C2 Skilled Working Skilled manual workers

D Working Semi and unskilled manual workers

State pensioners, casual and lowest grade E Non-Working workers, unemployed with state benefits only

The Pro-environment class has the highest proportion of middle- and upper middle-class members, while the Neutral class is mostly comprised of middle- and lower middle-class members. The Disengaged and Paradoxical classes share similar proportions of individuals from each social grade, both largely comprised of lower middle-class and working-class workers. The Paradoxical class also has the highest proportion of non-workers, though these are most likely retired individuals given the comparatively higher proportion of older individuals within this class.

Overall, these findings highlight a socio-economic divide. Pro-environment class members generally occupy a higher social grade and report higher levels of household income and education. The Paradoxical and Disengaged classes seem to exist at the opposite end of the social spectrum. Though it is natural that comparisons should be drawn between these two classes, it is also

201

important to recognise the social differences that distinguish these two groups.

5.5.5 Climate change belief

Sibley and Kurz (2013) and Maibach et al. (2011) suggest that negative environmental attitudes were comprised of general climate change sceptics and anthropocentric sceptics. To determine whether there is a difference between the two negative environmental classes – the Paradoxical and

Disengaged – with regard to anthropocentric climate change scepticism, the proportion of participants who agreed that climate change was caused by energy usage was examined. As Table 21 shows, among respondents who thought climate change was not caused by human energy consumption, the highest proportion was found in the paradoxical class (62.5%).

202

Table 21: Climate change belief and most likely class membership

Class

Climate Change Belief Pro- Environment Neutral Paradoxical Disengaged

I don't believe that climate change is 10 (8.5%) 12 (8.9%) 79 (62.5%) 22 (20.1%) caused by human energy consumption.

I am unsure about the reality of climate 21 (15.0%) 30 (19.0%) 56 (36.2%) 45 (29.7%) change.

I believe that climate change is caused by 790 (31.5%) 992 (39.0%) 364 (13.9%) 354 (15.5%) human energy consumption.

Chi2 p =.00 Cramer’s V = .23

203

5.6 Part three – Regression analysis

Regression analysis is used here to model the association between SES and class membership. The highest qualification, household income and social grade are again used as measures of SES. Age and gender are also accounted for within the model.

5.6.1 Missing data

Both household income and highest qualification variables have a high proportion of item non-response (31.7% and 30%, respectively). Listwise deletion is a potential solution to this problem, as it discards all observations with missing values. Unfortunately this means that participants with missing responses will be removed from the analysis. The inclusion of the household income and highest qualification variables within the model variables would therefore restrict sample size (to 1423), resulting in a model with larger standard errors, wider conﬁdence intervals and less power.

Further problems become apparent when the potential nature of the missing data is considered. Missing data are said to be missing completely at random

(MCAR) if the probability that data are missing does not depend on observed or unobserved data. Under MCAR, the missing data values are a simple random sample of all data values, and so any analysis that discards the missing values remain consistent (though inefﬁcient). Missing data are said to be missing at random (MAR) if the probability that data are missing does not

depend on unobserved data but may depend on observed data. Under

204

MAR, the missing data values do not contain any additional information given observed data about the missing data mechanism. When missing data are

MAR, listwise deletion may lead to biased results. To ensure that this regression analysis is both robust and representative of the UK population, multiple imputation (MI) is performed on both the household income and highest qualification variables. Multiple imputation provides a useful strategy for dealing with datasets with missing values. Instead of ﬁlling in a single value for each missing value, Rubin’s (1987) multiple imputation procedure replaces each missing value with a set of plausible values that represent the uncertainty about the right value to impute. Multiple imputation is a Monte Carlo technique where missing values are replaced by m > 1 simulated versions, where m is typically small (e.g. 3-10). Multiple imputation does not attempt to estimate each missing value through simulated values, but rather to represent a random sample of the missing values.

Multiple imputation inference involves three distinct phases (Yuan 2010):

• The missing data are filled in m times to generate m complete data sets.

• The m complete data sets are analysed by using standard procedures.

• The results from the m complete data sets are combined to produce

estimates and confidence intervals that incorporate missing data

uncertainty.

This process results in valid statistical inferences that properly reflect the uncertainty due to missing values.

205

5.6.2 Regression

Having imputed missing values for household income and highest qualification the association between SES and class membership is analysed using multinomial regression. Table 22 shows the relative risk ratios of SES indicators on most likely class membership. The Disengaged class is used as the reference category for this analysis, as these particular class members are apathetic and detached from environmental concern. Ratios have low standard errors, suggesting that MI was successful.

The ratios are consistent with the exploratory analysis in Section 4.2, while accounting for age, gender and SES simultaneously. As can be seen in Table

22, social grades C1, B and A are significant predictor of Pro-environment class membership over Disengaged class membership. There are significantly more women in all other classes compared to the Disengaged class. The

Paradoxical class also has significantly more 60+ members than the

Disengaged.

206

Table 22: Multinomial regression of SES on most likely class membership

Pro-Environment Neutral Majority Paradoxical Variable Category RRR SE t RRR SE t RRR SE t

Gender Female 1.52* 0.19 3.33 1.59* 0.19 3.81 1.56* 0.22 3.17

25-34 1.80* 0.48 2.21 1.36 0.32 1.31 1.25 0.32 0.87

35-44 2.28* 0.59 3.20 1.69* 0.39 2.29 1.04 0.26 0.15

Age 45-54 2.24* 0.57 3.17 1.36 0.31 1.34 0.87 0.22 -0.53

55-59 3.68* 1.16 4.14 1.96* 0.58 2.30 0.80 0.28 -0.63

60+ 1.53 0.36 1.80 0.94 0.20 -0.27 0.43* 0.10 -3.60

D 1.11 0.25 0.45 1.28 0.27 1.17 0.94 0.22 -0.28

C2 1.09 0.23 0.40 1.27 0.25 1.19 0.74 0.17 -1.33 Social C1 1.72* 0.36 2.56 1.82* 0.36 3.01 0.96 0.22 -0.17 Grade B 2.54* 0.66 3.59 2.10* 0.52 3.01 1.12 0.33 0.38

A 2.67* 1.12 2.33 1.24 0.53 0.50 1.42 0.72 0.70

A levels / O 1.05 0.23 0.20 0.99 0.21 -0.06 1.05 0.23 0.21 Education levels

Degree level 1.49 0.35 1.73 1.26 0.29 1.00 0.58* 0.17 -1.91

Household £20k - £39k 1.10 0.21 0.48 1.45* 0.25 2.12 1.29 0.27 1.26

Income £40k - £50+ 1.22 0.35 0.69 1.55 0.40 1.72 1.20 0.36 0.62

*p<.05

207

5.7 Discussion

Pro-environment

This class has the highest proportion of upper-class individuals10 who report higher levels of education have the highest household incomes. Respondents in this class are predominantly middle aged. Pro-environmentals are named as such because they have the highest probabilities for agreeing with pro- environmental statements, in combination with the highest average score for human-centric and biospheric attitudes and the lowest average score for denial.

Neutral Majority

This class is primarily comprised of participants in their late 30s and early 40s and has the highest proportion of middle-class workers. While having a similar pattern of item probabilities as the Pro-environment class, factor score for

Neutral Majority class members are not as high. This suggests that although

Neutral class members possess a positive cognitive evaluation of the environment, it is weak in strength.

Disengaged

The Disengaged are primarily young respondents (most are under 30) from the middle and lower social classes. This class has the lowest levels of education, with most not educated past GCSE level. However, this is likely to be at least

10 Upper class is based on membership of NRS social grades a and b.

208

in part due to the age distribution for this class. The disengaged have the lowest probability of agreeing with pro-environmental statements and the second highest probability of agreeing with the negative-environmental statements. These respondents also have the lowest average score for human centric and biospheric attitudes, as well as a high score for denial. This is interpreted as apathy towards environmental issues.

Paradoxical

This class of respondents produced a contradictory collection of item probabilities, earning them the title of Paradoxical. These respondents have, on average, a very high average score for denial, suggesting that their odd responses to statements may be as a result of an inability to process or accept information regarding climate change. A large proportion of this class do not accept that climate change is due to energy consumption, making them similar to the doubtful group discovered by Maibach et al. (2011), and distinguishing them from the Disengaged class. This class has the highest proportion of older participants, as well as low-income and low social-class workers. This class, unlike the others, also has a higher proportion of men.

209

5.8 Conclusion

This analysis produced a four-class model, consisting of distinct clusters of environmental attitudes. Participant item response probabilities informed the initial interpretation and titles for the classes. The dimensional model of environmental concern developed in Chapter 4 was also used to aid interpretation of the environmental attitudes held by these classes.

The largest class (Neutral Majority, which is 35% of the sample) is apathetic towards the national environment. The Paradoxical and Disengaged classes hold negative / inconsistent attitudes towards the environment. The age division and belief in the anthropocentric cause of climate change are the primary distinguishing factors between these two classes.

Overall, important socio-economic distinctions exist between the classes. The multinomial model seems to suggest that pro-environmental attitudes are positively correlated with class, income and education. Age does not seem to be positively associated with pro-environmental attitudes, as some previous research has suggested. The two sceptical classes – the Disengaged and the

Paradoxical – have high proportions of young and old respondents, respectively. Both classes are unable to accept or simply do not believe in climate change; whether these beliefs are due factors relating to their age warrants further inquiry.

The analysis performed in this chapter expands upon the existing DEFRA

segmentation studies by using robust methods, focusing specifically on

210

pro-environmental attitudes and examining the association between social characteristics and class membership. Results supports the dual scepticism finding from other segmentation studies performed outside the UK, and shows strong links between class membership and key socio-economic and demographic factors.

211

Chapter 6 The relationship between environmental attitudes and behaviour

212

6.1 Introduction

The purpose of this chapter is to study the relationship between environmental concern and behaviour. Chapters 4 and 5 examined UK concern for the natural environment. Two models have been produced from these analyses: one outlining the attitudinal structure of environmental concern, and one showing the four forms of environmental concern that exist in the UK population. Thus, this thesis has treated environmental concern as both continuous (Chapter 4) and categorical (Chapter 5). A decision must now be made regarding which of these models of environmental concern is the most suitable for this analysis in this chapter.

Treating environmental concern (EC), as categorical, and using latent class analysis to divide EC variance into categories, is to examine concern as it exists amoungst a given population. It is a ‘people based’ way of categorising a latent concept, where the emphasis is place on participant response patterns, rather than the responses alone. Alternatively, to divide EC variance up into continuous components using factor analysis, is to examine concern at a conceptual level. The problem with this latter technique is that factors are correlated and free to vary across scales (and in various combinations). This is a problem because Denial is incompatible with the other two components of environmental concern found in this thesis. Latent class analysis is not bound by this assumption, and allows for a negative denial score, combined with positive ecocentric and human-centric scores, as illustrated in Figure 19

213

Figure 19: Mean factor scores within the Pro-environment class

Eco-centric Human-centric Denial

-.5

Neutral Pro-environment Paradoxical Disengaged Majority

When examining environmental concern amoungst a given population, it may not be best captured by 3 correlated structures. Classes have more explanatory power regarding the resistance towards climate change, producing a partially nominal, partially ordinal model of environmental concern, as illustrated by Figure 20. This not only shows a hierarchy in level of environmental concern but also distinguishes between two forms of environmental scepticism (Paradoxical and Disengaged). This model best captures the variability in environmental concern amoungst the British public, and will be used to assess how UK environmental concern is associated with pro-environmental behaviour.

214

Figure 20: Environmental classes in order of increasing environmental concern

Pro-environment

Increasing Neutral Majority Environmental Concern

Paradoxical Disengaged

This chapter will examine the relationship between environmental classes and behaviour. More specifically, the aim of this assessment is to test two hypotheses:

2. Environmental class is significantly associated with pro-environmental

behaviour.

3. Socio-economic status (SES) indirectly affects the relationship between

class membership and behaviour.

The purpose of hypothesis 1 is to assess the predictive power of environmental class membership; i.e. how environmental concern affects the probability that an individual will engage in (or increase their level of engagement in) pro-environmental behaviour. The detailed picture of environmental concern developed in previous chapters provides the

215

foundation to test this hypothesis. Previous research which has examined the relationship between concern and behaviour has yielded results inconsistent with one another, but overall, the median finding is a weak relationship: environmental concern is usually associated with environmental behaviour but not strongly so. This inconsistency has led some studies to focus on behaviour-specific attitudes, and dismiss general attitudes towards the environment as weak predictors of behaviour. This practice began in the early

1980s, when Ajzen and Fishbein (1980) argued that attitudes and behaviours should be measured at a comparable level of specificity, otherwise correlations between them are likely to be modest. This has arguably lead to an eager pursuit of statistical significance and a setting aside of deeper study of the relationship between general environmental concern and environmental behaviour. This line of inquiry should not be overlooked. If anything, weak or inconsistent results highlight the need for further research to determine whether this weak relationship occurs for all people, across all measures of behaviour.

Several studies have explored the direct relationship between value-based forms of environmental concern (such as though unearthed in Chapter 4) and behaviour. For example, Thompson and Barton (1994) find that ecocentric and anthropocentric value orientations independently contribute to explanations of conservation behaviours, membership in environmental organisations and apathy toward the environment. Schultz and Zelezny (1998) explore whether the relationship between values and behaviour continues in countries and cultures other than the United States. Examining survey data from Mexico,

216

Nicaragua, Peru, Spain, and the US, Schultz and Zelezny (1998) find a positive relationship between biospheric values and behaviour, and a negative relationship between egoistic values and behaviour. However, as yet there has been little research assessing the relationship between categorical measures of environmental concern (i.e. groups/classes of individuals defined by their form of concern for the environment) and their relationship with private sphere environmentalism in the UK.

Hypothesis 2 is derived from research suggesting that environmental attitudes are influenced by socio-economic factors. Chapter 5 indicated that environmental class membership is associated with income, education and social grade. It is plausible, therefore, that variations in past attitude-behaviour studies are due to socio-economic differences amongst research participants.

Indeed, it could be that well-educated individuals have a greater knowledge and awareness of environmental issues, and that poorer individuals prioritise more pressing social issues, such as health and wealth, or that simply performing frequent acts of pro-environmental behaviour is easier for wealthy individuals.

Both hypotheses are tested through three stages of analysis. First, the direct effects of environmental concern on behaviour are assessed using logistic regression analysis. Following this, two forms of indirect effects are analysed: mediation and moderation. The second stage of analysis examines the extent to which SES moderates the attitude-behaviour relationship – that is, whether attitudes are more strongly associated with behaviours at high or low levels of

217

SES. The third stage examines if environmental concern mediates (accounts for) the relationship between SES and behaviour through the use of path analysis.

Measures of behaviour are taken from the EAS dataset; these behaviours are what Stern (2000) refers to as private sphere environmentalism, exhibited by the UK public.

218

6.2 Past studies of the environmental attitude behaviour relationship

Mounting scientific evidence suggests human-induced climate change may pose a significant threat to humans and the wider environment. In the 1970s the revelation that environmental degradation is the consequence of

‘maladaptive human behaviour’ (Maloney & Ward, 1973, p. 583) motivated social scientists to analyse individual motives underlying this behaviour. Such environmental studies concentrated primarily on environmental concern or attitudes as predictors of environmental behaviour. Often the measured attitudes have been broader in scope than the measured actions; assessing how an individual cares about the environment and how this effects their recycling frequency for example (see Rajecki 1982). Many cite this broad scope as the reason for subsequent inconsistences between findings. The following quotes are taken from two environmental attitude-behaviour studies, only four years apart, both published in peer reviewed journals.

“Many studies establish attitudes as predictors of behaviour and behavioural intentions”

(Clark, Kotchen, & Moore, 2003)

“The majority of attitudinal studies have shown that environmental concern or attitudinal variables fail to correspond to behaviour”

(Tanner 1999)

219

That one publication can make such a statement and, a few years later, an equivalent study makes such an opposing statement, is a testament to the inconsistent results produced from such attitude-behaviour studies. Of the studies that did find an association, it was usually low to moderate (Eckes &

Six, 1994; Fuhrer 1995; Hines et al., 1987; Six 1992; Spada 1990; Weigel

1983). Hines et al. (1987), for example, report an average correlation of 0.35 in their meta-analysis of 128 studies; Eckes and Six (1994) found only an average correlation of 0.26 in their meta-analysis (17 studies) and that environmental concern explains, at most, 10% variance of specific environmental behaviours.

Scepticism regarding the explanatory power of environmental concern ensued and some abandoned the claim that general environmental concern is a direct predictor of specific environmental behaviour all together. Instead behaviour- specific attitudes have been tested as predictors of behaviour, adhering to the correspondence principle developed by Ajzen and Fishbein (1980) which posits that only when the attitudinal and behavioural measures correspond to each other concerning the relevant action, context and time, is there a substantial relationship.

Focusing upon specific attitude-behaviour relationships, rather than general environmental concerns, has come at a cost. The reason the attitude concept received so much attention in psychology was in part due to its assumed function as predictor of multiple behaviours (e.g. Rokeach & Regan, 1980).

Bamberg (2003) points out that specific attitudes do not fulfill this function, and are only able to predict the behaviour they are specific to. Furthermore, attitudes towards recycling are arguably attitudes towards recycling behaviour

220

but not attitudes towards the environment.

6.2.1 Low-cost behaviour

Diekmann and Preisendöerfer (1992) explain the lack of a consistent relationship between environmental attitudes and pro-environmental behaviour by using a low-cost/high-cost model.

This model suggests that people choose to engage in pro-environmental behaviours that demand the least cost. ‘Cost’ is not only defined in an economic sense but also in a broader psychological sense that includes, among other factors, the time and effort needed to undertake a particular behaviour. Diekmann and Preisendöerfer (1992) suggest that environmental attitudes and low-cost pro-environmental behaviour (e.g. recycling) do correlate significantly and therefore, people who care about the environment tend to engage in activities such as recycling but do not necessary engage in activities that are more costly and inconvenient such as driving less.

They conclude that positive environmental attitudes can directly influence easy, low-cost pro-environmental behaviour such as recycling, but that people with high levels of environmental awareness might not be willing to make bigger lifestyle sacrifices. Whitmarsh (2012) supports this, finding recycling to be the most common mitigating response to environmental problems.

Whitmarsh also found resistance to changing travel habits. When provided with a list of alternative mitigation strategies, most British people claim they

221

would recycle household waste and improve home energy efficiency, but few would change their transport habits or pay more to travel. Researchers in the

United States have found a similar resistance to changing driving habits, while there is a greater willingness to adopt domestic energy conservation practices

(Bord et al., 2000; Fortner et al., 2000; O'Connor, Bord, Yarnal & Wiefek,

2002).

6.2.2 Is the environment a low priority

A recurring explanation for environmental attitude–behaviour discrepancy is that the environment is viewed as a low priority, compared to other social issues. Health, security, and finances are found to be more important than environmental issues for the public (Bord et al., 2000; Norton & Leaman, 2004;

Poortinga & Pidgeon, 2003). So, although attitudes are likely to influence behaviour, they lose their predictive power, as pro-environmental behaviours are de-prioritised in favor of actions that directly increase income, happiness and wellbeing. This low ranking of the environment reflects a widespread perception that the consequences of environmental problems are neither severe nor imminent. Individuals may consider environmental problems to be genuine, but most do not consider such problems to pose a prominent personal threat (Lorenzoni & Pidgeon, 2006). In the UK, 52% of people believe that climate change will have ‘little’ or ‘no effect’ on them personally (bbc

2004; see also Poortinga 2003, and Hillman 1998). The Energy Savings Trust

(2004) found that 85% of UK residents believe the effects of climate change will not be seen for decades. This lack of urgency and removal of personal

222

threat allows other social issues to be prioritised. If circumstances arrive that competing priorities are satisfied, then environmental issues may be prioritised more highly. Consequently, those in good health, with wealth and security may have a higher probability of engaging in pro-behaviour.

6.2.3 The effect of socio-economic status

Research suggests that SES affects the attitude-behaviour relationship. It has been shown, for example, that individuals who conserve energy may do so for financial or health reasons rather than for environmental ones (DEFRA 2002).

Whitmarsh (2009) finds that energy reduction is more often motivated by economic self-interest and other tangible benefits than by environmental concern. Given that SES is linked to environmental action and concern, further research into the dynamics of the relationship between these three components is required. Does SES act as a modifier for the relationship between attitudes and behaviours? Or does SES influence behaviour indirectly, via attitudes? Few studies have attempted to answer such questions, though Cottrell (2003) is one exception. This study hypothesised that general environmental attitudes directly relate to behaviour but that the strength of the relationship increases with income. This hypothesis was tested on a sample of 230 recreational boaters from the Netherlands. This group of individuals was chosen because they have direct interest and exposure to environmental issues. Using regression analysis, it was found that income, age, and stand on political issues predicted environmental concern. Two of the three environmental attitude variables used in the study were predictors of

223

behaviour. The assumption that controlling for socio-demographic characteristics would increase the combined influence of the attitudinal variables was weakly supported. Overall, concern was moderately correlated with behaviour. SES increased the variance captured in the model. No mediating effect was found.

6.2.4 Summary

It seems counterintuitive that positive attitudes towards the environment do not significantly affect the frequency of pro-environmental behaviour. But this is what past studies have often found. Some studies have found a significant relationship between strong pro-environmental attitudes and some behaviours

(particularly low cost ones), but overall, the findings from environmental attitude-behaviour studies have been highly inconsistent.

In response to this discrepancy between environmental attitudes and environmental behaviours, and upon the advice of Ajzen and Fishbein (1980), studies have tended to focus upon the link between specific attitudes and specific pro-environmental behaviours, thus ensuring a significant relationship.

Others have suggested that attitudes influence only low cost pro- environmental behaviours (e.g. Diekmann & Preisendöerfer, 1998), suggesting previous inconsistencies were due to the chosen measures of pro- environmental behaviours. Bamberg (2003) summarises the study of the environmental attitude-behaviour relationship perfectly: “either the concept of environmental concern is completely substituted by behaviour-specific

224

attitudes or environmental concern is now viewed more as an ideology, which influences only symbolic ‘low-cost’ environmentally related behaviours” (2003, p. 22).

Overall, there is sufficient evidence to suggest that examination of the attitude- behaviour relationship can still yield valuable results (without substituting concern for the natural environment with behaviour-specific attitudes, or focusing only on easy forms of pro-environmental engagement). What may yield more consistent (and possibly more accurate) results is to not just account for SES, but to assess it as an indirect effect. Indeed, it may be possible that focusing solely on the direct effects of attitudes on behaviour may be erroneous. This assertion is based on a large body of research that links SES to both environmental concern (Givens & Jorgenson, 2011; Pampel

2014; Xiao & McCright, 2007), and pro-environmental behaviour (Lorenzoni,

Nicholson-Cole & Whitmarsh, 2007). Research suggests that SES is important regarding the execution of pro-environmental behaviour, as environmental issues are often de-prioritised behind other social issues such as health and wealth. It is also possible that some high-cost pro-environmental behaviours are easier to engage in when a member of a higher social grade. Cottrell

(2003), using path analysis, found SES produced no indirect effect on the attitude behaviour relationship. However, Cottrell’s (2003) research was performed on only a small sample of recreational boaters in the Netherlands, and indirect effects may exist in the wider (UK) population.

225

6.3 Methods

Wegener and Fabrigar (2000) posit that indirect effects are categorised as either mediating or moderating effects.

Moderator variables “affect the direction and/or strength of the relation

between an independent or predictor variable and a dependent or

criterion variable. Within a correlational analysis framework, a moderator

is a third variable that affects the correlation between two other variables.

A moderating effect can be represented as an interaction between a focal

independent variable and a factor that specifies the appropriate

conditions for its operation”

(Baron & Kenny, 1986, p. 1174)

Mediator variables “account for the relation between the predictor and

the criterion. Mediators explain how external physical events take on

internal psychological significance. Whereas moderator variables specify

when certain effects will hold, mediators speak to how or why such

effects occur”

(Baron & Kenny, 1986, p. 1174)

226

In short, a moderator influences the strength or direction of a relationship between two variables, whereas a mediator explains the relationship between the two other variables. The following subsections will discuss mediating and moderating effects, and the most suitable methods with which these effects can be tested empirically.

6.3.1 Measuring moderation

Figure 21: Moderating effect (adapted from1986)

X Y C

A third variable (Mo) in an analysis can change the relationship between a predictor (X) and an outcome (Y). When moderation or interaction is present, the slope to predict Y from X differs across scores on the Mo control variable; therefore, the nature of the X-Y relationship differs depending on scores for

Mo. This analysis aims to determine how the relationship between concern (X) and behaviour (Y) varies across social grade (Mo) (an indicator of class, income and job type).

227

6.3.2 Interaction effect

Figure 22: Moderation through interaction (adapted from Warner 2012)

Mo Y

X * Mo

An effective method of assessing moderating effects is by including the product of X and Mo as an additional predictor variable in a regression model.

Figure 22 illustrates this approach. The unidirectional arrows toward the outcome variable Y represent a regression model in which Y is predicted from

X, Mo, and an interaction term that represents the interaction between X and

Mo. The three predictors are correlated with each other11. An interaction between X and Mo predictors in a linear regression can be assessed by forming a new variable that is the product of X × Mo and including this product

11 These correlations are represented by the double-headed arrows. 228

term in a regression, along with the original predictor variables. Interaction effects represent the combined effects of variables on the criterion or dependent measure. When an interaction effect is present, the impact of one variable depends on the level of the other variable.

Equation 1: Moderating effect of X on Y

� = �� + �� + �� + ��(�×��)

Where Y is the outcome variable, and X and Mo are predictor variables. If the b3 coefficient for the X × MO product term in the regression is statistically significant, this is interpreted as a statistically significant interaction between X and MO as predictors of Y. Though there are often correlations among these predictor variables (X, MO, and X × MO).

This chapter will examine the product of an interaction between environmental class membership and SES, and how this interaction is associated with levels of pro-environmental behaviour. Studying this association will show how environmental concern, when conditional on SES, can influence behaviour. It may be that only when SES is sufficiently high, that pro-environmental concern can be allowed to translate into high levels of pro-environmental concern.

229

6.3.3 Measuring mediation

Figure 23: Mediated effect (adapted from Baron & Kenny, 1986)

Me a b

C X Y

Figure 23 illustrates the mediating effect of X on Y via Me. To calculate this effect, different combinations of the parameters a, b and c are used. The product of the a and b parameters, ab, is the mediated effect. The mediated effect is also equal to the difference between the c and ab. As a result, the total effect can be partitioned into a direct effect (c), and an indirect effect (ab).

230

Equation 2: Direct effect of X on Y

� = �� + �� + ��

Equation 3: Mediating effect of X on Y

� = �� + �� + �� + ��

�� = �� + �� + ��

Equation 2 defines the total direct effect of X on Y. Equation 3 defines the mediated effect of Me where Y is the dependent variable, X is the independent variable, Me is the mediating variable or mediator, c represents the relation between the independent variable to the dependent variable, b is the parameter relating Me to Y adjusted for the effects of the X, a is the parameter relating X to the mediating variable, e1, e2, and e3 represent unexplained or error variability, and the intercepts are i1, i2, and i3. The rationale behind this calculation is that mediation depends on the extent to which X affects Me (a) and the extent to which the Me affects the Y (b). The ab quantity reflects how much a 1 unit change in X affects Y indirectly through Me. Similarly, the change in c when adjusted for the mediator, reflects how much of the relation between Y and X is explained by Me (MacKinnon 2008).

This study posits that attitudes influence behaviours, but that SES affects this relationship. Proposition is based on the results of studies such as Franzen and Meyer (2010) and Givens and Jorgenson (2011) who document positive

231

associations between environmental concern and education, income and social class. As well as Marquart-Pyatt (2008), who, using data for 19

European countries, consistently found education and income to be positively associated with both pro-environmental attitudes and behaviour. Indeed, as discussed in Section 6.2.3, environmental issues tend to be prioritised lower than other issues such as health and wealth. At higher social grades, such issues are likely to be satisfied, thus allowing secondary priorities, such as the environment, to influence decision-making and behaviour.

Since environmental attitudes are unlikely to affect social grade, this chapter will examine how environmental class membership mediates the relationship between social grade and general level of pro-environmental behaviour. In other words, high SES increases the probability of an individual engaging in pro-environmental behaviour, environmental attitudes then account for this relationship; the direct relationship between SES and behaviour weakens without the presence of pro-environmental attitudes. This hypothesis is tested using path analysis; a statistical technique used to examine the comparative strength of indirect effects among variables (Lleras 2005). This method estimates the magnitude and significance of hypothesised connections between sets of variables. A series of parameters are estimated by solving one or more structural equations in order to test the fit of the correlation matrix between two or more connections (2005).

232

6.3.4 Indirect effects in cross-sectional data

It is acknowledged that this study suffers from what Roe (2012) refers to as

‘temporal illusion’, a term used to denote the belief among researchers that the flow of time is present when it is not. Cross-sectional data such as the

EAS does not permit a causal conclusion, however both mediating and moderating effects, are by nature, causal models as the underlying theories suggest directional inferences that are intrinsically causal (Rose, Holmbeck,

Coakley & Franks, 2004). It is common among researchers to theorise about events and processes, which by definition unfold over time, to gather and analyse cross-sectional data from which time is lacking, and interpret results in terms of events and processes (e.g. Wang & Takeuchi, 2007). This practice is typically accompanied by the ritualistic statement, ‘the cross-sectional findings should be confirmed by longitudinal research’. However, given the potentially valuable information on the attitude-behaviour relationship that examining indirect effects could yield, it is arguably better to examine indirect effects in an imperfect way (using cross-sectional data), than to overlook them entirely. Indeed, “sequence is in the eye of the beholder” (Roe 2012, p. 7).

6.3.5 EAS measures of environmental behaviour

The EAS contains both single and multiple measures of environmental behaviour. Single measures of behaviour are the reported level of generally pro-environmental behaviour, and the reported satisfaction with this level. The

233

EAS also contains multiple measures of specific behaviours.

Participants are asked to indicate on a five-point scale the extent to which they engage in pro-environmental activities. The response categories and distribution of this variable are shown in Figure 24.

Figure 24: Response categories and distribution for single measure of behaviour

47%

a) I don't really do anything that is environmentally friendly

b) I do one or two things that are environmentally friendly 25% 23% c) I do quite a few things that are environmentally friendly

d) I'm environmentally friendly on most things I do

e) I'm environmentally friendly in everything I do 3% 2%

a b c d c 12

Further to this, participants are asked to assess their satisfaction with their current level of behaviour. The three response categories and the variable’s distribution are displayed in Figure 25.

12 Figure counts: a = 88, b = 674, c = 1377, d = 732, e = 58

234

Figure 25: Response categories and distribution for participant assessment of behaviour

47% 45% a) I'm happy with what I do b) I'd like to do a bit more to help the environment c) I'd like to do a lot more to help the environment 8%

a b c 13

The multiple measures of pro-environmental behaviour in the EAS are organised into four behaviour categories: recycling, travel, food and household. The majority of specific behavioural measures are captured on a five-point scale (from always to never) indicating the frequency with which the respondent says that they engage in that type of pro-environmental behaviour.

Questions on travel, some on food and household energy conservation, were captured with DEFRA’s standard scale or repeat purchasing scale. These scales are nominal, complex and capture other concepts as well as behaviour, making results difficult to interpret. As such, measures captured by these

DEFRA scales have been recoded as binary variables (shown in Table 23 and

Table 24), so as to provide meaningful results in the present analysis.

13 Figure Counts: a = 234, b = 1377, c = 1318

235

Table 23: Dichotomous recode of the DEFRA ‘Standard’ scale

Dichotomous DEFRA ‘Standard’ scale Recode

I haven’t heard of this

I don’t really want to do this

I haven’t really thought about doing this I am not doing this I’ve thought about doing this, but probably wont do it

I’m thinking about doing this

I’ve tried doing this, but I’ve given up

I’m already doing this, but I probably wont manage to keep it up I am doing this

I'm already doing this and intend to keep it up

Table 24: Dichotomous recode of the ‘Regular Purchasing’ scale

Dichotomous DEFRA ‘Regular Purchasing’ scale Recode

I haven’t heard of this

I don’t really want to do this

I haven’t really thought about doing this I haven’t done this I’ve thought about doing this, but probably wont do it

I’m thinking about doing this

I’ve tried doing this, but I’ve given up

I’ve done this, but I probably wont do it again I have done this I’ve done this and intend to do it again

236

Some measures of behaviour were omitted from the analysis in this chapter as they were not applicable to a large portion of participants, resulting in a high proportion of missing responses. This included behaviours that were conditional on wealth or property ownership, such as the installation of solar panels and home insulation. To avoid introducing bias against low-income respondents and the exclusion of a large portion of respondents, these items were excluded. Table 25 shows the behavioural measures used for analysis in this chapter, and indicates which have been dichotomised.

237

Table 25: Measures of specific pro-environmental

behaviours used in this analysis

Category Measure of behaviour

Recycling items rather than throwing them away Recycling Reuse items like empty bottles, tubs, jars, envelopes or paper

*Taking fewer flights

*Switching to public transport instead of driving for regular journeys

Travel *Switching to walking or cycling instead of driving for short, regular journeys

*Driving in a fuel efficient way

*Wasting less food

*Buying fresh food that has been grown when it is in season in

Food the !country where it was produced

Take your own shopping bag when shopping

Decide not to buy something because it has too much packaging

*Cutting down on the use of gas and electricity at home

*Turning down thermostats (by 1 degree or more)

Washing clothes at 40 degrees or less Household Making an effort to cut down on water usage at home

Cut down on the use of hot water at home

Leave your TV or PC on standby for long periods of time at home

* Dichotomous variables

238

6.3.6 Measure of social grade

As in the previous chapter, social grade is used as a proxy measure of socio- economic status. The NRS (National Readership Survey) social grades system of socio-economic classification is used, based on the occupation of either the individual being interviewed, or head of the household. Categories for this measure are displayed in Table 20.

239

6.4 Direct effects of environmental class on behaviour

The association between environmental class membership, and specific measures of environmental behaviour is assessed using logistic regression.

The Disengaged class is used as the reference category for this analysis, due to their apathy and detachment from environmental concern. Class membership is regressed on 16 measures of pro-environmental behaviours separately, each adjusting for age and gender. The odds ratios for these 16 regressions are displayed in Table 26.

240 Table 26: Logistic regression of environmental class on environmental behaviour

Class F Measure of Behaviour Pro- Neutral Paradoxical Environment Majority

Taking fewer flights 2.87* 1.72* 1.41 6.60**

Driving in a fuel efficient way 1.47 1.06 0.65 7.13**

Switching to public transport instead of Travel 2.11* 1.65* 1.23 17.12** driving for regular journeys

Switching to walking or cycling instead of 1.66* 1.28 0.88 3.20** driving for short, regular journeys

Cutting down on the use of gas and 2.41* 1.57* 0.94 7.80** electricity at home

Turning down thermostats (by 1 degree or 2.19* 1.62* 0.94 9.48** more)

Wash clothes at 40 degrees or less 1.61* 1.17 0.91 5.39** Home Make an effort to cut down on water usage 2.66* 1.60* 1.23 17.99** at home

Cut down on the use of hot water at home 2.56* 1.44* 1.30* 16.07**

Leave your TV or PC on standby for long 0.58* 0.72* 0.87 6.41** periods of time at home

Checking whether the packaging of an 3.78* 2.44* 1.55* 18.16** item can be recycled, before !you buy it

Take your own bag when shopping 2.23* 1.52* 1.06 31.04**

Buying fresh food that has been grown Food when it is in season in the !country where it 3.43* 2.15* 1.43* 16.28** was produced.

How much effort do you and your household go to in order to minimize the 2.91* 1.73* 1.40* 25.05** amount of uneaten food you throw away?

Recycle items rather than throw them away 2.50* 1.39* 1.17 15.37**

Recycling Reuse items like empty bottles, tubs, jars, 2.84* 1.48* 1.12 13.76** envelopes or paper *p < 0.05 **prob > F

Table 26 indicates that the odds of engaging in environmental behaviour are higher among those holding pro-environmental attitudes (compared to those who are disengaged from environmental issues). Pro-environment class membership is a significant predictor of all but one measures of behaviour.

Though, this measure (driving in a fuel efficient way) is not significantly associated with other environmental classes either, suggesting perhaps than environmental attitudes do not influence such behaviour. Membership of the

Neutral Majority class is significantly associated with the majority of behaviour

(13/16). Though of these significant relationships, odds ratios are smaller than those obtained from the Pro-environment class. This provides further evidence that greater pro-environmental concern leads to a higher probability of engaging in pro-environmental behaviour. Paradoxical class membership is only significantly associated with four measures of behaviour. Membership of this class is significantly associated with cutting down on hot water in the home, minimising food waste, checking product packaging, and buying fresh local produce (compared to membership of the disengaged class). Though the odds of Paradoxical class members engaging in these behaviours are lower than for members of the Pro-environment and the Neutral classes.

Table 26 displays the f ratio for each regression. This is an indication of between and within class variance. The majority of measures in this table have a high f ratio, indicating distinct classes that capture a large proportion of the variance.

242

6.4.1 The relationship between class membership and general

level of pro-environmental behaviour

Reported levels of general pro-environment behaviour made by members of each environmental class are shown in Figure 26. As shown in this figure, a greater proportion of the Pro-environment class report engaging in high levels of pro-environmental activity, compared to the other classes, where as the

Paradoxical and Disengaged classes have the highest proportion of individuals who engages in nothing or very little to help the environment. However, only a minority of people in each class does nothing to help the environment; the highest proportion belonging to the paradoxical class with 7%.

Figure 26:Within-class reported level of environmental behaviour

.60 Pro-environment Neutral Majority

n .40

op .20

.00

.60 Paradoxical Disengaged

o .40

r .20

.00

I don't do I do quite a few I am I don't do I do quite a few I am anything that is things that are environmentally anything that is things that are environmentally environmentally environmentally friendly in environmentally environmentally friendly in friendly friendly everything I do friendly friendly everything I do

Which of these would you say best describes your current lifestyle?

243

What is noticeable in Figure 26 is the clustering around the median, as well as a lack of substantial skew. This shows that within each class, the majority of people claim to do “quite a few things that are environmentally friendly” rather than doing very little or very much. Such a distribution suggests that environmental concern does not strongly influence general levels of pro- environmental behaviour. This is supported by the finding of a Cramers V =

.193, indicating a weak relationship between environmental class and general level of pro-environmental behaviour. However, findings in Table 26 indicate that environmental class is a strong predictor of specific behaviour. Indeed, when examining both Table 26 and Figure 26, it can be concluded that broad level, general environmental concern is not a particularly good predictor of general behaviour, but is a good predictor of specific behaviours (such as the behavioural measures used in Table 26). This finding is in contrast to the Ajzen and Fishbein (1977) suggestion that attitudes and behaviour should be studied at equal levels of specificity, as well as their recommendations of studying the relationship between behaviour specific attitudes and the relevant behaviour.

Figure 27 shows the level of satisfaction class members have with their level of general pro-environmental behaviour. This figure indicates that members of the

Pro-environment class are the least satisfied with their level of activity, and wish to do more. This is despite the fact that they reportedly engage in the most pro-environmental behaviour compared to other classes. In contrast to this, the Paradoxical and Disengaged classes are the most satisfied with their level of behaviour, though they reportedly do the least.

244

Figure 27: Within-class satisfaction with environmental behaviour

Pro-environment Neutral Majority .60

n .40

tio

opo .20

.00

Paradoxical Disengaged

.60

n .40

tio

opo

r .20

.00

I'd like to do I'd like to do a I'm happy I'd like to do I'd like to do a I'm happy a lot more bit more with what I do a lot more bit more with what I do

How do you feel about your current lifestyle and the environment?

This shows a strain between what people are doing, and what they want to do.

The high level of satisfaction combined with such low levels of behaviour amoungst the Paradoxical and the Disengaged is likely a reflection of what is considered to be normative behaviour in the UK. Current social norms allow members of the Paradoxical and Disengaged groups to do little for the environment, while at the same time, be satisfied with such inaction. That is to say that it is not normative to frequently engage in many forms of pro- environmental behaviour. Such norms inhibit the behaviour of Pro-environment class members, resulting in a form of cognitive dissonance; this class cares

245

greatly about the environment but are unable to engage in a level of pro- environmental behaviour that they are satisfied with. Ideally, to promote high levels of such behaviour, norms should shift in a way that leaves the

Paradoxical and Disengaged dissatisfied, and allows the Pro-environment group to engage in their desired level of behaviour.

246

6.5 The moderated association between environmental concern and behaviour

This section examines the moderating influence of social grade on the attitude- behaviour relationship. Only measures of general behaviour, and the reported satisfaction with that behaviour are used in this analysis, since the focus is upon whether or not a moderating effect exists, rather than how this effect varies across multiple measures of behaviour. To determine how SES moderates the relationship between environmental class and pro- environmental behaviour, an interaction between environmental class membership and social grade is used. As described in section 6.3.2, interaction effects represent the combined effects of variables on the criterion or dependent measure. When an interaction effect is present, the impact of one variable depends on the level of the other variable. The product of this interaction is included as an variable in a regression model. Figure 22 illustrates this approach. A comparison is made between the Pro-environment class and the Disengaged class. This comparison is made between these, the two most opposing classes, to illustrate the difference between those who are strongly concerned about the environment, and those who are apathetic to environmental issues. Figure 28 shows linear predictions of the interaction between these two classes and social grade, and how this interaction affects both the level of, and satisfaction with pro-environmental behaviour. These measures of behaviour are treated as continuous for these linear predictions.

247

Figure 28: Pro-environment and Disengaged classes on level and satisfaction with behaviour, moderated by social grade.

Level of pro-environmental behaviour

3.5

3.0 Level of Behaviour (linear predication) 2.5 e d c2 c1 b a Social Grade

Satisfaction with level of pro-environmental behaviour

2.8

2.6

2.4

2.2

2.0 Level of satisfaction (linear predication) e d c2 c1 b a Social Grade

Pro-Environment Disengaged

248

Figure 28 confirms findings from section 6.4.1, that Pro-environment class members report engaging in high levels of pro-environmental behaviour, while being dissatisfied with this level (wanting to do more). The Disengaged report the opposite; low levels of behaviour combined with high levels of satisfaction.

For the Pro-environment class, level of behaviour is significantly higher for the c1 social grade than social grade d. This suggests that an increase in SES can allow attitudes to influence behaviour. However, this only appears to be the case when comparing social grade d (lower class workers) with social grade c1

(middle class workers). However, when all social grades are considered, behaviour does not significantly alter with the increase in social grade. For the

Disengaged, there is a significant increase in behaviour between the middle classes (social grades c1 and c2) and the upper class (social grade a). Overall, is a significant difference between these two classes regarding their reported level of behaviour, except for member social grades a and d. The significant effect of environmental concern on behaviours disappears at social grades a and d, indicating that the predictive ability of environmental concern disappears for members of these social grades. Regarding the reported level of behavioural satisfaction, the affect of environmental concern disappears for social grade a for both classes. Further to this, Pro-environment class satisfaction significantly decreases when comparing social grade d to social grade b. This may be because, as SES increases for members of this group, they feel able to do more for the environment, and so grow dissatisfied with their current level of pro- environmental behaviour.

249

6.6 The mediating effect of environmental class on the relationship between social grade and behaviour

This section uses path analysis to study the mediating effect of environmental class on the relationship between social grade and behaviour. This will determine how much of the relationship between social grade (a proxy for SES) and pro-environmental behaviour is accounted for by environmental concern.

Table 27 shows the, the direct effects of social grade on behaviour, social grade on environmental class, and environmental class on behaviour. This table also shows the indirect effect social grade on behaviour via environmental class14.

14 The probability of belonging to the relvant class is used for this analysis (a continuous variable) as opposed to class membership (categorical) as used in previous analysis.

250

Table 27: Direct and indirect effect between environmental class, level of

pro-environmental behaviour and social grade

Class

Pro- Neutral Paradoxical Disengaged Environment Majority

Social Grade

Level of Pro- 0.02 0.04* 0.04* 0.03* environmental Behaviour

Environmental Class

ON 0.55* -0.01 -0.32* -0.56* Level of Pro- Direct Effect Direct

environmental Behaviour

Social Grade

ON 0.04* 0.01* -0.03* -0.02*

Environmental Class

Social Grade

Level of Pro- Effect environmental 0.02* 0.00* 0.01* 0.01* Behaviour

Indirect VIA

Environmental Class

* = p<.05

251

The results shown in Table 27 suggest that social grade has a significant but very weak affect on pro-environmental behaviour for all classes except Pro- environment for which the relationship is non-significant. Suggesting perhaps

SES does not strongly encourage pro-environmental behaviour. Further to this, social grade has a significant but very weak relationship environmental class membership, suggesting that SES does not affect what form of environmental concern one ascribes to. Table 27 does indicate that environmental concern is significantly associated with behaviour. The probability of belonging to the Pro- environment class is positively associated with level of behaviour, while the

Paradoxical and Disengaged group are negatively associated (the latter class more so). Probability of belonging to the Neutral class is not associated with behaviour, which unsurprisingly indicates that neutrality does not influence levels of behaviour.

Aside from these direct effects, very weak (though significant) indirect effects are found, suggesting that environmental class does not mediate the relationship between social grade and behaviour.

252

6.7 Discussion

Results in section 6.4 strongly suggest that attitudes do significantly affect behaviour. Both Neutral and Pro-environment class membership is positively associated with the majority of behavioural measures (though the latter group more so). Comparatively, Paradoxical class membership is only significantly associated with two measures of behaviour. Pro-environment class members are less satisfied with the level of behaviour that they engage in, despite their comparatively higher levels of pro-environmental behaviour. The Disengaged and the Paradoxical classes report the lowest levels of behaviour but the highest level of satisfaction. In short, the more people do for the environment, the less satisfied they are.

Pro-environment class membership is found to be a significant predictor of taking fewer flights, as well as walking and using public transportation over driving. Membership of the Neutral class is only a significant predictor of taking fewer flights, and using public transport, while Paradoxical class membership does not predict any measures pro-environmental travel behaviour. These findings do not support previous research conducted by Whitmarsh (2012) and

Diekmann and Preisendöerfer (1992), which suggests that transport behaviours are not affected by attitudes towards the environment. Whitmarsh (2012) proposed that travel behaviours are driven by habit, and difficult to alter, making them high-cost behaviours. Diekmann and Preisendöerfer (1992) also suggested that environmental attitudes are only able to influence low-cost behaviours such as recycling, because they are easy to do in terms of time, effort and financial cost. As such, attitudes are not strong enough to influence

253

high-cost behaviour, i.e. a dramatic change in one’s use of transportation. For an individual to give up the use of their car or start walking to work everyday requires more than a concern for the natural environment. However these results suggest that concern is enough to motivate such high-cost behaviour.

Environmental concern was not however significantly associated with whether an individual drives in a fuel efficient way, suggesting that not all pro- environmental transport behaviours are predicted by environmental concern.

Previous research suggests that SES affects the relationship between concern and behaviour. More specifically, those those who have a high SES are more likely to engage in high levels of pro-environmental behaviour (this has been found by Whitmarsh, 2009 and Cottrell, 2003). Maslow’s hierarchy of needs

(Maslow et al., 1970), indicates that people with higher SES have satisfied their basic needs (regarding physiology and safety), and are thus able to focus on satisfying other ‘higher' needs including environmental improvement/conservation. Another explanation indicates that those with higher income and higher education are able to assimilate environmental information more quickly. This increased the probability of these individuals possessing more knowledge about environmental problems. Furthermore, some studies suggest that high-income and well-educated people are more likely to have post-materialist views emphasising quality of life and environmental sustainability instead of economic growth and material possessions (see Inglehart 1995; Van Liere & Dunlap, 1980).

254

However, results of this thesis15 support those found by Dietz et al. (1998) and

Kanagy et al. (1994) found income and job type to be weak predictors of EC.

However this chapter found SES has a weak affect on the concern – behaviour relationship. For example, social grade was not found to strongly affect behaviour directly nor indirectly via environmental concern, and social grade does not strongly moderate the concern-behaviour relationship.

Furthermore, results in this chapter also indicate that the higher levels of environmental behaviour the lower the level of satisfaction with such behaviour, and vice versa. This is thought to be a result of current social norms that allow members of the public to engage in low levels of behaviour while feeling satisfied with such levels. Norms also prevent members of the Pro- environment class from engaging in the high levels of behaviour that they would be satisfied with.

15 Results obtained in section 5.6.2 and 6.6 find that social grade and income to be weak predictors of concern.

255

6.8 Conclusion

This chapter began by positing two hypotheses. First, that concern for the natural environment is significantly associated with environmental behaviour.

This has been supported by the analysis in this chapter. The Paradoxical class is not significantly associated with specific measures of pro-environmental behaviour, while the Pro-environment class and Neutral class is significantly associated with the majority of behaviour measure. Indeed, pro-environmental concern is able to affect recycling, food behaviour and home energy conservation and travel behaviour. This latter finding does not support previous findings but does suggest that broad level environmental concern is a stronger predictor of pro-environmental behaviour that the literature would suggest.

The second hypothesis tested in this chapter was that socio-economic status affects the association between concern and behaviour. Unexpectedly, social grade was not found to strongly affect behaviour directly nor indirectly via environmental concern. Social grade has a mild moderating effect on environmental class membership and level of, and satisfaction with, pro- environmental behaviour.

In conclusion, broad level pro-environmental concern is a strong significant predictor of specific pro-environmental behaviours. The relationship between such concern and general level of, and satisfactions with, pro-environmental behaviour is moderated, to a limited extent, by social grade.

256

Chapter 7 Discussion

257

7.1 Introduction

This discussion chapter summarises the findings from the preceding analytical chapters (4-5) before discussing the thesis’ contribution to the study of environmental concern, its relationship with pro-environmental behaviour, while also addressing the wider implications within the current public policy framework in England. The limitations of the analysis and potential for future research are also discussed.

7.2 Thesis summary

Chapter 1 of the thesis proposed three broad research objectives:

1. To explore environmental attitudes in one of the few large scale national

samples of environmental attitudes and behaviour in the UK;

2. To identify the nature of environmental concern in the same sample;

3. To investigate how environmental concern is associated with pro-

environmental behaviours in the same sample.

In order to achieve these objectives, the thesis reviewed previous studies of attitudes and behaviour, concern for the natural environment and environmental action. A new empirical investigation into environmental concern and its relationship with pro-environmental behaviour was then conducted across three analytical chapters, culminating in an understanding of how the

British public perceives the natural environment (particularly in relation to environmental problems) and how this perception affects levels of pro- environmental behaviour.

258

The literature review presented a brief history of attitudinal research to demonstrate: a) how attitudes have long been defined by their influence on behaviour, b) how researchers should use caution when defining attitudes in relation to behaviour, and c) why the present research maintains an ontological distinction between attitude and behaviour. Following on from this, attitudes towards the natural environment were discussed and the term environmental concern (EC) was introduced. This term is used throughout the thesis to denote broad-scope, general attitudes towards the natural environment. One of the most common conceptualisations of environmental concern is the values beliefs norms (VBN) theory (Stern et al., 1999), shown in Figure 29.

Figure 29: New Environmental Paradigm (NEP) portion of the Theory of Values Beliefs Norms

Values Beliefs Biospheric Adverse Perceived Ecological consequences ability to worldview for valued Altruistic reduce threat (NEP) objects (AR) (AC) Egoistic

This theory argues that environmental concern is comprised of value-based attitudes. Value is placed on the self, other people or nature and these values then influence and shape environmental attitudes. Placing high value on society for example (referred to in the VBN as a social-altruistic value orientation) forms attitudes regarding the earth’s ability to sustain the human population. Highly valuing nature (biospheric value orientation) forms attitudes

259

towards the natural environment and how it is at risk from environmental problems such as climate change. Finally, valuing the self (egoistic value orientation) produces attitudes focusing on how the individual should interact with, and will be personally affected by, the environment. This theory however has not been tested on a large, representative UK sample, so it is uncertain whether these components of environmental concern exist amongst the UK population. Further to this, little to no studies have considered that these components may exist in combination with each other in varying strengths, potentially producing multiple forms of environmental concern.

Studies since the 1970s have found that concern for the environment varies according to socio-demographic and economic factors. Though age and gender are, overall, found to be inconsistent in their predictive abilities, socio- economic variables such as education and income do consistently correlate with high levels of concern. Finally, the complex relationship between environmental concern and environmental behaviour was reviewed. There is evidence to suggest that environmental concern is affected by socio-economic status (SES). There is also evidence to suggest that behaviour too is affected.

Research generally suggests that members of higher social grades (those with high household incomes, who are well educated and possess a high managerial or professional job) are more likely to be concerned with environmental issues, and are better able to engage in actions that mitigate the effects of such issues. However very few studies have examined the role that socio-economic status plays in detail in the relationship between environmental attitudes and behaviour. For example, it is uncertain whether high SES increases the probability of possessing strong EC, which then 260

translates into increased rates of EB, or whether high SES enables EC to lead to pro-environmental action.

The findings from the literature reviewed in this chapter, when considered in context with the research aims of this study, led to the postulation of the following research questions, each of which builds on the foregoing general objectives for the thesis:

1) What are the components of UK environmental concern?

2) What ‘forms’ of environmental concern exist amongst the UK public?

3) What social characteristics are associated with different forms of

environmental concern?

4) How is environmental concern associated with pro-environmental

behaviours and what is the role of socio-economic status in this

relationship?

Chapter 3 provides a discussion for how to best answer these research questions, focusing on data and methods. Potential datasets were shortlisted and discussed, outlining reasons for using the DEFRA EAS survey data over other data sources. This chapter concludes with a discussion of what statistical methods are best to answer the research questions with these data.

Chapter 4, the first of three analytical chapters, investigated the structure and dimensions of environmental concern by studying attitudes towards the environment. Appropriate measures were selected from the attitudes portion of the EAS and exploratory factor analysis was used to uncover the components of environmental concern. Parallel analysis, K1 testing and scree plot

261

examination were used to establish the optimal number of factors. A three- factor model was found to be the optimum. These factors captured the following components of environmental concern: ecocentric (attitudes towards plants and animals) human-centric (attitudes towards the earth’s ability to sustain the human race), and denial of environmental problems. The latter is interpreted as a result of egoistic value orientation. This model was specified and tested using confirmatory factor analysis (CFA), and a promising new method known as Bayesian structural equation modelling (BSEM). BSEM uses information acquired from exploratory analysis to specify variable parameters, thus removing the unnecessary restrictions placed upon the model parameters by CFA, which prohibit any crossloading. CFA goodness-of-fit indices indicate that the model fits the data well. Unfortunately BSEM goodness-of-fit indices penalised the model because of the large sample size, indicated poor fit for the full sample, but excellent fit on a smaller sub-sample. The three-component model strongly reflected the VBN value-based model of environmental concern.

Analysis in this chapter suggests that valuing others leads to concern over the use of natural resources, and for the earth’s ability to sustain humanity. Valuing nature leads to concern for plants and animals, and how these forms of life are harmed by environmental problems. When value is placed on the self, people deny the existence of environmental problems or downgrade their severity.

Chapter 5 examined how environmental concern exists amongst the UK population. This was done by grouping respondents according to homogeneity in their responses towards measures of environmental attitudes. This produced groups (or classes) of people with similar forms of environmental concern. A 262

four-class model was produced. Classes were interpreted by examining within- class item probabilities (response patterns of class members) and mean factor scores of the environmental concern components found in Chapter 4. From their interpretation, these classes were labelled Pro-environment, Neutral

Majority, Disengaged and Paradoxical. A description of these classes is shown in Table 28.

These class members demonstrate a counterintuitive item response pattern.

Like the Disengaged class, item responses do not differ/change/vary between positive and negative statement, but unlike the Disengaged, item response probabilities are reasonably high. Members of this class therefore agree with both positive statement and opposing negative statement. Upon further analysis, class members acknowledge that climate change may be real, but if it is, humans are not believed to be the cause.

263

Table 28: Environmental classes discovered in Chapter 5

Environmental % Description Class

High item probabilities for pro-environmental statements combined with low probabilities for negative statements suggests that members of this class posses strong, pro- Pro- 29% environmental attitudes. This interpretation was supported when environment the factor scores of these class members were examined. These members have a low mean score of denial, but a high means score for human and ecocentric attitudes.

Item probabilities for this class (the largest) suggest that members hold a favourable appraisal of the natural environment. Neutral Majority 35% However, members also hold low scores on all environmental attitudes, indicative of apathy.

Interpretation suggests that these class members are detached from environmental issues. Item probabilities were low across all Disengaged 18% measures (both positive and negative environmental statements16). Further analysis reveals that Disengaged class members are sceptical of the reality of climate change.

These class members demonstrate a counterintuitive item response pattern. Like the Disengaged class, item responses do not differ/change/vary between positive and negative statement, but unlike the Disengaged, item response probabilities are Paradoxical 18% reasonably high. Members of this class therefore agree with both positive statement and opposing negative statement. Upon further analysis, class members acknowledge that climate change may be real, but if it is, humans are not believed to be the cause.

16 i.e. The probability that class members register agreement with statements is low.

264

Age, gender, income, education and social grade of environmental class members were examined. Pro-environment class members are largely middle class and well educated. Neutral Majority class members are average in all respects. Disengaged and Paradoxical class members are poorly educated and have low household income. However, these two groups differ in their age distribution; Paradoxical members are old17 and Disengaged members are young. Chapter 6 examined the relationship between environmental concern and level of pro-environmental behaviour, while also examining the potential indirect effect of social grade on this relationship.

Environmental class membership was regressed onto 16 measures of pro- environmental behaviour. Membership of the Pro-environment class is significantly associated with all but one measure of pro-environmental behaviour. Membership of the Neutral class is significantly associated with

13/16 measures, though odds ratios are weaker than those obtained by members of the Pro-environment class. Finally, membership of the Paradoxical classes is not significantly associated with any measure of behaviour.

How SES the relationship between concern and behaviour was also examined.

This was done by assessing how concern mediates the relationship between social grade and behaviour, and by assessing how social grade moderates the concern – behaviour relationship. When examining mediating effects, the

17 One can speculate that they have lived long enough to perceive environmental change (generally milder winters).

265

indirect effects of social grade on behaviour via environmental concern, were extremely weak (ranging from 0.00 – 0.02). The direct effect of concern on behaviour was comparatively much stronger. SES was however found to have a modest moderating effect on the relationship between environmental concern and behaviour.

266

7.3 Answers to research questions

Analyses conducted in this thesis provide answers to the research questions posited in Chapter 2. These answers are presented throughout this section.

7.3.1 What are the components of environmental concern?

The purpose of Chapter 4 was to answer this research question by examining the structure and components of environmental concern. Analysis in this chapter was exploratory in nature. This is not to say that this chapter did not engage with theory; indeed, this research drew comparisons between results with VBN theory. However, results were not restricted by the parameters of this theory. The NEP portion of the VBN theory posits that environmental concern is formed of three value-based attitudes towards the environment. Many studies have engaged with this theory but have done so by using it as a tool of measurement. For example, Snelgar (2006) and Schultz (2000) used questionnaires that were explicitly designed to reflect components of EC specified by the VBN (egoistic, biospheric and social altruistic). However data collection according to theoretically-based categories precludes analysis in other terms, limiting the conceptualisation of environmental concern to one theoretical perspective. This thesis however, uses DEFRA’s EAS, which contains 24 measures of environmental attitudes, designed without a priori commitment to one specific theoretical framework, reducing the likelihood that participants are conditioned to a particular response pattern.

267

Nine measures of environmental concern were selected from the environmental attitudes portion of the DEFRA EAS. Items were excluded because they captured attitudes towards environmental behaviour instead of the natural environment itself, or were felt to be poor measures of environmental concern.

The remaining nine measures were examined using three forms of factor analysis, producing a three-factor model shown in Figure 13. These components were labelled as Denial, Human-centric concern, and Ecocentric concern, and appear to overlap strongly with VBN components.

Human-centric concern

The welfare of humanity is highly valued, making this form of concern

similar to the social-altruistic component of the VBN. Environment is

valued for its ability to sustain humanity, alongside a concern for the

diminished capacity of the earth to do this, expressed as a negative

attitude towards a high people/resource ratio and an acknowledgement of

the imminent threat of an environmental disaster.

Ecocentric concern

The natural environment is highly valued, but unlike human-centric

concern, nature is appreciated for its intrinsic value. This is expressed in a

negative attitude towards changes made to the countryside and a concern

for the loss of animal species.

268

Denial

Denial is an interpretation of the lack of belief in environmental damage,

resulting in low environmental concern and sceptical views towards

environmental problems. It is viewed here as mapping onto the VBN’s

egoistic component. Denial as a result of egoism has also been found by

Hanlsa et al. (2008) and includes a high valuation of the self relative to

others or the natural environment. It is possible that this value orientation

results in denial, as when value is placed on the self, people deny the

existence of environmental problems or downgrade their severity. This may

be done as a means of coping, to mitigate the individual’s fear of being

harmed or changed by environmental problems. Another possibility is that

the individual does not wish to alter their environmentally detrimental

behaviour and denial is used to justify inaction. Further research is required

to confirm these possibilities.

7.3.2 How does environmental concern exist amongst the UK

public?

Chapter 5 built on Chapter 4’s investigation into the structure of environmental concern by examining the degree to which the UK public experiences such concern. It is thought to be unlikely that environmental attitudes exist in a vacuum, but rather that they exist in combination with each other in varying quantities. Individuals therefore experience different forms of environmental concern depending on this combination of components.

269

To investigate what forms of environmental concern exist amongst the UK public, the EAS sample was grouped according to homogenous response patterns for the nine selected measures of environmental concern, producing a typology. Using latent class analysis, four environmental classes were discovered (this being the optimal number of classes for the model). The class names and sizes are shown in Table 29.

Table 29: Environmental classes discovered in Chapter 5

Class Label N %

Class 1: Pro-environmental 837 28.59

Class 2: Neutral Majority 1036 35.38

Class 3: Paradoxical 513 17.53

Class 4: Disengaged 542 18.50

Total 2,928 100.00

Respondents were grouped into four categories of environmental concern.

Class 1 consists of members who hold extremely positive perceptions of the environment and are concerned about environmental problems. Classes 3 and

4 are by contrast extremely negative. Though these negative classes are distinguished from each by their socio-demographic characteristics, members of both classes are uncaring of environmental problems and deeply sceptical about climate change. Finally, Class 2, the largest at 35% of the sample, hold very minor concern for the environment.

Mean factor scores of class members revealed that Pro-environment class members care strongly about nature, mildly about humanity, and have low

270

levels of denial. Members of the negative classes (Paradoxical and

Disengaged) both have high levels of denial and extremely negative ecocentric and human-centric scores. Neutral Majority class members however have scores of close to 0 across all measures, indicating neutral or even non- existent attitudes towards the environment. This is interpreted as a form of environmental apathy.

7.3.3 What social characteristics are associated with different

forms of environmental concern?

In Chapter 5, following the classification and interpretation of environmental classes, socio-demographic information of class members was assessed.

Pro-environment

The probability of being a member of this class increases with level of

education and social grade. The Pro-environment class has the highest

proportion of middle and upper-middle class members Therefore the

majority of Pro-environment members are well educated and according to

the NRS social grade classification, are largely middle or upper class. This

class has a relatively equal gender ratio, though slightly more women then

men, and most class members are middle aged (40-55).

Disengaged and Paradoxical

Both of these classes are primarily comprised of people with low social

grade, low household income and have the lowest level of education (high

271

proportions of people with no formal qualifications combined with low

proportions of people with degree level qualifications). SES appears to be

a major distinguishingPro-environment characteristic between peopleModerate with negative Mainstream forms .03 of concern and those with positive forms.

.02

.01 Figure 30: Age distribution of Paradoxical and Disengaged class members .00 Paradoxical Disengaged

.03 Proportion

.02

.01

.00 20 40 60 80 100 20 40 60 80 100 Age

There are two primary differences between the Disengaged and the

Paradoxical group, however: the Disengaged class has the highest

proportion of young people (16-30) and the Paradoxical has the highest

proportion of older people (70+). Age distributions of these class members

are shown in Figure 21.

Those who possess negative forms of environmental concern are mostly

either the youngest or the oldest members of the UK public, which

certainly warrants further research. It should be highlighted of course that it

is highly unfortunate that young British people are disengaged from

272

environmental issues, given that they are likely to be the worst affected by

environmental degradation.

Neutral Majority

This class is primarily comprised of participants in their late 30s and early

40s and has the highest proportion of middle and lower-middle class

workers. The characteristics of this class are similar to Pro-environment

class members, though they are slightly younger and are of a slightly lower

class.

Overall there does appear to be a link with SES and environmental attitudes.

Income, education and social grade are all positively linked with the group who are most concerned about the environment, while the opposite can be said regarding the groups who possess rather negative attitudes towards the environment and are sceptical of environmental problems.

7.3.4 How is concern for the environment associated with

pro-environmental behaviour, and what is the role of

socio-economic status in this relationship?

Ajzen and Fishbein (1980) have argued that attitudes and behaviours should be measured at a comparable level of specificity; otherwise correlations between them are likely to be modest. Evidence for this assertion has not been found in this thesis. Indeed, broad level attitudes towards the natural environment have been found to be strong predictors of both specific and general level behaviour.

273

Chapter 6 examined the relationship between environmental concern and behaviour, as well as how SES may affect this relationship. 16 measures of environmental behaviour relating to transport, food, recycling and home energy reduction were used to make this assessment, as well as a single measure of overall pro-environmental behaviour. Pro-environmental attitudes are shown to be strongly predictive of pro-environmental behaviours. Neutral concern also predicts behaviour, but for fewer measures of behaviour. Disengaged and

Paradoxical forms of concern are not significant predictors of behaviour.

Members of the Pro-environment class were found to engage in higher levels of behaviour compared to other classes, though they are the least satisfied with what they do. In contrast to this, though who do the least (Disengaged and Paradoxical classes) are the most satisfied with what they do.

On examination of the influence of socio-economic, it was found that social grade did not mediate the relationship between environmental concern and level of pro-environmental behaviour. It was also found that the predictive power of environmental concern disappears for the lowest and highest working social grades (a and d).

274

7.4 Key messages

This thesis has three key messages: that environmental attitudes do influence levels of pro-environmental behaviour, that social grade moderates but does not mediate this relationship between environmental attitudes and behaviours, and that environmental apathy is a significant barrier to pro-environmental behaviour.

7.4.1 Attitudes do influence behaviour

As discussed in Chapters 1 and 2, it was assumed in early social research that attitudes were direct predictors of behaviour, being conceptualised as behavioural tendencies by Allport (1954). Indeed, the field of social psychology was originally defined as the scientific study of attitudes (Thomas & Znaniecki,

1918; Watson 1925) because it was assumed that attitudes were the key to understanding human behaviour. By the mid-1930s however, several studies had begun to identify a large discrepancy between attitudes and behaviour, challenging their synonymity. This was the start of much debate surrounding the attitude-behaviour relationship, leading to many arguing that attitudes were not linked to behaviour at all. By the 1970s some studies took a more level- headed approach, proposing that though attitudes do not always translate to behaviour, there is still a link between the two concepts. However, studies continued to produce weak evidence for such a link, overall suggesting a relationship that at best is inconsistent and at worst is none-existent. These findings extend to the study of environmental attitudes and behaviour. Findings in this field largely suggest that caring about the environment does not make

275

people engage in pro-environmental behaviour. Therefore there is a discrepancy between what people think about the environment and what they actually do. The mismatch between environmental attitudes and behaviours remains an unresolved problem that poses a substantial challenge when it comes to designing publicly acceptable, climate-related policy. Many researchers have suggested, as a solution, that only the relationship between behaviour-specific attitudes and behaviour be examined (for example, the relationship between attitudes towards recycling and level of recycling).

Though this line of enquiry continues the study of environmental attitudes and behaviours, it detracts from the study of why attitudes towards the natural environment do not appear to strongly and consistently affect behaviour. This thesis has sought to unpick attitude-behaviour disparity, to determine if it occurs for all of the UK, across all environmental attitudes, and all measures of behaviour. Chapter 6 found a strong, significant relationship between pro- environmental concern and high levels of pro-environmental behaviour.

Membership of the Pro-environment class is significantly associated with all but one measure of pro-environmental behaviour. The probability of being a member of this class increases considerably with the general level of environmentally-friendly behaviour as shown in Figure 31, and further crosstabulation of attitudinal and behavioural measures produced a chi2 = p<0.05 and Cramers V =.18, confirming the relationship. This strongly suggests that the probability of possessing strong concern for the wellbeing of the natural environment is associated with a high level of pro-environmental activity.

276

Figure 31: Bar chart depicting the relationship between Pro-environment class membership probability and level of pro-environmental behaviour

Pro-environment class membership probability

I don't do anything I do quite a few things Everything I do is that is environmentally that are environmentally environmentally friendly friendly friendly

It was also found that concern predicts changes in travel habits and the substitution of driving for more environmentally friendly method of transportation. Contradicting Diekmann and Preisendöerfer’s (1992) low-cost hypothesis, which predicts that the strength of the effect of environmental concern on environmental behaviour diminishes with increasing behavioural costs.

Overall, results suggest that not only do attitudes predict both low-cost behaviour (recycling, home energy conservation) but also high-cost behaviour

(using other methods of transport rather than driving).

277

7.4.2 The problem of apathy

Social research on human interaction with the natural environment has largely endeavoured to answer the question ‘do environmental attitudes predict for environmental behaviour?’ The results of this study, strongly suggest that the answer to this question is yes, they do. Now it is time to change the line of inquiry. Findings in this thesis suggest that it is not that attitudes do not influence behaviour; it is that a large portion of the population does not possess environmental attitudes strong enough to do so.

As shown in Figure 32, the Paradoxical and Disengaged classes (which combined consist of 36% of the sample) have extremely low average scores on the Ecocentric and Human-centric. Research now needs to be focused on how to address this form of environmental apathy.

Figure 32: Within-class mean environmental attitude scores

Eco-centric Human-centric Denial

-.5

Neutral Pro-environment Paradoxical Disengaged Majority

278

This study posited that an attitude does not exist in a vacuum, and that the three environmental attitudes unearthed in Chapter 4 exist in combination with each other. The relationship that exists between these environmental attitudes produces a form of environmental concern, and given that there are many possible attitude combinations, there are potentially numerous forms of concern. Using latent class analysis, EAS participants were classified into groups defined by their form on concern.

These classes, and specifically their apathy, should be targeted by future policy, with the aim of encouraging them to be more concerned with environmental problems. Paradoxical class members are heavily in denial, old, and have an inconsistent perception of the environment. Item response probabilities for this class do not decrease for negative statements, nor increase for positive statements. Instead, probabilities remain between 0.4 and

0.7 for all nine items. Members of this class therefore are reasonably likely to agree that an environmental crisis has been exaggerated, that it is a low priority and too far in the future to be of concern. However, class members are also reasonably likely to be concerned about the countryside and animal species, the planet’s ability to sustain an ever-growing human population, and acknowledge that if things continue there will be a major environmental disaster. This is clearly illogical, making this perception difficult to engage with.

In short, converting them will be difficult and may not yield long-term benefits.

Disengaged class members are predominantly young, are sceptical of climate change and have the lowest probability for positive items. Probability for negative items is also low, although not as low as Neutral Majority and Pro- environment classes. The low probability for positive items is indicative of 279

scepticism or denial, however item probability for negative items is too low to support this interpretation. Therefore, given the low item probability for all measures of environmental concern, it is likely that class members are disengaged with environmental issues entirely.

The key message here is that 36% of the UK public appear to be unconcerned/ disengaged/ apathetic to the environment. Such views do not lead to pro-environmental behaviour. More should be done to increase the strength of their attitudes towards the environment, in the hope that they may transcend to the Neutral or Pro-environment class.

7.4.3 Moderation not mediation

It is clear from the previous research that socio-economic factors influence levels of environmental concern and behaviour (site Poortinga & Pidgeon,

2003, and Cottrell, 2003). Further to this, there has also been some evidence to suggest that SES is correlated with levels of environmentally friendly activity.

Little research however has examined how SES can indirectly affect this relationship between environmental attitudes and behaviour.

Results in Chapter 6 suggest that environmental concern does not mediate the effects of SES on behaviour. This is to say that the indirect effect between social grade and behaviour via environmental concern was found to be extremely weak, and there appears only to be direct effects between these three concepts.

280

There are weak but significant direct effects between social grade and environmental concern, and between social grade and behaviour. The relationship between environmental class membership and general level of pro-environmental behaviour is positive for the Pro-environment class, and negative for the Paradoxical and the Disengaged classes. There is not a significant relationship between membership of the Neutral class, and general level of behaviour.

When a comparison was made between Pro-environment and the Disengaged class, social grade was found to have a modest moderating effect on the relationship between concern and behaviour. This suggests that the influential effect of environmental concern on relevant behaviours is partially dependent on social grade. There is some limited evidence to suggest that the effects of environmental concern increase with social grade, but the most important moderating effect found was that the influence pro-environmental attitudes on behaviour disappears for members of the social grades a and d.

281

Figure 33: Pro-environment and Disengaged classes on general level of behaviour, moderated by social grade.

Level of pro-environmental behaviour

3.5

3.0 Level of Behaviour (linear predication) 2.5 e d c2 c1 b a Social Grade

As shown inSatisfaction Figure 33, confidence with level intervals of pro-environmental overlap for members behaviour of social grade2.8 a and d, indicating that the predictive ability of environmental concern is non-significant for the very poor, and for the very wealthy. It is concluded that 2.6 attitudes do influence behaviour, but that this relationship appears to be moderated2.4 , to a limited extent, by social grade. Environmental concern also seems to lose its ability to predict levels of behaviour for members of social 2.2 grade d and social grade a. Social grade e are none workers, therefore, grades 2.0 a andLevel of satisfaction (linear predication) d represents and both ends of the employment spectrum. e d c2 c1 b a Social Grade

In termsPr ofo -Epossiblenvironmen explanations,t it is plausible if not likely, that both poverty Disengaged and affluence condition behaviour through disabling and enabling effects.

Those with severely limited resources are likely to have the least capacity to act, the reasons for which may include reduced access to the necessary

282

physical means, be this through their location, ability to physically reach or engage with infrastructure, or high demands on their time and attention.

Conversely, those with a high level of affluence will almost inevitably have high levels of environmental consumption, afforded by their resources. It would not be surprising if they held pro-environmental attitudes inconsistent with high consumption lifestyles. For example, the wealthy of industrialised countries use air transport several times more than the population average, while the poor may never fly (Hares, Dickinson & Wilkes, 2010).

The key message here is that the environmental attitude-behaviour relationship is moderated, not mediated by social grade (an indicator or SES), but that this moderation is complicated and warrants further research.

283

7.5 Further research

7.5.1 Global environmental concern

This thesis has examined environmental concern and behaviour in one country: a natural extension would be to extend this research to examine multiple countries. Such a study would reveal how UK environmental concern exists in a global setting, by making comparisons between countries. Data from the

Eurobarometer survey or the International Social Survey Programme (ISSP)

(both were discussed in Chapter 3) would be suitable for such analysis.

In 2008, 2011 and recently in 2014, a special Eurobarometer survey was conducted to capture the attitudes of European citizens towards the environment, with a specific focus on the perceived role of the EU in tackling environmental problems. The survey is conducted across 28 European Union

Member States. The sample for each country is representative of the national population 15+. In addition, the ISSP survey is a global survey conducted across 32 countries that contains a rotating module on environmental concern.

The ISSP Environment module consists of three surveys from 1993, 2000 and

2010. The 2010 survey is a partial replication of the 2000 study and the 2000 survey is a partial replication of the 1993 study. The ISSP Environment module focuses on attitudes to the environment, environmental protection, respondents' behaviour and respondents' preferences regarding governmental measures on environmental protection.

284

Data obtained at a global level over multiple time points could be analysed using multilevel or multiple-group repeated cross-sectional analysis. Results would demonstrate how environmental concern and apathy has evolved over time across multiple countries and how other countries compare to the UK. A potential method for this line of enquiry is a new technique for implementing multiple-group factor analysis using Bayesian estimation known as the alignment method. Standard multiple-group factor analysis aims to compare latent variable means, variances, and covariances across groups while holding measurement parameters invariant. For factor means to be comparable, invariance of both factor loadings and measurement intercepts is required and is referred to as scalar invariance (Millsap, 2012). A model with such strict invariance is often rejected. This is typically followed by the use of modification indexes (Sörbom, 1989) to relax some of the invariance restrictions. With many groups, the usual multiple-group CFA approach is too cumbersome to be practical due to the many possible violations of invariance, and the modification index exploration could well lead to the wrong model due to the scalar model being far from the true model. The alignment method can be used to estimate group-specific factor means and variances without requiring exact measurement invariance and thus can conveniently estimate models for many groups. The method also provides a detailed account of parameter invariance for every model parameter in every group.

7.5.2 The three-step method

In Chapter 6, results of the latent class analysis conducted in Chapter 5 were regressed onto 16 measures of environmental behaviour. This latent class 285

regression analysis was done separately, after the LCA. Though this method is adequate if latent classes are stable18, it is both time consuming (having to save and export result from the LCA to then be analysed separately) and susceptible to measurement error. With latent class analysis and other forms of mixture modelling, indicator variables are used to identify an underlying latent categorical variable. Researchers are often interested in using the latent categorical variable for further analysis and exploring the relationship between that variable and other, auxiliary observed variables. The standard way to conduct such an analysis is to combine the latent class model and the latent class regression model into a joint model that can be estimated with the maximum-likelihood estimator, referred to as the one-step method. Such an approach, however, can be flawed because the secondary model could affect the latent class formation and the latent class variable could lose its meaning as the latent variable measured by the indicator variables. Vermunt (2010) pointed out several disadvantages of the one-step method:

“The first is that it may sometimes be impractical, especially when the

number of potential covariates is large, as will typically be the case in a

more exploratory study. Each time that a covariate is added or removed

not only the prediction model but also the measurement model needs to

be re-estimated. A second disadvantage is that it introduces additional

model building problems, such as whether one should decide about the

18 Stability is indicated by high entropy (>.6), which shows that classes are separate, as well as adequate goodness-of-fit indices.

286

number of classes in a model with or without covariates. Third, the

simultaneous approach does not fit with the logic of most applied

researchers, who view introducing covariates as a step that comes after

the classification model has been built. Fourth, it assumes that the

classification model is built in the same stage of a study as the model

used to predict the class membership, which is not necessarily the case. It

can even be that the researcher who constructs the typology using an LC

model is not the same as the one who uses the typology in a next stage of

the study”

(Vermunt, 2010, p. 451)

For these reasons, particularly three and four, the one-step method was not used in this thesis, and instead, latent class regression was used.

Asparouhov and Muthén (2014) have recently developed an alternative approach to studying the relationship between latent classes and auxiliary variables referred to as the three-step method, expanding ideas presented in

Vermunt (2010). In this approach the latent class model is estimated first. In the second step, the most likely class variable S is created using the latent class posterior distribution obtained during the LCA estimation. In the third step, S is used as latent class indicator variable with uncertainty rates prefixed by the class membership probabilities obtained in step two. This way the measurement error in the most likely class S is taken into account in the third step model estimation. In this final stage, auxiliary variable(s) are also included.

287

Initial use of this approach suggests it is as efficient as the one-step approach, while also outperforming it when analysing the relationship between a latent class variable and an auxiliary variable independently of the latent class model estimation.

7.5.3 Newly available data

Not only are there additional and new methods with which to study environmental attitudes and behaviour, there will soon be new data. The

European Social Survey (ESS) is an academically driven cross-national survey that has been conducted every two years across Europe since 2001. The survey measures the attitudes, beliefs and behaviour patterns of diverse populations in more than 30 nations. The ESS was primarily designed as a time series that could monitor changing attitudes and values across Europe. For this reason its questionnaire comprises one core module, containing items measuring a range of topics of enduring interest to the social sciences as well as the most comprehensive set of socio-structural variables of any cross- national survey. In each round of the ESS, multi-national teams of researchers are selected to contribute to the design of two rotating modules for the questionnaire. From ESS Round 8 there will be a module entitled ‘Public

Attitudes to Climate Change, Energy Security, and Energy Preferences’, alongside an additional module on attitudes towards welfare. This module covers four broad areas of (1) beliefs on climate change, (2) concerns about climate change and energy security, (3) personal norms, efficacy and trust, and

(4) energy preferences. The module is specifically designed to fit within the core ESS questionnaire to create a comprehensive dataset that directly 288

contributes to a better understanding of environmental attitudes. Rory

Fitzgerald, Director of the ESS has said: “Climate change is a key issue across

Europe and the world whilst the future role and scope of the welfare state is a challenge in many European countries. The ESS will address these two issues using the high methodological standards in both fieldwork and questionnaire design for which it is well known. We will spend the next two years carefully designing the modules ready for fieldwork in 2016. The first data from these modules is expected to be available towards the end of October

2017”. Analysing this data will help to make robust comparisons of Europeans’ perceptions of climate change, energy security, and energy preferences. This will in turn increase understanding of how environmental concern is shaped by national socio-political factors, examine the role of socio-political values and engagement, and examine the relative importance of individual-motivational versus national-contextual variables in public energy preferences.

289

7.6 From a qualitative perspective

A clear limitation of this study is that analysis is restricted to the realm of quantitative research. Though this research has studied environmental concern and its association with pro-environmental behaviour, it is unknown why the

UK public holds the environmental attitudes found in this research or the reasoning attributed to possessing a particular form of concern.

Further qualitative research could provide such insights, complementing the quantitative findings obtained in this thesis. As noted by Madey (1982), combining quantitative and qualitative research can be advantageous when exploring complex research questions; statistical analysis can provide detailed assessment of patterns of responses, and additional qualitative analysis can provide a deep understanding of survey responses.

7.6.1 Combining qualitative and quantitative research

Rossman and Wilson (1985) describe three major schools of thought on combining qualitative and quantitative perspectives: purists, situationalists and pragmatists. These three camps can be conceptualised as lying on a continuum, with purists and pragmatists lying on opposite ends, and situationalists lying somewhere between purists and pragmatists.

Purists posit that quantitative and qualitative methods stem from different

ontological assumptions about the nature of research (Tashakkori &

Teddlie, 1998). Moreover, for purists, the assumptions associated with

290

both paradigms are incompatible regarding how the world is viewed and

what it is important to know. Purists, such as Smith (1983) and Smith and

Heshusius (1986), contend that quantitative and qualitative approaches

cannot and should not be mixed. As such, they advocate mono-method

studies.

Situationalists maintain the mono-method stance held by purists, while

also arguing that both methods have value. However, they believe that

certain research questions lend themselves more to quantitative

approaches, whereas other research questions are more suitable for

qualitative methods.

Pragmatists, unlike purists and situationalists, contend that a false

dichotomy exists between quantitative and qualitative approaches

(Newman 1998). As such, pragmatists advocate integrating methods within

a single study (Creswell 1999).

It is the perspective of this researcher that the mono-method stance adopted by both purists and situationalists is unnecessarily restrictive. Researchers who subscribe to epistemological purity disregard the fact that research methodologies are tools designed to aid understanding of the social world.

Indeed, ‘epistemological purity doesn’t get research done’ (Miles & Huberman,

1984, p. 21). By having a positive attitude towards both techniques, researchers are in a better position to use qualitative research to inform the quantitative portion of research studies, or to confirm that qualitative findings apply to a wider population by using quantitative research. Alternatively, a

291

committed pragmatist is likely to believe that studies should use a mixture of quantitative and qualitative methods in order to be effective. That said, researchers should not feel the need to combine these approaches. The use of both quantitative and qualitative methods should be optional, not compulsory.

It is perfectly acceptable that a body of work be either entirely qualitative or quantitative (as this thesis is).

Figure 34: Spectrum of perspectives on combining qualitative and quantitative research

Purist Situationalist Pragmatist

This Research

This perspective exists towards the pragmatic end of the spectrum (shown in

Figure 34), representing the rejection of epistemological purity, while also embracing the notion that excellent research does not need to be mixed methods.

7.6.2 Further qualitative research

Further qualitative research is required to understand some of the findings obtained in this thesis. Specifically, more should be done to better understand:

• Climate change denial as a result of an egoistic value orientation.

Could this be self-preservation, cognitive dissonance or selfishness?

292

• The inability of concern to significantly influence behaviour for

specific social grades.

• Ways to engage the Disengaged and the Paradoxical groups.

Interviews should be conducted with those who dismiss, de-value or are in denial of climate change, or the severity of environmental problems. This line of enquiry should aim to find if these people engage with this line of reasoning as a coping mechanism, or to justify inaction. This can be achieved by discussing their perception of the environment and environmental problems. Inquiries regarding behaviour should also be made; for example, do these individuals see pro-environmental behaviour as something that is both accessible and achievable? Do they believe that they are able to make a difference at an individual level?

Do they engage in a particular form environmentally detrimental behaviour that they cannot / will not discontinue? Further to this, to try and understand the odd moderating effect of social grade on the attitude-behaviour relationship, interviews should be conducted on people across all social grades, and inquiries made about any need they feel to engage in environmental action, and how easy and achievable such action is. This will help to determine what the constraints are that prevent environmental attitudes from transforming into actions. Finally, qualitative analysis of apathetic views towards the environment and environmental issues would be extremely valuable. Why are people apathetic to such issues? Possible explanations that should be explored include:

293

Weak locus of control: The individual believes that they are unable to

control or alter the events affecting them.

Environmental issues are removed from space and time: The individual

believes that they are immune to the effects of environmental problem and

will continue to be for the foreseeable future.

Free-rider problem: The individual believes they there is no point in doing

anything environmentally friendly if others continue to harm the

environment.

294

7.7 Limitations of this research

This section discusses the limitations of the research conducted in this thesis, and how this thesis should be extended not only with additional quantitative research, but also qualitative. Further to this, newly available data and methods pertinent to the study of environmental attitudes and behaviour are discussed.

7.7.1 The advantages and disadvantages of cross-sectional

research

This study analyses cross-sectional data to understand environmental concern as it exists across the UK, and how this concern influences levels of pro- environmental behaviour. Cross-sectional data has some valuable attributes. It can provide researchers with information about age group differences or inter- individual differences (Miller 2007). It is also useful for generating and clarifying hypotheses, piloting new measures or technology, and can lay the groundwork for decisions about future follow-up studies (Kraemer 1994). Cross-sectional studies are however subject to some methodological concerns and limitations.

For example, cross-sectional research cannot easily separate the effects of age from the effects of belonging to a particular cohort. In Chapter 5 of this thesis, when examining the social characteristics of environmental classes,

Paradoxical class members were found to be mainly comprised of old respondents and Disengaged class members were primarily young respondents. Because of the cross-sectional nature of the data however, it cannot be said for certain whether this is a generational effect rather than an age effect. Only follow up studies with these specific participants can indicate

295

whether these forms of environmental concern remain with this cohort throughout their life course, or if these concerns are simply associated with the relevant age groups.

The most prominent shortcoming of cross-sectional data is its inability to provide definite information about cause-and-effect relationships. This is because such data only offers a snapshot of a single moment in time, and does not consider what happens before or after this snapshot is taken.

Therefore, cross-sectional data does not permit a causal conclusion. However, analysis in Chapter 6 studies both mediating and moderating effects which are by nature, causal models as the underlying theories suggest directional inferences that are intrinsically causal (Rose et al., 2004). Using these methods in conjunction with cross-sectional data creates what Roe (2012) refers to as

‘temporal illusion’; a term referring to research that suggests the flow of time is present when it is not. Given the potentially valuable information on the attitude-behaviour relationship that examining indirect effects could yield, it is arguably better to examine indirect effects in an imperfect way, than to discount them entirely in this research. Nevertheless, findings from this thesis should be confirmed with longitudinal research.

Aside from the general problems with cross-sectional research, there are addition problems with the specific data used throughout this analysis. Data for the EAS was gathered in March 2009, such that results of the present analysis, which begun three years ago, now relate to data that is five years old. Enquires have been made to DEFRA regarding the possibility of continuing the EAS.

Further environmental attitudes and behaviour studies from DEFRA would

296

open the possibility for repeated cross-sectional data analysis. Unfortunately funding for DEFRA has been severely cut by the current government and as a result there are no plans to continue the survey.

7.7.2 Subjective measures of behaviour

An obvious limitation of this research is that it uses subjective, rather than objective measures of environmental behaviour. The behavioural measures in the EAS are self-reported by respondents, rather than being observed independently, and as such are subject to bias. It is entirely likely that levels of behaviour have been exaggerated by respondents or are simply inaccurate. As discussed in Chapter 3, the use of secondary data means in part inheriting any problems these data may have. Subjective measures of behaviour are a limitation of the EAS data that could not be avoided.

297

Chapter 8 Conclusion

This thesis has provided a thorough inquiry into UK environmental concern and an assessment of its relationship with level of pro-environmental behaviour. As discussed in Chapter 1, research into UK environmental attitudes and behaviour is sorely needed. The mean surface temperature of the earth has been steadily increasing since the industrial revolution to the detriment of the natural environment. Reports generated from the IPCC have been consistent in their findings: climate change is happening and human activity is the cause

(IPCC 2007). Unfortunately the UK is amongst the top 10 global contributors of

climate change, producing more CO2 per capita than China (see data from the

The World Bank n.d.). The public plays a critical role in tackling climate change by reducing their direct consumption of fossil fuels and through their support for political leadership. At a national level, public choices collectively have enormous impacts on the earth. As such, individual-level environmental concern is important and needs to be fully understood. However, relatively little is known about the detailed nature of environmental concern in relation to UK national attitudes and behaviour, specifically regarding climate change. What is known about UK public opinion is that public consensus on climate change is out of alignment with scientific consensus, despite the fact that scientific understanding of climate change is now sufficiently convincing to justify taking prompt action. The opinion gap between scientists and the public in 2009 stood at 84% to 49%, that global temperatures are increasing because of human activity, and the 2011 Angus Reid Public Opinion poll found that 20%

298

of Britons thought “global warming is a theory that has not yet been proven”

(Reid n.d.).

Climate change science is arguably well developed, relatively coherent in terms of theory and method, and capable of measuring, analysing, and assessing what we do and do not know about the environmental consequences of climate change. Levin et al. (2012) argue that by comparison, social scientific research on climate change is more recent, far less coherent, and lacks consensus on either epistemological or substantive grounds. But such social science research is required in order to make effective policies aimed as reducing the UK contribution to this global problem. “Scientists need to know how the public is likely to respond to climate impacts or initiatives, because those responses can attenuate or amplify the impacts” (Bord et al., 1998, p.

75). How the public perceive the natural environment and subsequently behave are thus vital matters.

Given the importance of public co-operation in environmental improvement, the mismatch between scientific and public consensus on climate change, and the lack of research on the subject in the UK, it is clear that further research is needed. This thesis this aimed to acquire an improved and nuanced understanding of public engagement with environmental problems.

To do this, this thesis drew upon DEFRA’s 2009 ‘Survey of Public Attitudes and Behaviours Towards the Environment’, a nationally representative sample of the UK, and accomplished the three following objectives:

299

1. Explore environmental attitudes in the DEFRA sample.

Environmental concern is comprised of multiple attitudes towards the

environment. While past studies have sought to map out the structure of

environmental concern, studies testing such models have yielded largely

inconsistent results. Chapter 4 explored the structure of environmental

concern through the use of multiple forms of factor analysis; results

were discussed in relation to established theories.

2. Identify the types of environmental concern that exist in the UK.

The environmental attitudes that comprise environmental concern co-

exist in varying quantities, producing different forms of environmental

concern. In Chapter 5, respondents of the 2009 DEFRA survey were

grouped according to their form of environmental concern using latent

class analysis. This produced a typology of environmental concern for

the UK. The attitudes of environmental class members were examined

and the social determinants of membership investigated.

3. Learn how different forms of environmental concern are associated

with pro-environmental behaviours.

It has often been assumed that concern for the environment is positively

associated with pro-environmental behaviours, yet empirical findings

remain less forthcoming. Building on past evidence showing socio-

economic status to be associated with both environmental attitudes and

behaviour, Chapter 6 not only thoroughly assessed the direct effects of

300

environmental concern on behaviour, but also examined the mediating

and moderating influences of social grade using path analysis and

regression models incorporating an interaction effect.

8.1.1 Summary of findings

This discussion chapter has reviewed the analytical work in this thesis and provided answers to the research questions proposed in Chapter 2.

What are the components of environmental concern?

This thesis concludes that environmental concern is the culmination of

three environmental attitudes: (a) a cognitive appraisal of plant and animal

welfare (ecocentric attitude); (b) welfare of the human race (human-centric

attitude); and (c) a prioritisation of the self alongside dismissal of

environmental problems (labelled ‘Denial’).

How does environmental concern exist amongst the UK public?

The British public is categorised as belonging to one of four groups defined

by their form of environmental concern.

Pro-environment: Strong pro-environmental attitudes and great

concern for the environment.

Neutral Majority: View the environmental favourably, but are not

overly concerned. This group possess a moderate form of concern.

301

Disengaged: Disengaged with environmental issues and sceptical

of climate change.

Paradoxical: Illogical response pattern and heavily in denial, also

sceptical of the anthropocentric causes of climate change.

What social characteristics are associated with different forms of environmental concern?

Results suggest that pro-environmental attitudes are more common

amongst those who are educated to at least degree level, have a high

social grade and household income. Disengagement with environmental

issues appears to be more common amongst the very young, and

paradoxical views (most likely due to denial) are more common amongst

the very old.

How is concern for the environment associated with pro-environmental behaviour, and what is the role of socio-economic status in this relationship?

An assessment of the direct effects of pro-environmental concern on

behaviour revealed a positive association. This association occurs for pro-

environmental behavioural relating to food, recycling, home energy

consumption and travel. Travel behaviours are often considered to be high-

cost behaviours (see Whitmarsh, 2009 and Diekmann and Preisendöerfer’s

1992), which have typically been unaffected by attitudes. Results in this

thesis however, suggest that attitudes are powerful enough to incite

behaviour change even for such difficult / high-cost behaviours.

302

The public often consider matters of health, wealth and security to be more

important than environmental issues (Bord et al., 2000; Norton & Leaman,

2004; Poortinga & Pidgeon, 2003). As such, though attitudes are likely to

influence behaviour, they lose their predictive power as pro-environmental

behaviours are de-prioritised in favour of actions that directly increase

income, happiness and wellbeing. Consequently, those in good health,

with wealth and security may have a higher probability of engaging in pro-

behaviour. This thesis examined how SES can affect the environmental

concern – behaviour relationship, and found it has only a weak affect.

8.1.2 How these findings relate to the broad debates in the

environmental social sciences

The findings in this thesis contribute to the on-going academic debate on attitude-behaviour relationships in four ways. First, this thesis has argued for the conceptualisation of broad scope attitudes (such as environmental concern) as ontologically distinct from behaviour. Second, the findings in this thesis support the assertion that attitudes are able to positively and significantly influence levels of pro-environmental behaviour, even ‘high-cost’ behaviours (such as reduction in driving and increase in use of public transportation) which have previously been found to be unaffected by attitudes. Third, broad scope environmental attitudes have been found in this thesis to predict both specific and general measures of such behaviour, thus providing evidence against the use of behaviour specific attitudes when investigating the environmental attitude-behaviour relationship. Finally, this

303

research has also gained insight into the structure and dimensions of environmental concern. Through the use of factor analysis and LCA not only have the components of environmental concern been examined, but also how such concern exists amongst the British public. Indeed this is one of the very few studies that has grouped the British public according to their form of concern.

Further quantitative research into the problem of apathy has also been suggested in this thesis, in an appeal to abandon the traditional line of enquiry that focuses on how attitudes predict behaviour. In addition to this recommendation, it has also been suggested that the research in this thesis be conducted in a global setting across multiple time points. This can potentially determine whether apathetic groups exist across Europe, and how these groups change over time. Such research can be done using the new alignment method and / or the three-step method. New data will also soon become available from the ESS that aims to thoroughly capture environmental concern across.

A contribution to knowledge alone however, should not be the only one made when the potential for real-work impact is so great. Social science research is meant to provide a contribution to knowledge regarding the social world in an academic setting, but recently there has been much discussion regarding the extent to which the contribution of such research should be extended further.

“Research impact”, a term which has been commonly used throughout the discussion is defined by the Research Councils UK (RCUK) as the contribution that excellent research makes to society and the economy. This includes

304

“increasing the effectiveness of public services and policy”. The research in this thesis has the potential to provide such a contribution.

8.1.3 Pro-environmental behaviour change as a policy tool

Findings in this thesis suggest that 36% of the British public are not concerned with the environment, and engage in very little pro-environmental behaviour.

This is compounded by the finding that those who do less are more satisfied with this level of behaviour than those who engage in high levels of activity.

This is clearly something that needs to be addressed, not only within the realm of academia through further research, but beyond it through the use of public policy. Such policy could be used to address the low levels of concern and the inability for some attitude to influence behaviour.

Pro-environmental behaviour change is a potentially valuable policy tool that can be used to engage the public with environmental issues, increase rates of pro-environmental behaviour throughout the country, and ultimately reduce UK

CO2 emission. UK public policy and campaigns should aim at reducing levels of apathy, and shifting social norms as a precursor to tackling behaviour change. Results obtain in Chapter 6 suggest that current social norms allow members of the Paradoxical and Disengaged groups to do little for the environment, while at the same time, be satisfied with such inaction. It is currently not normative to frequently engage in many forms of pro- environmental behaviour, and such norms inhibit the behaviour of Pro- environment class members. This results in a form of cognitive dissonance; the

Pro-environment class care greatly about the environment, but are unable to

305

engage in a level of pro-environmental behaviour that they are satisfied with.

Ideally, to promote high levels of such behaviour, norms should shift in a way that leaves the Paradoxical and Disengaged dissatisfied, and allows the Pro- environment group to engage in their desired level of behaviour. A public campaign might, for example, inform the people of the strong evidence for climate change, how climate change will negatively affect the UK, and what mitigating behaviour the public can engage in. However, this tactic of informing the public in an attempt to encourage attitude / behaviour change may likely fall into an information deficit trap. As discussed in Chapter 2, more information does not guarantee behaviour change (indeed, information deficit models lack empirical and theoretical support; see Marteau, Sowden &

Armstrong, 1998). A way of avoiding such a trap is by considering the most effective way to frame climate problems when engaging with the public. How environmental messages are presented has important ramifications for the way that they are perceived. Spence and Pidgeon (2010) found that framing climate change impacts as distant resulted in climate change impacts being perceived as more severe. Additionally, attitudes towards climate change mitigation were found to be more positive when participants were asked to consider social rather than personal aspects of climate change, and when the gains produced through climate change mitigation were discussed (as opposed to the losses of not mitigating climate change). Further to this, an IPCC report (Field & Van

Aalst, 2014) proposes that climate change be framed as a food issue in an attempt to mobilise people; “The public connects with these issues through food better than through any other issue in a way that we haven't been able to

306

mobilise people by just telling them to drive a hybrid or switch the light off”

(Goldenberg n.d., p. 1).

A potentially effective means of encouraging behaviour change is to have climate change taught in schools. Here, children should be made aware of the findings obtained by the IPCC and DEFRA, and the curricula periodically updated to ensure what is being taught is both accurate and up to date. If the high proportion of young people making up the Disengaged class is an age effect rather than a cohort effect, then teaching children in school about the importance of the natural environmental and about the reality / evidence of climate change will reduce this proportion. This strategy is also likely to encourage younger people to become more engaged in environmental issues, and additionally may also help to reduce levels of apathy across the UK.

Furthermore, the inclusion of climate change in the education curricula is the first step towards allowing frequent engagement of pro-environmental behaviour to be seen as normative behaviour. This is because widespread inclusion of climate change studies in schools sends a message to the general public that action against climate change should involve everyone and is encouraged.

8.1.4 Final message

This thesis examines public perceptions towards a major, modern day problem and has gained information regarding such perceptions valuable to the field of environmental social sciences and for the development of environmental policy. The results are clear: attitudes matter.

307

Social research on human interaction with the natural environment has largely endeavoured to answer the question ‘do environmental attitudes predict for environmental behaviour?’ The results of this study, strongly suggest that the answer to this question is yes, they do. But in answering this question, this thesis has uncovered the environmentally detrimental problem of apathy. It is not that the environmental attitudes of individuals do not influence their behaviour, it is that a large proportion of the population, do not hold environmental attitudes strong enough to do so.

Now is the time to change the line of inquiry. Environmental social science research must move away from the traditional ‘do attitudes predict behaviour’ approach, and instead focus on the problem of apathy. Why are people apathetic? How can they be engaged with? What increases the probability of these people transitioning out of an apathetic group? These are all important questions that should be answered by future research and considered by policy makers. Through such ‘apathy focused’ research and policy implementation, the UK may begin to more successfully mitigate the effects of climate change, building a stronger public consensus for action.

308

Bibliography

Abdalla, C. W., Roach, B. A., & Epp, D. J. (1992). Valuing environmental quality changes using averting expenditures: An application to groundwater contamination. Land Economics, 163-169.

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and

Human Decision Processes, 50(2), 179 - 211. doi:10.1016/0749-

5978(91)90020-T

Ajzen, I. (2005). Attitudes, personality, and behavior. McGraw-Hill International.

Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and review of empirical research. Psychological Bulletin, 84(5), 888.

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social.

Behaviour. Englewood Cliffs, NJ: Prentice-Hall.

Ajzen, I., & Fishbein, M. (2005). Understanding attitudes and predicting social behaviour. Retrieved from Common.

Albarracín, D., Zanna, M. P., Johnson, B. T., & Kumkale, G. T. (2005). Attitudes:

Introduction and scope.

Alibeli, M., & White, N. (2011). The structure of environmental concern.

International Journal of Business and Social Science, 2(4), 1-8.

309

Allport, F. H. (1924). Response to social stimulation in the group. Social

Psychology, 260-291.

Allport, G. W. (1954). The historical background of modern social psychology.

Handbook of Social Psychology, 1, 3-56.

Antonak, R. F., & Livneh, H. (1988). The measurement of attitudes toward people with disabilities: Methods, psychometrics and scales. Charles C

Thomas, Publisher.

Arber, S., Dale, A., & Proctor, M. (1988). Secondary data analysis. London:

UnwinHyman.

Arcury, T. A., & Christianson, E. H. (1990). Environmental worldview in response to environmental problems kentucky 1984 and 1988 compared.

Environment and Behavior, 22(3), 387-407.

Arcury, T. A., & Christianson, E. H. (1993). Rural-urban differences in environmental knowledge and actions. The Journal of Environmental

Education, 25(1), 19-25.

Arcury, T. A., Scollay, S. J., & Johnson, T. P. (1987). Sex differences in environmental concern and knowledge: The case of acid rain. Sex Roles, 16(9-

10), 463-472.

310

Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta-analytic review. British Journal of Social Psychology, 40(4),

471-499. doi:10.1348/014466601164939

Asparouhov, T., & Muthén, B. (2014). Multiple-group factor analysis alignment.

Structural Equation Modeling: A Multidisciplinary Journal, 21(4), 495-508.

Attfield, R. (1981). The good of trees. The Journal of Value Inquiry, 15(1), 35-54.

Bagozzi, R. P. (1978). The construct validity of the affective, behavioral, and cognitive components of attitude by analysis of covariance structures.

Multivariate Behavioral Research, 13(1), 9-31.

Bagozzi, R. P., & Burnkrant, R. E. (1979). Attitude organization and the attitude-

-behavior relationship. Journal of Personality and Social Psychology, 37(6),

913.

Ball-Rokeach, S. J., & Loges, W. E. (1994). Choosing equality: The correspondence between attitudes about race and the value of equality.

Journal of Social Issues, 50(4), 9-18.

Bamberg, S. (2003). How does environmental concern influence specific environmentally related behaviors? A new answer to an old question. Journal of Environmental Psychology, 23(1), 21-32.

Banister, D. (2002). Transport planning. Taylor & Francis.

311

Banister, D., Watson, S., & Wood, C. (1997). Sustainable cities: Transport, energy, and urban form. Environment and Planning B, 24, 125-144.

Barkan, S. E. (2004). Explaining public support for the environmental movement: A civic voluntarism model*. Social Science Quarterly, 85(4), 913-

937.

Baron, R. M., & Kenny, D. A. (1986). The moderator--mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6),

1173.

Barr, S. (2003). Strategies for sustainability: Citizens and responsible environmental behaviour. Area, 35(3), 227-240

Barr, S. (2008). Environment and society. Sustainability, Policy and the Citizen.

Ashgate, London

Barr, S., Gilg, A., & Shaw, G. (2006). Promoting sustainable lifestyles: A social marketing approach. London: University of Exeter and Defra (www. Defra. Gov.

Uk/science/project_data/DocumentLibrary/SD14005/SD14005_3524_frp. Doc,

Accessed December 8, 2008).

BBC. (2004). Poll for Climate Change Special. Retrieved from

/http://news.bbc.co.uk/1/hi/ sci/tech/3934363.stm and http://news.bbc.co.uk/nol/shared/bsp/hi/ pdfs/28_07_04_climatepoll.pdfS

312

Bechtel, R. B., Corral-Verdugo, V., Asai, M., & Riesle, A. G. (2006). A cross- cultural study of environmental belief structures in USA, japan, mexico, and peru. International Journal of Psychology, 41(02), 145-151.

Becker, B. W., & Connor, P. E. (1981). Personal values of the heavy user of mass media. Journal of Advertising Research.

Bentler, P. M. (1980). Multivariate analysis with latent variables: Causal modeling. Annual Review of Psychology, 31(1), 419-456.

Best, H., & Kneip, T. (2011). The impact of attitudes and behavioral costs on environmental behavior: A natural experiment on household waste recycling.

Social Science Research, 40(3), 917-930.

Bird, C. (1940). Social psychology . New York: Appleton-Century.

Black, C., Collins, A., & Snell, M. (2001). Encouraging walking: The case of journey-to-school trips in compact urban areas. Urban Studies, 38(7), 1121-

1141.

Black, J. S., Stern, P. C., & Elworth, J. T. (1985). Personal and contextual influences on househould energy adaptations. Journal of Applied Psychology,

70(1), 3.

Blake, J. (1999). Overcoming the value-action gapin environmental policy:

Tensions between national policy and local experience. Local Environment,

4(3), 257-278. 313

Blocker, T. J., & Eckberg, D. L. (1997). Gender and environmentalism: Results from the 1993 general social survey. Social Science Quarterly, 841-858.

Blumer, H. (1955). Attitudes and the social act. Social Problems, 3(2), 59-65.

Bord, R. J., Fisher, A., & O'Connor, R. E. (1998). Public perceptions of global warming: United states and international perspectives. Climate Research,

11(1), 75-84.

Bord, R. J., & O'Connor, R. E. (1997). The gender gap in environmental attitudes: The case of perceived vulnerability to risk. Social Science Quarterly,

830-840.

Bord, R. J., O'Connor, R. E., & Fisher, A. (2000). In what sense does the public need to understand global climate change? Public Understanding of Science,

9(3), 205-218.

Braithwaite, V. (1998). The value balance model of political evaluations. British

Journal of Psychology, 89(2), 223-247.

Breckler, S. J. (1984). Empirical validation of affect, behavior, and cognition as distinct components of attitude. Journal of Personality and Social Psychology,

47(6), 1191.

Bryman, A., & Hardy, M. (2004). Handbook of data analysis. Sage.

314

Burgess, J., Harrison, C. M., & Filius, P. (1998). Environmental communication and the cultural politics of environmental citizenship.

Burt, C. L. (1940). The factors of the mind. An Introduction to Factor Analysis in

Psychology. London: University of London Press.

Buttel, F. H., & Flinn, W. L. (1974). The structure of support for the environmental movement, 1968-1970. Rural Sociology.

Buttel, F. H., & Flinn, W. L. (1978). Social class and mass environmental beliefs

A reconsideration. Environment and Behavior, 10(3), 433-450.

Campbell, D. T. (1963). Social attitudes and other acquired behavioral dispositions.

Cattell, R. B. (1966). The scree test for the number of factors. Multivariate

Behavioral Research, 1(2), 245-276.

Chau, H., & Hocevar, D. (1995). The effects of number of measured variables on goodness-of-fit in confirmatory factor analysis.

Christie, I., & Jarvis, L. (2002). How green are our values? In: J.A. Park, J.

Curtice, K. Thomson, L. Jarvis, and C. Bromley, eds. British social attitudes:

The 18th report. London: Sage, 113-152.

315

Clark, C. F., Kotchen, M. J., & Moore, M. R. (2003). Internal and external influences on pro-environmental behavior: Participation in a green electricity program. Journal of Environmental Psychology, 23(3), 237-246.

Cluck, R. (1998). The multi-dimensional basis of environmental attitudes.

Thesis.

Cnossen, C. (1997). Secondary research: Learning paper 7. School of Public

Administration and Law, the Robert Gordon University.

Conway, J. M., & Huffcutt, A. I. (2003). A review and evaluation of exploratory factor analysis practices in organizational research. Organizational Research

Methods, 6(2), 147-168.

Cooper, C. L., & Pervin, L. A. (1998). Personality : Critical concepts in psychology. London ; New York: Routledge. Retrieved from Google Books.

Corey, S. M. (1937). Professed attitudes and actual behavior. Journal of

Educational Psychology, 28(4), 271.

Corral-Verdugo, V. (1997). Dual 'realities' of conservation behavior: Self-reports vs observations of re-use and recycling behavior. Journal of Environmental

Psychology, 17(2), 135-145.

316

Cottrell, S. P. (2003). Influence of sociodemographics and environmental attitudes on general responsible environmental behavior among recreational boaters. Environment and Behavior, 35(3), 347-375.

Creswell, J. W. (1999). Mixed-method research: Introduction and application.

Handbook of Educational Policy, 455-472.

Cronbach, L. J. (1950). Further evidence on response sets and test design.

Educational and Psychological Measurement.

Dalton, R. J., Gontmacher, Y., Lovrich, N. P., & Pierce, J. C. (1999).

Environmental attitudes and the new environmental paradigm. Critical Masses:

Citizens, Nuclear Weapons Production, and Environmental Destruction in the

United States and Russia, 195-230.

Danziger, K. (1997). Naming the mind: How psychology found its language.

Sage.

Davidson, D. J., & Freudenburg, W. R. (1996). Gender and environmental risk concerns a review and analysis of available research. Environment and

Behavior, 28(3), 302-339.

Defra. (2002). Survey of public attitudes to quality of life and to the environment: 2001. Department for Environment, Food and Rural Affairs

London.

317

Defra. (2008). Framework for pro-environmental behaviours. Department for

Environment, Food and Rural Affairs, London.

Deutscher, I. (1965). Words and deeds: Social science and social policy. Soc.

Probs., 13, 235.

Diekmann, A., & Preisendörfer, P. (1992). Persónliches umweltverhalten:

Diskrepanzen zwischen anspruch und wirklichkeit. Koelner Zeitschrift Fuer

Soziologie Und Sozialpsychologie.

Diekmann, A., & Preisendörfer, P. (1998). Environmental behavior discrepancies between aspirations and reality. Rationality and Society, 10(1),

79-102.

Dietz, T., Stern, P. C., & Guagnano, G. A. (1998). Social structural and social psychological bases of environmental concern. Environment and Behavior,

30(4), 450-471.

Dillehay, R. C. (1973). On the irrelevance of the classical negative evidence concerning the effect of attitudes on behavior.

Dolan, P., Hallsworth, M., Halpern, D., King, D., Metcalfe, R. & Vlaev, I. (2010).

Mindspace: Influencing behaviour through public policy. Cabinet Office and

The Institute for Government.

Donthu, N., & Cherian, J. (1992). Hispanic coupon usage: The impact of strong and weak ethnic identification. Psychology & Marketing, 9(6), 501-510. 318

Duhem, P. (1906). La théorie physique: Son objet et sa structure. Paris:

Chevalier and Rivière.

Dunlap, R. E., & Dillman, D. A. (1976). Decline in public support for environmental protection evidence from a 1970-1974 panel study. Rural

Sociology, 41.

Dunlap, R. E., & Gale, R. P. (1972). Politics and ecology: A political profile of student eco-activists. Youth and Society, 3(4), 379-397.

Dunlap, R. E., & Heffernan, R. B. (1975). Outdoor recreation and environmental concern: An empirical examination. Rural Sociology.

Dunlap, R., & Jones, J. (2002). Environmental concern: Conceptual and measurement issues. In D. Dunlap & M. Michelson (Eds.), Handbook of environmental sociology (pp. 482 - 542). London: Greenwood Press.

Dunlap, R. E., & Jones, R. E. (2003). Environmental attitudes and values.

Encyclopedia of Psychological Assessment, 1, 364-369.

Dunlap, R. E., & Michelson, W. (2002). Handbook of environmental sociology.

Greenwood Publishing Group.

Dunlap, R. E., & Van Liere, K. D. (1978). The new environmental paradigm. The

Journal of Environmental Education, 9(4), 10-19.

319

Dunlap, R. E., Van Liere, K. D., Mertig, A. G., & Jones, R. E. (2000). New trends in measuring environmental attitudes: Measuring endorsement of the new ecological paradigm: A revised NEP scale. Journal of Social Issues, 56(3), 425-

442.

Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Harcourt Brace

Jovanovich College Publishers.

Eagly, A. H., & Kulesa, P. (1997). Attitudes, attitude structure, and resistance to change. Environment, Ethics, and Behavior: The Psychology of Environmental

Valuation and Degradation, 122-53.

Eckes, T., & Six, B. (1994). FACT AND FICTION IN ATTITUDE-BEHAVIOR

RESEARCH-A METAANALYSIS. ZEITSCHRIFT FUR SOZIALPSYCHOLOGIE,

25(4), 253-271.

Energy Savings Trust. (2004). ‘‘The day after tomorrow’’ is happening today— but 9 out of 10 people in UK don’t believe it!. Retrieved from http://www.est.org.uk/myhome/newscentre/ pressreleasesarchive/index.cfm?mode=view&press_id=116S

Environmental Transport Association. (2015). Air travel’s impact on climate change. Retrieved from https://www.eta.co.uk/environmental-info/air-travels- impact-on-climate-change

Exley, S., & Christie, I. (2002). Off the buses.

320

Fabrigar, L. R., MacDonald, T. K., & Wegener, D. T. (2005). The structure of attitudes. The Handbook of Attitudes, 80.

Fabrigar, L. R., & Wegener, D. T. (2011). Exploratory factor analysis. Oxford

University Press.

Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999).

Evaluating the use of exploratory factor analysis in psychological research.

Psychological Methods, 4(3.272-299), 272-299.

Festinger, L. (1964). Conflict, decision, and dissonance (Vol. 3). Stanford

University Press.

Field, C., & Van Aalst, M. (2014). Climate change 2014: Impacts, adaptation, and vulnerability (Vol. 1). IPCC.

Finch, W. H., & Bronk, K. C. (2011). Conducting confirmatory latent class analysis using M plus. Structural Equation Modeling, 18(1), 132-151. doi:10.1080/10705511.2011.532732

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research.

Fishbein, M., & Ajzen, I. (1981). Acceptance yielding and impact: Cognitive processes in persuasion.

321

Fisher, A., McClelland, G. H., Schulze, W. D., & Doyle, J. K. (1991).

Communicating the risk from radon. Journal of the Air & Waste Management

Association, 41(11), 1440-1445.

Fisher, R. A. (1922). The goodness of fit of regression formulae, and the distribution of regression coefficients. Journal of the Royal Statistical Society,

597-612.

Fisher, R. A. (1935). The design of experiments.

Fong, T. C. T., & Ho, R. T. H. (2013). Factor analyses of the hospital anxiety and depression scale: A bayesian structural equation modeling approach.

Quality of Life Research, 22(10), 2857-2863.

Fortner, R. W., Lee, J. -Y., Corney, J. R., Romanello, S., Bonnell, J., Luthy, B., .

. . Ntsiko, N. (2000). Public understanding of climate change: Certainty and willingness to act. Environmental Education Research, 6(2), 127-141.

Francis, B. (2006). What is latent class analysis? Retrieved from http://www.ccsr.ac.uk/methods/festival/programme/wiwp/francis.pdf

Franzen, A., & Meyer, R. (2010). Environmental attitudes in cross-national perspective: A multilevel analysis of the ISSP 1993 and 2000. European

Sociological Review, 26(2), 219-234.

Fuhrer, U. (1995). Sozialpsychologisch fundierter theorierahmen fć1/4r eine umweltbewuć tseinsforschung. Psychologische Rundschau. 322

Gagnon Thompson, S. C., & Barton, M. A. (1994). Ecocentric and anthropocentric attitudes toward the environment. Journal of Environmental

Psychology, 14(2), 149-157.

Gardner, G. T., & Stern, P. C. (1996). Environmental problems and human behavior. Allyn & Bacon.

Garson, G. D. (1998). Neural networks: An introductory guide for social scientists. SAGE Publications Limited.

Gatersleben, B., Steg, L., & Vlek, C. (2002). Measurement and determinants of environmentally significant consumer behavior. Environment and Behavior,

34(3), 335-362.

Gelissen, J. (2007). Explaining popular support for environmental protection A multilevel analysis of 50 nations. Environment and Behavior, 39(3), 392-415.

Gelman, A., & Meng, X. -L. (1996). Model checking and model improvement. In

Markov chain monte carlo in practice (pp. 189-201). Springer.

Gelman, A., Meng, X. -L., & Stern, H. (1996). Posterior predictive assessment of model fitness via realized discrepancies. Statistica Sinica, 6(4), 733-760.

Genge, E. (2014). A latent class analysis of the public attitude towards the euro adoption in poland. Advances in Data Analysis and Classification, 8(4), 427-

442.

323

Givens, J. E., & Jorgenson, A. K. (2011). The effects of affluence, economic development, and environmental degradation on environmental concern: A multilevel analysis. Organization & Environment, 24(1), 74-91.

Global Food Security. (n.d.). UK threat. [Web page] Retrieved from http://www.foodsecurity.ac.uk/issue/uk.html

Golay, P., Reverte, I., Rossier, J., Favez, N., & Lecerf, T. (2013). Further insights on the french WISC--IV factor structure through bayesian structural equation modeling. Psychological Assessment, 25(2), 496.

Goldenberg, S. (n.d.). Frame climate change as a food issue, experts say. The

Guardian. Retrieved from http://www.theguardian.com/environment/2014/apr/01/climate-change-food- issue-ipcc-report

Goodmonson, C., & Glaudin, V. (1971). The relationship of commitment-free behavior and commitment behavior: A study of attitude toward organ transplantation. Journal of Social Issues, 27(4), 171-183.

Gorsuch, R. L. (1990). Common factor analysis versus component analysis:

Some well and little known facts. Multivariate Behavioral Research, 25(1), 33-

39.

Greeley, A. (1993). Religion and attitudes toward the environment. Journal for the Scientific Study of Religion, 19-28.

324

Grob, A. (1995). A structural model of environmental attitudes and behaviour.

Journal of Environmental Psychology, 15(3), 209-220.

Grundy, C., Steinbach, R., Edwards, P., Green, J., Armstrong, B., & Wilkinson,

P. (2009). Effect of 20 mph traffic speed zones on road injuries in london,

1986-2006: Controlled interrupted time series analysis. Bmj, 339.

Guttman, I. (1967). The use of the concept of a future observation in goodness- of-fit problems. Journal of the Royal Statistical Society. Series B

(Methodological), 83-100.

Haapio, A. (2012). Towards sustainable urban communities. Environmental

Impact Assessment Review, 32(1), 165-169.

Hakim, C. (1998). Developing a sociology for the twenty-first century:

Preference theory. British Journal of Sociology, 137-143.

Hakim, C. (2000). Research design: Successful designs for social and economic research. Psychology Press.

Hansla, A., Gamble, A., Juliusson, A., & Gärling, T. (2008). Psychological determinants of attitude towards and willingness to pay for green electricity.

Energy Policy, 36(2), 768-774.

Hares, A., Dickinson, J., & Wilkes, K. (2010). Climate change and the air travel decisions of UK tourists. Journal of Transport Geography, 18(3), 466-473.

325

Hausman, C. R. (1997). Charles S. Peirce's evolutionary philosophy. Cambridge

University Press.

Hayes, B. C. (2001). Gender, scientific knowledge, and attitudes toward the environment: A cross-national analysis. Political Research Quarterly, 54(3),

657-671.

Hayton, J. C., Allen, D. G., & Scarpello, V. (2004). Factor retention decisions in exploratory factor analysis: A tutorial on parallel analysis. Organizational

Research Methods, 7(2), 191-205.

Heinen, T. (1996). Latent class and discrete latent trait models: Similarities and differences. Sage Publications, Inc.

Hempel, C. (1965). Aspects of scientific explanation and other essays in the philosophy of science.

Hillman, M. (1998). Why climate change must top the agenda. TOWN AND

COUNTRY PLANNING-LONDON-TOWN AND COUNTRY PLANNING

ASSOCIATION-, 67, 287-288.

Hines, J. M., Hungerford, H. R., & Tomera, A. N. (1987). Analysis and synthesis of research on responsible environmental behavior: A meta-analysis. The

Journal of Environmental Education, 18(2), 1-8.

Howell, S. E., & Laska, S. B. (1992). The changing face of the environmental coalition a research note. Environment and Behavior, 24(1), 134-144. 326

Hu, L. -T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural

Equation Modeling: A Multidisciplinary Journal, 6(1), 1-55.

Inglehart, R. (1977). The silent revolution (Vol. 8). Princeton: Princeton

University Press.

Inglehart, R. (1990). Culture shift in advanced industrial society. Princeton

University Press.

Inglehart, R. (1995). Public support for environmental protection: Objective problems and subjective values in 43 societies. PS: Political Science & Politics,

28(01), 57-72.

IPCC, A. R. (2007). Climate change 2007: Synthesis report. IPCC Secretariat

Geneva.

Jackson, T. (2005). Motivating sustainable consumption. Sustainable

Development Research Network, 29, 30

Jackson, S. E., Ones, D. S., & Dilchert, S. (2012). Managing human resources for environmental sustainability (Vol. 32). John Wiley & Sons.

Johnson, F. R., & Luken, R. A. (1987). Radon risk information and voluntary protection: Evidence from a natural experiment. Risk Analysis, 7(1), 97-107.

327

Joireman, J. A., Van Lange, P. A., & Van Vugt, M. (2004). Who cares about the environmental impact of cars? Those with an eye toward the future.

Environment and Behavior, 36(2), 187-206.

Jöreskog, K. G., & Sörbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS command language. Scientific Software International.

Jorgensen, B. S., & Stedman, R. C. (2001). Sense of place as an attitude:

Lakeshore owners attitudes toward their properties. Journal of Environmental

Psychology, 21(3), 233-248.

Kaiser, H. F. (1960). The application of electronic computers to factor analysis.

Educational and Psychological Measurement.

Kaiser, F. G., Ranney, M., Hartig, T., & Bowler, P. A. (1999). Ecological behavior, environmental attitude, and feelings of responsibility for the environment. European Psychologist, 4(2), 59-74.

Kals, E., Schumacher, D., & Montada, L. (1999). Emotional affinity toward nature as a motivational basis to protect nature. Environment and Behavior,

31(2), 178-202.

Kanagy, C. L., Humphrey, C. R., & Firebaugh, G. (1994). Surging environmentalism: Changing public opinion or changing publics? Social

Science Quarterly.

328

Kelley, S., & Mirer, T. W. (1974). The simple act of voting. American Political

Science Review, 68(02), 572-591.

Kelly, George. The psychology of personal constructs. 2 vols. New York:

Norton, 1955.

Kelman, H. C. (1974). Attitudes are alive and well and gainfully employed in the sphere of action. American Psychologist, 29(5), 310. doi:10.1037/h0037623

Klineberg, S. L., McKeever, M., & Rothenbach, B. (1998). Demographic predictors of environmental concern: It does make a difference how it's measured: Research on the environment. Social Science Quarterly, 79(4), 734-

753.

Kluckhohn, C. (1951). The study of culture. Cambridge, MA: Harvard University

Press.

Knott, M., & Bartholomew, D. J. (1999). Latent variable models and factor analysis.

Kohut, A. (2009). Public praises science; scientists fault public, media; scientific achievements less prominent than a decade ago. Pew Research

Center for the People and the Press, Washington, DC: Http://www. People- press. Org/2009/07/09/public-praises-science-scientists-fault-public-media.

329

Kollmuss, A., & Agyeman, J. (2002). Mind the gap: Why do people act environmentally and what are the barriers to pro-environmental behavior?

Environmental Education Research, 8(3), 239-260.

Kothandapani, V. (1971). Validation of feeling, belief, and intention to act as three components of attitude and their contribution to prediction of contraceptive behavior. Journal of Personality and Social Psychology, 19(3),

321.

Kraemer, H. C. (1994). Special methodological problems of childhood developmental follow-up studies: Focus on planning. Developmental Follow-up

Concepts, Domains, and Methods, Academic Press, New York, 260-276.

Kraft, P., Rise, J., Sutton, S., & Røysamb, E. (2005). Perceived difficulty in the theory of planned behaviour: Perceived behavioural control or affective attitude? British Journal of Social Psychology, 44(3), 479-496.

Laidley, T. M. (2013). The influence of social class and cultural variables on environmental behaviors municipal-level evidence from massachusetts.

Environment and Behavior, 45(2), 170-19

LaPiere, R. T. (1934). Attitudes vs. Actions. Social Forces, 13(2), 230-237.

Leiserowitz, A. A. (2005). American risk perceptions: Is climate change dangerous? Risk Analysis, 25(6), 1433-1442.

330

Leiserowitz, A. (2006). Climate change risk perception and policy preferences:

The role of affect, imagery, and values. Climatic Change, 77(1-2), 45-72

Leonard, T. C. (2008). Richard H. Thaler, cass R. Sunstein, nudge: Improving decisions about health, wealth, and happiness. Constitutional Political

Economy, 19(4), 356-360.

Levin, K., Cashore, B., Bernstein, S., & Auld, G. (2012). Overcoming the tragedy of super wicked problems: Constraining our future selves to ameliorate global climate change. Policy Sciences, 45(2), 123-152.

Littig, B. (2000). Eco-efficient services for private households: Looking at the consumer's side. Summer Academy on Technology Studies.

Lleras, C. (2005). Path analysis. Encyclopedia of Social Measurement, 3, 25-30.

Lorenzoni, I., Nicholson-Cole, S., & Whitmarsh, L. (2007). Barriers perceived to engaging with climate change among the UK public and their policy implications. Global Environmental Change, 17(3), 445-459. doi:10.1016/j.jenvp.2006.09.003

Lorenzoni, I., & Pidgeon, N. F. (2006). Public views on climate change:

European and USA perspectives. Climatic Change, 77(1-2), 73-95.

Lynn, J., Straube, B. M., Bell, K. M., Jencks, S. F., & Kambic, R. T. (2007).

Using population segmentation to provide better health care for all: The bridges to health model. Milbank Quarterly, 85(2), 185-208. 331

MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and determination of sample size for covariance structure modeling.

Psychological Methods, 1(2), 130-149.

MacCallum, R. C., Roznowski, M., & Necowitz, L. B. (1992). Model modifications in covariance structure analysis: The problem of capitalization on chance. Psychological Bulletin, 111(3), 490.

MacKinnon, D. P. (2008). Introduction to statistical mediation analysis.

Routledge. Englewood, NJ: Erlbaum.

Madey, D. L. (1982). Some benefits of integrating qualitative and quantitative methods in program evaluation, with illustrations. Educational Evaluation and

Policy Analysis, 223-236.

Maibach, E. W., Leiserowitz, A., Roser-Renouf, C., & Mertz, C. K. (2011).

Identifying like-minded audiences for global warming public engagement campaigns: An audience segmentation analysis and tool development. PloS

One, 6(3), e17571.

Maloney, M. P., & Ward, M. P. (1973). Ecology: Let's hear from the people: An objective scale for the measurement of ecological attitudes and knowledge.

American Psychologist, 28(7), 583.

332

Maloney, M. P., Ward, M. P., & Braucht, G. N. (1975). A revised scale for the measurement of ecological attitudes and knowledge. American Psychologist,

30(7), 787.

Marquart-Pyatt, S. T. (2008). Are there similar sources of environmental concern? Comparing industrialized countries*. Social Science Quarterly, 89(5),

1312-1335.

Marsh, H. W., Muthén, B., Asparouhov, T., Lüdtke, O., Robitzsch, A., Morin, A.

J., & Trautwein, U. (2009). Exploratory structural equation modeling, integrating

CFA and EFA: Application to students' evaluations of university teaching.

Structural Equation Modeling: A Multidisciplinary Journal, 16(3), 439-476.

Marteau, T. M., Sowden, A., & Armstrong, D. (1998). Implementing research findings into practice: Beyond the information deficit model. Getting Research

Findings Into Practice, 36-42.

Martin, M., Williams, I. D., & Clark, M. (2006). Social, cultural and structural influences on household waste recycling: A case study. Resources,

Conservation and Recycling, 48(4), 357-395

Maslow, A. H., Frager, R., Fadiman, J., McReynolds, C., & Cox, R. (1970).

Motivation and personality (Vol. 2). Harper & Row New York. doi:coping

333

Melis, G., Elliot, M., & Shryane, N. (2014). Environmental concern over time:

Evidence from the longitudinal analysis of a british cohort study from 1991 to

2008*. Social Science Quarterly, 95(4), 905-919.

Meng, X. -L. (1994). Posterior predictive p-values. The Annals of Statistics,

1142-1160.

Merchant, C. (1992). Radical ecology. New York: Routledge.

Miles, M. B., & Huberman, A. M. (1984). Drawing valid meaning from qualitative data: Toward a shared craft. Educational Researcher, 20-30.

Milfont, T. L., & Duckitt, J. (2010). The environmental attitudes inventory: A valid and reliable measure to assess the structure of environmental attitudes.

Journal of Environmental Psychology, 30(1), 80-94. doi:10.1016/j.jenvp.2009.09.001

Milfont, T. L., & Gouveia, V. V. (2006). Time perspective and values: An exploratory study of their relations to environmental attitudes. Journal of

Environmental Psychology, 26(1), 72-82.

Miller, S. A. (2007). Developmental research methods. (3rd ed.). Thousand

Oaks, CA: Sage.

Millsap, R. E. (2012). Statistical approaches to measurement invariance. New

York: Routledge.

334

Minton, A. P., & Rose, R. L. (1997). The effects of environmental concern on environmentally friendly consumer behavior: An exploratory study. Journal of

Business Research, 40(1), 37-48.

Monroe, M. C. (2003). Two avenues for encouraging conservation behaviors.

Human Ecology Review, 10(2), 113-125.

Muthén, B. (2010). Bayesian analysis in mplus: A brief introduction.

Unpublished Manuscript. Www. Statmodel. Com/download/IntroBayesVersion,

203.

Muthén, B., & Asparouhov, T. (2011). Bayesian SEM: A more flexible representation of substantive theory. Psychological Methods.

Muthén, B., & Asparouhov, T. (2012a). Bayesian structural equation modeling:

A more flexible representation of substantive theory. Psychological Methods,

17(3), 313-35. doi:10.1037/a0026802

Muthén, B., & Asparouhov, T. (2012b). Bayesian structural equation modeling:

A more flexible representation of substantive theory. Psychological Methods,

17(3), 313.

Naess, A. (1984). A defence of the deep ecology movement. Environmental

Ethics, 6(3), 265-270.

335

Newhouse, N. (1990). Implications of attitude and behavior research for environmental conservation. The Journal of Environmental Education, 22(1),

26-32.

Newman, I. (1998). Qualitative-quantitative research methodology: Exploring the interactive continuum. Carbondale, IL: SIU Press.

Neyman, J., & Pearson, E. S. (1928). On the use and interpretation of certain test criteria for purposes of statistical inference: Part I. Biometrika, 175-240.

Nord, M., Luloff, A. E., & Bridger, J. C. (1998). The association of forest recreation with environmentalism. Environment and Behavior, 30(2), 235-246.

Norton, A., & Leaman, J. (2004). The day after tomorrow: Public opinion on climate change. MORI Social Research Institute, London.

Oberheim, E. and P. Hoyningen-Huene. (2009) "The Incommensurability of

Scientific Theories". E. N. Zalta (ed.): The Stanford Encyclopedia of

Philosophy (Spring 2013 Edition)

O'Connor, R. E., Bord, R. J., Yarnal, B., & Wiefek, N. (2002). Who wants to reduce greenhouse gas emissions? Social Science Quarterly, 83(1), 1-17.

O'Hagan, A., Forster, J., & Kendall, M. G. (2004). Bayesian inference. Arnold

London.

336

Olli, E., Grendstad, G., & Wollebaek, D. (2001). Correlates of environmental behaviors bringing back social context. Environment and Behavior, 33(2), 181-

208

Olofsson, A., & Öhman, S. (2006). General beliefs and environmental concern transatlantic comparisons. Environment and Behavior, 38(6), 768-790. doi:10.1177/0013916506287388

Olson, J. M., & Zanna, M. P. (1993). Attitudes and attitude change. Annual

Review of Psychology, 44(1), 117-154.

Oskamp, S., Harrington, M. J., Edwards, T. C., Sherwood, D. L., Okuda, S. M.,

& Swanson, D. C. (1991). Factors influencing household recycling behavior.

Environment and Behavior, 23(4), 494-519

Oskamp, S. (2000). A sustainable future for humanity? How can psychology help? American Psychologist, 55(5), 496.

Ostrom, T. M. (1969). The relationship between the affective, behavioral, and cognitive components of attitude. Journal of Experimental Social Psychology,

5(1), 12-30.

Pampel, F. C. (2014). The varied influence of SES on environmental concern.

Social Science Quarterly, 95(1), 57-75.

337

Patil, V. H., Singh, S. N., Mishra, S., & Donavan, D. T. (2008). Efficient theory development and factor retention criteria: Abandon the eigenvalue greater than onecriterion. Journal of Business Research, 61(2), 162-170.

Peirce, C. S. (1974). Collected papers of charles sanders peirce (Vol. 5).

Harvard University Press.

Peirce, C. S. (1998). The essential peirce: Selected philosophical writings (Vol.

2). Indiana University Press.

Pendergraft, C. A. (1998). Human dimensions of climate change: Cultural theory and collective action. Climatic Change, 39(4), 643-666.

Perrin, J. L., & Benassi, V. A. (2009). The connectedness to nature scale: A measure of emotional connection to nature? Journal of Environmental

Psychology, 29(4), 434-440.

Pierce, J. C., Dalton, R. J., & Zaitsev, A. (1999). Public perceptions of environmental conditions. Critical Masses: Citizens, Nuclear Weapons

Production, and Environmental Destruction in the United States and Russia,

97-129.

Poortinga, W., & Pidgeon, N. F. (2003). Exploring the dimensionality of trust in risk regulation. Risk Analysis, 23(5), 961-972.

Poortinga, W., Spence, A., Whitmarsh, L., Capstick, S., & Pidgeon, N. F.

(2011). Uncertain climate: An investigation into public scepticism about 338

anthropogenic climate change. Global Environmental Change, 21(3), 1015-

1024.

Poortinga, W., Steg, L., & Vlek, C. (2004). Values, environmental concern, and environmental behavior a study into household energy use. Environment and

Behavior, 36(1), 70-93.

Prakash, V., & Munson, J. M. (1985). Values, expectations from the marketing system and product expectations. Psychology & Marketing, 2(4), 279-296.

Rajecki, D. W. (1982). Attitudes: Themes and advances. Sunderland, MA.

Sinauer Associates, Inc.

Reid, A. (n.d.). Britons question global warming more than americans and canadians. [Web page] Retrieved from http://www.angusreidglobal.com/wp- content/uploads/2011/09/2011.09.12_Climate.pdf

Roe, R. A. (2012). What is wrong with mediators and moderators. The

European Health Psychologist, 14(1), 4-10.

Rokeach, M. (1973). The nature of human values (Vol. 438). Free press New

York.

Rokeach, M., & Regan, J. F. (1980). The role of values in the counseling situation. The Personnel and Guidance Journal, 58(9), 576-582.

339

Rose, B. M., Holmbeck, G. N., Coakley, R. M., & Franks, E. A. (2004). Mediator and moderator effects in developmental and behavioral pediatric research.

Journal of Developmental & Behavioral Pediatrics, 25(1), 58-67.

Rossman, G. B., & Wilson, B. L. (1985). Numbers and words combining quantitative and qualitative methods in a single large-scale evaluation study.

Evaluation Review, 9(5), 627-643.

Rubin, D. B. (1984). Bayesianly justifiable and relevant frequency calculations for the applies statistician. The Annals of Statistics, 12(4), 1151-1172.

Rubin, D. B. (1987). The calculation of posterior distributions by data augmentation: Comment: A noniterative sampling/importance resampling alternative to the data augmentation algorithm for creating a few imputations when fractions of missing information are modest: The SIR algorithm. Journal of the American Statistical Association, 543-546.

Russo, F., & Comi, A. (2012). City characteristics and urban goods movements:

A way to environmental transportation system in a sustainable city. Procedia -

Social and Behavioral Sciences, 39, 61-73. doi:10.1016/j.sbspro.2012.03.091

Sandell, R., Lazar, A., Grant, J., Carlsson, J., Schubert, J., & Broberg, J. (2006).

Therapist attitudes and patient outcomes. III. A latent class analysis of therapists. Psychology and Psychotherapy: Theory, Research and Practice,

79(4), 629-647.

340

Sanne, C. (2002). Willing consumers—or locked-in? Policies for a sustainable consumption. Ecological Economics, 42(1), 273-287

Schahn, J., & Holzer, E. (1990). Studies of individual environmental concern the role of knowledge, gender, and background variables. Environment and

Behavior, 22(6), 767-786.

Schreiber, J. B., Nora, A., Stage, F. K., Barlow, E. A., & King, J. (2010).

Reporting structural equation modeling and confirmatory factor analysis results: A review. The Journal of Educational Research, 99(6), 323-338. doi:10.3200/JOER.99.6.323-338

Schultz, P. W. (2000). Empathizing with nature: Toward a social-cognitive theory of environmental concern. Journal of Social Issues, 56, 391-406.

Schultz, P. W. (2001). The structure of environmental concern: Concern for self, other people, and the biosphere. Journal of Environmental Psychology, 21(4),

327-339.

Schultz, P. W., Oskamp, S., & Mainieri, T. (1995). Who recycles and when? A review of personal and situational factors. Journal of Environmental

Psychology, 15(2), 105-121.

Schultz, P. W., Shriver, C., Tabanico, J. J., & Khazian, A. M. (2004). Implicit connections with nature. Journal of Environmental Psychology, 24(1), 31-42.

341

Schultz, P. W., & Zelezny, L. C. (1998). Values and proenvironmental behavior

A five-country survey. Journal of Cross-cultural Psychology, 29(4), 540-558.

Schultz, P. W., & Zelezny, L. (1999). Values as predictors of environmental attitudes: Evidence for consistency across 14 countries. Journal of

Environmental Psychology, 19(3), 255-265.

Schwartz, S. H. (1977). Normative influences on altruism1. Advances in

Experimental Social Psychology, 10, 221-279

Schwartz, S. H. (1992). Universals in the content and structure of values:

Theoretical advances and empirical tests in 20 countries. Advances in

Experimental Social Psychology, 25(1), 1-65.

Schwarz, M., & Thompson, M. (1990). Divided we stand: Re-defining politics, technology and social choice. University of Pennsylvania Press.

Shove, E. (2010). Beyond the ABC: Climate change policy and theories of social change. Environment and Planning. A, 42(6), 1273

Sibley, C. G., & Kurz, T. (2013). A model of climate belief profiles: How much does it matter if people question human causation? Analyses of Social Issues and Public Policy, 13(1), 245-261.

342

Six, B. (1992). Neuere entwicklungen und trends in der einstellungs-verhaltens- forschung. Einstellung Und Verhalten. Braunschweiger Studien Zur Erziehungs- und Sozialarbeitswissenschaft, Bd, 32, 61-92.

Slimak, M. W., & Dietz, T. (2006). Personal values, beliefs, and ecological risk perception. Risk Analysis, 26(6), 1689-1705.

Smith, E. A. (1965). Social welfare: Principles and concepts. Association Press.

Smith, E. R. & Mackie, D. M. (2007). Social Psychology. London: Psychology

Press.

Smith, H. N. (1932). A scale for measuring attitudes about prohibition. The

Journal of Abnormal and Social Psychology, 26(4), 429.

Smith, J. K. (1983). Quantitative versus qualitative research: An attempt to clarify the issue. Educational Researcher, 6-13.

Smith, J. K., & Heshusius, L. (1986). Closing down the conversation: The end of the quantitative-qualitative debate among educational inquirers. Educational

Researcher, 15(1), 4-12.

Smith, M. B. (1963). Personal values in the study of lives. APA PsycNET. doi:10.1037/12238-014

343

Snelgar, R. S. (2006). Egoistic, altruistic, and biospheric environmental concerns: Measurement and structure. Journal of Environmental Psychology,

26(2), 87-99.

Smith, R. (2015). How reducing food waste could ease climate change.

National Geographic. Retrieved from http://news.nationalgeographic.com/news/2015/01/150122-food-waste- climate-change-hunger

Sörbom, D. (1989). Model modification. Psychometrika, 54(3), 371-384.

Spada, N. (1990). Observing classroom behaviors and learning outcomes in different second language programs. Second Language Teacher Education,

293-310.

Spearman, C. (1904). General intelligence,” objectively determined and measured. The American Journal of Psychology, 15(2), 201-292.

Spence, A., & Pidgeon, N. (2010). Framing and communicating climate change:

The effects of distance and outcome frame manipulations. Global

Environmental Change, 20(4), 656-667.

Stagner, R. (1942). Some factors related to attitude toward war, 1938. The

Journal of Social Psychology, 16(1), 131-142.

344

Steg, L., Dreijerink, L., & Abrahamse, W. (2005). Factors influencing the acceptability of energy policies: A test of VBN theory. Journal of Environmental

Psychology, 25(4), 415-425.

Steg, L., Geurs, K., & Ras, M. (2001). The effects of motivational factors on car use: A multidisciplinary modelling approach. Transportation Research Part A:

Policy and Practice, 35(9), 789-806.

Stern, P. C. (1992). What psychology knows about energy conservation.

American Psychologist, 47(10), 1224.

Stern, P. C. (1999). Information, incentives, and proenvironmental consumer behavior. Journal of Consumer Policy, 22(4), 461-478.

Stern, P. C. (2000). New environmental theories: Toward a coherent theory of environmentally significant behavior. Journal of Social Issues, 56(3), 407-424.

Stern, P. C., & Aronson, E. (1984). Energy use: The human dimension.

Stern, P. C., & Dietz, T. (1994). The value basis of environmental concern.

Journal of Social Issues, 50(3), 65-84. doi:10.1111/j.1540-4560.1994.tb02420.x

Stern, P. C., Dietz, T., Abel, T., Guagnano, G. A., & Kalof, L. (1999). A value- belief-norm theory of support for social movements: The case of environmentalism. Human Ecology Review, 6(2), 81-98.

345

Stern, P. C., Dietz, T., & Guagnano, G. A. (1995). The new ecological paradigm in social-psychological context. Environment and Behavior, 27(6), 723-743.

Stern, P. C., Dietz, T., & Kalof, L. (1993). Value orientations, gender, and environmental concern. Environment and Behavior, 25(5), 322-348.

Stern, P. C., & Gardner, G. T. (1981). Psychological research and energy policy.

American Psychologist, 36(4), 329.

Stern, P. C., Kalof, L., Dietz, T., & Guagnano, G. A. (1995). Values, beliefs, and proenvironmental action: Attitude formation toward emergent attitude objects1.

Journal of Applied Social Psychology, 25(18), 1611-1636.

Stoll-Kleemann, S., ORiordan, T., & Jaeger, C. C. (2001). The psychology of denial concerning climate mitigation measures: Evidence from swiss focus groups. Global Environmental Change, 11(2), 107-117.

Stradling, S., Anable, J., Anderson, T., & Cronberg, A. (2008). Car use and climate change: Do we practise what we preach? BRITISH SOCIAL

ATTITUDES, 24, 139.

Tanner, C. (1999). Constraints on environmental behaviour. Journal of

Environmental Psychology, 19(2), 145-157.

Tashakkori, A., & Teddlie, C. (1998). Mixed methodology: Combining qualitative and quantitative approaches (Vol. 46). Sage.

346

Terry, D. J., & Hogg, M. A. (2000). Attitudes, behavior, and social context. New

Jersey.

Tesser, A., & Shaffer, D. R. (1990). Attitudes and attitude change. Annual

Review of Psychology, 41(1), 479-523.

Thagard, P., & Shelley, C. (1997). Abductive reasoning: Logic, visual thinking, and coherence. In Logic and scientific methods (pp. 413-427). Springer.

The World Bank. (n.d.). [Data file] The Climate Change Knowledge Portal

(CCKP) is a central hub of information, data and reports about climate change around the world. Retrieved from http://data.worldbank.org/topic/climate- change

Thomas, W. I., & Znaniecki, F. (1918). The polish peasant in europe and america: Monograph of an immigrant group. Vol. I primary group organization.

Boston: Richard G. Badger. The Gorham Press.

Thomson, S. (1951). The factorial analysis of human ability, by g. Thomson.

University of london Press.

Thompson, S. C. G., & Barton, M. A. (1994). Ecocentric and anthropocentric attitudes toward the environment. Journal of Environmental Psychology, 14(2),

149-157.

Thompson, M., Ellis, R., & Wildavsky, A. (1990). Cultural theory. Westview

Press. 347

Thurstone, L. L. (1935). The vectors of mind. Chicago: University of Chicago

Press.

Thurstone, L. L. (1947). Multiple-factor analysis; a development and expansion of the vectors of mind. In APA psycnet. Chicago: University of Chicago Press.

Thurstone, L. L., & Chave, E. J. (1929). The measurement of attitude: A psychophysical method and some experiments with a scale for measuring attitude toward the church. University of Chicago Press. doi:10.1037/11574-00

Thøgersen, J. (1996). Wasteful food consumption: Trends in food and packaging waste. Scandinavian Journal of Management, 12(3), 291-304.

Tindall, D. B., Davies, S., & Mauboules, C. (2003). Activism and conservation behavior in an environmental movement: The contradictory effects of gender.

Society & Natural Resources, 16(10), 909-932.

Triandis, H. C. (1971). Attitude and attitude change. Wiley New York.

Triandis, H. C., Brislin, R., & Hui, C. H. (1988). Cross-cultural training across the individualism-collectivism divide. International Journal of Intercultural Relations,

12(3), 269-289.

Tyler, T. R., Orwin, R., & Schurer, L. (1982). Defensive denial and high-cost prosocial behavior. Basic and Applied Social Psychology, 3(4), 267-281.

348

Uyeki, E. S., & Holland, L. J. (2000). Diffusion of pro-environment attitudes?

American Behavioral Scientist, 43(4), 646-662.

Van Liere, K. D., & Dunlap, R. E. (1980). The social bases of environmental concern: A review of hypotheses, explanations and empirical evidence. Public

Opinion Quarterly, 44(2), 181-197.

Velicer, W. F., Eaton, C. A., & Fava, J. L. (2000). Construct explication through factor or component analysis: A review and evaluation of alternative procedures for determining the number of factors or components. In Problems and solutions in human assessment (pp. 41-71). Springer US. doi:10.1007/978-

1-4615-4397-8_3

Vermunt, J. K. (2010). Latent class modeling with covariates: Two improved three-step approaches. Political Analysis, 18(4), 450-469.

Vermunt, J. K., & Magidson, J. (2004). Latent class analysis. Encyclopedia of

Social Science Research Methods.--Sage Publications.--2003.

Vickers, D., & Rees, P. (2007). Creating the UK national statistics 2001 output area classification. Journal of the Royal Statistical Society: Series A (Statistics in Society), 170(2), 379-403. doi:10.1111/j.1467-985X.2007.00466.x

Vining, J., & Ebreo, A. (1990). What makes a recycler? A comparison of recyclers and nonrecyclers. Environment and Behavior, 22(1), 55-73.

349

Vining, J., & Ebreo, A. (1992). Predicting recycling behavior from global and specific environmental attitudes and changes in recycling opportunities1.

Journal of Applied Social Psychology, 22(20), 1580-1607.

Wang, M., & Takeuchi, R. (2007). The role of goal orientation during expatriation: A cross-sectional and longitudinal investigation. Journal of

Applied Psychology, 92(5), 1437.

Warner, R. M. (2012). Applied statistics: From bivariate through multivariate techniques: From bivariate through multivariate techniques. Sage.

Watson, J. B. (1925). Behaviorism. New Brunswick, New Jersey: Transaction

Publishers.

Watson, R. K., Murphy, M. H., Kilfoyle, F. E., & Moore, S. M. (1999). An opportunistic field experiment in community water conservation. Population and Environment, 20(6), 545-560.

Wegener, D. T., & Fabrigar, L. R. (2000). Analysis and design for nonexperimental data: Addressing causal and noncausal hypotheses.

Handbook of Research Methods in Social and Personality Psychology, 412-

450.

Weigel, R., & Weigel, J. (1978). Environmental concern the development of a measure. Environment and Behavior, 10(1), 3-15.

350

Weigel, R. H. (1983). Environmental attitudes and the prediction of behavior.

Environmental Psychology: Directions and Perspectives, 257-287.

Werner, P. (2004). Reasoned action and planned behavior. In: Peterson SJ,

Bredow TS, editors. Middle range theories: Application to nursing research.

Philadeliphia, PA: Lippincott, Williams, & Wilkins, 125–147.

Wesley Schultz, P. (2001). The structure of environmental concern: Concern for self, other people, and the biosphere. Journal of Environmental Psychology,

21(4), 327-339.

West, S. G., Finch, J. F., & Curran, P. J. (1995). Structural equation models with nonnormal variables: Problems and remedies.

Whitmarsh, L. (2009). Behavioural responses to climate change: Asymmetry of intentions and impacts. Journal of Environmental Psychology, 29(1), 13-23.

Whitmarsh, L. (2012). How useful is the multi-level perspective for transport and sustainability research? Journal of Transport Geography, 24, 483-487. doi:10.1016/j.jtrangeo.2012.01.022

Whitmarsh, L., Lorenzoni, I., & O'Neill, S. (2012). Engaging the public with climate change: Behaviour change and communication. Routledge.

Wicker, A. W. (1969). Attitudes versus actions: The relationship of verbal and overt behavioral responses to attitude objects. Journal of Social Issues, 25(4),

41-78. 351

Williams Jr, R. (1968). Values. U: Sills david L.(Ur.) International encyclopedia of the social sciences. London-New York, Itd: Macmillan Publishing.

Xiao, C., & Dunlap, R. E. (2007). Validating a comprehensive model of environmental concern cross-nationally: A us-canadian comparison*. Social

Science Quarterly, 88(2), 471-493.

Xiao, C., & McCright, A. M. (2007). Environmental concern and sociodemographic variables: A study of statistical models. The Journal of

Environmental Education, 38(2), 3-14.

Yin, J. (1999). Elite opinion and media diffusion exploring environmental attitudes. The Harvard International Journal of Press/Politics, 4(3), 62-86.

Yu, C. H. (1994, April). Abduction? Deduction? Induction? Is there a logic of exploratory data analysis? Retrieved from http://www.creative- wisdom.com/pub/Peirce/Logic_of_EDA.htm

Yu, C. -Y. (2002). Evaluating cutoff criteria of model fit indices for latent variable models with binary and continuous outcomes. Thesis.

Yuan, Y. C. (2010). Multiple imputation for missing data: Concepts and new development (version 9.0). SAS Institute Inc, Rockville, MD.

Zelezny, L. C., Chua, P. -P., & Aldrich, C. (2000). New ways of thinking about environmentalism: Elaborating on gender differences in environmentalism.

Journal of Social Issues, 56(3), 443-457. 352

Zimbardo, P. G. (1985). Psychology and life (11th ed.). Glenview, IL: Scott

Foresman.

353

Appendix a)

DEFRA Attitudes and Behaviour Tracker

I am carrying out a survey about people’s lifestyles and their attitudes towards topical issues. I work for a company called TNS and we have been commissioned to carry out this research on behalf of the Government. We are offering a £5 voucher as a thank you to anyone who takes part in the research. If you take part, any information you give will be treated in strict confidence. No information that can identify you or your household will be passed to any other organisation.

354

A. Household and respondent characteristics

QUESTIONS 1-4 ARE ASKED HERE TO ESTABLISH QUOTAS FOR THE SURVEY

ASK ALL INTERVIEWER ONLY - DO NOT READ OUT 1. CODE RESPONDENT’S GENDER

Male Female

ASK ALL

Please could you tell me your age last birthday?

Numeric range (16 – 99) Refused

ASK ALL

INTERVIEWER CODE WORKING STATUS

SINGLE CODE ONLY

Full-time paid work (30+ hours per week) Part-time paid work (8-29 hours per week) Part-time paid work (under 8 hours per week) Retired

Still at school In full time higher education Unemployed (seeking work) Not in paid employment (not seeking work) Refused

ASK ALL

Do you have any children under 16 in the household?

SINGLE CODE ONLY

Yes No Refused

ASK ALL

Including yourself, how many people usually live here?

Please include all adults (aged 16 and over) and children (aged under 16).

Enter number (range 1-30)

ASK IF 2 OR MORE AT Q5

And how many adults aged 16 years and over are there in the household in total including yourself?

355

Enter number (range 1-30)

FOR EACH ADULT AT Q6

And what age are the adults in the household (excluding you)? ENTER AGE LAST BIRTHDAY

INTERVIEWER COLLECT AGE FOR EACH ADULT (IN ANY ORDER)

Enter number (range 16-99)

FOR EACH CHILD (IF Q5 > Q6)

And what age are the children in the household? ENTER AGE LAST BIRTHDAY, IF LESS THAN 1 YEAR OLD, ENTER 0.

INTERVIEWER COLLECT AGE FOR EACH CHILD (IN ANY ORDER)

Enter number (range 0-15)

ASK ALL SHOW SCREEN

Can I just check, which of these applies to you at present? Please choose the first on the list that applies SINGLE CODE ONLY

Married Living in a civil partnership Living with a partner Separated (after being married) Divorced Widowed Single (never married) Refused

ASK IF MARRIED OR LIVING WITH PARTNER (CODES 1,2 OR 3) SHOW SCREEN

Which of the statements applies to your spouse or partner? SINGLE CODE ONLY

Full-time paid work (30+ hours per week) Part-time paid work (8-29 hours per week) Part-time paid work (under 8 hours per week) Retired

Still at school In full time higher education Unemployed (seeking work) Not in paid employment (not seeking work) Refused

ASK ALL SHOW SCREEN 11.Which of these statements best applies to you? SINGLE CODE ONLY

Living with husband, wife or partner Living with friends Living with parents or other relatives I am the only adult in the household Refused

356

ASK ALL SHOW SCREEN 12.Which of these best describes your home? SINGLE CODE ONLY

Detached house Semi-detached house Terraced house Bungalow

Flat (in a block of flats) Flat (in a house) Maisonette Other (specify)

Don’t know Refused

ASK ALL SHOW SCREEN 13.Which of these best describes how you occupy your accommodation? SINGLE CODE ONLY

Own it outright Buying with a mortgage Pay part rent part mortgage (shared ownership) Rented Other Don’t know Refused

ASK ALL WHO RENT (IF Q13=3 OR 4) SHOW SCREEN 14.Who is your landlord? SINGLE CODE ONLY

Local authority/council/new town development A housing association, charitable trust or local housing company Private landlord Not applicable – living rent-free with parents Other Don’t know Refused

ASK ALL SHOW SCREEN

Do you have at least some responsibility for the physical upkeep of your home? SINGLE CODE ONLY

Yes No

ASK ALL SHOW SCREEN 16.Do you know roughly when your home was built? SINGLE CODE ONLY

1929 or earlier 1930-1965 1966-1994 1995 or later Don’t know Refused

ASK ALL SHOW SCREEN 17.How long have you lived in your current home? SINGLE CODE ONLY

Up to 1 year More than 1 year, up to 2 years More than 2 years, up to 5 years More than 5 years, up to 10 years More than 10 years, up to 20 years More than 20 years Don’t know Refused

ASK ALL

Do you have a water meter, so that you are billed based on how much water you use?

357

ADD IF NECESSARY: A water meter is a device, installed by the water company, that records the amount of water being used in your home. It may be installed underground where the mains water supply comes into your property.

Yes No Don’t know

ASK ALL SHOW SCREEN 19.Do you have a garden? SINGLE CODE ONLY

Yes – own garden Yes – garden shared with others No Don’t know

ASK ALL SHOW SCREEN

All things considered, how satisfied are you with your life as a whole nowadays? Please answer on a scale of 0-10, where 0 means extremely dissatisfied and 10 means extremely satisfied.

0 (Extremely dissatisfied) 1 2 3

4 5 6 7 8 9 10 (Extremely satisfied) Don’t know Refused

B. Environmental and energy behaviours

ASK ALL SHOW SCREEN

I am now going to read out some changes that people might make to their lifestyles. For each one tell me what answer on the screen applies to you personally at the moment. Remember there are no right or wrong answers – we’re just interested in what you personally do at the moment, not what you think you should or shouldn’t be doing.

INTERVIEWER: READ OUT BEHAVIOUR AND PROMPT AS NECESSARY: ‘Which of the options on this list best describes what you personally think about this’

Answer codes (vary by behaviour):

Standard (S):I don't really want to do this I haven't really thought about doing this I've thought about doing this, but probably won't do it I'm thinking about doing this I'm already doing this, but I probably won't manage to keep it up I'm already doing this and intend to keep it up I've tried doing this, but I've given up I haven’t heard of this Don’t know (SPONTANEOUS ONLY) Not applicable (SPONTANEOUS ONLY)

358

Regular Purchasing (RP): I don't really want to do this I haven't really thought about doing this I've thought about doing this, but probably won't do it I'm thinking about doing this I've done this, but I probably won't do it again I've done this and intend to do it again I've tried doing this, but I've given up I haven’t heard of this Don’t know (SPONTANEOUS ONLY) Not applicable (SPONTANEOUS ONLY)

One-Off purchasing (OO) I don't really want to do this I haven't really thought about doing this I've thought about doing this, but probably won't do it I'm thinking about doing this I've already done this I haven’t heard of this Don’t know (SPONTANEOUS ONLY) Not applicable (SPONTANEOUS ONLY)

Statements:

RANDOMISE

Cutting down on the use of gas and electricity at home S

Buying energy efficient (‘A’ rated or better) appliances excluding energy saving light bulbs RP

Washing clothes at 40 degrees or less S (QUANT)

Turning down thermostats (by 1 degree or more) OO

Making an effort to cut down on water usage at home S (QUANT)

Cutting down on the use of hot water at home S (QUANT)

Q22. ASKED FOR ALL – ONLY FOR OPTIONS MARKED ‘QUANT’ AT Q21 SHOW SCREEN

Please tell me how frequently you personally... REPEAT FOR:

RANDOMISE

Wash clothes at 40 degrees or less Make an effort to cut down on water usage at home Cut down on the use of hot water at home

PLUS Leave the heating on when you go out for a few hours (QUANT) Leave your TV or PC on standby for long periods of time at home (QUANT)

Leave lights on when you are not in the room

ANSWER OPTIONS:

359

(QUANT)

Always Very often Quite often Sometimes Occasionally Never Don’t know (SHOULD BE SEEN ON SCREEN) Not applicable / cannot do this (SHOULD BE SEEN ON SCREEN)

INTRODUCTION TO Q23 SHOW SCREEN Cavity walls: In most houses and parts of houses built after the 1920s, the external walls are made of two layers with a small air gap or 'cavity' between them. These are known as ‘cavity walls’.

ASK ALL SHOW SCREEN

Which of these best describes the outside walls of your home?

CODE ALL THAT APPLY

Brick cavity walls Brick or concrete solid walls Brick – some solid and some cavity Concrete cladding Timber frame Other(specify) Don’t know

ASK ALL READ OUT

Do you have a loft – one that is not inhabited - as part of your house?

Yes No Don’t know

ASK ALL SHOW SCREEN

I am now going to read out some other changes that people might make to their lifestyles. For each one, please tell me which answer on the screen applies to you personally at the moment.

INTERVIEWER: READ OUT BEHAVIOUR AND PROMPT AS NECESSARY: ‘Which of the options on this list best describes what you personally think about this’

Answer codes (vary by behaviour):

Standard (S): I don't really want to do this I haven't really thought about doing this I've thought about doing this, but probably won't do itI'm thinking about doing this I'm already doing this, but I probably won't manage to keep it up I'm already doing this and intend to keep it up I've tried doing this, but I've given up I haven’t heard of this Don’t know (SPONTANEOUS ONLY) Not applicable (SPONTANEOUS ONLY)

360

Regular Purchasing (RP): I don't really want to do this I haven't really thought about doing this I've thought about doing this, but probably won't do it I'm thinking about doing this

I've done this, but I probably won't do it again I've done this and intend to do it again I've tried doing this, but I've given up I haven’t heard of this

Don’t know (SPONTANEOUS ONLY) Not applicable (SPONTANEOUS ONLY) One-Off purchasing (OO) I don't really want to do this I haven't really thought about doing this I've thought about doing this, but probably won't do it I'm thinking about doing this I've already done this I haven’t heard of this Don’t know (SPONTANEOUS ONLY) Not applicable (SPONTANEOUS ONLY)

Statements:

RANDOMISE

Installing cavity wall insulation (SKIP IF CODES 2,5,7 AT Q23) 1 OO / barriers

Installing solid wall insulation1 OO / barriers

Installing loft insulation or top-up loft insulation 1 OO / barriers (EXCLUDE IF NO LOFT –IF Q24=NO)

Installing double glazing1 OO / none

Installing draught exclusion OO/none

Installing solar panels for electricity at home1 OO / levers

Installing solar water heating at home1 OO / levers

Installing a wind turbine to generate electricity at home1 OO / levers

Installing a condensing boiler 1 OO / none

10.Installing a ground source heat pump 1OO / levers 11.Installing biomass heating1 OO / levers 12.Having thermostat controls fitted on individual radiators 1 OO / none 13.Volunteering with a conservation group (or other group helping the environment) S / barriers NOTE: Levers and barriers – Options above lead into Q26/Q27. Not all options get these. Each statement is marked with levers / barriers / both / none 1[EXCLUDE [IF RENTING – IF Q13=CODE 4] OR IF NOT RESPONSIBLE

FOR PHYSICAL UPKEEP OF HOME IF Q15 = NO]] 361

Q26 = BARRIERS Q27 = LEVERS

ASK Q26 FOR STATEMENTS 1,2,3,13 AT Q25 ONLY – IF CODES 1,2,3,4 FOR EACH STATEMENT

INTRODUCTION I’m now going to ask you the reasons you have not done some of the things I just asked you about...

SHOW SCREEN

What would you say are the main reasons you have not done this? CODE ALL THAT APPLY

Never thought about it Don’t know if I have it or not I cannot afford it Takes too long to get costs back through lower energy bills Causes too much disruption It is too much hassle Waiting until we do major renovations Don’t know how to go about it – or who to ask Would not look right Other reason (please specify)

ANSWER CODES ARE DIFFERENT FOR Q26 FOR STATEMENT 13 ONLY

Not enough time / too busy Don’t know where to find out about what I can do Not interested Rather do a different type of volunteering Tried to volunteer but it was too difficult to sort out Other (specify)

ASK Q27 FOR THOSE INSTALLED (6,7,8,10,11 AT Q25) ONLY – IF CODE 5 FOR EACH BEHAVIOUR

INTRODUCTION I’m now going to ask you the reasons you have done some of the things I just asked you about...

SHOW SCREEN

What would you say are the main reasons you have already done this? CODE ALL THAT APPLY

It saves money It is easy to do/install It helps the environment It prevents waste It reduces your carbon dioxide emissions Doing a refurbishment anyway Have or can get a grant (loan) for the work It was a legal requirement

Saw / knew other people had done it Makes the home a warmer / nicer place to be Other reason (specify)

362

Able to sell surplus electricity OPTIONS 6 & 8 ONLY Able to generate my own heat and/or power OPTIONS 6, 7 & 8 ONLY It is a reliable energy supply OPTIONS 6 & 8 ONLY

IF ANSWER CODE 5 FOR LOFT INSULATION AT Q25. SHOW SCREEN 28.B7. How thick is the insulation in your loft?

Don’t know thickness 1

50mm (2”) thick 2

100mm (4”) thick 3 GO TO Q29 150mm (6”) thick 4

200mm (8”) thick 5

250mm (10”) thick 6

270mm (11”) thick or more 7 GO TO Q30

IF LOFT THICKNESS UNKNOWN OR LESS THAN 270MM (ANSWER CODE 1- 6 AT Q28) SHOW SCREEN 29.When did you install loft insulation or last top up your loft insulation?

1 year ago 2 years ago 3-5 years ago 6-10 years ago 11-15 years ago 16-20 years ago more than 20 years ago Don’t know

ASK ALL SHOW SCREEN 30.I am now going to read out some other changes that people might make to their lifestyles. For each one, please tell me which answer on the screen applies to you personally at the moment.

INTERVIEWER: READ OUT BEHAVIOUR AND PROMPT AS NECESSARY: ‘Which of the options on this list best describes what you personally think about this’

Answer codes (vary by behaviour):

Standard (S): I don't really want to do this

I haven't really thought about doing this I've thought about doing this, but probably won't do it I'm thinking about doing this I'm already doing this, but I

363

probably won't manage to keep it up I'm already doing this, though I’d like to do it more I'm already doing this and intend to keep it up I've tried doing this, but I've given up I haven’t heard of this Don’t know (SPONTANEOUS ONLY) Not applicable (SPONTANEOUS ONLY)

Regular Purchasing (RP): I don't really want to do this I haven't really thought about doing this I've thought about doing this, but probably won't do it I'm thinking about doing this I've done this, but I probably won't do it again I've done this before, though not as much as I’d like I've done this and intend to do it again I've tried doing this, but I've given up I haven’t heard of this Don’t know (SPONTANEOUS ONLY) Not applicable (SPONTANEOUS ONLY)

RANDOMISE STATEMENTS

Recycling items rather than throwing them away S (QUANT)

Wasting less food S

Buying fresh food that has been grown when it is in season in the country where it was produced S

Checking whether the packaging of an item can be recycled, before you buy it RP

Deciding not to buy something because it has too much packaging RP (QUANT)

‘INTERVIEWER: IF ASKED ‘This isn’t just about buying food but all types of products’

Reusing items like empty bottles, tubs, jars, envelopes or paper RP (QUANT)

Taking your own shopping bag when shopping RP (QUANT)

Buying plants that especially encourage wildlife in the garden RP

Growing your own fruit and vegetables RP

Taking fewer flights

364

Installing a water butt to collect rainwater

Buying fish from sustainable sources (such as certified by the Marine

Stewardship Council)

Buying wood and wood products from certified sustainable sources (such as certified by the Forest Stewardship Council)

Composting your household’s food and/or garden waste

Only boiling the kettle with as much water as you need

1[EXCLUDE [IF RENTING – IF Q13=CODE 4] OR IF NOT RESPONSIBLE FOR PHYSICAL UPKEEP OF HOME IF Q15 = NO]]

ASKED FOR ALL – ONLY FOR OPTIONS MARKED ‘QUANT’ AT Q30 SHOW SCREEN

Please tell me how frequently you...

REPEAT FOR:

Recycle items rather than throw them away

Decide not to buy something because it has too much packaging

Reuse items like empty bottles, tubs, jars, envelopes or paper

Take your own shopping bag when shopping

Compost your household’s food and/or garden waste – ONLY FOR THOSE WITH GARDEN (IF YES AT Q19)

Only boil the kettle with as much water as you need

ANSWER OPTIONS:

Always Very often Quite often Sometimes Occasionally Never Don’t know (SHOULD BE SEEN ON SCREEN) Not applicable / cannot do this (SHOULD BE SEEN ON SCREEN)

365

C. Food behaviours

INTRODUCTION SHOW SCREEN

There are many different types of uneaten food that people throw away. These might include the types of things on this list:

Inedible food waste (e.g. peelings, bones) Fruit, vegetables or salad Processed meat & fish (e.g. sandwich meats) Bread and cakes

Food left on the plate after the meal Food you cooked or prepared too much of but didn’t serve up Raw or home-cooked meat & fish Ready meals or convenience foods Cheese and yoghurt

ASK ALL SHOW SCREEN

How much uneaten food, overall, would you say you generally end up throwing away? SINGLE CODE ONLY

Quite a lot A reasonable amount Some A small amount Hardly any None Don’t know (SPONTANEOUS ONLY)

ASK ALL SHOW SCREEN

Thinking about when you have to throw uneaten food away, to what extent, if at all does it bother you personally? SINGLE CODE ONLY

A great deal A fair amount A little Not very much Not at all Don’t know

(SPONTANEOUS ONLY)

ASK ALL SHOW SCREEN

Thinking about the different types of food waste on the list I just showed you, how much effort do you and your household go to in order to minimise the amount of uneaten food you throw away? SINGLE CODE ONLY

A great deal A fair amount A little Not very much None at all Don’t know (SPONTANEOUS ONLY)

ASK ALL SHOW SCREEN

How much do you agree or disagree with the following statement?

STATEMENTS

366

Food production contributes to climate change

ANSWER CODES

Strongly agree Tend to agree Neither agree nor disagree Tend to disagree Strongly disagree Don't Know

ASK ALL SHOW SCREEN

Please read the statement below and tell me which of the options best applies to you.

If I had a better understanding of the environmental impacts of how food is produced...

I would still buy the food I usually buy I would be willing to make changes to the food I buy to reduce my impact on the environment I already make changes to the food I buy to reduce my impact on the environment I already make changes to the food I buy to reduce my impact on the environment and I’d like to do more Don't Know

367

D. Recycling behaviours

ASK ALL SHOW SCREEN

As far as you know, which of these can you put outside for a council recycling or composting collection?

CODE ALL THAT APPLY

Paper/Newspapers/magazines Glass bottles/jars/glass Tins/Cans/Foil Cardboard

Clothes Shoes Plastic bottles/plastic packaging Food waste Garden waste Other items (Specify) None of these Don’t know

SHOW SCREEN

Which of these do you normally put outside for recycling or composting collection?

CODE ALL THAT APPLY

Paper/Newspapers/magazines Glass bottles/jars/glass Tins/Cans/Foil Cardboard

Clothes Shoes Plastic bottles/plastic packaging Food waste Garden waste Other items (Specify) None of these Don’t know

ASK ALL

Is there a bottle bank or recycling bank in your area where you can take things like bottles, cans or paper to recycle?

Yes No Don’t know

IF YES AT Q39

Do you [or your household] ever use these facilities?

Yes No Don’t know

IF YES AT Q40 DO NOT SHOW SCREEN 41.What things do you take to recycle?

DO NOT PROMPT.

CODE ALL THAT APPLY.

368

PROBE FULLY: What else? Anything else?

Paper/Newspapers/magazines Glass bottles/jars/glass Tins/Cans/Foil Cardboard

Clothes Shoes Plastic bottles/plastic packaging Other items (Specify) None of these Don’t know

369

E. Energy in the home

ASK ALL

Approximately, how many light bulbs do you have in your home?

WRITE IN NUMBER BELOW.

INTERVIEWER - IF NECESSARY: You do not need to count the number of bulbs. Please give your best estimate.

Don’t know

ASK ALL SHOW SCREEN WITH IMAGE OF ENERGY SAVING LIGHT BULB These are energy saving light bulbs.

Approximately, how many of the light bulbs in your house, if any, are energy saving light bulbs?

Don’t know

ASK IF ALL BULBS ARE NOT ENERGY SAVING (Q43 IS LESS THAN Q44) DO NOT SHOW SCREEN

What are the main reasons stopping you fitting/fitting more energy saving light bulbs in your home?

DO NOT PROMPT - CODE ALL THAT APPLY PROBE FOR SPECIFIC REASONS, e.g. REASONS FOR NOT LIKING THEM RATHER THAN ‘I don’t like them’.

Don’t like the way they look Do not fit my light fittings Not as bright as ordinary bulbs/quality of light is poor Will replace as other bulbs blow Too expensive Takes too long to turn on Can’t use with a dimmer switch Don’t believe they save you money Not thought about it Other reason (no need to specify) Don’t know

ASK ALL READ OUT

Which of the following types of heating does your home have?

SINGLE CODE.

Central heating Warm air heating system (heating grates bringing heat from a communal building source) Electric storage heating None of these Don’t know

370

ASK ALL SHOW SCREEN

Thinking about your heating system at home, which of these statements best describes how you set the temperature during the winter? CODE ALL THAT APPLY

INTERVIEWER: SETTING THE TEMPERATURE CAN INCLUDE SETTING A CENTRAL THERMOSTAT OR THERMOSTATIC CONTROLS ON INDIVIDUAL RADIATORS (IT DEPENDS ON HOW THE RESPONDENT CONTROLS THE TEMPERATURE IN THEIR HOME)

I change it whenever it gets too hot or too cold, I don’t like to wear a lot of layers indoors I change it whenever it gets too hot or too cold, I’ll often wear a jumper indoors

I don’t change the setting often, but it can be too warm I don’t change the setting often, but it can be too cold I don’t change the setting often, it’s a comfortable temperature I don’t tend to use the central heating Don’t know

ASK ALL

DO NOT READ OUT

What proportion of the windows in your home are double-glazed?

SINGLE CODE ONLY

371

None 1

ASK Q48 Some (25%) 2

About half (50%) 3

Most (75%) 4

GO TO Q49 All (100%) 5

Don’t know (DO NOT READ) 6

ASK ALL EXCLUDING RENTERS [IF Q13=CODE 4] AND THOSE NOT RESPONSIBLE FOR PHYSICAL UPKEEP OF HOME [IF Q15 = NO]

READ OUT

Which of the following windows are you interested in getting/replacing?

CODE ALL MENTIONS

Replacing single glazing with double glazing Getting new double glazing for a new extension/renovation e.g. extension or loft extension Replacing old double glazing with new improved double glazing Don’t know None of these

ASK ALL READ OUT 49.What proportion of your single glazed opening windows and doors are draught-proofed?

None Some (25%) About half (50%) Most (75%) All (100%) Don’t know

ASK ALL EXCLUDING RENTERS [IF Q13=CODE 4] AND THOSE NOT RESPONSIBLE FOR PHYSICAL UPKEEP OF HOME [IF Q15 = NO]

SHOW SCREEN

372

Have you bought any of the following household appliances in the last year?

CODE ONE FOR EACH OWNED

SHOW SCREEN WITH ENERGY SAVING RECOMMENDED LOGO ASK FOR ONE APPLIANCE BOUGHT IN LAST 12 MONTHS (SCRIPT TO SELECT ONE AT RANDOM)

When you were looking for the (APPLIANCE), did you look for the Energy Saving Recommended logo on it?

SINGLE CODE ONLY

ASK FOR ONE APPLIANCE BOUGHT IN LAST 12 MONTHS (SCRIPT TO SELECT ONE AT RANDOM – SAME AS Q51)

And did the (APPLIANCE) you actually bought have the Energy Saving Recommended logo on it?

CODE ONE ONLY

373

Logo Bought Within the

past year Yes No DK Yes No DK

Washing machine 1 1 2 3 1 2 3

Tumble dryer 2 1 2 3 1 2 3

Washer-dryer 3 1 2 3 1 2 3

Dishwasher 4 1 2 3 1 2 3

Fridge-freezer 5 1 2 3 1 2 3

Fridge 6 1 2 3 1 2 3

Freezer 7 1 2 3 1 2 3

374

G. Travel

ASK ALL READ OUT

How many cars or vans are there in your household currently? SINGLE CODE

ASK ALL READ OUT

Do you drive?

ASK ALL WHO DRIVE (IF YES AT Q54) SHOW SCREEN 55.Approximately how many miles a year do you personally drive? SINGLE CODE

Less than 5000 miles 5000 – 7999 miles 8000 – 10,999 miles 11,000 – 15,999 miles 16,000 – 20,000 miles More than 20,000 miles Don’t know

ASK ALL WHO DRIVE (IF YES AT Q54)

SHOW SCREEN

Thinking of the car you drive the majority of the time, what size engine does it have?

Less than 1 litre 1.0 – 1.4 litres 1.5 – 2.0 litres 2.1 – 3.0 litres More than 3.0 litres Don’t know

1 1

2 2 ASK Q54

3 or more 3

None 4

Yes 1 ASK Q55

No 2

ASK ALL WHO DRIVE (IF YES AT Q54)

SHOW SCREEN

And what sort of fuel does it run on? SINGLE CODE ONLY

Petrol Diesel LPG Hybrid petrol/electric Electric

375

Other Don’t know

ASK ALL WHO DRIVE (IF YES AT Q54)

I am now going to read out some changes that people might make. For each one, please tell me which answer on the screen applies to you personally at the moment.

INTERVIEWER: READ OUT BEHAVIOUR AND PROMPT AS NECESSARY: ‘Which of the options on this list best describes what you personally think about this’

Answer codes (vary by behaviour):

Standard (S): I don't really want to do this I haven't really thought about doing this I've thought about doing this, but probably won't do it I'm thinking about doing this I'm already doing this, but I probably won't manage to keep it up I'm already doing this and intend to keep it up I've tried doing this, but I've given up I haven’t heard of this Don’t know (SPONTANEOUS ONLY) Not applicable (SPONTANEOUS ONLY)

STATEMENTS RANDOMISE

Switching to public transport instead of driving for regular journeys S 24

Switching to walking or cycling instead of driving for short, regular journeys S

Driving in a fuel efficient way S

Switching to an electric / hybrid / LPG car OO

Buying or using a more fuel efficient / smaller / diesel car OO

INTRODUCTION READ OUT Next I’d like to ask you about air travel. I’m just interested in air travel for leisure, holidays and visiting friends or family, not air travel for work or business purposes.

ASK ALL

376

Have you taken any flights in the last 12 months for leisure, holidays or visiting friends or family?

Yes No Don’t know

IF YES AT Q59

How many flights within the UK did you take in the last 12 months? Please count the outward and return flight and any transfers as one flight.

Numeric range (0 to 99)

Don’t know

IF YES AT Q59

How many flights to other European countries did you take in the last 12 months? Please count the outward and return flight and any transfers as one flight.

Numeric range (0 to 99)

Don’t know

IF YES AT Q59

How many flights to countries outside Europe did you take in the last 12 months? Please count the outward and return flight and any transfers as one flight.

Numeric range (0 to 99)

Don’t know

377

H. Environmental and energy attitudes

QUESTIONS Q63 – Q74 TO BE COMPLETED AS SELF-COMPLETION SECTION IN CAPI

ASK ALL

SHOW SCREEN

How much if anything would you say you know about the following terms?

SINGLE CODE ONLY

STATEMENTS

Climate change (to always appear first in the list)

Global warming

Carbon footprint CO2 (carbon dioxide) emissions

Biodiversity

ANSWER CODES

A lot A fair amount Just a little Nothing – have only heard of the name Nothing – have never heard of it

Don’t know

ASK ALL

SHOW SCREEN

Which of these best describes how you feel about your current lifestyle and the environment?

I’m happy with what I do at the moment I’d like to do a bit more to help the environment I’d like to do a lot more to help the environment Don’t know

ASK ALL

SHOW SCREEN

And which of these would you say best describes your current lifestyle? SINGLE CODE ONLY

378

I don’t really do anything that is environmentally-friendly I do one or two things that are environmentally-friendly I do quite a few things that are environmentally-friendly I’m environmentally-friendly in most things I do

I’m environmentally-friendly in everything I do Don’t know

ASK ALL

SHOW SCREEN

How much do you agree or disagree with these statements?

RESPONSE CODES

Strongly agree Tend to agree Neither agree nor disagree Tend to disagree

Strongly disagree Don't Know

STATEMENTS RANDOMISE

I don't really give much thought to saving energy in my home

People have a duty to recycle

I don't pay much attention to the amount of water I use at home

The so-called 'environmental crisis' facing humanity has been greatly exaggerated

I find it hard to change my habits to be more environmentally-friendly

We are close to the limit of the number of people the earth can support

It would embarrass me if my friends thought my lifestyle was purposefully environmentally friendly

It's not worth Britain trying to combat climate change, because other countries will just cancel out what we do

The Earth has very limited room and resources

The effects of climate change are too far in the future to really worry me

It's not worth me doing things to help the environment if others don't do the same

If things continue on their current course, we will soon experience a major environmental disaster

379

It's only worth doing environmentally-friendly things if they save you money

For the sake of the environment, car users should pay higher taxes

I would only travel by bus if I had no other choice 16.People who fly should bear the cost of the environmental damage that air travel causes

Being green is an alternative lifestyle it's not for the majority

ASK ALL

SHOW SCREEN

And how much do you agree or disagree with these statements:

RESPONSE CODES

Strongly agree Tend to agree Neither agree nor disagree Tend to disagree

Strongly disagree Don't Know

STATEMENTS RANDOMISE

The Government is doing a lot to tackle climate change

Any changes I make to help the environment need to fit in with my lifestyle

I need more information on what I could do to be more environmentally friendly

I sometimes feel guilty about doing things that harm the environment

I would be prepared to pay more for environmentally-friendly products

I do worry about the changes to the countryside in the UK and the loss of native animals and plants

‘Waste not want not' sums up my general approach to life

I often talk to friends and family about the things they can do to help the environment

I try to persuade people I know to be more environmentally friendly

I’ve suggested improvements at my workplace/the place where I study to make it more environmentally friendly

Climate change is beyond control – it’s too late to do anything about it

380

The environment is a low priority compared to other things in my life

If government did more to tackle climate change, I’d do more too

I don’t believe my everyday behaviour and lifestyle contribute to climate change

I make an effort to buy things from local retailers and suppliers

ASK ALL

SHOW SCREEN

Here are some things other people have said. For each one, please say how much you agree or disagree with the statement:

RESPONSE CODES

Strongly agree Tend to agree Neither agree nor disagree Tend to disagree Strongly disagree Don't Know

STATEMENTS RANDOMISE

I do worry about the loss of species of animals and plants in the world

It’s important to me that I can be proud of my local environment

We should all try and save water regardless of whether it rains or is sunny

There are many natural places that I may never visit, but I’m glad they exist

If business did more to tackle climate change, I would too

It really disappoints me when I see big offices and public buildings with their lights on when the building is empty

It really bothers me when I see people wasting energy or food

ASK ALL

SHOW SCREEN

And thinking now about your overall attitudes towards energy usage and climate change, which of these statements best reflects how you currently feel?

SINGLE CODE ONLY

381

I don’t believe there are climate change problems caused by energy use and I’m not willing or able to change my behaviour with 1 regards to energy use

Whether there are climate change issues or not, I am not willing or 2 able to change my behaviour with regards to energy use

Climate change is caused by energy use and I’m beginning to think 3 that I should do something

Climate change is caused by energy use and I’m doing a few small 4 things to help reduce my energy use and emissions

Climate change is caused by energy use and I’m doing quite a 5 number of things to help reduce my energy use and emissions

Climate change is caused by energy use and I’m doing lots of 6 things to help reduce my energy use and emissions

Don’t know (NOT ON CARD) 7

ASK ALL

SHOW SCREEN

How important is it for you to have public gardens, parks, commons or other green spaces nearby?

SINGLE CODE ONLY

Very important Fairly important Not very important Not important at all Don't Know

ASK ALL

SHOW SCREEN

And how often do you visit public gardens, parks, commons or other green spaces?

SINGLE CODE ONLY

6-7 days a week 3-5 days a week 1-2 days a week

382

Once a fortnight Once a month Several times a year Once a year Less often Never Don't Know

ASK ALL

SHOW SCREEN

What are the three most important reasons for you spending time in public gardens, parks, commons or other green spaces?

CODE ALL THAT APPLY – UP TO THREE

Tranquillity Scenery Open space Fresh air Plants and wildlife Leisure opportunities Way of life

Villages / historic buildings Nothing Don't Know Other (specify)

ASK ALL SHOW SCREEN

How would you judge the current situation in each of the following?

SINGLE CODE ONLY

RESPONSE CODES

Very good Good Neither good nor bad Bad Very Bad Don’t know

STATEMENTS RANDOMISE

The economic situation in the UK

The economic situation in the World

The financial situation in your household

ASK ALL

SHOW SCREEN

And, do you think that the general economic condition of the UK will improve, stay the same, or get worse over the next 12 months?

SINGLE CODE ONLY

Improve Stay the same Get worse Don't know

ASK ALL

DO NOT SHOW SCREEN

383

What do you think are the most important issues the Government should be dealing with?

DO NOT PROMPT. CHOOSE ALL THAT APPLY

Health/Social Services Education Crime Environment/Pollution Pensions and benefits Public transport Unemployment Economy in general Housing (including costs) Taxes

European Union Don’t know Other (specify)

END OF SELF-COMPLETION SECTION

VISUAL CHECK ASK ALL

Before I ask you the last few questions – I’d like to check one more thing

Could I ask what room temperature your heating is set to now?

[answer in centigrade]

[RANGE BETWEEN 0 AND 50]

INTERVIEWER – IF RESPONDENT DOESN’T KNOW PLEASE ASK THEM TO GO AND CHECK.

WRITE IN NUMBER BELOW.

INTERVIEWER: Enter 50 if they cannot answer – i.e. have no way of controlling the temperature in the home / thermostat does not have a temperature scale

Don’t know Refused

384

I. Household and respondent characteristics

INTRODUCTION

Finally, I’d just like to ask you a few more questions about your circumstances.

ASK ALL

SHOW SCREEN

Are you a member of, or do you make regular donations to, any of the organisations on this list? CODE ALL THAT APPLY

National Trust/The National Trust for Scotland Royal Society for the Protection of Birds (RSPB) WI (Women’s Institute) Civic Trust

Wildlife Trusts WWF The Woodland Trust Christian Aid Stop Climate Chaos Oxfam British Trust for Conservation Volunteers (BTCV) Greenpeace Ramblers Association Friends of the Earth Council to Protect Rural England None of these Another organisation concerned with the environment (specify)

ASK ALL

READ OUT

In the last 12 months, have you volunteered with, given time to or taken part in any groups?

Yes No Don’t know Refused

ASK ALL WHO HAVE VOLUNTEERED (YES AT Q78) SHOW SCREEN 79.Which of the following types of group have you volunteered with, given time to, or taken part in? CODE ALL THAT APPLY

Schools Youth / Children’s Activities (outside school)

Environment/Conservation Adult education Sports/Exercise – in team, coaching or organising Religion Politics Health, Disability, Counselling and support services, Advice on welfare Safety / First Aid Animal protection Justice and Human Rights Local Community or Neighbourhood Groups Hobbies/Recreation/Arts groups Trade Union Activity Other

ASK ALL

385

Do you read any daily newspapers at least 3 times a week? INTERVIEWER: This would include any regional or local daily paper

Yes No Don’t know

ASK ALL WHO READ DAILY NEWSPAPER (YES AT Q80) SHOWN SCREEN

Which one of the following daily newspapers do you read most often? SINGLE CODE ONLY

Daily Express Daily Mail Daily Mirror Daily Star Daily Telegraph Financial Times The Guardian The Independent The Sun

The Times Metro London Lite The London Paper Regional/local daily paper Other daily newspaper None of these

ASK ALL

Do you read any Sunday newspapers at least twice a month?

Yes No Don’t know

ASK ALL WHO READ SUNDAY NEWSPAPER (YES AT Q82) SHOWN SCREEN

And which of these Sunday newspapers do you read most often? SINGLE CODE ONLY

News of the World Mail on Sunday Sunday Express Sunday Mirror Daily Star Sunday The People Sunday Times The Observer Sunday Telegraph Independent on Sunday Regional/local Sunday newspaper Other Sunday newspaper

None of these

ASK ALL

SHOW SCREEN

To which of these groups do you consider you belong?

SINGLE CODE ONLY

White-British

White–Irish

White – other white background

386

Mixed – White and Black Caribbean

Mixed – White and Black African

Mixed – White and Asian

Mixed – any other Mixed background

Asian or Asian British – Indian

Asian or

Black or

Chinese

Other (specify)

Don’ know

Refused

Asian British – Pakistani Asian British – Bangladeshi Asian British – other Asian background Black British – Caribbean Black British – African Black British – other Black background

ASK ALL

SHOW SCREEN

I am now going to ask you about your household income. I only need to know an approximate amount, to see if this influences people’s views and experiences.

Please can you tell me your overall HOUSEHOLD income from all sources in the last year? This includes earnings from employment or self-employment, income from benefits and pensions, and income from other sources such as interest and savings.

387

Please look at the screen and tell me which option represents your TOTAL HOUSEHOLD INCOME in the last year from all sources BEFORE tax and other deductions.

Annual Monthly Weekly

1 Under £2,500 Under £200 Under £50

2 £2,500 - £4,999 £200 - £399 £50 - £99

3 £5,000 - £9,999 £100 - £199 £400 - £829

4 £10,000 - £14,999 £200 - £289 £830 - £1,249

5 £15,000 - £19,999 £1,250 - £1,649 £290 - £389

6 £20,000 - £24,999 £1,650 - £2,099 £390 - £489

7 £25,000 - £29,999 £490 - £579 £2,100 - £2,499

8 £30,000 - £34,999 £580 - £679 £2,500 - £2,899

9 £35,000 - £39,999 £680 - £769 £2,900 - £3,349

388

10 £40,000 - £44,999 £770 - £869 £3,350 - £3,749

11 £45,000 - £49,999 £3,750 - £4,149 £870 - £969

12 £50,000 -£59,999 £4,150 - £4,999 £970 - £1,149

13 £60,000 - £74,999 £5,000 - £6249 £1,150 - £1,449

14 £75,000 - £99,999 £6,250 - £8,299 £1,450 - £1,919

15 £100,000 or more £1,920 or more £8,300 or more

16 Don’t know (HIDDEN CODE)

17 Refused (HIDDEN CODE)

ASK ALL

SHOW SCREEN

Which, if any, of these state benefits are you currently receiving in your own right?

ADD IF NECESSARY: That is where you are the named recipient

CODE ALL THAT APPLY

Unemployment related benefits, or National Insurance Credits Income support (not as an unemployed person) Sickness or disability benefits (not including tax credits) State Pension

389

Family related benefits (excluding Child Benefit and tax credits) Child benefit Cold weather payment

Housing, or Council tax benefits Tax credits Other (specify) None of these

Don’t know Refused

ASK ALL

READ OUT

Do you have any qualifications...?

CODE ALL THAT APPLY

From school college or university Connected with work (e.g. on the job training, apprenticeship) From Government schemes/programmes Don't Know No qualifications

ASK ALL WITH QUALIFICATIONS (Q87=YES)

SHOW SCREEN

From this list, please tell me the highest academic qualification that you have obtained?

PROMPT AS NECESSARY. SINGLE CODE ONLY PRIORITY CODE: IF TWO OR MORE QUALIFICATIONS HELD, CODE THE ANSWER WHICH IS HIGHER UP THE LIST.

Higher degree, e.g. MSc, MA, MBA, PGCE, PhD First degree, e.g. BSc, BA, BEd, MA at first degree level GCE 'A'-level / SCE Higher Grades (A-C) GCSE grade A-C / GCE 'O'-level passes / CSE grade 1 / SCE O Grades (A- C) / SCE Standard Grades (1-3) / School Certificate / Matriculation GCSE grade D-G / CSE grade 2-5 / SCE O Grades (D-E) / SCE Standard Grades (4-7) / SCOTVEC National Certificate Modules Other academic qualifications (PLEASE DESCRIBE) None of these Refused

ASK ALL WITH QUALIFICATIONS (Q87=YES)

SHOW SCREEN

And please tell me the highest qualification on this list you have obtained?

390

PROMPT AS NECESSARY. SINGLE CODE ONLY PRIORITY CODE: IF TWO OR MORE QUALIFICATIONS HELD, CODE THE ANSWER WHICH IS HIGHER UP THE LIST.

Level 1 NVQ/SVQ / Foundation GNVQ/GSVQ Level 2 NVQ/SVQ / Intermediate GNVQ/GSVQ Level 3 NVQ/SVQ / Advanced GNVQ/GSVQ

Level 4 NVQ/SVQ Level 5 NVQ/SVQ NVQ/SVQ - not sure what level BEC (General) / BTEC (General) / City & Guilds Craft or Ordinary level / RSA Diploma ONC/OND / BEC (Higher) / TEC (Higher) / BTEC (Higher) / RSA Advanced Diploma Other vocational or pre-vocational qualification (PLEASE DESCRIBE) No, none of these Refused

ASK ALL

SOCIAL GRADING QUESTIONS

Who would you say is the chief income earner in this household?

IF RELATED: IF TWO EQUAL INCOMES TAKE THE ELDER PERSON IF LIVING AS MARRIED TREAT AS MARRIED AND THEREFORE RELATED IF UNRELATED: TAKE THE RESPONDENT AS CHIEF INCOME EARNER

Respondent Someone else

I would now like to ask you about your/their current or most recent job. STANDARD SOCIAL GRADING QUESTIONS INTERVIEWER: CODE RESPONDENT’S SOCIAL GRADE.

A B C1 C2 D E

ASK ALL

A certain number of interviews on any survey are checked by a supervisor to make sure that people were satisfied with the way the interview was carried out. In case my supervisor needs to contact you it would be helpful if we could have your telephone number.

Record number

Number refused

No phone

ASK ALL 391

It is possible that we may want to contact you again for additional information. Would you be willing to be contacted again?

Yes – willing to be re-contacted No – not willing to be re-contacted

ASK ALL

If additional information was being collected for Defra or the Energy Saving Trust by another research organisation, would you be willing for TNS to pass your name, contact details and information from this survey to another research organisation so they could contact you?

IF NECESSARY: This survey was on behalf of the Department for the Environment Food and Rural Affairs (Defra) and the Energy Saving Trust

Yes – willing for details to be passed on No – not willing for details to be passed on

392

Appendix b)

Assessing the Structure of UK Environmental Concern and its

Association with Reported Pro-Environmental Behaviour.

(Accepted by the Journal of Environmental Psychology)

Abstract

Understanding the structure and composition of environmental concern is crucial to the study of society’s engagement with environmental problems.

Here, we aim to determine if components of the VBN model emerge when applying a combination of exploratory and confirmatory factor analysis to a large UK dataset, one designed without a priori commitment to a theoretical model. A three-factor model was confirmed to be the most substantively and methodologically optimal. Two of the factors correspond to the VBN’s ecocentric and anthropocentric factors. However, the third factor does not routinely map onto the third factor of the VBN (ecocentric concern). We have called our factor ‘denial’, as high scorers tend to be responding positively to statements that would suggest inaction. The association between these factors and level of reported pro-environmental behaviour is assessed.

393

Introduction

As a psychological phenomenon, concern for the environment has been continuously investigated for four decades. Its study has provided a greater understanding of how individuals relate to their environment as well as the comprehension and possibly inclination towards pro-environmental behaviour.

When research began in the 1970s, the emphasis was on how environmental concern (EC) could spread from small groups of environmental activists, to the wider population (e.g. Buttel & Flinn, 1974, 1978; Dunlap, 1976 and Dunlap &

Gale, 1972 amongst others). Later, research on environmentalism turned inward, focusing on the examination and conceptualisation of environmental attitudes, driven by the need to understand attitudinal and behavioural engagement with environmental issues. In terms of practical application, understanding public engagement with ‘the environment’ might lead to more targeted and nuanced environmental policies, encouraging greater concern and increasing the frequency of pro-environmental behaviour. Yet achieving such benefits presumes an understanding of what environmental concern is – and therein lies the focus of this paper.

In the literature, EC is taken to broadly refer to the degree to which people are aware of problems regarding the environment, their support of efforts to solve such problems and a willingness to contribute personally to their solution (Dunlap & Jones 2002, p. 485). This definition rightly indicates that EC is a very broad concept covering a wide range of phenomena with multiple aspects and dimensions (see also Xiao & Dunlap 2007; Alibeli & White

2011). Both Dunlap and Jones (2002) and Klineberg et al. (1998) emphasise 394

that the broad definition of EC implicitly requires researchers to: “think clearly at the outset about what aspects or facets of environmental concern they want to measure, and then carefully conceptualize them prior to attempting to measure them” (Dunlap & Jones 2002, p. 515), thus avoiding further ambiguity in concept definition and variations or errors in variable measurement.

EC is largely considered to be attitudinal in nature. Minton and Rose

(1997) conceptualise EC as constructed from a broad range of environmental attitudes. Similarly, Vining (1992) treats EC and environmental attitudes as synonymous, defining EC as the development of an array of attitudes toward the environment. However, there is only weak consensus on the specific structure of these attitudes and as such the composition of EC varies across studies. Furthermore, on-going EC research is called for due to the continuously changing nature of both environmental problems and the relationship of the human population to these. As the effects of climate change are experienced and perceived in different ways by different people in different countries and mediated by a host of differing factors, attitudes are likely to change in unpredictable ways. To reiterate Stern et al. (1995), “Although it is safe to expect many newly described environmental conditions to take form as social attitude objects, it is not easy to predict what form they will take, what attitude will form about them, or whether public opinion will be of one mind or be fragmented” Stern et al. (1995, p. 1612). Without greater clarification of the structure and composition of EC in any given study, a clear understanding of attitudinal and behavioural engagement with current environmental issues is unlikely to emerge.

395

Many researchers such as Yin (1999), Cottrell (2003) Schultz et al.(2004), and Milfont and Duckitt, (2010) adhere to the orthodox three-component attitude model as an approach for specifying the broad structure of environmental attitudes. However, some contemporary attitude theorists hold that cognition, affect and behaviour form the basis of evaluations of particular psychological objects. For example, Albarracin et al. (2005) states “affect, beliefs and behaviours are seen as interacting with attitudes rather than as being their parts” (p.5). This contemporary approach suggests that attitudes should be conceptualised as evaluative tendencies that can both be inferred from and have an influence on beliefs, affect and behaviour.

A combination of these two theoretical perspectives is used in this study.

Here, EC is considered to be a concept that consists of cognitive and affective components, with which behaviour interacts but is not a part of. Our position is that with EC - as with many other attitudinal constructs - there are many mediating and moderating influences between the internal, latent concern and the outward environmental behaviour and therefore it seems most appropriate to treat behaviour and attitudes as theoretically distinct. However, we also want to make a theoretical statement about what EC is. Concern in its relational sense expresses beliefs about negative states or potential outcomes and is associated with specific affective states such as fear and worry. Schultz et al. (2005: 458) “use the term environmental concern to refer to the affect

(i.e., worry) associated with beliefs about environmental problems”. We too aim to incorporate this affect component in our definition of EC.

396

The NEP

There is an extraordinary number of measures of environmental attitudes, a fact that led Stern (1992) to describe the situation as an ‘‘anarchy of measurement’’ (p. 279). Three classic environmental attitude measures are the

Ecology Scale (Maloney & Ward, 1973; Maloney, et al., 1975), the

Environmental Concern Scale (Weigel & Weigel, 1978), and the New

Environmental Paradigm (NEP) Scale (Dunlap & Van Liere, 1978; Dunlap et al.,

2000). These three scales examine multiple phenomena or expressions of concern, such as beliefs, attitudes, intentions and behaviours, and they also examine concerns about various environmental topics, such as pollution and natural resources. Hence, according to Dunlap and Jones’ (2002) typology these measures are all multiple-topic/multiple-expression assessment techniques. Although widely used, both the ecology scale and the environmental concern scale include items tapping specific environmental topics that have become dated as new issues emerge (Dunlap and Jones,

2002, 2003). The NEP Scale avoids this issue by using only general environmental topics that do not become dated, at least in the short to medium term, to improve the psychometric properties of the scale.

The NEP Scale measures an ecocentric system of beliefs (i.e., humans as just one component of nature) as opposed to an anthropocentric system of beliefs (i.e., humans as independent from, and possibly superior to, other organisms in nature) (Bechtel, Corral-Verdugo, Asai, & Riesle, 2006; Dunlap et al., 2000), and is the most widely used measure to investigate environmental issues.

397

The original NEP Scale was published in 1978 by Dunlap and Van Liere, and consists of 12 items (8 pro–trait and 4 con–trait) responded to on a 4–point

Likert scale (anchored by strongly agree to strongly disagree). This was later updated in 2000 by Dunlap et al. (2000) who included additional items to make the scale more psychometrically sound, and updated the terminology used.

The items in both versions were developed based on both the literature on the emergence of the NEP worldview and consultation with environmental experts.

The items are intended to tap three main facets of environmental attitudes: a belief in (1) humans’ ability to upset the balance of nature, (2) the existence of limits to growth, and (3) humans’ right to rule over the rest of nature. The NEP

Scale measures the overall relationship between humans and the environment; higher NEP scores indicate an ecocentric orientation reflecting commitment to the preservation of natural resources, and lower NEP scores indicate an anthropocentric orientation reflecting commitment to exploitation of natural resources.

The VBN value frame

Since the late nineties, a second wave of EC study has asked fundamentally different questions. Rather than investigating general attitudes about environmental issues, this research seeks identify underlying values that provide the basis for environmental attitudes (e.g. Schultz & Zelezny 1999), thus moving towards a more differentiated conceptualisation of environmental attitude formation. Values are usually theorised as being relatively stable over the life course and allow individuals to subjectively judge what is important

(Slimak & Dietz 2006). By contrast, Stern et al. (2000) maintain that attitudes

398

are mutable; they can appear, disappear and change over time. One approach is to view relatively enduring value orientations interacting with more fluid contextual (and life course) factors to produce attitudes. A key theory that embodies this approach is the value-belief-norm theory described by Stern et al. (1995; 1999; Stern 2000).

The VBN, in an attempt to explain pro-environmental behaviour, links three theoretical models: norm-activation theory, the theory of personal values, and the NEP, into a unified explanation for environmentalism. It postulates that the consequences that matter in activating personal norms are those that are perceived as adverse with respect to whatever the individual values. Thus, people who value other species highly will be concerned about environmental conditions that threaten those valued objects, just as those who care about other people will be concerned about environmental conditions that threaten the other people’s health or well-being and so on. Therefore, while the VBN theory is intended to explain behaviour, embedded within it is a theory of environmental concern, specifically the NEP portion.

This NEP portion of the VBN theory postulates that values are at the core of environmental concern (Slimak & Dietz 2006) and that an individual’s value orientation is focused on the self, other people or nature. From these value orientations, corresponding attitudes of EC are then formed. More specifically – from this perspective – the EC attitude object is formed of the three components. Of these components, egoistic concerns are based on a person’s valuing himself or herself above other people and other living things.

As Stern & Dietz (1994) observe “Egoistic values predispose people to protect

399

aspects of the environment that affect them personally, or to oppose protection of the environment if the personal costs are perceived as high”.

Social-altruistic values lead to concern for environmental issues when a person judges environmental issues on the basis of costs to or benefits for other people. Biospheric EC is based on a value for all living things; regardless of any social benefits the natural environment may yield.

Much empirical research has been conducted utilizing the NEP portion of the VBN model as a theoretical framework to clarify EC composition. Results have, though, been inconsistent. For example, empirical support is mixed for a separate biospheric value orientation. Positive evidence comes from Steg et al.

(2005) who reported direct evidence for a distinct biospheric value orientation.

Social-altruism has also been distinguished from biospheric attitudes in some studies (Stern et al. 1993; Thompson & Barton 1994). However, in some factor analytic studies, social altruistic and biospheric value items have been found to load on the same factor (Schwartz 1992; Stern et al. 1995; Stern et al. 1999). A consolidation of biospheric concern with social-altruism might suggest a desire to preserve the natural environment because of the benefits this may potentially yield to society, or possibly, as Stern et al. (1995) suggest that the biospheric value orientation may be part of a more general altruistic orientation.

In another permutation, Schultz (2000; 2001) found a distinct biospheric concern, with egoistic and social-altruistic concerns combining into a single factor. This result is in line with Thompson and Barton’s (1994) proposition that environmental attitudes may be considered as having either an anthropocentric or ecocentric value focus.

400

These varied findings challenge the VBN model, in that they do not conform to the notion of three clearly separate and distinct value orientations.

Instead attitudes of EC seem to be derived from two possible dichotomised values sets as shown below in Figure 1. Both of these dichotomous value orientations allude to how individuals appreciate nature (i.e. for its intrinsic value or its potential benefits) and whether EC attitudes are based on an individual’s distinction between the individual self and the outside world, or between society and nature.

These contrasting findings and reflections raise the question of the veridical value/attitude structure for EC. In response to such inconsistencies, both Schultz (2000; 2001) and Snelgar (2006) have tested several different factor structures for EC. As shown in Table 1, Schultz (2000; 2001) tested one, two and three-factor measurement models for EC. The three-factor model

(highlighted below) was found to be both theoretically and statistically optimal: adhering to the VBN model and satisfying both the K1 and scree plot tests.

Snelgar’s (2006) later study tested a total of five models (shown in Table 2). Of the two-factor models examined, the one with a distinct biospheric component had the best fit to the data. Overall however, the best model was a four-factor structure, in which the biospheric attitude split into two separate biospheric concerns for plant and animal life.

Overall therefore, studies suggest that the biosphere is perceived to have an intrinsic value. However Snelgar’s (2006) study suggests that there is a distinction between concern for the welfare of species and the preservation of

401

the countryside, opening up the possibility of a fourth value orientation, or possibly that VBN value orientations form the basis for multiplicious attitudes.

A problem of theory driven survey design is that the instrument is not an independent tool for testing the theory. The survey instrument that has been used in many of the above studies is precisely designed to tease out the structure of EC, the likelihood of finding the NEP structure and no other is therefore greatly enhanced. Inductive, secondary data analysis of representative survey data provides at least a partial solution to this problem of circularity. If when using secondary data, which, while palpably about environmental concern is not theory-specific, one, then finds that the same structure emerges, then the evidence for theory is stronger. If it does not, then modifications to the theory should be considered.

This places the research emphasis on determining whether a population exhibit NEP / VBN components at all. Data generated without an a priori commitment to a specific theoretical framework places fewer limitations on participant responses, potentially reducing bias and allowing for results that are out with the theory. This approach thus has the potential to not only independently test theories of EC but also to possibly reveal alternative EC attitudes. This is not to argue for a purely inductive approach to research. Both inductive and deductive approaches have their strengths and weaknesses. The issue here is that research to date in this area has been heavily biased towards deductive theory testing with the inherent problem that and that the theory itself is (artificially) part of the data generating process. Some studies, have conducted exploratory research that is abductive in nature, such as Milfont

402

and Duckitt’s (2010) study on the validity of the environmental attitudes inventory. However the majority or environmental attitude research adopts a deductive approach, which we argue should be balanced with a more inductive research. Of course, there is unlikely to be no relationship between the EC-related survey items in a secondary dataset and those in that have been produced in VBN test set and indeed we are seeking to find specifically non-VBN items; the goal here is not to deliberately produce a different structure. However because the item construction is not primarily theory driven we allow differences and nuances of meaning to emerge and, as we shall see, that is precisely what happens.

A secondary problem of theory driven scale implementation is the burden placed on researchers to gather a suitable sample, ideally a representative one. Given the high demand on time and resources required to gather primary data, such a sample often cannot be obtained. For example, conclusions drawn by Schultz (2000) and (2001) cannot be generalised to their respective populations given their use of small and unrepresentative samples: both studies consisted of psychology undergraduate students from the United

States (sample of 400 and 1010 respectively). Stern seems to have specialised in idiosyncratic sampling. For example, in 1998 Dietz, Stern et al., actively dropped 10% of his sample who are in other or Jewish categories. Stern et al.1995 used random digit dialling to select 199 Virginia households. Snelgar

(2006) obtained a convenience sample of 368 participants. Of these participants, 296 were undergraduate students taking psychology modules at the University of Westminster. The remaining 72 participants were recruited with the use of snowball sampling. Snelgar acknowledges that due to these 403

sampling methods, conclusions about larger populations cannot be drawn.

Results that cannot be generalised to the wider population are diminished in value: it is uncertain whether the findings exist in the social world or if they are simply characteristics of the sample acquired.

Environmental concern and pro-environmental behaviour

For over 30 years much social research has explored the roots of direct and indirect environmental behaviour, specifically looking at the relationship between concern for the environment and pro-environmental behaviour. As mentioned in the previous section, pro-environmental behaviour is often defined as behaviour that minimizes an individual’s negative impact on the natural world (e.g. reducing energy consumption and waste production). The value-action gap, sometimes referred to as the attitude-behaviour gap (Blake,

1999), is the gap that can occur when the values or attitudes of an individual do not correlate to his or her actions. Though the extent to which attitudes affect behaviour is not as strong as logic would dictate, the disparity between the two concepts is particularly prominent when engaging with the natural environment (Kollmuss, 2002). The outcome is that there is a divergence between the high value people place on the natural environment and the relatively low level of action taken by individuals to counter environmental problems. Related research often focuses on cognitive theories of attitude formation and how this affects individuals’ behaviour, endeavouring to explain why high regard for environmental issues does not translate into behaviours to solve environmental issues (such as Cottrell, 2003). Results have thus far

404

suggested that there are many internal and external factors that affect behaviour making it difficult to identify the exact reasons why this gap exists.

While most commentators agree that there is no simple correspondence between attitudes and behaviour, different studies have posited various possible explanations for the discrepancy. Taken together, they suggest that the attitude-behaviour relationship is moderated by two primary sets of variables: external / situational constraints, and the formation of attitudes towards the environment (O'Riordan, 1981; Guagnano et al., 1995; Hallin,

1995; Baron & Byrne, 1997).

Aims

Given the above, this study aims to answer four main questions: first, can and do theoretically familiar EC constructs emerge from large scale environmental attitude and behaviour survey data without the use of strict EC scales?

Second, if so, are recognizable NEP / VBN components evident when using a nationally representative British sample, given the originally US basis of the above? As stated, data generated without an a priori commitment to a specific theoretical framework places fewer limitations on participant responses and more fully allows for results that are outwith the model. Exploratory, inductive research thus has the potential to not only independently test theories of EC but also to reveal if there are alternative EC attitudes. Thirdly, what is the value of an ontological distinction between attitudes and reported behaviours in this context? Fourthly, returning to a long-standing theme in the literature, how do

405

environmental attitudes relate to behaviour in such a dataset? Do environmental attitudes influence reported pro-environmental behaviour?

Analysis

The results are divided into two parts, with corresponding methods and analysis. First, the optimal number of factors is determined through examination of factor retention criteria, providing structure for the EC model.

The fit of the model is then confirmed before the model factors are interpreted.

Second, regression analysis is performed to examine how scores from model factors affect the frequency of reported environmentally friendly behaviour.

Part 1

Data

Analysis is conducted with data from DEFRA’s ‘Survey of Public Attitudes and

Behaviours towards the Environment’ (hereafter EAS – Environmental Attitudes

Survey). DEFRA (Department for Environment, Food and Rural Affairs) is the

UK government department responsible for policy and regulations on environmental, food and rural issues. The 2009 wave of EAS is used, with a sample size of 2929 British participants. Data was gathered using quota sampling via face to face interviews and a two stage stratified sample design.

Interviews were carried out using census output areas as sampling units.

Census output areas are small, homogeneous areas, comprising about 125 –

406

150 households (See Vickers and Rees, 2007 for a description of the creation of the Office for National Statistics output area classification). Output areas were also stratified by socio-economic variables within region, to ensure a representative sample of all areas. Finally, quotas were applied in each output area to control for likelihood of selected respondents being at home.

These quotas were set on sex, working status and presence of children in the household. Using demographic quotas effectively forms a second level of stratification. Interviewers worked between 2pm and 8pm on weekdays and at weekends to further minimise the response bias which is introduced by only working during standard working hours. Residual non-representativeness is dealt with through the use of population and design weights.

The EAS dataset is explicitly divided into three sections: Household and

Respondent Characteristics, Environmental Behaviours, and Environmental

Attitudes. Variables for this analysis were as such selected from those explicitly defined by the dataset as reflections of environmental attitudes. These items were developed to measure British public attitudes towards the environment, without commitment to one specific theoretical framework.

Selection was based on our theoretical understanding of EC, that it is as primarily a cognitive and affective state. Based on this understanding of EC, we independently reviewed the selection and excluded variables that were not compatible with this understanding of EC. Environmental attitude statements that were in part behavioral – that is, statement which commented on the execution, frequency or opinion of environmental behavior – were excluded, so maintaining an ontological divide between attitude and behavior. Statements

407

that remarked on the willingness of participants to incur a financial penalty for engaging in environmentally detrimental activities or pay an increased price for comparatively environmentally friendly products were also excluded.

Responses to such statements are indicative of participant willingness to dispense with monetary resources in order to achieve a positive effect (or avoid a negative effect) on the environment. Consequently, responses are potentially influenced by participant income or wealth. To include such variables would be to introduce additional variance into the analysis – constraining EC and potentially producing results relating to income or wealth.

It is possible that such variables do have a relationship with environmental concern but they are likely prior rather than constitutive. What remains are raw belief statements un-moderated by extraneous variables. These variables were derived from responses to the statements shown in

the analysis – constraining EC and potentially producing results relating to income or wealth. Undoubtedly such variables do have a relationship with environmental concern but they are almost certainly confounding. What remains are raw belief statements unmoderated by extraneous influences.

These items were derived from responses to the statements shown in Table 8 with which participants indicated levels of agreement on a five-point Likert scale.

409

Table 8Methods

A combination of Exploratory Factor Analysis (EFA) and Confirmatory Factor

Analysis (CFA) methods are used. The software package employed was

MPLUS. Deciding upon the optimal number of factors to be retained from EFA is crucial. It is important to distinguish between major and minor factors; specifying too few or too many can distort results. There is no clear consensus for factor retention criteria. The most commonly used method is known as the

K1 rule, which retains factors with eigenvalues greater than 1 (see Kaiser

1960). Another, less sophisticated method for retaining factors is through the examination of Cattells (1966) scree plot for breaks and discontinuities, only retaining factors above a significant inflection. This method suffers from subjectivity and ambiguity, particularly if there is no clear inflection.

A third method is Parallel Analysis (PA), which uses random data with the same number of observations and variables as the original data (see Fabrigar et al. 1999; Hayton et al. 2004). The correlation matrix of random data is used to compute eigenvalues; these eigenvalues are then compared to the eigenvalues of the original data. The optimum number of factors is the number of the original data eigenvalues that are larger than the random data eigenvalues. This method adjusts for sampling error and is a sample-based alternative to the K1 rule and scree plot examination. In most studies, one or two of these methods are used, however in this analysis all three are used to ensure the best possible model fit and accurate interpretation of retained factors.

410

The production of factors through the use of EFA is generally followed by their rotation so as to improve their interpretability and to simplify the factor structure (Thurstone 1935 p.1947). Oblique rotation is used as it allows factors to correlate and given that factors within this model form the EC attitude object, it is highly likely that they will correlate. The maximum likelihood EFA fitting procedure is used for this analysis. Though most research typically uses

Principal Components Analysis (PCA) or Primary Axis Factoring (PAF) methods of EFA, maximum likelihood allows researchers to test for the statistical significance of and correlations between factors, as well as generating goodness of fit statistics. Gorsuch (1990) has shown important differences between PCA and common factor solutions such as principal axis and maximum likelihood factoring. In such cases, the evidence favours the common factor model as the more accurate. Conway and Huffcutt (2003) therefore urge researchers to make greater use of common factor model approaches (maximum likelihood in particular due to the fit indices that can be used to help determine the number of factors).

Once the optimal number of factors is established and a factor model is generated, this factor structure is specified and tested through CFA.

Modification Indices are used to ensure that there are no additional cross loadings that should be accounted for in the model. Goodness of fit indices are also examined to determine how well this model fits the data. Various goodness of fit indices exist and reporting them all would be a hindrance to interpreting the validity of the model. The main index is the chi-square, which should always be reported as it shows the difference between expected and observed covariance matrices (Hu & Bentler 1999). According to various 411

studies (Hu & Bentler 1999; MacCallum et al. 1996; Yu 2002) the TLI, CFI, and

RMSEA indices should also be reported alongside the chi-square statistic.

Results

Three factor retention criteria are implemented to determine optimal number of factors. The scree plot shows no single point of inflection and appears to suggest the retention of two-four factors. According to K1 factor retention criteria, factors generated with an eigenvalue >1.0 are to be retained. Parallel analysis produces eigenvalues from randomly generated parallel data. If eigenvalues from this parallel data are smaller than those from the original data, then this is indicative of an optimal model. As shown in Table 4, both the

K1 and parallel analysis methods emphasise a three-factor model.

The rotated factor loadings of the three-factor model are displayed in Table 5.

Variables with a coefficient above minimum criteria of .3 are highlighted to indicate their contribution to that factor. CFA was performed to test this factor structure. High loading variables (coefficient >.3) were allowed to load freely onto their respective factors; all other loadings were restricted to 0. The final model displayed in Table 6 reports variable loadings for the CFA model as well as correlations between factors. Maximum likelihood method of parameter estimation was used to produce this Table 6. These factors are named and interpreted below.

412

Factor 1 – Denial

This factor contains the following key components:

• Scepticism (positive loading of the exaggerated crisis variable)

• Belief that there is no need to respond to environmental problems at

present (positive loading of both the low priority and too far in the future

variables)

• The belief that it is not too late to do something about the environment,

that problems can be controlled is necessary (negative loading of the

control variable)

Factor 2 – Human-Centric Concern

The Over Populated and Limited Resources variables together

indicate an EC with respect to the human population, specifically

their impact on the planet and its ability to sustain them.

Factor 3 – Ecocentric Concern

This factor demonstrates a distinct ecocentric component, capturing

concern for both animal species and countryside.

Figure 3 shows the final diagram and its goodness of fit statistics. The CFI and

TLI are both >0.9, the RMSEA is <.05, and SRMR is <.08, all of which indicate that the model is a good fit for the data.

413

Part Two

Data

The environmental behaviour portion of the EAS contains items relevant to four categories of behaviour; food, home, travel and recycling. From of these behavioural categories, 2 – 6 items are selected. The selected variables not only capture environmental behaviour but have a low portion of missing cases.

Some items have a high portion of missing cases as they attempt to capture a form of environmentally friendly behaviour that is conditional upon the participant owning property and / or owning land i.e. composting, growing vegetables and buying household appliances. For each question, participants indicate the level at which the behaviour in question is performed on a 5 point likert scale.

Results

Ordered logistic regression analysis was performed to determine how environmental attitudes are associated with reported environmental behaviour.

16 measures of pro-environmental behaviour are used to make this assessment. Factors scores for the environmental attitudes displayed in Figure

3 were entered simultaneously as independent variables into each regression, with one of the behaviour measures as the dependant variable. This produced a total of 16 regressions, the results of which are displayed in Table 7. Age and gender were accounted for in each model.

414

Table 8 shows that human-centric concern is not a significant predictor of concern, but that both ecocentric and denial attitudes are largely significant predictors of reported environmental behaviour. Thus greater ecocentricity is associated with an increase in the frequency of environmental behaviour, while higher denial is associated with a decrease in the frequency of reported environmental behaviour.

Discussion

The Model of Environmental Concern

We have been concerned here with uncovering latent components of EC from the EAS dataset, a large, nationally representative British dataset complied from a survey without an explicit, particular theoretical basis. Indicator variables corresponding to our specified theoretical understanding EC were selected from environmental attitudes section of the dataset. Both exploratory and confirmatory factor analysis were performed and a three-factor model of

EC was produced. This model has similarities with those produced by Stern and Dietz (Stern & Dietz 1994) as well as some important differences.

Regarding these differences, the denial factor could be interpreted as mapping onto the egoistic component of the VBN; indeed, previous research has suggested a relationship between egoistic value orientation and denial

(Hansla et al. 2008). It could be that the drivers of denial and those of concern co-occur, and it would certainly be an interesting study to establish if this was so. Here we simply suggest that those who score highly on this factor may be

415

exhibiting a form of denial or resignation, where an expressed lack of EC is used as a coping mechanism in the face of numerous environmental problems.

Factor two corresponds to the social altruistic component of the VBN model. The variable loadings of this factor suggest recognition of society’s environmental impact, though the focus is on Earth’s ability to continue meeting growing needs of this population. Due to the limitations of the data, this factor is not altruistic in the sense intended by Stern (1994): the variables that have loaded onto this factor appear to indicate a concern for the Earth’s ability to continue meeting the needs of human society rather than a concern for the welfare of society. Therefore due to the lack of solely altruistic variables in the EAS data, this factor has been labelled here as Human Centric. The final factor reflected an ecocentric concern in that it concerns the impacts on the non-human parts of the biosphere.

Overall therefore, the factors extracted do broadly align with the VBN model, though the denial factor does need to be considered in more detail.

What is of interest is the significant minority of respondents who record high scores on both the denial factor and one or both of the other factors. As

Table 8 shows, that whilst over 50% of the sample are consonant with how one might expect the factors to relate psychologically, 9.5% of the sample are high scorers (above the mean) on the denial factor whilst also being high scorers on both of the other factors. This appears paradoxical since such respondents are both expressing and denying concern. This would tend to route interpretations away from a simplistic equation of denial and egocentrism

416

suggested above. Though it is too soon to determine if this is an improvement on the VBN, further examination of combinations of the different varieties of EC and their relationship with the VBN theoretical model is required. From a policy making point of view, understanding the holders of such apparently contradictory beliefs might be important in achieving a shift in norms away from relative lethargy to proactive concern.

Reflections on The Data

Defra’s Environmental Attitude Survey is intended to measure environmental attitudes, norms, values and behaviours, including barriers to pro- environmental behaviour. The survey is not intended to embody a particular theoretical commitment but nonetheless does appear to be influenced by the dominant models. The results produced from the analysis of the EAS provide broad support for the VBN, in that the factors found could conceivable be attitudes of EC derived from the three value orientations outlined by the VBN.

Our analysis suggests that there is value to this dataset in terms of its ability to characterise EC in the UK. However, while the 2009 EAS is part of a series of public attitude surveys run by DEFRA, data from the majority of previous waves can no longer be obtained by the commissioning government department and those cohorts for which data is available data has been conducted rarely and infrequently. In light of this and the nuances in the results presented here and meriting follow-up work, we would recommend that serious consideration be given to longitudinal maintenance of the EAS.

Longitudinal methods of data analysis are particularly appropriate, given that

417

attitudes are subject to change, particularly environmental attitudes, as previously noted (Stern, 2000).

Further Research

There are initially two ways in which the work presented here could be extended. First, alternative statistical methods could be employed. Bayesian

Structural Equation Modelling (BSEM) is a new method of performing CFA, one more nuanced and reflective of the data. The approach uses Bayesian estimation and prior information from EFA to increase the variance of certain cross-loadings while keeping the mean at 0. However, factor analysis more broadly is only one possible method for analysing and understanding EC.

Using factor analysis imposes the assumption that attitudes are continuous in nature and exist on a scale. The strength of an individual’s attitude is dictated by the position on the scale. Individuals can therefore hold a combination of attitudes in varying quantities. An alternative approach is to assume that attitudes towards a particular concept have a higher level of mutual exclusivity.

Or that value, given their high level of stability, can be used as a classification system. In either case, individuals could potentially be segmented according to their attitudes and / or values. If this were the case, a better method of analysing EC may be Latent Class Analysis (LCA). LCA models identify a categorical latent variable measured by a number of observed response variables. The objective is to categorize people into classes using the observed items, and identify items that best distinguish between classes.

418

A second means of extending the work is through qualitative research. It is acknowledged that quantitative methods of analysis may not be able to fully capture all aspects of EC. Qualitative research could provide a greater level of insight into the mechanisms of EC and justifications for why sections of the population adhere to the EC components uncovered in the paper. In particular, it would seem of value to investigate the psychological processes leading to a high score on the denial factor.

Conclusion

In this paper we have examined the concept of environmental concern and its structure through inductive means. We have initially defined concern as based on a two-component attitudinal model based upon relevant affect, keeping behaviour theoretically distinct. Through a series of analyses of a representative sample of UK residents focusing particularly on those questions that express environmental concern defined as above, a three-factor solution emerges. That structure has some overlap with the VBN model of environmental concern in that two of the factors correspond to the VBN’s ecocentric and anthropocentric factors. To the extent that these emerge from a different set of items to those contained in the NEP questionnaire this can be interpreted as an affirmation of that component the VBN model. Though the third factor (denial) may not routinely map onto the third factor of the VBN

(egocentric concern), some previous research suggests that denial is related to an egoistic/self-enhancement value orientation (Hanlsa et al. 2008). While a psychological relationship between egocentrism and denial is intriguing, it is

419

not one that we are able to explore directly here, but which merits follow-up work.

We also found that ecocentric concern was predictive of increased reported pro-environmental behaviour, and that denial has a negative relationship with said behaviour. Human-centric concern is not a significant predictor of reported pro-environmental behaviour. Therefore, results indicate that those in denial of environmental problems are less likely to engage in pro- environmental behaviour, and of those who are concerned about the natural environment, it is only concern regarding plants and animal species (rather than the welfare of the human-race) which motivates pro-environmental behaviour. The different relationships between human-centric and ecocentric concern, with reported pro-environmental behaviour, merits further work. For example, this is somewhat suggestive that a more advanced moral development such as an individual at Kohlberg’s (1981) principles stage is required before belief is translated into action, but gain this is speculation and would require different data to what we have available here. Further work to address these questions directly is needed

420

References

Alibeli, M., & White, N. (2011). The structure of environmental concern.

International Journal of Business and Social Science, 2(4), 1-8.

Attfield, R. (1981). The good of trees. The Journal of Value Inquiry, 15(1), 35-54.

Baron, R. A., & Byrne, D. (1997). Social psychology. 8 utg. London: Allyn &

Bacon.

Blake, J. (1999). Overcoming the value-action gap in environmental policy:

Tensions between national policy and local experience. Local Environment,

4(3), 257-27

Buttel, F. H., & Flinn, W. L. (1974). The structure of support for the environmental movement, 1968-1970. Rural Sociology.

Buttel, F. H., & Flinn, W. L. (1978). Social class and mass environmental beliefs

A reconsideration. Environment and Behavior, 10(3), 433-450.

Cattell, R. B. (1966). The scree test for the number of factors. Multivariate

Behavioral Research, 1(2), 245-276.

Conway, J. M., & Huffcutt, A. I. (2003). A review and evaluation of exploratory factor analysis practices in organizational research. Organizational Research

Methods, 6(2), 147-168.

421

Cottrell, S. P. (2003). Influence of sociodemographics and environmental attitudes on general responsible environmental behavior among recreational boaters. Environment and Behavior, 35(3), 347-375.

Duhem, P. (1906). La théorie physique: Son objet et sa structure. Paris:

Chevalier and Rivière.

Dunlap, R., & Jones, J. (2002). Environmental concern: Conceptual and measurement issues. In D. Dunlap & M. Michelson (Eds.), Handbook of environmental sociology (pp. 482 – 542). London: Greenwood Press.

Dunlap, R. E., & Dillman, D. A. (1976). Decline in public support for environmental protection evidence from a 1970-1974 panel study. Rural

Sociology, 41.

Dunlap, R. E., & Gale, R. P. (1972). Politics and ecology: A political profile of student eco-activists. Youth and Society, 3(4), 379-397.

442.

Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999).

Evaluating the use of exploratory factor analysis in psychological research.

Psychological Methods, 4(3.272-299), 272-299.

422

Feyerabend, P. K. (1962). Explanation, reduction and empiricism.

Fox, N., & Ward, K. (2008). Health, ethics and environment: A qualitative study of vegetarian motivations. Appetite, 50(2), 422-429

Gagnon Thompson, S. C., & Barton, M. A. (1994). Ecocentric and anthropocentric attitudes toward the environment. Journal of Environmental

Psychology, 14(2), 149-157.

Gorsuch, R. L. (1990). Common factor analysis versus component analysis:

Some well and little known facts. Multivariate Behavioral Research, 25(1), 33-

39.

Guagnano, G. A., Stern, P. C., & Dietz, T. (1995). Influences on attitude- behavior relationships: A natural experiment with curbside recycling.

Environment and Behavior, 27(5), 699-718.

Hallin, P. O. (1995). Environmental concern and environmental behavior in foley, a small town in minnesota. Environment and Behavior, 27(4), 558-578.

Hansla, A., Gamble, A., Juliusson, A., & Gärling, T. (2008). Psychological determinants of attitude towards and willingness to pay for green electricity.

Energy Policy, 36(2), 768-774.

Hayton, J. C., Allen, D. G., & Scarpello, V. (2004). Factor retention decisions in exploratory factor analysis: A tutorial on parallel analysis. Organizational

Research Methods, 7(2), 191-205. 423

Hu, L. –T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural

Equation Modeling: A Multidisciplinary Journal, 6(1), 1-55.

Kaiser, H. F. (1960). The application of electronic computers to factor analysis.

Educational and Psychological Measurement.

Klineberg, S. L., McKeever, M., & Rothenbach, B. (1998). Demographic predictors of environmental concern: It does make a difference how it’s measured: Research on the environment. Social Science Quarterly, 79(4), 734-

753.

Kohlberg, L. (1981). The philosophy of moral development: Moral stages and the idea of justice (essays on moral development, volume 1). San Fancisco:

Harper & Row.

Kollmuss, A., & Agyeman, J. (2002). Mind the gap: Why do people act environmentally and what are the barriers to pro-environmental behavior?

Environmental Education Research, 8(3), 239-260.

Kuhn, T. (1962). The structure of scientific revolutions (2nd ed.). Chicago:

University of Chicago Press.

MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and determination of sample size for covariance structure 424wedish424r.

Psychological Methods, 1(2), 130-149.

424

Merchant, C. (1992). Radical ecology. New York: Routledge.

Milfont, T. L., & Gouveia, V. V. (2006). Time perspective and values: An exploratory study of their relations to environmental attitudes. Journal of

Environmental Psychology, 26(1), 72-82.

Milfont, T. L., & Duckitt, J. (2010). The environmental attitudes inventory: A valid and reliable measure to assess the structure of environmental attitudes.

Journal of Environmental Psychology, 30(1), 80-94

Minton, A. P., & Rose, R. L. (1997). The effects of environmental concern on environmentally friendly consumer 425wedish425r: An exploratory study.

Journal of Business Research, 40(1), 37-48.

Naess, A. (1984). A defence of the deep ecology movement. Environmental

Ethics, 6(3), 265-270.

O'Riordan, T. (1981). Environmentalism. In Environmentalism. Pinon limited.

Oberheim, E., & Hoyningen-Huene, P. (2009). The incommensurability of scientific theories.

Schultz, P.W., & Zelezny, L. (1999). Values as predictors of environmental attitudes: Evidence for consistency across 14 countries. Journal of

Environmental Psychology, 19(3), 255-265.

425

Schultz, P. W., (2000). Empathizing with nature: Toward a social-cognitive theory of environmental concern. Journal of Social Issues, 56, 391-406.

Schultz, P. W., (2001). The structure of environmental concern: Concern for self, other people, and the biosphere. Journal of Environmental Psychology,

21(4), 327-339.

Schwartz, S. H. (1992). Universals in the content and structure of values:

Theoretical advances and empirical tests in 20 countries. Advances in

Experimental Social Psychology, 25(1), 1-65.

Slimak, M. W., & Dietz, T. (2006). Personal values, beliefs, and ecological risk perception. Risk Analysis, 26(6), 1689-1705.

Snelgar, R. S. (2006). Egoistic, altruistic, and biospheric environmental concerns: Measurement and structure. Journal of Environmental Psychology,

26(2), 87-99.

Steg, L., Dreijerink, L., & Abrahamse, W. (2005). Factors influencing the acceptability of energy policies: A test of VBN theory. Journal of Environmental

Psychology, 25(4), 415-425.

Stern, P. C. (2000). New environmental theories: Toward a coherent theory of environmentally significant 426wedish426r. Journal of Social Issues, 56(3), 407-

424.

426

Stern, P. C., & Dietz, T. (1994). The value basis of environmental concern.

Journal of Social Issues, 50(3), 65-84.

Stern, P. C., Dietz, T., & Kalof, L. (1993). Value orientations, gender, and environmental concern. Environment and Behavior, 25(5), 322-348.

Stern, P. C., Dietz, T., Abel, T., Guagnano, G. A., & Kalof, L. (1999). A value- belief-norm theory of support for social movements: The case of environmentalism. Human Ecology Review, 6(2), 81-98.

Stern, P. C., Kalof, L., Dietz, T., & Guagnano, G. A. (1995). Values, beliefs, and proenvironmental action: Attitude formation toward emergent attitude objects1.

Journal of Applied Social Psychology, 25(18), 1611-1636.

Thurstone, L. L. (1935). The vectors of mind.

Thurstone, L. L. (1947). Multiple factor analysis.

Vickers, D., & Rees, P. (2007). Creating the UK national statistics 2001 output area classification. Journal of the Royal Statistical Society: Series A (Statistics in Society), 170(2), 379-403.

Vining, J., & Ebreo, A. (1992). Predicting recycling 427wedish427r from global and specific environmental attitudes and changes in recycling opportunities1.

Journal of Applied Social Psychology, 22(20), 1580-1607.

427

Wurzinger, S., & Johansson, M. (2006). Environmental concern and knowledge of ecotourism among three groups of 428wedish tourists. Journal of Travel

Research, 45(2), 217-226.

Xiao, C., & Dunlap, R. E. (2007). Validating a comprehensive model of environmental concern cross-nationally: A us-canadian comparison*. Social

Science Quarterly, 88(2), 471-493.

Yu, C. –Y. (2002). Evaluating cutoff criteria of model fit indices for latent variable models with binary and continuous outcomes. Thesis.

428

Figure 1: Representation of components within the VBN theory of

environmentalism (Stern et al. 1999)

Behaviours

Activism Values Beliefs Personal Biospheric Non-activist Adverse Norms Perceived public-sphere Ecological consequences ability to Sense of obligation behaviours worldview for valued Altruistic reduce threat to take pro- (NEP) objects (AR) environmental (AC) Egoistic action. Private-sphere behaviours

Behaviours in organizations

Figure 2: Dichotomous Value Orientations

Thompson and Barton (1994) Stern (1995) Ecocentric Anthropocentric General Altruistic Individualistic

egoistic biospheric biospheric egoistic social- social- altruistic altruistic

429

Figure 3: Diagram of the Final E nvironmental Concern Model and Goodness of

Fit Indices

e1 e2 e3 e4 e5 e6 e7

Too Far in Beyond Exageratted Major Limited Over Low Priority the Future Control Crisis Disaster Resources Populated

Human-Centric Denial Concern

e8 e9

RMSEA 0.037 Changes to Loss of Animal CFI 0.982 Countryside Species TLI 0.947

SRMR 0.015 Ecocentric Chi-2 61.041 Concern df 12

430

Table 1: Environmental Concern Models Tested by Schultz (2000, 2001)

Model 1 One-factor model: Uni-dimensional EC

Two-factor model: Biospheric items loading onto one factor

with both egoistic and altruistic items loading on another factor. Model 2 This is consistent with Thompson and Barton’s (1994)

classification of environmental attitudes.

Three-factor model: Egoistic, altruistic, and biospheric

Model 3 concerns fitted the data well, providing support for the notion

that three value-orientations underlie EC.

431

Table 2: Environmental Concern Models suggested by Snelgar (2006)

Model 1 One-factor model: Uni-dimensional EC.

Two-factor model: Egoistic items load onto one factor, both

altruistic and biospheric items load onto a second. This is based Model 2 on Stern et al.’s (1995) suggested that biospheric value may be

part of a general-altruistic cluster.

Two-factor model: Egoistic and altruistic items load onto one

factor, biospheric load onto a second. This provided a better fit of Model 3 the data than Model 2, supporting Thompson and Barton’s (1994)

dichotomous value orientation.

Three-factor model: Separate biospheric, egoistic and social- Model 4 altruistic components, as suggested by the VBN model.

Four-factor model: Distinct egoistic and social-altruistic

Model 5 components, as well as two separate biospheric components for

plant and animal life. This model provides the best fit to the data.

432

Table 3: Indicator Variables for Subsequent Latent Variable Analysis