University of Tennessee, Knoxville TRACE: Tennessee Research and Creative Exchange
Doctoral Dissertations Graduate School
12-2006
Understanding Environmental Concerns and Contexts: A Multilevel Exploration of Economic, Social and Natural Environment Contexts on Environmental Concerns
Sean Thomas Huss University of Tennessee, Knoxville
Follow this and additional works at: https://trace.tennessee.edu/utk_graddiss
Part of the Sociology Commons
Recommended Citation Huss, Sean Thomas, "Understanding Environmental Concerns and Contexts: A Multilevel Exploration of Economic, Social and Natural Environment Contexts on Environmental Concerns. " PhD diss., University of Tennessee, 2006. https://trace.tennessee.edu/utk_graddiss/4253
This Dissertation is brought to you for free and open access by the Graduate School at TRACE: Tennessee Research and Creative Exchange. It has been accepted for inclusion in Doctoral Dissertations by an authorized administrator of TRACE: Tennessee Research and Creative Exchange. For more information, please contact [email protected]. To the Graduate Council:
I am submitting herewith a dissertation written by Sean Thomas Huss entitled "Understanding Environmental Concerns and Contexts: A Multilevel Exploration of Economic, Social and Natural Environment Contexts on Environmental Concerns." I have examined the final electronic copy of this dissertation for form and content and recommend that it be accepted in partial fulfillment of the requirements for the degree of Doctor of Philosophy, with a major in Sociology.
Robert Emmett Jones, Major Professor
We have read this dissertation and recommend its acceptance:
Donald W. Hastings, J. Mark Fly, Mary Sue Younger, Susan Carlson
Accepted for the Council:
Carolyn R. Hodges
Vice Provost and Dean of the Graduate School
(Original signatures are on file with official studentecor r ds.) To the Graduate Council:
I am submitting herewith a dissertation written by Sean Thomas Huss entitled "Understanding Environmental Concerns and Contexts: A Multilevel Exploration of Economic, Social and Natural Environment Contexts on Environmental Concerns." I have examined the final paper copy of this dissertation for form and content and recommended that it be accepted in partial fulfillment of the requirements for the degree of Doctor of Philosophy, with a major in Sociolo y. (l,.�'-----...... --______,,�
We have read this dissertation and recommend its acceptance: /}
Dr.16:d/sigso/#
··,•...._ Ac r th Council: L- � �
Vice Chancellor� Graduate Studies
Understanding Environmental Concerns and Contexts: A Multilevel Exploration of Economic, Social and Natural Environment Contexts on Environmental Concerns
A Dissertation Presented for the Doctor of Philosophy Degree The University of Tennessee
Sean Thomas Huss December 2006 Copyright © 2006 by Sean Thomas Huss All Rights Reserved
ii DEDICATION
This dissertation is dedicated to Gail Huss, Kristen Huss Williamson, Sydney McCauley, and those who have inspired and supported me throughout my life that are now with us in spirit.
Although they saw me only part of the way, they are remembered.
Rev. Joseph W. Hunter Charles E. Huss Mary Huss Dr. Harold E. Hyder Annie Bell Curry
Most importantly, this dissertation is dedicated to my father, Charles T. Huss, who spent his life speaking out and protecting those who could not stand up for themselves. He has been not only a loving father, but also a beacon of light and has inspired me to do the same.
iii ACKNOWLEDGEMENTS
Any set of thanks in this dissertation would not and could not suffice as a response to the help I've received from a special few over the course of my time at the University of Tennessee.
First and foremost, Dr. Robert Emmett Jones has been a perpetual source of support throughout the dissertation process. His compliments coupled with his constructive criticism on both my work and my progress have not only increased my admiration and respect for him, but also have helped me to deal with the demons that inevitably confound the path to completion. In particular, the past few years have been very difficult for me and my family in terms of financial and human loss; a time when pressure to complete the dissertation often became a source of despair rather than motivation. Yet, Dr. Jones continued to push me forward through these dark times with a necessary, albeit at times unpleasant, unending chorus of "get it done." It was that single and, at times, harsh voice that caused me to see through the fog and concentrate on the light at the end of the tunnel. Likewise, Dr. Chip Hastings was instrumental in pushing me toward the goal line in his typical coach-like fashion. He has been a constant mentor and a voice of reason throughout my time at the University of Tennessee. As he now moves into retirement, I wish him all the best of life and I hope he recognizes how important he was to so many students during his tenure as a professor. Both of these men have become role models upon which I now seek to build my own academic future and I can never personally repay them for their patience and thoughtful tutelage.
In return, I hope to continue their respective legacies by working with and encouraging students in similar fashion.
I also owe a special debt of gratitude to several professors facilitated my desire to expand and apply the knowledge I had gained and skills I honed at the university. At the Energy,
Environment and Resources Center, Dr. Mary English was the first professor to take an interest in me and afford me the opportunity to work on funded research projects. (It is no coincidence that she did so at the recommendation of Dr. Jones). It was during my work with Dr. English that I came to realize, quite early on, the difference between applied and pure research. Dr. English proved an excellent mentor as she helped me to understand how to acquire funding and the distinction between the requirements in reporting information to state agencies and academic iv journals. Indeed, it was this work with Dr. English that cultivated in me a desire to move in two worlds, both as an academic and applied sociologist, in an attempt to not only create but also apply knowledge from research to benefit society. I am also indebted to Dr. Mark Fly of the
Human Dimensions Research Lab for hiring me to work on the Woodland Survey. Of all the applied projects on which I participated as a· graduate student, the Woodland Survey was, perhaps, the most challenging and the most fulfilling. I tend to see multiple research opportunities in almost any endeavor, which makes completion of tasks very difficult. Dr. Fly represents a clear exemplar of how to focus a research project, to bring tasks in on budget, and to find ways to accomplish seemingly difficult tasks in a short period of time. Dr. Fly also was instrumental in funding a portion of my advanced study in courses on hierarchical linear modeling at the University of North Carolina. In short, these experiences and the support I received from each of these two individuals made it possible for me to explore newer and more complex opportunities in research.
Some of my favorite professors at the University of Tennessee happen to be considered intimidating by many students due to the difficult subject matters they address, yet they deserve special thanks for making my pursuit of academics simultaneously complex and comprehensive.
Dr. James Schmidhammer, Dr. Mary Sue· Younger, and Dr. Fumiko Samajima have all helped me develop the skills necessary to not only know how to perform statistical analyses, but also to understand the intricacies and importance of statistical analyses. In particular, I thank Dr.
Younger for expending the extra time and effort to serve on my dissertation committee as the statistics representative and for serving as a role model as a teacher of statistics. Dr. Patrick
Curran and Dr. Ken Bollen both contributed to my interest while at the University of Michigan and
University of North Carolina. In fact, Dr. Curran was the first to comment on the general ideas for this dissertation and to assure me that the multilevel techniques were appropriate for the research problem I sought to investigate. Last but not least, Dr. Susan Carlson of Western Michigan
University has been the biggest influence on my desire to learn more advanced methods. Her ability to teach statistics almost painlessly has become the golden standard to which I aspire when passing on what I know to frightened students. In fact, it was Dr. Carlson that encouraged V me to begin my foray into the multilevel modeling techniques used in this dissertation and she deserves special thanks for examining this body of work, serving on my committee, and continually encouraging me to keep moving forward. To all of these professors, I thank you, individually and collectively, for challenging me to overcome the fears many experience when dealing with such complex social issues and research methods.
I also am very grateful to my family for their support throughout the completion of my graduate studies and this dissertation. My father, Charles T. Huss, has always been a guiding light in my life and provided constant support right up to the point of his recent death. Although I have not had time to fully grieve, I know that he is with me in spirit and much of the attention to history and context found in this dissertation arise from his influence over the years. By training and profession, my father was an activist seriously concerned about the well-being of people and the significant influence that context has on people's beliefs and decisions. While an undergraduate, I had the rare opportunity to take some of his classes at the University of
Arkansas. His teachings still influence my thinking to this day. I would indeed be proud to serve the academy and community in a like fashion. More importantly, my father's passion for learning, teaching and research guided me throughout much of the work presented in this dissertation. My mother, Gail Huss, perhaps the strongest person I ever have known, was a constant source of encouragement during the writing of this dissertation. Her own work in the environmental field, coupled with her comments on my work as well as those of her colleagues at the Centers for
Toxicological and Environmental Health, proved to be quite valuable resources. Beyond the professional help, my mother maintained a steady supply of encouragement throughout the process, even, and especially, in the wake of my father's passing. She is an inspiration to many, but perhaps the one single, constant force in my life that never wavered in her support of me. My sister, Kristen Williamson, also offered support through her continuous encouragement and the time she took to help me organize the dissertation in accord with requirements. Her work in the environmental field also proved valuable, especially by placing me in contact with individuals from the Environmental Protection Agency in the early phases of research. In short, much is owed to my family and I thank them for everything they have done for me during my research and my life. vi No acknowledgement section for this dissertation would be complete without mention of those whom I consider as close as family. Dr. Glenn Coffey may be ranked as first among equals in regard to his academic contribution to the completion of this work and in helping me maintain my mental well-being throughout the writing of this dissertation. His years as a prosecutor trained him well in finding ways to motivate me even in the darkest of times. Indeed, Dr. Coffey falls a very close second to Dr. Jones in terms of overall credit for prompting me and providing valuable insight throughout the process. Now a Sociologist by trade, he was one of the few to read the entire manuscript and to take the time to edit every sentence, word for word, and he is one of the few that truly understands my choices in this dissertation. He is one of my dearest friends in the world and has made it possible for me to forge my way across the goal line. Although a late comer to the dissertation process, Beth Wilson became an invaluable resource in facilitating the completion of this dissertation as well. Like Dr. Coffey, Beth's experience as an attorney provided her with the ability to help clarify what needed to be stated and how to state it. Her unique critical eye from years of policy analysis and time spent on the campaign trail in the nation's capital helped give shape to the discussions on environmental policy contained in this manuscript.
Beth's belief in me, her emotional support, and the practical advice she provided assisted me greatly in finding meaning and confidence in this work. She is an extraordinary person, a passionate advocate, and, more importantly, she allows me to be human.
In addition to Dr. Coffey and Beth Wilson, I consider several others to have become more family than friend to me during my time at the University of Tennessee. Dr. David Steele has been a constant source of support during the writing of this dissertation. He offered clear advice and substantial guidance on how to navigate through the academic arguments during my forays into the environmental literature. Although, admittedly, I found his folksy advice annoying at times, his constant badgering was much more important than I have told him until this point in time and it is greatly appreciated. I would also like to express gratitude to Dr. Brent Marshall.
Although we, by necessity, fell out of contact while I finished this work, our early discussions of this project helped give shape to the final product. Dr. Marshall is a close friend and compatriot in what now is a growing passion I have developed for understanding environmental issues. Dr. vii David Riesman has also had a grand influence not only on this dissertation, but also more generally in my graduate studies. In fact, Dr. Riesman was so important in helping me to become more philosophical in my thinking that our "philosophy of science" discussions gave shape to the approach taken in this dissertation and remain a part of my overall teaching strategy to this day.
Chuck Wright, another dear friend, also deserves mention for his constant support and encouragement, and I thank him for his faith in my abilities. I would· also like to recognize Dr.
Cory Blad and Jeremy Honeycutt as two of the best friends an ABO student could ever have, especially considering their efforts to get hard copies of this dissertation to Dr. Jones. All of these individuals are owed much and I hope one day to repay them in kind for their support and strong advocacy during my studies and the completion of this project.
The faculty, students, and staff at Arkansas Tech University, as well as state and national officials that have placed their faith in me in my short time as a professor have also been terribly supportive throughout this process. I thank Jordan Willis and Town Travis for helping me to organize, type, and format the bibliography for this dis·sertation. In their roles as "Napolean's
Idiots," both of these students helped me to transforrri complicated material into something more readable and comprehensible and I sincerely appreciate their work and effort. Dr. Terri Earnest and Dr. David Ward were helpful and kind enough to cover classes and serve on committees in my place so that I could find more time to dedicate to writing this manuscript. As a brand new member of the department, I found their friendship and support to be above and beyond what I ever expected from colleagues. Dr. Earnest, Shannon Flynt and Marti Wilkerson provided much needed additional relief during the final stages of writing by taking over the reigns of the River
Valley Methamphetamine Project when I was engrossed in the process. Although we collectively have been recognized by the President of the United States for our service on the Meth Project, it was Marti and Terri who stepped forward when work on my dissertation necessarily took priority.
Likewise, Terri, was always there to assist with my responsibilities on the Federal Drug Court review board while I worked long nights on this dissertation. Sheriff Jay Winters, Prosecutor
David Gibbons, and Judge Keith Rutledge similarly deserve my appreciation for their constant support and interest in my completion of this project. I am also grateful to Dr. Dan Martin for viii working with the administration to get the breathing room necessary to complete this work.
Taking on an academic position when one's priority necessitates writing a dissertation is nearly
an impossible task, which is something I had not realized until I found myself in the deep end of
the academic ocean without a flotation device and seemingly miles from shore. All of these
individuals have gone above and beyond the call of even the most giving of good Samaritans,
and I thank them for their comments, assistance, and belief in me.
Finally, I am grateful to many people at the University of Tennessee for helping me to
cultivate my understandings of society and human behavior. Early in my student career, several
professors became central to my learning process and I am thankful for what the knowledge and
critical thinking skills they imparted to me. In particular, I would like to thank Dr. Robert Perrin
and, later, Dr. Harry Dahms for encouraging me to pursue and enhance my understanding of
sociological theory. Dr. Perrin laid the foundation during my early years, and I was fortunate to
encounter Dr. Dahms during my last year at UT, which allowed me to re-engage in sociological
debates over a variety of theoretical issues. I do not believe that a full understanding of social
theory would be complete without the study of social psychology and am thankful to Dr. Thomas
Hood and Dr. Suzanne Kurth for encouraging my interest in how social structure, personality,
attitude, and behavior may all be linked as revealed in this dissertation. Simply put, my interest in
environmental sociology, the foundation of this project, would not have eventualized were it not for teachings of Dr. Hood and Dr. Jones, both of whom sparked a very real interest that I had not heretofore developed in the subject. While there are numerous others to thank in the Department
of Sociology, I must say that Scott Frey and the staff have been very helpful in promoting my
efforts and assisting me in navigating the tangled bureaucracy. Further, I am thankful to
professors outside of Sociology that helped me increase my knowledge of survey methods;
namely, Lillard Richardson, Bill Lyons, and Tony Nownes.
It would be difficult, if not impossible, to thank everyone that played a role in this
dissertation. My hope is that I have adequately thanked all those individuals important during my time at the University of Tennessee and this last year at Arkansas Tech University. In closing, I
can only say that if I have left someone out, it is an honest error of omission rather than a lack of ix appreciation. I have been a graduate student at the University of Tennessee for quite some time and during that time I have either helped or have been helped by many. Although my studies have taken me from Criminology to Social Psychology to Political Economy to Statistics and
Research Methods and, finally, to Environmental Sociology, I would not trade any of the experiences or any of the hardships I have faced in trade for an easier path. During these past several years, I have had the good fortune of developing friendships and garnering the support of more people than I can count. More importantly, during times of hardship, I have become more than someone that has merely completed a dissertation. I have found in Sociology a calling in its fullest sense, and none of this would have been possible had I been presented with an easier path or fewer significant individuals along the way. Some contend that a Ph.D. is a Ph.D., no matter the path. I disagree. Mine has been a difficult path, filled with experiences most others did not face. Yet, through it all, I have found not only what it is that I want to do with my life, but also my integrity and, thus, who I am. In sum, I am grateful to have this opportunity to thank everyone with whom I have ever argued, debated, commiserated, befriended, or just plain pissed off at the University of Tennessee.
Sean Thomas Huss
X ABSTRACT
That social, economic, and environmental context affect human attitude and agency is a central yet untested assumption in much of sociology. The effect of context on individual attitudes is particularly important when considering how and when environmental policy gains support from the public. Drawing on Ecological Modernization and Economic Contingency theories, this dissertation investigates the influence of economic, social, and pollution contexts on environmental concerns. These theoretical perspectives suggest that failure to protect the environment is a function of political processes that privilege economic concerns over environmental concerns. Drawing on literature indicating that periods of economic growth function as windows of opportunity to promote an environmental message, this dissertation tests assertions that lower levels of economic health may decrease public environmental concern. Data from a 1995 survey of Southern Appalachian residents on environmental attitudes were merged with county level data from the region, with county treated as a higher level unit within which individuals were embedded. A total of 95 counties and 1044 residents were used in single level and multi-level models. Using confirmatory factor analysis, a measure of environmental concern was specified and then introduced into single level and multi-level models. At the individual level, social class, migrant status, and cohort were used to explain environmental concern. Next, multi level models using economic distress, social distress, and pollution context were introduced to determine the degree to which contextual variables may explain cross-county variation in environmental concerns. Findings support the assertion that economic context is important in explaining the relationship between social class and environmental concern across counties.
Even when compositional effects of birth cohorts and migrants are controlled for, the relationship between economic context, social class, and environmental concern remains. Implications of findings are discussed in the final chapter, along with discussions on how to improve multi-level studies of this nature.
xi TABLE OF CONTENTS Chapter Section Page
CHAPTER ONE: Economi1 Environment1 and Public O�inion 1 1 Introduction and Overview 1 2 Environmental Attitudes and Environmental Policy 1 3 Ecological Modernization, Local Contexts, and Cultural Capacity 6 4 Purpose of Dissertation, Research Questions, and Outline 9
CHAPTER TWO: Economic Contexts1 Public Attitudes1 and Cultural Ca�aciti 14 1 Introduction and Overview 14 Trends in Environmental Concern and Ecological Modernization 2 in America 15 A Early Phases of American Environmental Concern 15 B Rise of the Golden Age of American Environmental Concern 16 C Deregulation and Globalization of Environmental Concern 19 Ecological Modernization in America and Importance of D Contextualization 21 3 Literature on Social Bases of Environmental Concern 25 A A Problem of Context 25 B Environmental Concern of Individuals 26 C Age and Environmental Concern 26 D Social Class and Environmental Concern 29 E Place of Residence and Environmental Concern 33 F Summarizing the Literature 38 4 Economic Contingency Hypothesis and Capacity Building 40 A Economic Contingency Hypothesis 40 B Cultural Capacity Building and Influence of Context 45 5 Discussion and Conclusions 49
xii TABLE OF CONTENTS Chapter Section Page
CHAPTER THREE: Conceetual Framework1 H�eotheses1 and Methods 52 1 Introduction and Overview 52 2 Conceptual Framework and Hypotheses 53 A Environmental Concern as Cultural Capacity Building 53 B General Framework and Assumptions 55 C Research Hypotheses and Hierarchical Linear Models 60 3 Data and Methods 63 A Survey Characteristics 63 8 General Discussion on SAA Variables 66 C SAA Data Caveats 68 4 Individual and Contextual Variables 70 A Dependent Variable Level-1--Environmental Concern 70 8 Independent Variables at Level 1 75 C Contextual Variables at Level 2 78 5 Statistical Models Used in Analysis 81 6 Discussion and Conclusion 82
CHAPTER FOUR: Testing H�eotheses and Exeloring Multilevel Models 85 1 Introduction and Overview 85 2 The Southern Appalachian Region 86 A A Brief History of the Region and Importance of Study 86 8 Contextual Indicators and Regional Maps 88 C Characterizing Contexts in Southern Appalachia 96 3 Level 1 Model of Environmental Concern 96 Expectations from the Literature, Distributions, and A 96 Correlations 8 Level 1 Regressions 99 4 Multi-Level Models 108 A Hypotheses and Unconditional Random Effects Model 108 B Level 2 Random Coefficients 111 C Random Coefficient Models and Cross-Level Interactions 115 D Explorations of Age and Migration in Fixed Effects Models 121 5 Discussion and Conclusions From All Analyses 126
xiii TABLE OF CONTENTS Chapter Section Page
CHAPTER FIVE: Summary, Conclusions, and Future Research 132 1 Introduction and Overview 132 2 Theories, Empirical Findings, and Caveats 132 A Overview of Theories 132 B Overview of Hypotheses and Findings 134 C Overviewof Specific Analyses Conducted and Findings 135 3 Caveats and Recommendations for Future Research 138 A Caveats and Methodological Recommendations 139 B Recommendations for Future Theory and Research 142 4 Discussion and Conclusion 148
BIBLIOGRAPHY 151 VITA 180
xiv LIST OF TABLES
Number Title Page
Polychoric Correlation Matrix Used as Input in Confirmatory Table 3.1 72 Factor Analysis
Table 3.2 First Order Factors, Attitudes, Awareness, and Behaviors 74
Second Order Factor Analysis on Attitudes, Awareness, and Table 3.3 76 Behaviors
Table 4. 1 Correlations Between Contextual Variables 95
Overall Summary of County Scores for Southern Appalachian Table 4.2 97 Region
Correlations Between Environmental Concern, Social Class, Table 4.3 99 Time in Residence, Age, and Migrant Status
Table 4.4 Regression of Environmental Concern on Social Class 100
Curve-fit Analysis for Regression of Environmental Concern on Table 4.5 100 Social Class
Regression of Environmental Concern on Social Class and Time Table 4.6 101 in Residence
Regression of Environmental Concern on Migrant Status and Table 4.7 103 Social Class
Table 4.8 Regression of Environmental Concern on Age 105
Curve-fit Analysis for Regression of Environmental Concern on Table 4.9 105 Age
ANOVA, Contrasts, and Post Hoc Tests for Mean Environmental Table 4.10 107 Concern of Cohorts
Table 4.11 Estimates for Unconditional Random Effects Model, County 110
Estimates for Hierarchical Linear Model, Economic Distress Table 4.12 112 Context
xv LIST OF TABLES
Number Title Page
Table 4.13 Estimates for Hierarchical Linear Model, Social Distress Context 113
Table 4.14 Estimates for Hierarchical Linear Model, Pollution Risk Context 113
Estimates for Hierarchical Linear Model, Economic Distress and Table 4. 15 116 Pollution Risk Interaction
Estimates for Random Coefficient Hierarchical Linear Model, Table 4. 16 117 Social Class
Estimates for Hierarchical Linear Model, Economic Distress and Table 4. 17 118 Social Class
Estimates for Random Coefficient Hierarchical Linear Model, Table 4. 18 120 Time in Residence
Estimates for Fixed Effectsfor Mixed Model, County and Table 4. 19 123 Cohorts
Estimates for Fixed Effects forMixed Model, County and Table 4.20 124 Migrants
Estimates for Full Fixed and Random Effects Model Social Table 4.21 125 Class, Pre-1942 Cohort, Non-Migrant, and Interaction
xvi LIST OF FIGURES
Number Title Page Figure 3.1 First Order CFA from Polychoric Correlation Matrix 74
Figure 3.2 Second Order CFA from Polychoric Correlation Matrix 76
Figure 4.1 Southern Appalachian Counties 89
Figure 4.2 Southern Appalachian Region Pollution Index Scores 91 Southern Appalachian Region Economic Distress Index Figure 4.3 93 Scores Figure 4.4 Southern Appalachian Region Social Distress Index Scores 94 Curve-fit Analysis for Social Class and Environmental Figure 4.5 100 Concern Figure 4.6 Migrant Effect on Social Class and Environmental Concern 103
Figure 4.7 Curve-fit Analysis for Age and Environmental Concern 105
Figure 4.8 Mean Environmental Concern Scores by Group 107
Figure 4.9 Economic Distress and Social Class Interaction 120
xvii
CHAPTER ONE
ECONOMY, ENVIRONMENT, AND PUBLIC OPINION
Introduction and Overview
Although a great deal of research has been conducted on the direction, strength, stability, and distribution of environmental attitudes, few researchers have worked to identify differences in such attitudes across economic, social, and environmental contexts. Recent innovations in the development of software packages have created the opportunity for researchers to identify links between economic or social variables at the macro/aggregate level and attitudes or opinions at the micro/individual level. This research uses the new multi-level techniques to explore and identify how individual level attitudes on the environment vary across and within economic, social, and environmental contexts. The focus of this dissertation is on the strength of pro-environmental attitudes within economically strong and weak contexts, to determine whether or not economic distress has an effect on the environmental attitudes of individuals living in these contexts. As the literature on environmental attitudes is large, this project focuses strictly on the relationship between individual-level environmental attitudes and economic, social, and environmental contexts. Findings from this research will enable one to elaborate on how cross-level interaction effects may play an important role in structur'ing individual level attitudes, as well as identify which contexts may be most relevant to fomenting environmental concerns. In addition, findings will have implications for policy-makers and land use decision strategies for county-level growth and environmental policy in the future, by linking environmental policy to both the people and places.
Environmental Attitudes and Environmental Policy
Pro-environmental attitudes and behaviors have become a part of the American culture; yet, the meaning of this type of support for environmental issues in the public mind is difficult to untangle. On the one hand, national polls consistently suggest that people are concerned with the natural environment and support efforts to protect the natural environment. This public expression of concern has remained important for more than thirty years. By reacting to these concerns, many pro-environmental goals have been achieved, and politicians have continued to include the "environmental issue" in debates and in policy. Indeed, environmental concern and 1 environmental activism has reshaped the nature of political discourse and education in America, mostly due to the broader environmental movement. On the other hand, public support for environmental issues has fluctuated over time and pro-environmental behaviors often require personal sacrifices individuals choose to forgo (Guber 2003). This reluctance does not mean that individuals avoid pro-environmental behaviors or don't express environmental concern. __ �
Americans are concerned with the environment and do engage in pro-environmental behaviors; however, support for the environment in either action or belief may be contingent on factors like the economy and personal situation. Thus, at the level of the nation, the issue appears to be what Guber (2003) has called a problem of "consensus," where national level agreement over the environmental issue belies the complexity of public environmental concerns and environmental policy.
Environmental policies targeted at improving conditions such as air and water have a peculiar quality to them. On the one hand, such policies are used to restrict pollutants introduced into an environment or to restrict certain forms of development, with the hope of reducing environmental impact and promoting equity in distribution of environmental harms (Vig and Kraft
1994). These policies are governmental action taken in response to what is perceived to be a social problem that rises to the level of "environmental crisis" and presumed to be based on some level of public concern that is open to public scrutiny and discourse (Vig and Kraft 1994). Ideally, the policy process in a modern democracy requires that the governmental bodies create policies in response to publicly expressed needs, while limiting governmental encroachment on individual liberties (Vig and Kraft 1994). On the other hand, environmental policies are typically constrained by the degree of importance placed on the profit potential for corporations. While all persons have the legal right to pursue profit, corporations are unique in that they are considered legal
"persons" separate from their shareholders, with the legal mandate of exclusively expressing economic values (Bakan 2004). Since corporations typically are the prime cause of environmental degradation and pollution, as well as employers and producers, the environmental policy process is very susceptible to privileging economic interests over environmental concerns even though the public desires both. In effect, environmental policy typically involves a contrast 2 between equity values such as employment opportunity, fair wages and environmental justice, and economic values such as individual profit, economic development and trade (Paehlke 1994).
Currently, economic values dominate the discourse on how to deal with a growing environmental crisis globally, nationally, and locally. Often, environmental problems are treated as economic problems with economic solutions, where costs and benefits are given economic valuations to be considered (Barry 2000). In this rational calculus, environmental harms such as pollution or environmental degradation from development are treated as "externalities" to be paid for by local citizens, the general public, and the world, as opposed to the corporation. For example, air pollution not only has a direct negative impact on human health but also reduces crop yields, thus increasing the need for additional resources (land, fertilizer, labor) to maintain yields (Kahn 1998). In light of this cycle, the problem with an environmental policy process guided by economic solutions is that it operates under the assumption of a limitless supply of natural resources, when the objective facts and economic processes dictate otherwise. Further, economic solutions to environmental problems tend to treat natural resources in terms of instrumental or use value such that these resources are to be used expressly for the purposes of human wants, which also are directed by market mechanisms. Finally, and more fundamentally, the dominance of economic solutions to environmental problems marginalize alternative strategies and "economize" the environment in public discourse, thus creating an impression in the public mind of "jobs vs. environment" choices (Barry 2000; Klein 2002).
Given the economic influence on environmental policy-making, protection of the natural environment at the national level lends itself to symbolic measures over substantive reform. As
Durr (1993), Stimson (1999), and Guber (2003) all note, much of American politics is dominated by a public that tends to be moderate in its beliefs, while often reacting aversely to dramatic policy change in either the right or left direction. This moderate tendency is reflected in the overall consensus on core issues and the reluctance to support major changes during, for example, economic downturns. When national economies are stable and growing, individuals are more financially secure and thus willing to accept efforts at protecting collective goods like the natural environment (Durr 1993; Stimson 1999). During times of economic downturn, individual 3 economic positions become less stable, thus leading to support for fiscal stability over the natural environment (Durr 1993; Stimson 1999). At the national level, politicians are limited in the degree to which they may take positions on the natural environment, especially in light of a moderating or undecided public (Guber 2003). National politicians- typically face numerous issues that must be addressed, with the environment as one among many. Given that a truly anti-environmental position would mean political death, politicians defer to the more pressing issues, typically economic, while addressing the environmental concerns of a public through symbolic measures.
Hence, at the national level one may expect a consensus that the environment should be protected, while the moderate public majority acquiesces to reduced national environmental protection.
At the state and local levels, the complexity of environmental concern and environmental policy increases because of the proximity to and effect on citizens' lives. It is at these levels that environmental concerns may become more important to political efforts and relevant to the individuals that directly experience such problems. For example, persons living in Phoenix, AZ, may have different attitudes about how to preserve or protect water than someone living in rural
Tennessee or Arkansas, where population density is lower and water resources are more abundant. Likewise, individuals in economically distressed areas like Cocke County, TN, may differ in their concerns over the importance of protecting the environment when compared to residents of Knox County, TN. At these local levels, protecting the natural environment or improving the economy is more directly experienced and more important to the lives of individuals who have a vested interest in changing conditions. This importance provides opportunities for politicians, citizens and natural resource professionals to be pro-active in environmental efforts.
In addition, as most investment and economic growth occur at the local l�vel, greater localization in policy will be the response to declining national sovereignty in the future of a globalizing world
(Mol 2001 ). Thus, understanding the possible influences on environmental concerns across various social, economic, and environmental contexts becomes a way to develop and tailor better policy and to encourage public participation in environmental decision-making so policy may fit with the needs of local contexts and citizens. 4 An important first step in creating a better fit between policy and local contexts is to identify the economic and social characteristics that influence human attitudes on the natural environment (Guerrin, Crete, Mercier 2001; Guagnano, Stern, and Dieta 1995; Honadle 1999;
Parker 1999; Peine et al. 1999). One starting point for the identification of such contextual factors may be found in the sociological literature on Ecological Modernization and public opinion on the environment, the Economic Contingency Hypothesis (discussed in Chapter Two). Ecological
Modernization theory is based on the assumption that economic growth leads to higher public support for ecological protection. While somewhat controversial, this theoretical approach points to the importance of economic and social contexts in influencing the potential for increased citizen support of state sponsored environmental protection. The Economic Contingency Hypothesis is based on the assertion that attitudes toward economic growth have become entrenched in the public sphere because of a structural and political need to meet the political demands of economically vulnerable groups (Buttel 1975, 1978a'; Jones and Dunlap 1992; Jones 2002).
Recognizing that economic stagnation highlights the state's inability to efficiently deal with problems of inequality, politicians promote economic growth and environmental deregulation to gain working class support. The implication of �his political process for environmental attitudes in public opinion polls is that environmental problems are treated as secondary to economic problems. As a result, times of economic prosperity have greater potential to function as windows of opportunity for the promotion of environmental issues.
Unfortunately, these theoretical approaches are relatively untested and often are treated as counter-intuitive to the relationship between economic development and ecological harm. For example, many environmental sociologists criticize Ecological Modernization theory as being overly supportive of a capitalist system that exploits the natural environment. While on the surface there is a great deal of merit in the assertion that ecological modernization is, perhaps, too optimistic and pro-capitalist, problems arise from such critiques in that they fail to address the complexity of the theory and the instances in which evidence supports its major claims. Likewise, the Economic Contingency Hypothesis has met with no clear support in environmental sociology, whereas a great deal of research from political science, using more sophisticated methods, has 5 yielded clear support for the assertion that public attitudes fluctuate with economic conditions
(Durr 1993). A more systematic approach to understanding the influences of contexts suggested
by these approaches would be highly informative on the issue of what influences public opinion
_ on the natural environment. More specifically, researchers must now move beyond "right" and
"wrong" positions and work toward identifying contextual conditions where economic growth may
have an outcome suited to ecological protection, as well as instances where the relationship does
not. Such an approach would better lend itself to the development and implementation of policies
where tailored approaches lead to greater ecological protection at local levels. To understand the
potential relationship between economic, social, and ecological contexts, it becomes necessary to
understand the relationship between contexts and attitudes at the individual level that lend
themselves to support for ecological protection, even if economic development is the necessary
first step.
Ecological Modernization, Local Contexts, and Cultural Capacity
Ecological Modernization theory is based on the assumption that economic growth leads
to ecological protection; environmental problems are a failure of governmental institutions to
adequately respond to public issues (Fisher and Freudenburg 2001, 2004; Mol 2001; Picou 1999;
White 2004). For these theorists, environmental problems will eventually require a change in
political and economic activities of modern industrial societies, making the environment central to
economic decision-making. This change in thinking occurs as a· result of unchecked economic
growth and higher levels of environmental harm. When harms increase, public awareness of the
environment and public environmental concerns will increase pressures for change in policy and
industrial practice. Changes in policy will result as the inefficiency of the state becomes more
apparent, and market changes will occur as a result of greater public demand for corporate
environmental responsibility and environmentally friendly products. While economic thinking
would still play an important role in environmental policy, the tensions and contradictions of
market solutions to environmental problems will, according to these thinkers, make the
environment an integral feature of any policy created. For example, innovations in business and
politics can only occur with increased economic growth and increased industrialization, both of · 6 which create conditions that point to the problems of modern social institutions (Fisher and
Freudenberg 2001; Hajer 1995; Mol 2001 ). From this perspective, the current processes of economic development are not antithetical to environmental protection; they are the driving force for recognition of environmental problems and institutional change. Thus, for the Ecological
Modernization theorist, the most efficient solution to growing environmental crises is not to abandon industrial development but to increase industrialization, to increase technological advancement, and to increase political responsiveness to environmental problems locally and globally.
While mostly a theory of globalization, the lynchpin for the Ecological Modernization position on the positive aspects of ecological transformation is "capacity building" (see Weinder
2002). Capacity building refers to the ability of a nation's institutions and public to identify, define, and solve environmental problems by drawing on material and non-material resources available
(Ohiorhenuan and Wunker 1995; Weinder 2002). More specifically, resources available may include cultural, political, and economic conditions, as well as environmental conditions and public awareness/concern over such conditions (Weidner 2002). As a process, capacity building relies on multiple factors for building environmental concerns to the level required for sustainable environmentally pro-active practices and institutional change (Weidner 2002). The most crucial factors in capacity building, according to Weidner (2002), are democratic institutions and an active public concern for the environment. In effect, capacity building is an interactive process between social, economic, and natural environmental conditions played out in the relationship between public concern and policy making (Weidner 2002). Central to capacity building as an approach to understanding ecological modernization is the notion that social, economic, and natural environmental conditions act as contexts within which individuals and groups express their environmental concerns at local, national, and global levels. In short, ecological modernization theory requires that attention be paid to factors that may influence individual-level environmental concerns as a precursor to institutional and policy reform (Weidner 2002).
Addressing similar problems in the environmental concern literature, many environmental researchers have begun to call for a greater attention to be paid to external factors that may 7 influence attitudes. For example, many researchers (Berger 1997; Guagnano et al. 1995; Guerrin et al. 2001; Margai 1997) note that external conditions such as physical contexts, financial contexts, social contexts, waste policies, availability of materials, and proportion of population participating in environmental groups all may have an effect on environmental attitudes. At the individual level, attachment to area and other normative beliefs directly affect the degree to which individuals express pro-environmental attitudes, and these individual-level beliefs are presumed to vary across contexts. More specifically, such researchers explicitly adopt the assumption that social, economic, and natural environment contexts may act to encourage or discourage both environmental attitudes and environmental behaviors, as well as structure the nature of the relationship between them (Guerrin et al. 2001). In these studies, external factors such as access to recycling bins (Berger 1997; Guagano et al. 1995), block leaders (Burn 1991 ), and structural constraints in buildings (Margai 1997) all shape pro-environmental attitudes and behaviors. In addition, Guerrin et al (2001) point out that the level of ecological mobilization, the level of environmental degradation, and the level of environmental consciousness are highly interrelated at the country and individual levels.
Although Ecological Modernization is not, in and of itself, a contextual theory, these elements provide opportunities to test hypotheses on environmental attitudes within specific contexts. For example, the importance of an active public concern for the environment under the auspices of "capacity building" point to the need for a clearer understanding of the dynamics of, and influences on, the environmental concern of individuals. In line with an active public concern is the importance of context in activating the concerns of the public, thus providing support for environmental reforms consistent with the process of Ecological Modernization. As local environs and economic conditions vary within nations, one may expect that environmental attitudes may also vary within and across these contexts. Variation in environmental attitudes, therefore, should correspond to economic distress, social distress, and environmental problems as contexts. Extending this logic, individual-level environmental concern should vary in the degree to which problems in these contexts are experienced. For example, in contexts where the economy is strong, individuals will experience their social class position much differently and the 8 effect of social class on environmental attitudes will vary within the context, thus affecting the capacity or support for changes in environmental strategies. Given the importance of local area and decentralization in terms of economies and the environment, information on the ways that environmental attitudes are shaped by contexts would be important to understanding support for and pressures to environmental policy reform.
Specifically, by better understanding relationships between local contexts and environmental attitudes, the literature on Ecological Modernization may be enhanced in three ways. First, ecological modernization is both a theory of industrialization and a theory of normative changes in public perceptions on environmental protection, with the former being the main focus and the latter often neglected in the theory. Greater information on the conditions or contexts under which public opinion on the natural environment may change would inform policy makers, political action groups, and natural resource professionals on the dynamics of public attitudes that create windows of opportunity for environmental protection. Second, one of the key assumptions of ecological modernization theory is that local areas will gain greater autonomy as globalization progresses. By understanding the relationship between local context and public opinion on the environment, a greater understanding of how local contexts may respond to growing environmental problems at the local level may be acquired. As local areas serve as the context where economic decisions are made at the expense of the environment, understanding of how public opinion corresponds to local economic conditions would be highly informative to the theory. Finally, the literature on Ecological Modernization would be clarified in terms of interactions between levels of analysis, thus drawing on the basic sociological assumption that behavior is not presumed to occur in a vacuum but, instead, in contexts influencing social attitudes. Thus, the theory would be enhanced by understanding the relationship between contexts and public attitudes along the dimension of "capacity building" and "ecological awakening" discussed above.
Purpose of Dissertation, Research Questions, and Outline
One gap in environmental sociology indirectly addressed by Ecological Modernization theory is that public concerns are a function of contexts of economic growth such that better 9 economic conditions lead to higher environmental concern in the general public and, subsequently, create pressures for environmental reform. Whether or not this holds to be true is one of the key empirical questions that must be answered to build greater understanding of public concern for the natural environment. The purpose of this dissertation is to identify and explore contextual variables that may be used to explain variation in environmental concerns within and across counties in the Southern Appalachian Region. Thus, this dissertation focuses on the relationship between economic distress, social distress, and pollution risk contexts and environmental concerns at the county level. Accordingly, this research began with the following questions:
• What, if any, contextual effects influence environmental atti(ude variables at the individual level?" • Do attitudinal characteristics vary across economic and social contexts? • Is the economic context more salient to environmental attitudes than other contexts? • To what extent may economic contexts be used to predict environmental attitudes? • To what extent may economic contexts be used to theorize on changes in environmental attitudes?
By addressing such questions, several contributions to the literature on environmental concern may be derived. First, bridging the gap between macro and micro oriented approaches will provide a fuller understanding of how context may play a role in shaping environmental attitudes and concerns. Second, by using multilevel statistical techniques, precision in the relationships between individual level correlates of environmental concerns may be acquired. Third, by bridging the gap between macro and micro and by adding precision to understandings of environmental concern, natural resource professionals and environmental sociologists will gain clarity on what types of macro effects may hinder or facilitate environmental action. Finally, by understanding how economic conditions, in particular, shape environmental concerns, the literature in Ecological Modernization theory on cultural capacity for environmental reform will be enhanced.
Before proceeding to hypotheses, measurement, and analyses presented in Chapters
Three and Four, Chapter Two of this dissertation is used to discuss the nature of environmental concern and explanations for patterns of public opinion on the environment. Starting with a brief overview of American environmentalism, the first section of this chapter helps to illuminate the 10 controversy over Ecological Modernization and discusses why such modernization may not have occurred. Next, the empirical literature on demographic correlates of environmental concern is presented. This section will focus on the relationship between age, social class, and place of residence as key variables in understanding the influence of context. Following a review of demographic correlates, the literature on economic influences over public opinion on the environment is presented. This section focuses on the empirical and theoretical literature suggesting that environmental attitudes are a function of economic context. This section also includes a discussion on cultural capacity building as a mechanism for institutional change and how economic growth may lead to higher concern and thus higher capacity for environmental reform at local levels. This section also presents a general discuss on expectations on relationships from the literature and a discussion on limiting definitions of context for this dissertation. Finally, this chapter presents a discussion and conclusion summarizing the literature reviewed.
Chapter Three is used to discuss conceptualization and measurement of contextual and individual-level variables. This chapter begins with a presentation of hypotheses and a general discussion on how measurements were determined for each variable. Next, the methods employed for this project are outlined, with brief descriptions on the use of such statistical techniques in contextual analysis. Following these initial sections, a description of the data sets used is presented. In particular, this section will discuss the use of individual level survey data and aggregate county-level economic, social, and environmental data. Following a discussion on sample characteristics and techniques, general descriptive statistics for the sample and factor analyses results are presented. Although more information on each of these topics is presented in greater detail in Chapter Three, it is important to note that all of the statistical analyses, variables, and additional techniques used in this research are well grounded in sociological literature, but rarely used in combination to test multi-level models on environmental concern.
Thus, this project fulfills many needs in the overall discipline of sociology and specifically in environmental sociology by taking some of the necessary preliminary steps toward improving understandings transformations that make up Ecological Modernization. 11 Chapter four will present the results of all analyses performed for this dissertation.
Beginning with an overview of the distribution of contextual variables across counties, GIS maps are included to demonstrate the variation across counties at the macro level. These presentations also include discussions on the variation in environmental concerns within counties across the region studied. Next, statistical models will be developed at the individual and aggregate levels, to identify relationships used in subsequent analyses. In particular, linear regression was used at the individual level to assess the relative importance of various individual characteristics on environmental concerns. Following discussions on Level 1 models, unconditional and random coefficient models are run to determine the degree of variation in environmental concerns across counties. These unconditional and random coefficient models were then used as a basis for comparison between models to identify increases or decreases in explained variance when including Level-2 contextual variables. In each section of Chapter Four, a discussion on findings is presented to help the reader understand the influence of context on environmental concerns. Further, multiple tables and figures are used to explain models, especially models in which interactions are identified. Finally, a discussion and summary of major findings related to hypotheses tested is included to help the reader understand how economic context increases or decreases environmental concern through social class.
Chapter Five will be used to summarize major points, statistical analyses, and conclusions from analyses conducted. The first section of this chapter focuses on the major issues and claims outlined in earlier chapters. In particular, the importance of capacity building and understanding environmental concern in context is discussed in this first section, with an emphasis on the relationship between context and social class. Following this first section, findings from analyses are discussed and placed in a broader context of environmental sociology and the issues surrounding ecological modernization. In this section, issues of environmental justice and issues of social justice also are discussed, along with how such issues may play a central role in the degree to which environmental concern may be affected by economic conditions at the individual and aggregate levels. Next, a discussion on the limitations of this study and avenues for future research is presented. Specifically, examples of application of 12 contextual approaches and measurement of context are discussed, with implications for natural resource management and ecological mobilization around environmental issues at the local level.
Finally, a section on the importance of understanding context influences in societies that provide citizen input is presented. This section focuses specifically on the possibility of environmental reform when taking into account economic context and economic thinking. In addition, this last section includes a discussion on the importance of current economic trends in possibly shaping the type of reform expected by ecological modernization theorists.
13 CHAPTER TWO
ECONOMIC CONTEXTS, PUBLIC ATTITUDES, AND CULTURAL CAPACITY
Introduction and Overview
Adopting a contextual approach for this study on environmental attitudes allows for a theoretical and empirical linkage between structural conditions and individual environmental concerns. To understand the relationship between structural contexts and individuals, one must understand how contexts may act to shape or constrain the beliefs and actions of individuals within such contexts. Unfortunately, most social science studies in the past have focused on the relationships between national economic performance and aggregate environmental concern
(macro-level studies) or individual-level sources of environmental attitudes such as age, income, or education (micro-level studies). Many macro-level studies often rely on differences over time, to the exclusion of variation within the public at any given time on environmental issues. These studies provide a great deal of information on the nature of environmental attitudes, yet rarely do the findings from one type of analysis translate into replicable findings in another type of study.
The gap between macro-level and micro-level studies has revealed problems in theorizing and understanding environmental attitudes. To address these problems and create a greater understanding of environmental concern, it is necessary to bridge the empirical and theoretical gap between contexts and individual environmental concerns, especially when considering the importance of such attitudes to policy implementation.
As the purpose of this dissertation is to identify context influences on environmental attitudes, it first becomes necessary to elaborate on the nature of public opinion, as well as what is known and not known about environmental public opinions. The initial section of Chapter Two briefly outlines a history of environmental concern in America, and discusses its role in environmental policy, especially in terms of ecological modernization. Following this history, the empirical literature and various explanations for environmental concern are identified. This section presents information, empirical findings, and explanations on individual-level studies addressing social demographic correlates of environmental concern, focusing on age, social class, and place of residence. Next, the empirical literature on economic contingency is 14 presented. This section focuses on research directly addressing the relationship between broad economic conditions and public opinion over time. This section elaborates on how economic conditions may influence individual attitudes differentially, thus manifesting as variation in environmental concerns across contexts. The next section is used to draw all the elements in the aforementioned sections together into a format that allows for identification of hypotheses tested in this dissertation. Finally, a discussion section is presented to outline how the research questions and hypotheses are linked in the broader field of environmental sociology and to introduce methods for hypothesis testing in Chapter Three.
Trends in Environmental Concern and Ecological Modernization in America Early Phases of American Environmental Concern
By the end of the 19th Century, conservationists in America began organizing on environmental issues and established the Sierra Club (1892) and National Audubon Society
(1896), with chapters in New York and Massachusetts (Dowie 1995). Although the membership of these organizations was made up of wealthy game hunters and outdoorsmen, many among the membership began to express concerns over the treatment of the wilderness. Chief among those concerns was the notion that natural resources . are commonly owned and thus all
Americans should work to efficiently manage and protect these resources. During this period, environmentalists such as Gifford Pinchot and John Muir expressed concerns over the use of natural resources, especially by businessmen that expressed no concern for future generations
(Dowie 1995). Pinchot, as head of the Agricultural Department's Forestry Division, would go on to spearhead initiatives to establish the U.S. Forest Service and to help set aside forestland for conservation and management under what he termed as "wise use" (Dowie 1995). These policies were considered an essential component to a progressive agenda of reform in terms of natural resources. While many would eventually accept Pinchot's notions of "wise use," a vocal group of dissenters under the leadership of John Muir would challenge Pinchot, thus giving rise to the Conservationist vs. Preservationist positions in environmental attitudes and giving birth to the early environmental movements in America.
15 This early environmental concern is unique from subsequent forms of environmental concern in that it focused mainly on the destruction of the environment because of greater industrialization and urban growth that potentially impacted recreational and aesthetic uses in only some habitats (Mol 2001 ). These early stirrings of modern environmental concern were a critique of the ethos of limitless resources and the need for expansion. Environmental reforms were typically targeted at protecting endangered animal species and ecological areas deemed important, thus curtailing any development in certain areas (Mol 2001). These concerns were mainly limited to well-educated and relatively affluent segments of society for some time (Mol
2001; Shabecoff 2001). In fact, as Shabecoff (2001) notes, much of the early American
Environmentalism was dominated by white Protestant males that wanted protections to be put in place for the purposes of hunting and outdoor recreation. Likewise, many of the early environmental organizations, like the Sierra Club, had the primary function of providing outdoor trips for members (Shabecoff 2001 ). Much of the early environmental concern also focused on environmental reforms, often sponsored by private interests, without seriously critiquing the foundations of industrial societies directly; that is, they focused on environmental problems but not the sources of environmental problems (see Mol 2001 ).
Rise of the Golden Age of American Environmental Concern
For many years, environmental concern remained on the periphery of the political agenda until a phase demographic, economic, and industrial growth created unparalleled environmental risks (Shabecoff 2001 ). Recognition of such problems marks a radical shift in thinking on the environment and a dramatic growth in public concerns for the environment, with special attention to the sources of environmental harms. This radical shift centered on global increases in population and the creation and use of chemicals that persist in ecosystems and are extremely dangerous to animal species (including humans). Attention to these issues arose when authors like Rachel Carson and Paul Ehrlich pointed to the collapse of ecosystems from population pressures, depletion of natural resources, and persistent toxins. As Carson (2002:187-188) noted in her book SilentSpri ng:
16 Today we are concerned with a diffe rent kind of hazard that lurks in our environment--a hazard we ourselves have introduced into the world as our modem way of life has evolved ...The new environmental health problems are multiple-created by radiation in all its forms, born of the never ending stream of chemicals of which pesticides are a part, chemicals now pervading the world in which we live, acting upon us directly and indirectly, separately and collectively. Their presence casts a shadow that is no less ominous because it is formless and obscure, no less frightening because it is simply impossible to predict the effects of lifetime exposure to chemical and physical agents that are not partof the biological experience of man.
Like Carson, Ehrlich and other Neo-Malthusians pointed to an inevitable environmental collapse because of increased human consumption (see Shabecoff 2001). Further, authors like Aldo
Leopold (1949) provided a set of ethical claims directly contrasting the dominant ideology of unchecked growth, with an emphasis on the importance of all life and the relationship of humans to nature in an overarching web of life. In short, a renewed environmental consciousness was born that helped to awaken the public and place environmental opinions directly into the fray of politics.
If the early debates between Conservationists _and Preservationists may be thought of as a "first-wave" of environmental concern, then the rise of public concern in the wake of Carson and others may be identified as the "second-wave" of environmental concern (Bell 1998; Duffy 2003;
Mol 2001; Shabecoff 2001). As Shabecoff (2001) notes, the rise in public expressions over environmental harm could no longer be marginalized from political processes and eventually manifested as a movement on the first Earth Day--April 22, 1970. This newer form of environmental concern arose in a historical, social, and political context of distrust, where movements were seen as a way to hold governments accountable on many social justice issues
(Shabecoff 2001). The environmental movement is no exception to the trend of accountability during this phase in that it not only promoted change but also served to "watchdog" federal agencies (e.g., Environmental Protection Agency) created to protect the environment and educate the public on environmental harms (Dowie 1995:18; Shabecoff 2001). The National
Environmental Policy Act (1969), the Occupational Safety and Health Act (1970), the Clean Air
Acts (1963, 1967, 1970), the Water Quality Act (1965), the Endangered Species Conservation
Act (1966) and the Federal Water Pollutions Act Amendments (1972) were a result of the tide of public opinion during this period. In addition, state and local governments gained control over
17 environmental protection processes during this period and worked to improve sub-national conditions through their own newly formed environmental agencies (Shabecoff 2001 ).
This "second-wave" of environmental concern and activism often is referred to as a
"golden age" of environmental concern (1 960s to early 1970s), where environmental issues became central to democratic government (Bell 1998; Dowie 1995; Mal 2001 ; Shabecoff 2001 ).
While some of the concerns just after World War II still focused on protection of land for recreational use, the battle waged on the policy front was over industrial growth and pollution as a consequence of this growth (Switzer and Bryner 1998). This was a period in American history in which citizens became active in knowing about their local environments and maintained a strong belief that action would result in substantive policy change (Shabecoff 2001 ). Consistent with this activity and belief, public opinion polls through the 1960s indicated that the public was concerned about the environment, with impressive increases up to 1970 (Dunlap 1991 ; Jones and Dunlap
1992). Environmental groups such as the Sierra Club Legal Defense Fund and the National
Resources Defense Council also get their start during this period; groups with the specific purpose of establishing legal precedents in environmental law (Duffy 2003). Already established groups expanded their agenda to include pollution related issues and environmental threats, and such groups actively created opportunities for legal suits and political action, as well as hired professional lobbyists (Duffy 2003). In short, this "second-wave" of environmental concern resulted in greater public outcry, institutional reform, and growth in the membership and organization capacity of environmental groups (Duffy 2003).
In contrast to the explosion of public environmental concern up to 1970, trends in public opinion polls indicate that public support for the environment declined from 1970 to 1980, yet this concern remained relatively important. Although data are limited on environmental concern for the early to mid-1 970s, Dunlap (1 995:80) points to a "significant decline" in public concerns over the early 1970s. The rapid decline in environmental concerns between 1970 and 1973 led
Downs (1 972) to contend that the environment was becoming less important to the public agenda
(Dunlap 1992). Downs (1972) and others pointed to the decline as a shift in the public attention cycle such that environmental concerns would naturally be expected to wane as new problems 18 emerged. Environmental concerns continued to decline from the mid-1970s to the 1980s, but not as rapidly as in the previous three years (Dunlap 1992; Dunlap and Scarce 1991; Jones and
Dunlap 1992). Although these declines were notable in comparison to all time highs in 1970, environmental concern remained important to the public during this period (with levels higher than the mid 1960s). Yet, concerns over the environment had become less a consensus for the public in that other concerns grew in importance during this period, partly due to the belief that the government was protecting the environment (Dunlap 1992). Thus, widespread public support for the environment waned during this period, but public environmental concerns exhibited a certain degree of staying power in the face of competing concerns (Dunlap and Scarce 1991; Jones and
Dunlap 1992). Deregulation and Globalization of Environmental Concern
By the 1980s, the "golden age" of environmental concern came to a close with the election of Reagan and a renewed emphasis on free market capitalism. Coupled with this renewed interest in free markets was the overarching policy strategy to limit governmental intrusion into the private sector and deregulation (Duffy 2003; Shabecoff 2001 ). One of the central targets of deregulation was environmental policy, thus reducing the capacity of national agencies formed in the 1970s to protect the environment. The guiding assumption of deregulation during this period was that stringent environmental enforcement would result in a negative impact on the economy; an economy that was in decline in the early 1980s (Kraft and
Vig 1994). Consistent with his agenda, Reagan appointees to the Environmental Protection
Agency and · similar agencies came directly from businesses and firms that opposed environmental protection in the past (Vig 1994 ). Industries and other capital interests responded to new opportunities opened by deregulation by pushing for greater deregulation and the opening of public lands (Dowie 1995; Shabecoff 2001 ). As a consequence, environmental groups were required to re-tool and deal with environmental protection through litigation or grassroots activism against specific issues (Dowie 1995; Shabecoff 2001 ). In addition, Dowie (1995) contends that this period is characterized by not only a stagnation of environmental reform but also a shift from
Washington to the local area as a target of concern among environmentalists. 19 Changes in environmental policy during this period often conflicted with a consistent public concern for the environment. Numerous polls indicate that the public was not only concerned about the environment but somewhat pessimistic about how well the government was protecting the environment. For example, Cambridge and · Gallup polls from 1980 to 1990 indicate that public belief in the quality of the environment declined, with a sharp increase in 1987
(Switzer and Bryner 1998). When asked specifically about concern over the natural environment,
66% of respondents to a Gallup survey in 1989 indicated extreme concern over water pollution,
50% indicated extreme concern over air pollution, and 41 % indicated concern over disposal of garbage (Switzer and . Bryner 1998). Likewise, public outrage over political appointments and budget cuts resulted in the resignations of many officials and increases to regulatory budgets under Reagan (Dunlap 1992; Szasz 1986). In fact, few environmental policy successes may be identified during this period and the few success stories that may be identified have to do with environmental scandals of the Reagan administration (Dewie, 1995). Despite these limited environmental successes, many point to the 1980s as essentially a period of environmental policy · setbacks, but a period of dramatic resurgence in public environmental concerns, particularly the late 1980s, possibly as a response to efforts on the part of the Reagan administration to curtail regulation (see Dunlap 1992; Jones and Dunlap 1992; Switzer and Bryner 1998).
Beginning in the late 1980s and early 1990s, attention to national environmental issues began to give way to global environmental issues. This change in attention from national problems to global environmental problems is identified as the starting point of the "third-wave" of environmental concern. During this "third-wave" two related trends began to emerge when dealing with environmental problems. First, many environmental sociologists noted a distinct shift in environmental discourse from arguments over growth and its limits to issues of global warming and global change (see Buttel, Hawkins, and Power 1990). As Mol (2001 :54) points out, the late
1980s and early 1990s are characterized by a reframing of environmental problems as issues of a "global commons," thus requiring international solutions. This reframing brought about new challenges to governments and an economic and environmental interconnection between countries. Second, facing fears that national environmental policies first adopted in the "golden- 20 age" of environmentalism were not working, many environmental groups also adjusted strategies to target local areas and to address problems with market mechanisms. By addressing problems at grassroots levels and shifting attention from command-and-control governmental intervention,
much of the work to protect the environment during this period was characterized as a "new environmentalism" (Shabecoff 2001 :9). Thus, environmental concerns became not only national, they became global and they became local in discourse and political action by the end of the 20th
Century.
Ecological Modernization in America and Importance of Contexualization
It is in this late 20th Century context that ecological modernization arose in European sociology as a way to frame the relationship between economic systems, public opinion on the environment, and environmental protection. Ecological modernization arose as recognition that national-level policies to protect the environment had failed to reduce environmental harm and, instead, spread environmental risks across the globe (Cohen 1998). Previous command-and control approaches to regulation no longer applied when environmental problems became global and trans-national corporations could move to areas with fewer environmental protections. Five general principles guided this early ecological modernization thinking (summarized from Cohen
1998:150):
• Super-industrialization will result in more efficient, less resource intensive strategies; • A change in industrial processes will require national governments to impose stricter regulations to forceindu stryto revise productive processes; • Policy and planning must anticipate negative environmental consequences; • Successful implementation requires organizational change from remediation to prevention; and • Successful implementation requires constructive relationships among government, industries, and the public.
Driving ecological modernization in its nascent formulations, and presently, is: What factors enhance the ability of countries to adopt an ecologically modern approach to the environment?
Confounding a great deal of the theoretical and empirical work from this perspective in its early phases, and now, is the simple fact that few nations were, or are, experiencing transformations consistent with ecological modernization, although some nations did experience alterations in how they regulated harm to the natural environment over the 20th Century.
21 The American history of environmental concern and policy change above is one of the instances in which ecological modernization was expected to occur but, as of yet, has failed to take complete hold. By the end of the 20th Century in America more than 100 pieces of legislation were enacted, environmental issues became even more salient to public discourse and political processes, and industries recognized that environmental concerns must be addressed
(Shabecoff 2001). Protections over the quality of water and air, protection of endangered species, and preservation of land all have also had demonstrable effects. Further, as Shabecoff
(2001 :1O) puts it, these shifts in environmental thinking have " ... provided an intellectual and ethical context that enabled Americans, and people in much of the rest of the world, to see the harm that human activity was inflicting on the natural world--and on their own bodies." Indeed, as
Mol (2001) notes, the American economy became more efficient in resource use and taking into account ecological harm. Despite these gains, environment�! problems not only continued but became more complex with the recognition of global interconnections. International solutions did not accomplish the level of outcomes expected, with countries like the United States refusing to participate. In the end, although some gains were made over the 20th Century, the general trend appeared to be a slowing of the kind of transformation expected in Ecological Modernization theory for the United States.
As to why the United States has slowed in its path toward becoming ecologically modern, many reasons are offered. First, political leadership in the United States differs from that in countries where institutions have transformed (Northern Europe, Japan), and thus policy outcomes differ. For example, in countries like the Netherlands and Japan political leaders play an active role in promoting "green" agendas that affect not only national and local governments but also businesses and even citizens (Fisher and Freudenberg 2001). Second, deregulation practices characterizing the United States have placed a greater emphasis on economic growth at the expense of protecting the environment. This trend characterized a great deal of American environmental policy through the 1980s and early 1990s, with a slight reprieve under Clinton, and this process now continues to characterize environmental policy (Devine 2004). Third, environmental refo m i'2_ th United Sta s ofte are reliant on he a'!vocac of nviro�_m�ntal � � ! � � _ _ ! _ y_ _� _ 22 groups and activating the citizenry, thus making reform reactionary. With deregulation and a political leadership that vacillates on the environment, environmental groups have had to seek funding for "capacity building" at local levels, often battling corporate interests that possess more resources (Duffy 2003). These issues have made political processes in the United States problematic for environmental protection.
Underpinning much of the decline in political commitment to environmental regulation in
American government may retrospectively be attributed to broader economic changes, starting in
1974 in the United States (Buttel 1978a). Economic downturns beginning in the mid-1 970s led to lower investments, increases in structural unemployment, and fiscal crises at all levels of government, all of which began to influence the degree of environmental commitment among political leaders (Buttel 1978a). Attempts to maintain or increase environmental regulation became untenable in an increasingly unstable global marketplace, possibly starting with the Arab
Oil-Embargo (Buttel 1978a}. The economic downturns of the mid-1 970s shifted attention to economic concerns, especially on a global economic recession, thus increasing pressures to reduce state spending on environmental protection and increasing pressures for deregulation.
These conditions have had the consequence of continued emphasis on economic growth at the expense of the environment· (Buttel 1978a; Duffy 2003; Devine 2004). In addition, many in political science suggest that these economic problems created conditions under which public opinion shifted toward economic concerns instead of the environment (see Durr 1993; Stimson
1999). Thus, explanations for changes fn environmental concern and a lack of ecological modernization may be found in contextual explanations, especially when considering that environmental concerns may be economically contingent. At this point in time, much of the research on how nations like the United States prepare for a wave of ecological modernization focuses on institutional or economic change, to the exclusion of cultural or knowledge-based aspects crucial to reform. These cultural or knowledge based changes are reflected in attitudes on sciences and environmental concerns (or an environmental consciousness), which become the basis for potential change. Nations and local areas are dependent on a strong public commitment to environmental reform and expressions of 23 environmental concern (or "capacity building") to support or push for any change or reform
(Cohen 1998). As Huber and Pedersen (1997) contend, experiences of the environment are
constrained by a variety of contexts, many of which may lead to inconsistencies or difficulties in consensus building. The net consequence of this variation is that environmental concerns may be treated as relatively uniform at the national level, yet at the same time low or high in specific local areas because of context. Hence, concern for the natural environment may be expected to vary from one context to another, thus creating difficulties in understanding that concern or
building consensus on environmental conditions as a problem.
The challenge to environmental groups and social theory on society-environment
interactions is to contextualize environmental concern so that variation in public support for the
environment and institutional change may be understood. Echoing the previous statement,
Shabecoff (2001) concludes that the future of environmental protection will rely on the degree to
which environmentalists are able to adapt to changing economic, political, and social contexts at
all levels (local, national, global). Acknowledging · the importance of context, Mol (2001 :224)
suggests that Ecological Modernization theory is " .. .in need of contextualization." He further
notes that theory and research must address how environmental reforms are "co-determined" by
national characteristics through a process called "environmental glocalization" (Mol, 2001 :224 ).
Although addressing mostly nation states, Mol (2001) also points to the importance of
understanding within-nation characteristics, especially in regard to environmental concerns at all
levels of analysis. One of these national characteristics is, as noted by Cohen (1998), the degree
to which citizens express and act on environmental concerns locally, nationally, and globally.
Hence, environmental concern is one of the key factors that promote the shift to an ecologically
modern state, and understanding the relationship between economic, social and environmental
contexts and individual environmental concern is needed.
24 Literature on Social Bases of Environmental Concern
A Problem of Context
The current state of the empirical literature on environmental concern suggests that the
research is complex, thus making it difficult to build a strong theoretical and conceptual
foundation (Dunlap and Jones 2001 ). Researchers may focus on a wide range of elements or
properties of the environment and do so at a variety of levels of specificity, and they have nearly
as many options in obtaining expressions of concern for it (Dunlap and Jones 2001).
Nonetheless, both Guber (1996) and Carman (1988) provide evidence to show that public
attitudes toward environmental issues are organized in a reasonably coherent fashion, and it is
therefore appropriate to treat environmental concern as a meaningful concept. Indeed, despite
the increased complexity of environmental issues, evidence suggests that environmental concern
may have become a more cognitively consistent, stable, and important phenomenon. Further,
Dunlap and Jones (2001) argue that with careful conceptualization and item selection,
researchers can create sets of measures of environmental concern that constitute reliable and
valid public expressions of environmental concern. In this sense, environmental concern is
assumed " ... a multi-dimensional theoretical construct that reflects the degree to which people are
aware of environmental problems, believe they are serious and need attention, are willing to
support efforts to solve them and actually do things that contribute to their solution" (Dunlap and
. Jones 2001 :485). Yet, as Reser and Bentrupperbaumer (2000:3) note, " ... we clearly need a
more useful and precise language for talking about environmental concern ... "
The lack of clarity in what factors influence environmental attitudes is partly due to
difficulties in adequately conceptualizing which contexts may influence individual level attributes,
how one might adequately measure such contexts, and how to specify models in a manner that
contextual effects may be clearly identified. Contexts are i�portant to sociological theorizing in
that the discipline take as an implicit assumption that contexts will influence the attitudes of
individuals. Indeed, even environmental sociology takes context as an implicit assumption in the
bulk of its research, thus leading many environmental sociologists (e. g. Brulle 1996; Canan 1996;
Daniels 1996; Goldman and Schurman 2000; MacNaghten and Urry 1999) to focus on contexts 25 that may lead to variation in environmental attitudes. In light of this implicit assumption,
researchers such as Guerrin et al. (2001) have suggested that future research focus on exploring
and elaborating on the manner and form of contextual influence on environmental attitudes
across levels of analysis, with an emphasis on developing theories to explain such effects. In
short, many environmental sociologists are now calling for a re-examination of implicit
assumptions in the discipline that focuses on a more critical, reflexive, and methodologically
sophisticated approach to theorizing on individuals' environmental attitudes within specified
contexts. Before discussing how contexts may affect environmental concerns, an overview of the
social bases of environmental concern literature will be presented.
En vironmental Concern of Individuals
Environmental concern has been the subject of a great deal of research in environmental
sociology. Such studies typically focus on the direction, strength, stability, and distribution of
attitudes, opinions, values, beliefs and behaviors. Given the vast nature of this literature, it is
prudent to limit this literature review to information on the social bases of environmental concern,
with greater attention to age, social class, and place of residence. The reasons for limiting the
bulk of this discussion are two-fold. First, a complete review of all literature on environmental
concern is not relevant to relationships tested in this dissertation. Although race, gender, and
political ideology are important to the literature, the focus of this dissertation is on those correlates
more heavily theorized. Second, the variables age, social class, and place of residence are most
directly linked conceptually to issues discussed in this dissertation. Although other demographic
characteristics are important to understanding environmental concern, the goal is to identify a
contextual effect, so the bulk of discussions on research will be limited to those variables that
best fit conceptually with the methods presented in the next chapter.
Age and Environmental Concern
The most consistent predictor of environmental concern in the literature is Age of
Respondent, with some exceptions (e.g., Derksen and Gartrell 1993; Greenbaum 1996;
Woodrum and Hoban 1994). Most often, age is treated as linear in its relationship to
environmental concern such that younger individuals are more likely to express higher levels of 26 environmental concern than older individuals. This treatment is based on the assumption that the younger individuals are less integrated into a dominant cultural framework that promotes dominant social values like economic success (Talley 2001 ). Yet, as cultural change occurs and environmental concerns become more central to social values systems, some suggest that this relationship will stabilize and become dispersed throughout the overall population as younger persons grow up. Much of the research focusing on the relationship between age and environmental concern rests on the notion that a cohort replacement effect is in process. For example, as Guber (2003) notes, two sources of change in attitudes may explain shifts in environmental concerns. On the one ha�d, aggregate change in environmental attitudes may occur because of policy preferences of individuals based on changing economic conditions
(Guber 2003). On the other hand, aggregate change in environmental attitudes may occur because of a gradual replacement of cohorts (Guber 2003). This gradual replacement of cohorts essentially reflects the aging of younger generations, particularly post-environmental movement formation, that maintain a high degree of environmental concern (Guber 2003).
Linked to these generational arguments is the work of Ronald lnglehardt, who suggests that a broad scale intergenerational cultural shift has been occurring in wealthier countries; values are shifting from material needs to postmaterial interests. From this perspective younger generations in wealthier countries are no longer as preoccupied with meeting material; they are shifting their attention toward values of freedom, self-expression, and quality of life (lnglehardt
1977). lnglehardt (1977, 1990, 1995) contends younger generations in wealthy nations do not experience historical periods in which money, natural threats, and health are a major concern; these younger generations instead experience historical periods that allow individuals to focus on self-actualization and quality of life needs. One of the postmaterial values that directly addresses quality of life issues is environmental concern (lnglehardt 1977, 1990, 1995). Individuals in wealthy countries should have higher environmental concern because of the relative affluence of the country, where pro-environmental attitudes and behaviors are consistent with postmaterial values. In addition, this pattern translates into public opinion as a relatively stable set of attitudes where economic concerns or contexts should not have much, if any, effect on the stability of such 27 attitudes for specific cohorts over time; that is, a cohort is static in terms of its formative economic experiences during adolescence, because early economic experiences shape attitudes later in life.
Most studies indicate an inverse relationship between environmental concern and age
(see Dunlap 1992; Jones and Dunlap 1992; lnglehardt 1990; Talley 2001; Van Liere and Dunlap
1980, 1981). In almost every study on environmental concern, younger individuals express
higher degrees of environmental concern or pro-environmental attitudes than older individuals
(see Dunlap 1992; Greenbaum 1996; lnglehardt 1990; Jones and Dunlap 1992; Talley 2001; Van
Liere and Dunlap 1980, 1981). These findings tend to hold across state level surveys (Buttel and
Flinn 1974; Howell and Laska 1992; Van Liere and Dunlap 1981 ), as well as when using various
statistical techniques (Buttel and Flinn 1974; Guber 2003). Further, Jones and Dunlap (1992)
point to a consistency across time in this relationship such that between 1973 and 1990 age is
the strongest predictor of environmental concern. They also note that age becomes the strongest
predictor between 1977 and 1982 and subsequently less important to models by the late 1980s.
In particular, Jones and Dunlap (1992:176) conclude that social bases for environmental concern
like age have remained important over time.
Theoretically driven studies testing the shift to postmaterial values also suggest a great
deal of support for the relationship between age and environmental concern. For example, using
a system of categorizing individuals based on responses to materialist and postmaterialist values,
lnglehardt (1995) found that 9% of individuals in the United States reported postmaterialist values
prior to the 1990s but by 1992 nearly 18% expressed postmaterialist values, with materialist
values down 16%; this trend also is apparent in Western and Northern European countries.
lnglehardt (1995) then linked this postmaterial trend to environmental concern as approval for the
environmental movement, with findings indicating that individuals expressing postmaterial values
strongly supported the movement. Further empirical support for the postmaterial thesis is found
in cohort studies that address the potential impact of economic conditions on specific
generations. For example, if a cohort has common economic experiences during adolescence,
then one should expect environmental concern to vary across cohorts experiencing different 28 historical economic conditions. Findings from many studies (e.g., Bakvis and Nevitte 1992;
George and Southwell 1986; Mohai and Twight 1987) indicate that specific cohorts do vary in environmental concern. In particular, cohorts born between 1942 and 1964 tend to maintain higher environmental concern; cohorts born in the mid-1960s (a time of economic downturn) tend to support economic issues over the environment (Bakvis and Nevitte 1992; George and
Southwell 1986; Mohai and Twight 1987).
Overall, Age of Respondent does - appear to be the most consistent predictor of environmental concern. Most studies indicate a moderate, inverse relationship between age and environmental concern such that younger individuals tend to be more concerned about the natural environment than older individuals. Yet, when thinking about the age-environmental concern relationship, it is apparent from the postmaterialist studies that early economic experiences may structure how strongly one expresses concern over the natural environment.
Some of this influence of cohorts may be related to a "waste not, want not" set of values adopted during economically bad situations experienced when young that characterize a cohort over time.
But, additional empirical evidence suggests that different types of environmental beliefs or even environmental organizational membership may reflect differences in both age cohort and economic situation. More formally, the age of an individual may also reflect the degree to which that individual has accrued resources or the relative affluence of that individual's experiences in life. As Cotgrove (1982) has observed in his own studies, members of conservative preservation organizations are disproportionately from cohorts born prior to World War 11, and members of less conservative and more proactive conservationist organizations tend to be younger. In both instances, relative affluence tends to be a clear indicator of how concerns are adopted, expressed, and acted upon; that is, social class may play a role in the age-environmental concern relationship. Social Class and Environmental Concern
The second major predictor of environmental concern is also the most controversial; that is, Social Class. The controversy over social class and environmental concern arises, in part, over the claims that environmentalism may be inherently elitist. This charge· of elitism in 29 environmentalism is one based on the belief that you have to be wealthy to be concerned with the environment, as wealthy people are not concerned with the basic material needs of more vulnerable populations like the poor (see Guber 2003). When discussing age and environmental concern, a shift from material concerns to postmaterial concerns for quality of life may well be linked to greater economic success experienced at different points in life. Similarly, a lack of concern for issues like the environment may be linked to negative economic experiences, because basic economic needs are not met. On the whole, the literature on the relationship between indicators of social class and environmental concern is unclear on how this relationship manifests in opinions and attitudes. On the one hand, a great deal of evidence reveals that higher incomes are positively associated with activism and pro-environmental behaviors (see
Cotgrove 1982; MacDermid and Stevenson 1991; Mohai 1985; Mohai and Twight 1987). On the other hand, many studies indicate only a weak relationship obtains between environmental concern and social class, particularly income (Jones and Dunlap 1992). The literature is unclear on much of the relationship between social class and environmental concern, but the importance of economic position in theories of environmental concern makes social class highly important to understanding environmental concern.
The problematic character of the relationship between social class and environmental concern is a function of how one parses out environmental concern and how one measures social class. For example, when asking about environmental activism or willingness to pay for certain items as measures of environmental concern, social class should be a good predictor of environmental concern because these measures reflect political participation and ability to pay for additional services both of which are related to social class (see Guber 2003; Klineberg,
McKeever, and Rothenbach 1998). Yet, when dealing with attitudes or general environmental concerns, the relationship between social class and levels of concern may vary or not achieve statistical significance. Further, social class is a complex measure, and that often is neglected in research on environmental concerns. Typically, researchers will use income, education, and occupational prestige separately. But, social class conceptually is a combination of at least education and income, with ideally all three measures used if possible. When using only income 30 with measures of environmental concern, a different type of relationship is being tested than presumed. More importantly, social class often can be a function of contextual factors like job availability in area or cost of living in a specific area; that is, economic health. For these reasons, the f�llowing literature review includes discussions on education, income, and occupation as indicators of social class.
As Greenbaum (1996) notes, the literature on the relationship between education and environmental concern is unclear. Most studies on the relationship between environmental concern and education indicate that persons with higher education are more likely to express higher levels of environmental concern (e.g., lnglehardt 1990; Jones and Dunlap 1992;
MacDermid and Stevenson 1991; Mohai 1990; Olsen, Lodwick, and Dunlap 1992; Van Liere and
Dunlap 1980). Some, such as Bakviis and Nevitte (1992) contend that education is the best predictor of environmental concerns, when controlling for all other variables. Likewise, Jones and
Dunlap (1992) indicate that education was a much stronger predictor of support for environmental spending in the late 1980s than in previous years. Howell and Laska (1992), using longitudinal data collected in Michigan, found that education was not significantly related to attitudes on environmental spending until the late 1980s when it became the most important demographic predictor. Yet, as indicated above, some. studies indicate that education may be more closely associated to a general level of concern than to pro-environmental behaviors (e.g. Van Liere and
Dunlap 1981). Still other studies (Nibert 1994; Nord, Luloff, and Bridger 1998; Woodrum and
Hoban 1994) suggest that education has no effect on various environmental attitudes. In many instances, research suggests that education may only have an indirect effect on environmental attitudes (mediated by income) and that the relationship between education and environmental concern may be nonlinear such that those at the extremes of environmental concern tend to be better educated (Hamilton 1985a, 1985b; Kanagy and Willits 1993; Milbrath 1984).
Income and occupational prestige also are complex and contradictory in the empirical literature. For the most part, higher incomes and occupational statuses are identified as positively associated with environmental concern in many studies (e.g., Mohai 1985; Kanagy and
Willits 1993), but many other studies demonstrate no effect of either on concern (e.g., Mohai and 31 Twight 1987; Kanagy, Humphrey, and Firebaugh 1994). As Greenbaum (1996) notes, the most consistent findings from studies that address the relationship between income and environmental concerns are that income is more closely associated with activism and pro-environmental
behaviors (see Cotgrove 1982; MacDermid and Stevenson 1991; Mohai 1985; Mohai and Twight
1987). Some of the literature suggests that higher incomes or higher occupational prestige also
are positively associated with a more environmentally sensitive worldview (Arcury and
Christianson 1990; Kanagy and Willits 1993). When dealing with general environmental attitudes
in national surveys, however, the relationship between income or occupational prestige and
environmental concern is inconsistent and quite often non-existent (e.g., Mohai 1985; Mohai and
Twight 1987). More importantly, lnglehardt (1990) observes that higher educated persons tend to
have higher occupational prestige jobs with lower pay and also tend to have high environmental
concerns. Evidence from studies focusing on this relationship have led many researchers to
conclude that the relationship between income/occupational status and environmental concern
may be curvilinear; that is, the middle class tends to be more environmentally concerned (see
Bakvis and Nevitte 1992; MacDermid and Stevenson 1991; Olsen et al. 1992).
Explanations for the relationship between social class and environmental concern also
draw on postmaterialism as a framework. For example, many of the explanations for the possible
relationship between social class and environmental concern focus on the emergence of a "new
middle class" that is well educated and working in the service sectors of the economy (Buttel
1992; Greenbaum 1996; Offe 1985). The assumption guiding such explanations is that those
more active on environmental issues are disproportionately drawn from the new middle class
(Greenbaum 1996). These members of the new middle class are influenced in part by their
occupational positions, but more fundamentally influenced by the economic conditions of their
upbringing (Greenbaum 1996). Such individuals also tend to have more time and social
resources that allow them to engage in environmental action or to acquire information on
environmental problems (Mohai 1985). In effect, the new middle class is a set of cohorts raised
in economic situations that allow for higher education, higher incomes, and higher occupational
prestige, all of which reflect conditions necessary for adopting postmaterial values (lnglehardt 32 1990). Likewise, derived from the influence of Maslow on intergenerational explanations, individuals in the new middle class have a greater capacity to meet their basic material needs, when compared to persons in lower social class positions (lnglehardt 1990; Maslow 1954; Talley
2001 ). Thus, social class explanations also tend to draw on broader intergenerational arguments much like explanations of age and environmental concern.
Overall, Social Class is an important but complex predictor of environmental concerns.
As noted above, it may depend on the dimension of concern tapped in measurement, or it may depend on how one uses components of social class in models. Despite these problems, some consistency is present in the relationships such that higher educated individuals tend to have higher general environmental concerns, and higher income and higher occupational prestige individuals tend to vary in the degree to which they express their concern (through action or behavior). One interesting thread through the discussion on these components of social class is the possibility that the relationships may be curvilinear. For education, the relationship may be that higher educated persons tend to be at both extremes of environmental concern (high and low). Likewise, evidence suggests that the relationship between education and environmental concern may be mediated by income, which is closely associated with occupational prestige. The relationship between income and environmental concern also may be curvilinear such that persons of higher and lower income are more likely to express less environmentally sensitive attitudes. Without attention to the potential that these components of social class are, first and foremost, best used as one measure and, second, possibly curvilinear, part of the confusion in the literature may derive from a failure to identifying relationships that are present.
Place of Residence and Environmental Concern
Another important factor in understanding environmental concern is ·Place of Residence.
Literature on place of residence and environmental concern typically focuses on either the distribution of environmental harms or differences in urban and rural · attitudes on the natural environment. As with social class, studies that focus on the distributions of environmental risks typically draw on explanations that deal with the relative vulnerability of a population and its exposure to various forms of pollution. The general relationship found in much of this research is 33 that groups living below poverty and/or minorities tend to be disproportionately exposed to air pollution, water pollution, and toxic releases from manufacturing (see Anderton et al. 1994;
Bullard 1994; Davies 1972; Gould 1993; Ringquist 1997). Studies on the rural-urban differences in environmental concern tend to focus on how and why rural populations may express lower environmental concerns or broader shifts in demographic patterns of migration to rural areas.
Past studies have indicated that urban residents tend to express higher levels of concern than rural residents, but recent studies suggest that this difference may no longer be true (see Jones et al 2003). Currently, studies on the relationship between rural residence and environmental concern suggest that rural residents are concerned about the natural environment (Jones, Fly, and Cordell 1999; Jones et al. 2003).
Most studies on the distribution of environmental risks indicate that minorities and the poor are disproportionately affected by toxic waste. Findings from a General Accounting Office study indicate that the majority of the population living in areas near these waste facilities was black (U.S. Government Accounting Office 1983). In a more systematic study, the United Church of Christ's Commission for Racial Justice (1987) concludes that, although household income and mean household value were not significant, percent minority in areas with toxic facilities was nearly twice that of areas with no toxic waste facilities (United Church of Christ Commission for
Racial Justice 1987). Focusing on the Detroit area, Mohai and Bryant (1992) conclude that race is far more important than income in exposure to toxic waste hazards, with 3% of whites and 11% of blacks living within one mile of a facility. Other studies on the location of waste facilities indicate that race is less important than certain economic or occupational characteristics.
Anderton et al. (1994), using census tract data for the United States, conclude that race and ethnicity are weakly associated with waste facility sites, while a 15% increase in proportion of the population in manufacturing doubled the odds of a waste facility being in the area. Krieg (1998), in a study of Massachusetts waste facilities, notes that communities with higher proportions of poor more likely to be selected as production and disposal sites.
Explanations for this relationship are far more complex and economic than racial or class prejudices, but such prejudices do play a role in waste facility locations. As Been ( 1994) notes, 34 residential housing costs and neighborhood quality are factors in determining where individuals move. The cost of housing is affected by the proximity of factories or waste facilities, where the property values of homes in neighborhoods close to these facilities is devalued (Been 1994 ). The location of a waste facility, therefore, makes some areas less desirable to more affluent families and more affordable for poorer families. Whether a waste facility is moved into an area or a neighborhood is built near such a facility, such areas are or become poorer because of this proximity. As ethnic minorities are disproportionately harmed by economic vulnerability and as discrimination does exist in the practice of selling homes, neighborhoods near toxic facilities become home to ethnic minorities over time (Been 1994). Krieg (1998) extends this explanation to suburban and rural areas near major cities, where poorer communities may be willing to accept factories to promote economic growth or may not have the political power to keep waste facilities out. Further, as suburbs begin to experience their own economic growth, this growth brings with it additional environmental costs (Krieg 1998). Thus, as Krieg (1998) concludes, historically urban areas have one type of problem with waste facility location; suburban areas are beginning to experience their own.
Related studies on the distribution of environmental risks and awareness of environmental conditions point to the importance of proximity in explaining environmental attitudes. For example, some researchers (Hannon 1994; Hannon and Norton 1997) point to a relationship between proximity and willingness to protect or be concerned over an environmental amenity or issue. Gobster (1998) found that neighborhood residents living near the Chicago
River expressed higher levels of concern for the quality of the water when compared to other
Chicago residents. Brown et al. (2002) demonstrate that attachment to community and proximity to environmental amenities are moderately important in explaining environmental attitudes. For environmental risks, studies on the proximity to nuclear facilities (Gawande and Jenkins-Smith
2001) and proximity to areas of high air pollution (Elliot, Regens, and Seldon 1995) indicate that proximity not only affects property values but also environmental concerns. Likewise, Bush,
Moffatt, and Dunn (2001) found that residents living near industrial areas had higher concern for air quality and expressed a need for more information on air quality. As Gould (1993} notes, the 35 visibility of environmental problems influences how individuals experience concern. Visibility also influences how communities respond to environmental problems. Hence, some areas are more environmentally vulnerable than others, and individuals living in these areas tend to express higher levels of concern.
The rural-urban divide on environmental concern typically focuses on the potential that urban residents are more concerned with environmental issues than are rural residents. Early on, studies on this relationship indicate that urban residents were more concerned over environmental issues than rural residents (see Buttel and Flinn 1987; Freudenberg 1991; Mohai and Twight 1986; Van Liere and Dunlap 1981). Jones and Dunlap (1992) found that individuals raised in urban areas tended to have higher environmental concern than those raised in rural areas. This residence variable was one of the strongest predictors of environmental concern in their study (Jones and Dunlap 1992). Although some studies could identify no relationship between rural and urban environmental concerns (Lowe, Pinhey, and Grimes 1980; Milbrath
1975), many studies suggest that the urban-rural division on environmental concern was important. Yet, recent studies indicate that the rural-urban divide is no longer as significant as it was in the past. These studies point to specific patterns of urban to rural migration that have
. broadened the base of environmental concern in rural areas (see Jones et al. 1999; Jones et al.
2003). For example, focusing on Southern Appalachia, Jones et al. (2003; Talley 2001) found that nearly 1 in 3 residents surveyed had moved to the region. Although these newer residents were different from long term residents on behavioral indicators of environmental concern, long term rural residents of the area did express environmental concerns (Jones et al. 2003; Talley
2001 ).
Explanations for the differences in environmental concern in early studies and reductions in differences in later studies focus on in-migration and economic changes. Early explanations
for these rural-urban differences in environmental concern rested on the notion that urban
residents experienced environmental risks at a higher rate and, therefore, were more likely to express concerns over the environment (Tremblay and Dunlap 1978). Coupled with these explanations was the assertion that rural residents may express lower environmental concern 36 because the primary source of employment for many rural resldents was in natural resource extraction (Van Liere and Dunlap 1980; Mohai and Twight 1986). Studies conducted by Jones et al (1999) indicate that employment in rural areas dependent on natural resource extraction is on the decline, thus undermining the claim (see also Xu and Bengston 1997). Likewise, increases in the migration of urban residents to rural areas may be reducing the previous differences. Such migrants move to rural areas for the natural amenities and bring with them higher environmental concerns, which essentially changes the composition of rural areas (Jones et al. 2003; Talley
2001). Further, patterns of economic growth, the relative affluence of in-migrants, and the environmental values in-migrants bring with them to rural areas may have awakened previously silent environmental concerns in long term residents (see Fortmann and Kusel 1990; Jones et al.
2003). In short, the rural-urban differences in environmental concern may be diminishing due to the influence of in-migration.
Overall, place of residence is a highly important feature of the environmental concern literature and the clearest link to possible contextual effects. As noted above, studies on the distribution of risk clearly indicate that minorities and the poor are disproportionately exposed to environmental harms. Of particular interest in this body of literature is the notion that some vulnerable communities may not be able to resist facility location or that they may attract facilities to their areas for the purposes of economic growth. Although racial and class prejudices play a role in the process, housing markets and economic conditions appear to drive how environmental risks are disproportionately experienced. The literature on proximity to environmental risks or environmental amenities and environmental concerns clearly indicates that concerns may be higher in areas of environmental degradation, if context is considered. Residents living in or near environmental problem areas do express higher levels of concern, especially when considering attachment to those communities. Although the rural-urban differences found in prior studies may be diminishing, the primary reasons offered tend to include changes in population composition and economic development, where new residents are more affluent and occupational changes play a role in environmental concern changes. On the whole, the literature on place of residence appears to imply that economic conditions are important to understanding environmental concern, 37 an interaction between economic context and that concern. Thus, as with age and social class, economic conditions may serve as another set of possible explanations for the relationship between place of residence and environmental concern.
Summarizing th.e Literature
One clear set of problems in studying environmental concern is the relatively weak and contradictory associations between socio-demographic correlates and attitudes. For example, the variable "age" typically is identified as the strongest and most consistent predictor of environmental concerns in most studies (e.g., Greenbaum 1996; Guber 1996; Mohai and Twight
1987; Van Liere and Dunlap 1980, 1981 ), where younger respondents demonstrate higher levels of concern than that of older respondents. Yet, other studies (e.g., Derksen and Gartrell 1993;
Woodrum and Hoban 1994) serve as exceptions to this rule, and international studies suggest that this relationship does not obtain in the same manner in European industrialized countries like
Great Britain and Germany (e.g., lngelhart 1990). Likewise, indicators of social class, such as education or income, are even more inconsistent. For example, many studies (e.g., Van Liere and Dunlap · 1980; lngelhart 1990; Mohai 1985, 1990) suggest that environmental concern increases with higher levels of education. Others (e.g., Hamilton 1985; Kanagy and.Willits 1993;
Nibert 1994) find that education has only an indirect and modest effect on environmental attitudes or concerns. Further, higher incomes and occupational status are identified as positively associated with environmental concern in many studies (Mohai 1985; Kanagy and Willits 1993), but many other studies find no effect (Mohai and Twight 1987; Kanagy et al. 1994). In addition, most studies focusing on the "social bases" of environmental concern indicate that demographic correlates tend to explain a modest amount of the variance (between 1 O and 15%) in environmental concern (Van Liere and Dunlap 1980; Guber 1996).
Despite these problems, the literature on factors like age, social class, and place of residence does enhance our understandings of how environmental concern may operate in individuals. Yet, as noted above, one recommended solution to clarifying the literature on environmental concern is to consider that other factors may play a role in the formation of environmental concerns. In particular, contextual factors like local economic health, local social 38 vulnerability, and local environmental risks may play a highly important role in shaping environmental attitudes. While age may operate independent of context, factors like social class and place of residence lend themselves well to contextual explanations. For example, as noted above, the relationship between social class and economic context may operate such that persons that are more economically vulnerable will express lower concerns when the local economy is bad. Likewise, when persons move to areas that are socially vulnerable to health risks, environmental risks, and lower educational opportunities, these factors may have a combined effect of suppressing environmental concern in certain areas and under certain economic conditions. By being poor in contexts where economic and social vulnerability to environmental risks are high may lead to a shift in priorities such that the environment is less important. Thus, identifying those contexts in which environmental concern may be suppressed or supplanted by other concerns would further inform our understanding of environmental concern.
Based on the literature review above, it is apparent that certain types of individual-level relationships are important, but it also is apparent that context is implied in much of the research.
For example, the age-environmental concern relationship asserting that age may be one of the more important predictors of environmental concern, with possible historical economic causes from intergenerational differences. Although not as stable, the social class-environmental concern relationship indicates that individual socio-economic characteristics are important, but more research is needed. When considering place of residence, the contextual factors become more apparent. Responding to environmental risks or the location of environmental risks appears to be a function of broader economic conditions experienced by vulnerable populations.
Economic growth and change in occupational structure may also have worked to decrease the previous differences between urban and rural residents on environmental concern. Further, the contextual studies outlined above indicate that the relationships among individual level variables may be more complex than individual models allow. In both contextual studies, the use of economic, natural environment, and social contexts allowed for a greater degree of precision in understanding the effect of context on environmental attitudes or behaviors. Thus, as Guerrin et 39 al (2001) note, individual level explanations may have run their course and multilevel explanations are now required to better understand how, for example, context like economic conditions may affect individuals and their environmental concerns.
Economic Contingency Hypothesis and Capacity Building Economic Contingency Hypothesis
Untangling environmental issues from economic issues is a complex task, especially when considering the possible influence of context. Given that much of decision making at the local level is oriented to economic growth and that many of the trade-offs between environment and economy are made at this level, much of the decision making in pragmatic and results in compromise of environmental protection. The consequence of these compromises often is acquiescence of the population toward economic growth over environmental protection without the required transformation of ecological modernization. If local populations give in to economic interests, environmental issues may be overlooked or treated as less important. By way of an example, a corporate lobbyist, Robert Monks, tells the story of his experiences with this type of compromise (cited from Bakan 2004: Pp. 70-71 ): ...Lodged in a motel room in a small town where he had stopped for the night during an early 1970s election campaign, he awoke with a start in the middle of the night, his eyes aflame with irritation. When he got up to look out the window, he was shocked by what he saw-mounds of white foam floating down the river on whose banks the motel was perched. Monks went back to sleep and the next morning asked a clerk what had happened during the night. Well, look,' the clerk told him, 'every night the paper company sends the stuff down the river .. .Don't you understand, that's how we get rid of the effluent fromthe paper mills.' Monks knew a lot of people in the town-the mayor, the people who worked in the mills, the mill owners. 'And' he says, 'I knew that there wasn't a person in there who wanted to have the river polluted, not a person. And yet here we're living in a world where it's happening every night.'
Such local conditions and local decision making are not uncommon, and often result in the types of environmental trade-offs that prevent ecological modernization from occurring. These types of
trade-offs are contingent upon the lack of a capacity to foment environmental concerns that hold local, state, and national governments accountable to environmental protection. Hence, greater attention to local levels and context influenced environmental concerns is required for a clearer understanding of human-environment relationships and policy.
40 Identifying economic contextual effects on environmental concern is particularly important
in that often public opinion is treated as homogenous, when contextual conditions may vary or be
experienced in much different ways. For example, bringing jobs to a city or improving the overall
economic conditions in a county does not mean that all persons benefit. Economic growth
operates on a "zero-sum" principle, where markets may open in one area but are closed in
another. Increased opportunities for employment do not translate into actual employment. As
Summers and Branch (1984) note, bringing jobs into a community often has no real effect on the
vulnerable populations; that is, the unemployed or poor rarely are employed. Many studies
indicate that, despite popular opinion or policy intent, fewer than 20% of the vulnerable
populations gain employment when economic growth occurs (Summers and Branch 1984). More
often, economic development in an area benefits persons just outside of a community or persons
willing to move to the community (Summers and Branch 1984). While per capita incomes may
increase in an area experiencing economic growth, this does not equate to widespread individual
economic improvement. Such effects are felt unequally in a community, leaving many of the
more vulnerable citizens relatively impoverished or out of work. In short, economic growth is not
experienced uniformly in an area and vulnerable populations may remain vulnerable independent
of local economic health, hence the importance of identifying individual-economic context
relationships.
One key to identifying potentially salient contexts that may influence environmental
attitudes and behaviors may be found in the relatively untested economic contingency
hypothesis. The economic contingency hypothesis is a term coined by Jones and Dunlap (1992;
see also Jones 2002 for a more detailed discussion) referring to the assumption that
economically vulnerable individuals are more likely to have lower environmental concern in
periods of economic stagnation or crisis. Originally derived from the work of Buttel ( 1975, 1978a),
this hypothesis is based on the assertion that public support for economic growth is a highly
important component of public belief and one that has implications for environmental concern. As
Buttel (1978a) notes, attitudes toward economic growth have crystallized in the public sphere because of a structural and political need to meet the political demands of economically 41 vulnerable groups. Recognizing that economic stagnation highlights the state's inability to efficiently deal with problems of inequality, politicians promote economic growth and environmental deregulation to gain working class support; that is, politicians publicly claim that growth is necessary, even at the expense of the environment, if working class individuals want relative economic benefits. The implication of this for environmental attitudes in public opinion polls is that " ... environmental problems must wait upon solutions until more basic economic problems have been handled" (Buttel 1978a:55).
Note that the assertion that economic contexts influence attitudes is not new. Much of public opinion is presumed to operate in concert with overall economic conditions. Durr (1993)
demonstrates this pattern quite well through an intensive time-series analysis of public opinion through research on expectations of economic downturns or growth and shifts in domestic policy for sixty (60) different issues, including the environment. Very much in line with Buttel's (1975,
1978a) claims, but dealing with the general public as a whole, Durr (1993) identifies a trend in public opinion that domestic policy sentiment takes a conservative turn when economic downturns are present or expected and a liberal turn when economic prosperity/growth is present or expected. Durr (1993) suggests the 1980s as an exemplar of this trend by pointing to the most conservative shift in 1980 and the most liberal shift in 1989 on issues like the environment, health care and gun control. Durr (1993) further notes that these changes operate independent of political party or political ideology identification over the time interval studied (1968 to 1989), and that this trend is apparent when controlling for historical events that may have influenced public opinion (Vietnam War, Watergate, Iran-Contra). Further, focusing on political alignment of the
middle class during a similar period, and using multiple correspondence analyses, Gerteis (1998)
demonstrates a clear class segmentation on issues between conservatives and liberals from
1974 to 1978 than in the 1989 to 1994 period; that is, when economic growth occurred, class divisions between conservatives and liberals on major issues decreased.
Many others note similar patterns when testing for the effects of economic conditions on issues like environmental concern using time series techniques. For example, Guber (2003) points to periods of environmental rejuvenation during times of natural disasters or scientific 42 discovery of additional environmental harms, but, during times of economic downturn, environmental concern decreases. Guber (2003) notes that a one point increase in unemployment decreases public support for the environment by as much as 4%, with differences as much as twenty percentage points between 1982 and 1998. Stimson (1999) finds similar results when analyzing General Social Survey and other survey data from 1954 to 1996.
Conclusions from his study suggest that not only is government responsive to the demands of public opinion, but public opinion sways in concert with beliefs about economic conditions
(Stimson 1999). In particular, Stimson (1999) concludes that when the public perceives that the nation cannot afford a certain policy agenda like environmental protection, the public will lose interest in that agenda and express less support, thus altering government response. Further,
Stevenson (2001 ), Weatherford (1983), and Weatherford and Sergeyev (2000) all find that economic conditions shape public opinion not only in terms of policy choice but also that, in aggregate, these economic conditions push public opinion to the left or right depending on the perceived health of the economy. In addition, Weatherford (1983) concludes that local economic conditions have an effect on an individual's personal economic situation in determining opinions on policy.
While many environmental sociologists contend that cohort replacement accurately depicts the relationship between, for example, age or income and e11vironmental attitudes, lnglehardt's thesis is may be incorrect when considering a great deal of additional evidence. For example, in Durr's (1993) work, public opinion fluctuates within cohorts depending on the economic context. Durr (1993) actually makes this claim in a footnote, where he contends that postmaterial thesis is a static notion of intergenerational change such that one would expect a generation experiencing a period of economic prosperity during formative years to maintain postmaterial attitudes over time. Durr (1993) and many others above demonstrate that attitudes actually are dynamic over time and within or across generations. Many cross-cultural investigations also indicate that individuals living in poverty or relatively poor countries express postmaterialist values like environmental concern (Bell 1998). lnglehardt {1990, 1995) contends that such concerns are based on material needs or traditional beliefs and therefore qualitatively 43 differ from similar concerns in wealthy nations (Bell 1998). More fundamentally, lnglehardt's approach suggests that environmental concerns in wealthier countries are de facto not materially
based, thus contradicting a great deal of research ·on environmental social movements and
environmental justice (e.g., Cable and Cable 1995). Confusion in environmental sociology is
further compounded when considering that the differing results of environmental sociologists
drawing on such explanations often come from the same publicly available data sets used by
those, like Durr ( 1993), finding support for the economic contingency hypothesis using more
sophisticated statistical methods (e.g., General Social Survey).
In sum, the literature on economic contingency suggests that times of economic
prosperity function as windows of opportunity for the promotion of liberal policy agenda, like
environmental issues, and times of economic downturn lead to resistance of liberal policy
agenda. In effect, public opinion is not exogenous to policy making, and real or perceived
changes in economic context serve as ways to identify when the public is more likely to
acquiesce to a particular agendum (Durr 1993). This condition is particularly true of
environmental movements and those promoting environmental policies. As Elliott et al. (1995)
observe in their studies on shifts in public support for environmental initiatives, economic
downturns create contexts that may harm the political goals of environmental groups, and they
add that the reality of an economic condition is far more harmful than perception of an economic
condition. Elliot et al. (1995) note that gains made during the environmental movement of the
1970s are placed in jeopardy when economic growth is even modest and, in that context,
environmental protection is pitted against policies oriented toward economic growth. For
example, as noted above, Dewie (1995) and Duffy (2003) both point to a "Golden Age" of
environmentalism from the 1960s to the 1970s, with a culmination in the early 1970s, but by the
1980s economic downturns resulted in the adoption of a mainstreaming of political activity and
"professionalizing" environmental organizations, in part, as a response to changing social
conditions.
This effect of economic context on public opinion is particularly salient to environmental
sociology, especially in conceptualizing and hypothesizing on environmental concerns and their 44 potential relationships. Much of the work on environmental public opinion points to relative
stability in attitudes over time in the aggregate. To explain this stability, environmental
sociologists typically draw on the work of lnglehardt (1977, 1990, 1995) and intergenerational
shifts from materialist to postmaterialist values. While the intergenerational argument is a type of
contextual explanation, evidence above suggests that environmental concerns are more fluid and
subject to specific contextual conditions as experienced. Intergeneration arguments tend to
conflate the individual and public, without specifically addressing contexts, and, at best, may
provide only a partial picture (see Brechin 1999; Dunlap and Mertig 1995). By drawing on
additional studies focusing on economic conditions and by drawing on the limited multilevel
studies in environmental sociology, the notions of capacity building central to Ecological
Modernization may be informed. In fact, Ecological Modernization provides only a partial picture
in that it tends to focus on institutional factors and structural change, with limited attention to
cultural factors like environmental concerns. Hence, these diverse bodies of literature point to the
importance of economic context in possibly influencing attitudes and even behaviors of
individuals in terms of environmental concerns. Cultural Capacity Building and Influence of Context
Ecological Modernization theory is controversial because it claims that economic growth
leads to environmental responsibility. But, this is only one aspect of Ecological Modernization
theory. Much of the literature on capacity building does focus on institutional changes that
include scientifically identifying problems, pro-active policy on problem solutions, and market
mechanisms responsive to environmental harms (Wiedner 2002). Yet, cultural factors such as
attitudes toward science and environmental concerns are becoming more important to the theory
(Cohen 1998; Wiedner 2002). In particular, cultural components such as public opinion reflect a
general commitment in the public requiring institutional change (Cohen 1998). This commitment to environmental issues forms the basis for what Ecological Modernization theorists call "cultural capacity building," where higher concern for the environment will not only lead to policy change but also to production and market reforms. Further, the awakening of an environmental consciousness is a "necessary precondition" for modernizing and most of the gains made in 45 countries like the United States have come from public pressures (Cohen 1998:151). But, like intergeneration explanations of environmental concern, Ecological Modernization on cultural capacity building is partial. It fails to link improved economic conditions to higher or lower environmental concern and it fails to identify variation within and across areas. Despite such problems with Ecological Modernization theory, research and theorizing on the importance of cultural capacity does suggest that the link between economic context and environmental concerns is important.
The link between context and individual level attitudes or behaviors in Ecological
Modernization theory is clearest in the research on national character. Drawing on the work of
th anthropologists and sociologists of the mid-20 century, Ecological Modernization theory points to national institutional and economic characteristics as laying the foundation for greater levels of industrialization and economic growth through various phases. In the earliest phases of development, societies are more reliant on agricultural based forms of production, eventually moving toward industrial production (Cohen 1998). During this industrial phase, tensions in modern forms of resource management begin to highlight the inability of political institutions to effectively regulate natural resource depletion and pollution (Cohen 1998). At this point during the process of social change, concerns may arise among citizens of a nation, thus incorporating a more emotive reaction to environmental protections that facilitate change (Cohen 1998; Cohen
2000). In effect, environmental concerns of individuals increase when tensions between industrial practice, governmental regulation and desires for greater protection become more apparent in public discourse. Given the centrality of industrial production and concern for economic growth during this phase, political goals of economic expansion and environmental protection begin to conflict, thus resulting in greater pressures for change (Cohen 2000). The final phase of the ecological modernization project takes hold when a re-alignment of economic growth and environmental protection occurs such that industries, the public and governmental regulations all move toward efficient production while addressing environmental issues (Cohen
2000). Hence, ecological modernization is a socio-historical force that results in a greater public
46 concern, greater attention to less harmful production through scientific developments and a more central role of the environment in policy making.
The most crucial component for ecological modernization to occur is a growth in the environmental consciousness of the public; that is, cultural capacity (Cohen 1998). An increased concern for the environment, as demonstrated in attitude and behavior, sets the stage for the types of transformations required in the process of ecological modernization (Cohen 1998). Yet, cultural capacity may vary across and within nations, as the experiences of the environment vary in relation to differing contextual conditions. For example, Cohen (1998) suggests that factors such as economic conditions may decrease the potential for building the required cultural capacity for ecological modernization. Pointing to industrialized nations, he suggests that periods of high unemployment and recession may explain why certain institutional transformations have not occurred (Cohen, 1998}. He also suggests that a strict focus on the national level has limited understandings of variation in environmental concerns within nations (Cohen 1998). Others, such as Spaargaren and Van Vliet (2000), suggest that social psychological models often used in research are excellent for identifying beliefs and values but very weak when linking such beliefs or values to broader contexts. Drawing on both structural and social psychological theories to examine environmental consumption, Spaargaren and Van Vliet (2000} contend that much of the past research on "green behaviors" like environmental purchasing have been treated as if independent of context, thus failing to address how consumers make choices. Researchers of this sort point to a growing economic rationality in consumption patterns heavily influenced by product availability and life-style preferences. Yet, while culture is apparent in this work, the links to broader contexts are implied and tentatively tested but never fully explored.
Like Ecological Modernization studies on cultural capacity, many of the multilevel studies conducted in sociology point to the importance of context in shaping attitudes, opportunities, and behaviors. For example, Garner and Raudenbush (1991) contend that deprivation in an area can lead to lower employment prospects and educational outcomes. In their study, Garner and
Raudenbush (1991) use hierarchical linear modeling on neighborhood and individual level data to demonstrate variation in educational attainments across and within neighborhoods. Findings 47 from this study clearly indicate that neighborhood deprivation (measured in terms of low income
and poor housing} explains 37% of the variation in individual educational attainment within
neighborhoods and 85% of the variance in · educational attainment across neighborhoods.
Likewise, Ross and Mirkowsky (2001} focused on the effect of neighborhood characteristics on
health. They found that neighborhood disadva�tage did affect health, but that individual
disadvantage increased the likelihood of poor health (Ross and Mirkowsky 2001}. They conclude
that individuals living in disadvantaged neighborhoods suffer poor health because of their
environment and that poorer individuals in these neighborhoods tend to have poorer health
overall. These types of studies are becoming more common in sociology, although they tend not
to focus strictly on context and environmental concern.
As of the writing of this dissertation, two multilevel studies do focus on the relationship
between contextual factors and environmental attitudes and behaviors. First, Guerrin et al.
(2001} use Eurobarometer data to identify an individual level model assessing the degree to
which individual level concerns over the global environmental problem and participation in
environmental initiatives are related to recycling behavior. Next, they developed a multilevel
model to assess the degree to which contextual factors such as degree of ecological mobilization
and level of deforestation at the country level influenced recycling behaviors. Results indicate
that individuals who participate in environmental initiatives and have high levels of environmental
concern are more likely to recycle (Guerrin et al. 2001 }. In the multilevel model, Guerrin et al
(2001} find that not only do contextual factors like ecological mobilization influence recycling
behaviors but that ecological mobilization explains nearly 44% of recycling between European
countries. Engle and Potschke (1998} use a similar technique to test the relationship between
perceptions of risk, �nvironmental concern, and socio-demographic variables at the individual
level and economic characteristics (mean income and Gini coefficient} at the regional and country
level. Using International Social Survey Data and a three-level model (country, region, and
individual}, they conclude that both individual and contextual variables are important to explaining
pro-environmental choices such as cutting· back in driving (Engle and Potschke 1998}. They find
that the relationship between individual social class and aggregate economic conditions 48 significantly explain pro-environmental choices. Thus, although limited in number, such contextual studies point to the importance of context in explaining environmental concerns through attitude and action.
Consistent with the work on ecological modernization and the multilevel studies above, this dissertation focuses on how context may play a role in influencing environmental concerns.
A great deal of work is required to test the central claims of Ecological Modernization theory before one can claim with confidence that economic conditions lead to higher environmental concern in the public and improved · environmental conditions overall. Indeed, no single study would be sufficient to verify all of the empirical claims made, thus requiring a number of substantively specific studies addressing levels of analysis and the influence of contexts is required. This dissertation is such a study in that it focuses on the relationship between higher level contexts and individual attitudes on the natural environment. By paying attention to economic, social, and environmental factors that may influence environmental attitudes the relationship between structural conditions like economic distress may be linked individual-level environmental attitudes. This type of information will help clarify notions of "capacity building" so central to Ecological Modernization. Theoretically, Ecological Modernization is unclear on the presumed links between structure and belief on the environment, although this link is believed to be crucial to changes in policy and structure. By understanding how phenomena at lower levels of analysis (individual environmental concern) may be constrained by variables at higher levels of analysis (economic conditions), the theory is specified and improved. Further, in a more practical vein, such information will provide public officials, natural resource professionals, and environmental action groups on the types of structural condition lead to individual expressions of greater concern, thus helping to shaping policy, resource management, and political action as capacity building.
Discussion and Conclusions
Ecological Modernization theory suffers from many weaknesses, including a lack of attention to levels below the nation state and a lack of attention to cultural factors like public attitudes on the environment. With information of this sort missing in Ecological Modernization, a 49 necessary precondition for ecological modernization as a process is missing from theoretical arguments and empirical research (Cohen 1998). Few dispute the fact that local areas are
becoming more important in protecting the environment, because much of the economic
development and environmental protection occurs at this level. This is particularly true in the
United States, where deregulation has essentially forced state and local governments to act on
their own behalf, instead of relying on national protections. Given the growing importance of local
areas in a globalizing world, contextualizing attitudes is the next logical step for understanding
environmental concerns as a "capacity building" factor. As noted above, many environmental
groups are aware of the importance of local area and aware of the problems in attempting to
foment greater public concern at the local level. Assuming homogeneity of environmental
concern nationally lends itself to problems of consensus, where it becomes a cultural norm with
little political substance beyond rhetoric. By understanding contextual influences on
environmental attitudes, newer strategies (as dictated by Ecological Modernization) for political
action may be informed, as will theory and empirical literature on environmental concern.
Drawing on the previous literature reviewed, it is apparent that Americans are generally
concerned about the natural environment, but more information is needed. As discussed in the
first section of this chapter, beginning in the mid-to-late 1960s environmental concern grew
rapidly, culminating in an all time high in the early 1970s. In fact, Americans continue to be
concerned about the natural environment. Despite these levels of public environmental concern,
the institutional changes required for the development of an ecologically modern state have not
yet been achieved. Many suggest that the reasons for a lack of institutional reform are
economically based. Historically, economic downturns may have resulted in less action oriented
interest in the environment issue and a shift in political emphasis toward the economy in
American attitudes. Contextual studies discussed above indicate that this shift may be an
interaction between economic contexts and individual level economic conditions such that lower
economic inequality and higher mean incomes do affect individual's experiences of economic
conditions and thus affect their environmental concerns. The real issues in identifying contextual
effects are: 1) identifying measures of contexts that are proximate enough to individuals to detect 50 an effect; and 2) identifying interactions between the individual and context levels. While the previous contextual studies inform the literature, attention to local situations and environmental concerns is required before links to context may be fully understood. Thus, in the next chapter, a general theoretical framework will be presented and linked to county level indicators of economic distress, social vulnerability, and environmental risks, then these contexts will be linked to individual level environmental concerns in measurement.
51 CHAPTER THREE
CONCEPTUAL FRAMEWORK, HYPOTHESES, AND METHODS
Introduction and Overview
Few dispute that Americans are concerned with their natural environment, but whether or not context may be used to explain variation in environmental concerns across local areas is unclear and untested. For example, much of the literature suggests that persons who have higher levels of concerns tend to be younger, well-educated, relatively affluent, and have some attachment to their place of residence. The literature also suggests that place of residence does play a role in how environmental risks are distributed and how people perceive these risks or express environmental concern. In these general trends is the implication that context may play a role in the degree to which individuals express concerns. The most often cited context that may influence environmental concerns is economic, such that poorer economic conditions lead to lower environmental concerns. Yet, these studies mostly focus on the national level, with environmental concern measured in terms of an overall support for environmental policy.
Studying relationships at the national level fails to address that the local levels may vary in their own degree of economic health, thus influencing environmental concerns at the local level. To identify such contextual relationships, specific data sets and methods must be employed that allow for the identification of variation within and across counties. By using Hierarchical Linear
Modeling to include not only traditional explanatory variables but also contextual variables, variations in levels of concern across and within local areas may better be understood.
This chapter focuses on the assumptions, theories, and contextual methods used to guide the analyses conducted for this project. The first major section of this chapter outlines the development of a conceptual framework by first linking cultural capacity building and environmental concern as similar concepts. Next is presented a discussion of the limited nature of contextual theories and the importance of drawing on existing work to provide a useful framework within which assumptions and hypotheses may be derived. Following this theoretical discussion are research questions and research hypotheses. The next major section of this chapter presents methods for data collection and measurement methods used for key variables. 52 In this section, details about how the original data set was generated as part of the Southern
Appalachian Assessment and survey characteristics are outlined. A discussion of problems found in the original data set, along with remedies for these problems, also is presented in this section. The next section includes a detailed discussion of how Level-1 and Level-2 variables were selected, transformed, and calculated using Confirmatory Factor Analysis and existing formulas from government sources. Next, a brief discussion of Hierarchical Linear models, along with examples of the types of models used in analysis is presented. Finally, a discussion and conclusion is presented to illustrate the utility of multi-level approaches and the analyses to be presented in Chapter Four.
Conceptual Framework and Hypotheses
Envi(onmental Concern as Cultural CapacityBui lding
Many Ecological Modernization theorists studying cultural capacity building point to
"environmental consciousness" or environmental concern as a necessary condition for institutional transformation. A great deal is known about environmental concern more generally, but problems stem from the fact that environmental concern is a complex phenomenon.
Environmental concern may include not only attitudes or environmental knowledge, but also behavioral intentions and self-reported behaviors. Dunlap and Jones (2001) provide a very detailed discussion on the complexity of measuring environmental concern and offer a sound conceptual definition, which includes the following components: 1) an awareness of environmental problems; 2) support efforts to solve them; and 3) a willingness to personally work to solve environmental problems. From this perspective, awareness of environmental problems includes beliefs and knowledge on environmental problems and possible solutions, with an affective component that addresses perceptions on problems or protection of the environment
(Dunlap and Jones 2001). Environmental concern also includes conative expressions of concern such as willingness to act on behalf of the environment or specific actions on behalf of the environment (Dunlap and Jones 2001). Hence, environmental concern includes both attitudes on environmental problems and behavioral aspects in dealing with environmental problems.
53 This definition and the attitudinal/behavioral components are highly consistent with definitions offered by Ecological Modernization. theorists on cultural capacity building. Cultural capacity building is based on a societal endorsement of environmental protection. This endorsement may be expressed through pro-environmental attitudes (or environmental concerns) and pro-environmental consumption practices (Cohen 1998; Wiedner 2002). These individual level tendencies in a population provide the basis for environmental groups to foment political support and action collectively at all levels (local, regional, national, global). More specifically, cultural capacity for ecological reform is based on a general awareness of environmental problems, knowledge of environmental problems, and a willingness to act or current action on behalf of the environment. These aspects of cultural capacity fit well conceptually with definitions of environmental concern often reported by experts in the field. The main difference between current environmental concern research and research on ecological modernization is that ecological modernization tends to neglect this cultural component by focusing on institutional change. By focusing on environmental concern as a dimension of cultural capacity, both bodies of literature are clarified, especially when integrating them through a focus on contextual factors.
Such contextual factors play an important role in either creating opportunities for political action or reducing these opportunities, thus altering the capacity for ecological modernization to occur.
This project draws on the definition of environmental concern offered by Dunlap and
Jones (2001) in identifying measures that may be used to assess individual levels models within contexts. This approach is consistent with a great deal of theorizing and research on environmental concern, which implicitly adopt the assumption that contexts are important. For example, the work of many environmental concern researchers drawing on lnglehardt (1990) for explanations of environmental concern are drawing on cohort effects, which are historical and economic context effects. Researchers focusing on environmental harms or perceptions of environmental problems draw on contexts to describe environmental perceptions and concerns.
Researchers using time series analyses have identified patterns that indicate swings in the levels of environmental concern that correspond to economic changes within nations. Likewise, studies on national character point to the importance of both culture and structure in shaping perceptions, 54 attitudes and behaviors, even environmental behaviors. _ Based on this evidence, this project begins with the assumption that contexts will influence levels of pro-environmental attitudes and behaviors as expressions of an underlying trait, environmental concern. In addition, given the conceptual similarity between environmental concern and cultural capacity building, context effects on environmental concern would, therefore, affect the degree of cultural capacity in a given context. It follows that contextual conditions affect the degree to which environmental mobilization may occur and this may prevent ecological modernization uniformly within or across nations. Thus, understanding the effect of context on environmental concern informs the literature on Ecological Modernization.
General Framework and Assumptions
The relationship between context and the individual is nothing new to sociology. For example, in addition to the classic theorists mentioned, the Human Ecology School and the
National Character Studies all point to the importance of context in shaping personality. For them, personality includes relatively stable attitudes, beliefs, and motives for action that are the consequence of social structure and culture (House 1981). These schools of thought would later become the foundation for more current theorizing on the relationship between structure and action found in the work of theorists like Giddens, Habermas, and Bordieu. The model most often used by research of this sort suggests that social systems and the economy are structured by the natural environment, and these social systems structure attitudes and behaviors of individuals to create social stability. This model served as a framework for some time until it was displaced by divisions in sociology that arose over methods for measurement and theoretical fragmentation into macro and micro camps (House 1981 ). Techniques for quantitative measurement at multiple levels have renewed an interest in understanding context, but specific theories on contextual effects are rare. Yet, common practice informs researchers that contexts like local economies, natural environment conditions and social vulnerability of populations are important and research on the context-individual link are now gaining greater theoretical importance.
The growth in importance of contextual studies is well demonstrated in the calls for greater attention to context in environmental sociology. Brand (1 997), in particular, calls for a 55 greater attention to contexts, because environmental concerns can manifest in different ways for different people along dimensions of experienced context. As Brand (1997} points out, most national studies on environmental concern cannot be generalized cross-culturally because of the role of context in shaping environmental behaviors and attitudes. He further notes that environmental concerns are tied to broader patterns of politicar culture and responsibilities for the environment, as well as priorities of environmental concerns (Brand 1997}. To account for these types of differences, Brand (1997} suggests that researchers must focus on relevant contexts
(social, economic, environmental} that shape environmental attitudes and behaviors. Drawing on the work of many modern social theorists like Bordieu, Giddens, and Habermas, Brand suggests that the research agenda for environmental sociologists should be to find links between contexts and environmental concerns, so that social psychological investigations into the main individual level determinants of environmental concern may be understood. In addition, research of this sort should help to identify the relative importance of certain facets of concern (knowledge, behavior, attitudes} in various contexts.
The work of Brand (1997} was used as a model for this project, with an emphasis on dimensions of culture, of context and ontological insecurity, and of capacity building, to develop a general theoretical framework and hypotheses. First, many contend that environmental concern has become an integral part of the public consciousness, even to the point of consensus; thus making environmental concern a part of American culture. If environmental concern is part of
American culture, then environmental concern is part of a cognitive structure that shapes attitudes and creates knowledge, as well as guides habits, skills and strategies for action
(Swidler, 1986}. Such attitudes and knowledge may manifest in different ways during different historical periods and within different contexts (Swidler 1986}. Further, the psychological experiences of individuals in terms of their objective economic conditions (social class} will vary relative to their economic position in an overall context (Wright 1997}. This relative economic position affects decisions, knowledge, and alternatives for action (Wright 1997}. Some individuals are more economically vulnerable and have fewer options within specific contexts, so their knowledge, attitudes, and actions are constrained relative to persons more affluent and powerful. 56 Thus, attitudes and concerns may manifest differently based on social class position and options for action differ accordingly.
Second, if environmental concern includes practices and attitudes that are culturally produced and reproduced, then we must take into account context. As Giddens (1991} notes, most social practices and attitudes are not directly motivated, rather they are situated in a context of social action where individuals draw on the knowledge structures available to them. Most attitudes and actions are unconsciously adopted from broader culture and knowledge structures, where individuals are oriented to adopt institutionalized patterns of behavior (Giddens 1991 }.
Such attitudes and actions are drawn on as responses to contexts, when the context makes the attitude or action relevant. For example, most people may not think about the environment on a daily basis, yet, when asked they will express an opinion or act. Asking a question has made the environment pertinent to the context at hand. If an individual is economically vulnerable and economic position is related to personal well-being, then living in a vulnerable condition remains more pertinent to various contexts than, for example, the environment. The primary motive for action, therefore, is to maintain ontological security, which is a tendency to maintain consistency in experience by habit, customs or routines in the more general pursuit of valued ends (Bordieu
1999; Giddens 1991 }. What may or may not be a valued end is contingent on context and life circumstance (Bordieu 1999; Giddens 1991 }.
Contextual influence on environmental concern, therefore, may be viewed as cognitive and behavioral expressions that are contingent upon a given context. From this perspective, environmental concern is a phenomenon that is shared by others culturally but varies in the degree to which it may be expressed. The literature suggests contextual factors that may influence the expression of environmental concern include economic conditions, social vulnerability to environmental harms, and natural environmental conditions. Environmental concern may, therefore, vary within contexts depending on the experiences of individuals in that context. For example, given that relative economic position affects knowledge, skills and actions, the more vulnerable an individual is the lower that person's environmental concern will be,
57 contingent on the relative economic health of the area in which they live. In effect, at the aggregate level one may find that high economic health corresponds to higher environmental concerns across all individuals, on average. But, when examining the context effect, the aggregate picture may obscure an interaction between economic health and social class such that persons of lower social class, on average, are less environmentally concerned even in contexts of relative economic growth. Socially vulnerable individuals in areas with high economic health also may have, on average, higher concerns than socially vulnerable individuals living in areas with low economic heath. Theoretically and methodologically, these would be examples of cross-level interactions between contexts and individuals, thus the effect of context may be identified.
To gain greater clarity on the relationship between contextual factors and attitudes or behaviors, Brand (1997) proposes a theoretical model for analyzing relationships between contexts and environmental concerns. As Brand (1997:212-213) notes, each level of context should be treated as interconnected "filter systems" that act as constraints or opportunities for the development of an "environmental consciousness" and participation in "environmental behaviors."
In this model, he delineates between the following contexts in levels of generality and influence
(Brand, 1997:213):
1. Structural and Cultural Context--lncluding level of affluence, degree of industrialization, cultural traditions, social heterogeneity, and political order.
2. Environmental Problems--lncluding media accounts, problem frames, problem solving strategies, and normative standards for defining problems.
3. Life-World Milieus--lncluding health considerations, understandings of nature, lifestyles, environmental values, and environmental behavioral norms.
4. Ecological Norms--lncluding milieu-specific interpretations of reality and action routines with concomitant political orientations
5. Constraints and Situational Contexts--lncluding profession, leisure, or any other context that may influence and structure expectations for environmental action and options for environmental action
Yet, as Brand (1997) notes, this model only provides a general framework for analysis and does not provide specific i�formation on directionality, strength or overall effect across levels. Brand
58 (1997) suggests that greater attention be paid to the possible relationship between macro-level contexts and micro-level (individuals') attitudes, beliefs or behaviors in theory testing. Thus, although this model suffers from some weaknesses, it does serve as a useful heuristic and starting point for exploring multi-level relationships to understand environmental concern in the public (Guerin et al. 2001).
One way to deal with some of the short-comings of Brand's model is to identify in the literature potential context-individual links, since not all contexts will influence individual level phenomena in a direct manner (House 1981). House (1981) proffers that one must begin by defining those contexts that are theoretically expected to have an effect at the individual level; what House terms the component principle. House (1981) continues that a proximity principle must be addressed such that the context must be conceptually identified at a level where contexts will directly impact individuals; that is, where contexts more directly influence individuals.
Further, House (1981) argues that one must hypothesize and test relationships among attitudes, behaviors, and socio-demographics at the individual level so that a fuller understanding of the connection between context and individual experiences (the psychological principle) can be identified. In this way, as House (1981) contends, various contexts are treated as analytically distinct, well-defined, clear in effect, and linked to individual-level processes. Hence, by keeping context as a separate conceptual component in theorizing, measurement and testing, the potential links to individual-level phenomena may be identified and clarified through cross-level interactions. House (1981) asserts that failure to address such issues in sociology results in an over-reliance on post hoc explanations, failure to identify context effects, and an inability to derive coherent explanations of relationships.
Contextual factors that may influence individual level environmental concern, and thus capacity building, include economic health, urbanization or other social factors, and relative environmental harms or risks in an area. As Cohen (1998) notes, the distribution of environmental hazards or poor economies coupled with little or no social cohesion can prevent opportunities for the cultural capacity necessary to bring about ecologically modern transformations. Such conditions at the local levels may manifest as: 1) low economic health; 2) 59 high population density; 3) high urbanization; 4) high proportions of vulnerable populations; and
5) environmental health risks. Variation in these local conditions creates situations in which residents are made socially vulnerable. Individuals living in these areas are, however, not uniformly vulnerable to the conditions at higher levels. Some individuals are more vulnerable due to low income, low education, or age (Cutter, Boruff, and Shirley 2003). In these situations, broader contextual patterns may influence the degree to which an individual expresses environmental concerns or is willing to make trade-offs for protection of the environment. In effect, contextual conditions at the local level essentially result in a vulnerability of place, which may be experienced differently by vulnerable individuals or groups. Thus, economic distress, social vulnerability, and environmental vulnerability contexts meet the component, proximity, and psychological principles recommended by House (1981) above and inform on the degree to which cultural capacity and environmental concern may vary from context to context.
Research Hypotheses and Hierarchical Linear Models
If contexts affect levels of environmental concern differently for individuals living within different contexts, then understanding how, for example, the economic health of an area may impact cultural capacity becomes an important avenue of research. By paying attention to economic, social, and environmental factors that may influence environmental attitudes, the relationship between structural conditions like economic distress may be linked to individual-level environmental attitudes. This type of information will help clarify notions of capacity building and
"ecological awakening" central to Ecological Modernization theory, which is unclear on the presumed links between structure and belief on the environment. By understanding how phenomena at lower levels of analysis (individual environmental concern) may be constrained by variables at higher levels of analysis (economic conditions), the theory is specified and improved.
Further, such information will provide public officials, natural resource professionals, and environmental action groups on the types of structural conditions that lead to individual expressions of greater concern, thus helping to shaping policy, resource management, and political action as capacity building. In short, knowing something about the relationship between context and environmental concern helps to understand environmental issues, target specific 60 problems within specific contexts, generate public debate and interest, and possibly tailor environmental protections to specific areas based on the concerns of local residents.
Economic growth may lead to higher environmental concerns because of a shift to quality of life concerns (a change in public priorities). If environmental concerns fluctuate with economic conditions such that better economies lead to higher concern, then economic growth may lead to greater pressure on institutions to reform environmental practice. This relationship is, in effect, the very type of cultural capacity building referred to in the Ecological Modernization literature. As people become more affluent and their concerns shift from economic to social justice or quality of life issues, there will be greater pressure on institutions and corporations to innovate or respond to these concerns. If, on the other hand, economic conditions worsen and priorities return to the economy, the capacity to pressure institutions on the environmental declines because of reductions in environmental concern. Likewise, individuals respond to more than just economic conditions, although economic conditions may have the most direct effect. Social vulnerability in the form of population density and rural-urban differences also has an impact environmental concern. Further, natural environmental conditions also play a role in how individuals report environmental concerns, with local policy responding to these concerns. By understanding the relationship between these contexts, the relationships between the social bases of environmental concern may be clarified, helping to overcome some of the limitations discussed previously. In short, identifying these types of relationships is an important research agenda in understanding environmental concerns for the 21st Century.
To identify the effect of context on environmental concern, Hierarchical Linear Modeling
(HLM) was used to test the hypotheses presented below. HLM is based on Ordinary Least
Squares (OLS) regression in that at the most fundamental level (individual or Level 1) the outcome variable is predicted as a linear function of one or more predictors and an intercept
(Raudenbush and Bryk 2002; Raudenbush et al. 2002; Singer 1998; Snijders and Bosker 2002).
At Level 2 (count y-level), the coefficients defined in the Level 1 model become outcome variables in the Level 2 model, as a function of one or more Level 2 predictors. The utility of these types of models, as an extension of the General Linear Model (GLM), is that one may account for non- 61 independence among sampled units and the residuals of statistical models (see Snijders and
Bosker 2002). More specifically, lower-level units that may be nested into higher-level units, as well as residuals, are presumed independent in OLS (Raudenbush and Bryk 2002; Raudenbush et al. 2002; Singer 1998; Snijders and Bosker 2002). Violation of this assumption produces standard errors that may be too small, unless design effects are modeled. Since it is reasonable to assume that lower-level units (individuals) share characteristics reflective of the higher-level units into which they may be aggregated (county of residence), observations at level-1 are not fully independent of one another thus the HLM technique is appropriate to this project. In addition to the basic model structure, several options are available for modeling both random and fixed effects (identifying differences among individuals within and across contexts) (Raudenbush and
Bryk 2002; Raudenbush et al. 2002; Singer 1998; Snijders and Bosker 2002). In short, using multi-level approaches allows for the prediction of variances in the slopes and intercepts of a
Level 1 relationship with Level 2 variables.
Based on the literature discussed in previous chapters, the theoretical framework outlined above, and the techniques outlined in the previous paragraph, the hypotheses to be tested are derived from the more general research questions presented in Chapter One:
• What, if any, contextual effects influence environmental attitude variables at the individual level?" • Do attitudinal characteristics vary across economic and social contexts? • Is the economic context more salient to environmental attitudes than other contexts? • To what extent may economic contexts be used to predict environmental attitudes? • To what extent may economic contexts be used to theorize on changes in environmental attitudes? Consistent with these questions, specific hypotheses are derived from findings of previous research and used to guide the possible effects of economic contexts on individual level relationships among socio-demographics and environmental concerns. Three hypotheses are tested in this project:
1. Individuals will have lower environmental concerns in counties with higher economic distress
2. Social class and economic distress will interact across levels such that the relationship between social class and environmental concerns will vary across counties, where economic distress explains some of this cross-county variation
62 3. Time of residence and economic distress will interact across levels such that the relationship between time of residence and environmental concerns will vary across counties, where economic distress explains some of this cross-county variation
Environmental concerns are expected to vary across counties such that higher environmental concerns will correspond to better economic conditions. Further, when allowing for differences between counties, the effect of the economic or social distress on relationships between social class, time in residence, and environmental concerns will be present. In addition, social distress and pollution risks were included in models to identify additional context effects for future research.
Data and Methods
Survey Characteristics The individual-level dataset (Level 1) used for this project was collected as part of the
Southern Appalachian Assessment (SAA) conducted by the Southern Appalachian Man and
Biosphere (SAMAS) project. The SAA was performed in 1995 as a "check up" of one of the most biologically diverse regions in the country: the Southern Appalachian Region (SAR) (Cordell,
Helton, and Peine 1996:4 ). The SAR covers approximately 37 million acres across seven
Southeastern states (Alabama, Georgia, North Carolina, South Carolina, Tennessee, Virginia, and West Virginia) and 135 counties. At the time of the assessment, forests, pastures, and cropland covered 70%, 17.4%, and 3.4% of the area in the region respectively, and developed areas covered 3.1 % of the region. Over the 20 years prior to this assessment, development of relatively untouched parts of the region grew rapidly at the expense of cropland, forests and pastures (Cordell et al. 1996). Although the SAR includes the largest block of publicly owned land in the Eastern United States, the bulk of the land in the region (83.7%) is privately owned.
As the decisions of these landowners play a central role in the health of the region, particularly in light of development trends, representatives from multiple universities, governmental agencies and various scientific disciplines were contacted to participate in a comprehensive assessment of the biological, cultural, social and economic characteristics of the region, to be coordinated through SAMAS. In particular, the goal of the SAA was to identify issues salient to residents of
63 the region, issues salient to effective management practice, and issues salient to biological threats to the region.
As part of this larger project, the SAA included a "Human Dimensions" component meant to identify social, economic and cultural aspects of the " ...pervasiveness of human influences in
Southern Appalachian natural ecosystems in which they live... " and the extent to which humans
" ... are being influenced by and are closely tied to those ecosystems" (Cordell et al. 1996:17).
Secondary data from governmental agencies or other scientific investigations were used to map the relationship between human residents and the ecosystems of the region. In two specific instances natural resource data were not available and data to. be collected on impacts of natural resource management and knowledge, attitudes, opinions, values and behaviors of residents
(Cordell et al. 1996). In one of these instances, focus groups in five communities were used to determine impacts of natural resource management near public lands (Cordell et al. 1996). In the other instance, a telephone survey was conducted by the Human Dimensions Survey Research
Laboratory in the Department of Forestry, Wildlife and Fisheries at the University of Tennessee.
In particular, the survey was conducted for the purpose of generating descriptive information on the importance of natural resources, general attitudes toward development, general awareness of environmental problems, overall environmental knowledge, general attitudes on protecting the environment of the region, and specific pro-environmental behaviors (Cordell et al. 1996). As part of the human dimensions component, results from the sample drawn for the survey were used to portray the general attitudes, beliefs, knowledge, and behaviors of the human residents of the region.
The resulting data set from the telephone survey are used as the individual-level data set to be used in the present study. Data were collected by telephone from August 11, 1995 to
September 21, 1995 by the Human Dimensions Research Laboratory (Cordell et al. 1996).
Sampling was based on a quota technique in an attempt to acquire equal sample sizes across four sub-regions (Northern Ridge and Valley, Blue Ridge, Southern Ridge and Valley, and
Southern Mountain and Piedmont) and two residence types (urban and rural) of the SAR (Cordell et al. 1996). All attempts to reach individuals in a household continued until target sizes were 64 reached for each sub-region and residence type (Cordell et al. 1996). Further, respondents were selected within households by asking for the number of residents that were 18 or over and then asking for the resident 18 or over who had the most recent birthday (Cordell et al. 1996). The selected individuals were then asked to respond to survey items, taking approximately 15 minutes per call (Cordell et al. 1996). A total of 6,000 telephone numbers were used and 2,829 individuals were contacted using a random-digit dialing method during the survey period (9. 7% were disconnects, businesses, or fax numbers) (Cordell et al. 1996). Of the 2,829 individuals contacted, 54.4% refused to participate, 2.4% terminated prior to survey completion or other problems presented, and a final sample size of 1,220 (a response rate of 43.8%) was included in the data set acquired (Cordell et al. 1996). This final sample gave equal representation across sub-regions and an even distribution of residence types (Cordell et al. 1996). Based on multi level model requirements for this study, a total of 97 counties and 1,044 cases were used in analysis.
The survey instrument included 168 items focusing on a variety of topics. These topics and number of questions may be summarized as follows:
1. Residence Status-Questions focused on county of residence, year round residency, length of residency for family and individual, and factors influencing decisions to stay or move to region (a total of 7 questions)
2. Community Attachment-Questions focused on the extent to which individuals are emotionally attached to their community, these items included a series of skips to the next question if attachment was low at any point (a total of 7 questions).
3. Rural Land Ownership--Questions focused on whether or not the respondent owns land in the region, counties in which the land is owned, amount of land owned, and intent for land use (a total of 6 questions).
4. Environmental Knowledge-Questions focused on whether or not the respondent feels he or she is knowledgeable about the environment and specific items were used to test the knowledge of respondents (a total of 18 questions).
5. Environmental Attitudes-Southern Appalachia-Questions focused on level of concern with environmental issues that affect Southern Appalachia (a total of 20 questions).
6. General Environmental Values-Questions focused on attitudes and opinions on local environment change and relative importance of natural environment as an important issue (a total of 3 questions)
65 7. Environmental Behaviors-Questions focused on specific behaviors such as recycling, purchases, reading books, writing politicians, contributing money, and participating in clean-ups (a total of 6 questions).
8. Recreational Use-Questions focused on outdoors or recreation behaviors such as picnicking, backpacking, hunting, and visits to facilities, as well as willingness to pay for outdoor activities in the region (a total of 16 questions).
9. Demographic/Background Characteristics-Questions included the standard demographic indicators (e.g., age, sex, race, education, and household income), as well as political orientation and respondent description of place of residence (a total of 9 questions).
Thus, this dataset included numerous items consistent with extant research on environmental attitudes/concern and behavior, as well as identifiers for aggregate contexts at the county level, where data may be acquired for the year of the survey. General Discussion on SAA Variables
Variable selection began by identifying questions from the Environmental Attitudes in
Southern Appalachia (EATSA) questions and behavioral indicators that would reflect a general environmental concern that is consistent with definitions of cultural capacity related to the region studied. Taking into account that cultural capacity includes knowledge, attitudes, and behaviors that may lead to greater pressures on social institutions for reform, the selection of variables to include in the creation of a dependent variable was specific. Of the many issues faced in the
SAR are patterns of pollution and patterns of development that encroach on natural habitats of various species living in the region. Hence, knowledge and attitude variables derived from the knowledge and EATSA indicators that focused on habitat protection and development, industrial responsibility in paying for pollution, and responses to knowledge items on pollution were selected. Behavioral indicators were selected based on actions that reflect a pro-active stance on the natural environment such as contributing money to environmental groups or subscribing to environmental publications. Awareness indicators focused on those related specifically to pollution issues in the region (air and water). These indicators all reflect facets of environmental concern noted by Dunlap and Jones (2001) above in that they included affective and conative aspects of environmental concern. Further, these indicators fit well with ecological modernization definitions of cultural capacity in that they reflect attitudinal and behavioral characteristics
66 consistent with an "ecological consciousness." It is important to note that these indicators also are consistent with policy approaches in the literature (see Dunlap and Jones 2001 ).
Relying heavily on the advice offeredby Dunlap and Jones (2001 ), the following variables were selected out of the master data set:
1. EATSA Attitudes (measured on Likert Scale Strongly Disagree to Strongly Agree; 1 to 5)
a. Land that provides critical habitat for plant and animal species should not be developed. b. Industries which pollute the waters and air should pay for the clean-up even if it means the loss of jobs or profits c. More public lands should be set aside as wilderness
2. Behaviors
a. Contributed money to an environmental group (measured dichotomously O=No and 1 =Yes) b. Subscribed to an environmental publication (measured dichotomously O=No and 1 =Yes)
3. Knowledge/Awareness
a. The most common pollutant of water is carbon monoxide b. Southern Appalachia is one of the few areas in the U.S. that does not experience air pollution.
4. Length of Time Living in Area (measured in years)
5. Income (measured in income ranges with mid-points used to approximate interval measurement)
a. Less than $15k b. $15k to $24,999 C. $25k to $34,999 d. $35k to $44,999 e. $45 to $54,999 f. $55k to $74,999 g. $75k to $100,000 h. Over $100k
6. Education (measured in levels with years of education for each degree used to approximate interval measurement)
a. Less than High School b. High School Graduate c. Some College d. Trade/TechnicalNocational Training e. College Graduate f. Post Graduate Work/Degree
7. Age (measured in years) 67 8. Time in Residence (measured in years)
9. Migration Status (measured dichotomously as 1 =Native and O=Migrant)
Some attitudinal variables were reverse coded from original data (Strongly Agree= 1 to Strongly
Disagree=5) and then recoded to operate on a similar metric to other variables prior to polychoric correlations, to measure strength of agreement. The attitudinal and behavioral variables were used in confirmatory factor analysis to create an environmental concern variable, following the advice of Dunlap and Jones (2001 ), as well as Tarrant and Cordell (1997) that used similar methods on this same data set. The education and income variables were used to create a composite social class measure, and the time in residence was used as given. In addition, age and time in residence were used as given, although age was subsequently dummy coded to identify cohort differences suggested in the literature (Pre 1942 and 1942 or Later).
SAA Data Caveats
Although this data set was well suited to test relationships between contexts and individuals, problems arose in that some commonly used attitude questions were not asked of all individuals, missing values for some variables created measurement issues, and some counties included only one individual. First, although five separate sets of environmental attitude questions were asked of respondents, such questions were not asked of all respondents (some scales were randomly selected to reduce survey time). This made it difficult to find common measurement on the standard environmental attitude scales for all individuals. To remedy this problem, items from the EAT SA questions were used to tap environmental attitudes (all behavioral questions were asked of all respondents). Second, Hierarchical Linear Modeling requires at least two lower level units per context to efficiently produce estimates; when the number of contexts exceeds 70 a minimum of two individuals within contexts may be used
(Raudenbush and Bryk 2002; Raudenbush et al. 2002; Singer 1998; Snijders and Bosker 2002).
Thus, since some counties included only a single individual, such counties were deleted from analysis.
68 In addition to measurement and estimation issues, missing values presented a problem in that a missing value on one item would result in a missing value for composite measures.
Although the dataset was, for the most part, complete (i.e., low missing values), missing values presented a problem in creating scales for independent and dependent variables in the survey data. To deal with this problem, the SM data file (Level 1) was subjected to a missing values analysis and subsequently to Multiple Imputation (Ml) by county, which is an important step in multi-level analyses. Imputation Posterior and Expectation Maximization algorithms were used to estimate the value of the missing cell based on the values of all other variables entered into the program (see Allison 2000). Multiple data sets are produced from this process and the coefficients from models run on each new data set are then combined into an interpretable model, so that a single set of estimates may be provided for interpretation. Further, SAS statistical software was used to create five imputed datasets. The statistical software packages
SPSS and HLM were used for multilevel analyses (Allison 2000; Raudenbush et al. 2001 ).
Finally, as different numbers of individuals were drawn from each county and, since corrections for this must be made to generalize at Level 1, weighting was used. A general rule in multi-level analysis is to include as many lower-level units per higher-level unit as possible. Yet, typically one must default to maximizing higher-level units because multi-level analyses will work with at least two individual-level units per higher-level unit (Level 2; counties in this study) (see
Snijders and Bosker 2002). If one wishes to generalize to the lower-level unit, however, weighting is required of the individual-level units. In effect, weights allow for better precision in estimating models when data is collected without multilevel analysis in mind, which was the case with the data used. Individuals were sampled unequally but have known probabilities for inclusion in the sample, thus a sample design weight may be calculated that assigns individuals weights inversely proportional to the probability for selection (see Raudenbush et al. 2002). To weight the lower-level units (individual cases within counties) in this study, the following formula was used (adapted from Raudenbush et al. 2002):
69 N!Pu
J nj L L 1/Pr j=1 i=1 y
Where:
W;j is the weight given to the ith individual in county j
N is the total number of individuals in the overall sample;
Pij is the probability of selecting individual i in county j;
and where: J nj LLw r = N . .= 1= 1 t 1 y
In the instance of this project, the may be presented as follows:
= (Total Sample Size I (1/County Population) (# counties*#cases in county)* (1/County Population)
Individual and Contextual Variables
Dependent Variable Level-1-Environmental Concern
The purpose for using Confirmatory Factor Analysis (CFA) was to generate both First
Order and Second Order Factors that would adequately capture the facets of environmental concern across different concepts and measures. First order factor analysis is essentially the same as a general factor analysis, where factor scores are produced from a confirmatory model on the Level-1 SAA data set. Second order factor analysis is a process of taking the first order factor scores and determining the extent to which all initial 'factors may be used as a single measure; that is, a factor analysis on factor scores on the Level-1 data. A polychoric correlation matrix for knowledge, attitude, and behavior variables was generated to account for potential 70 problems in correlations between variables measured on dichotomies or ordinal scales (Kline,
1998). Dichotomous and ordinal variables are not normally distributed, which violates maximum likelihood estimation in CFA and results in the rejection of theoretically sound models (Kline,
1998). Solutions to these kinds of problems include generating a polychoric correlation matrix that produces estimates which may be used create First Order Factors. The First Order Factors were then used as variables in the Second Order Factor Analysis. The scores generated by the
Second Order Factor Analysis were then used to generate a single measure of environmental concern that included pro-environmental behaviors and attitudes at Level 1. In addition, First
Order factors generated from CFA were conceptually organized in terms of types of behaviors and within scales, to account for possible respondent biases in answering similar questions on similar response measures (see Kline 1998 or Dunlap and Jones 2001 ).
Once created, the polychoric matrix was used as input into Analysis of Moment
Structures software (AMOS) to perform confirmatory factor analysis. The matrix is included as
Table 3.1. First Order Factors included Behaviors, Awareness and Attitudes on Protection of
Habitat and Industry. Each of the First Order Factors was included in the CFA model to confirm that the models fit the data well. This step was performed per requirements of the technique and to determine the degree to which error variances of items may be required to covary with one another in the CFA model. In instances where indicators may measure something in common that is not in the model, error must be modeled to account for this unknown (Kline 1998). For the purposes of the analyses performed in this project, error variances within and across factors were used under the assumption that attitudes and behaviors on the natural environment may all be conceptually linked to an underlying construct, environmental concern. Hence, error variances between behavioral indicators on different factors and error variances between behavioral indicators and attitudes were to be included in models, if necessary, to identify a model that allows for Second Order Factor Analysis. All error variances are included underneath the item as a circle in subsequent figures.
71 Table 3.1--Polychoric Correlation Matrix Used as Input in ConfirmatoryFactor Analyses (Variables Relevant to Subsequent Models)
Aware of Aware of Subscribe Contribute Industries Protect Protect Water Air to Env. Money Env. Pay Clean Habitat Wilderness Pollution Pollution Pubs. Group Up
Aware of Water 1.00 Pollution Aware of Air 0.36 1.00 Pollution Subscribe to Env. 0.24 0.23 1.00 Pubs. Contribute Money Env. 0.23 0.16 0.66 1.00 Group
Protect 0.18 0.07 0.20 0.23 1.00 Habitat
Industries Pay Clean 0.17 0.14 0.26 0.29 0.31 1.00 Up
Protect 0.18 0.13 0.26 0.29 0.32 0.37 1.00 Wilderness
N 1044 1044 1044 1044 1044 1044 1044 Mean 0.39 0.19 0.28 0.41 3.73 4.03 3.58 Std. Dev. 0.49 0.39 0.45 0.49 1.12 0.9 1.02
72 The Full Factor Analysis is presented in the following graphic (Figure 3.1 ), with model fit indicators and standardized coefficients included. Fit measures for this model indicate that the model fits the data well. The Chi Square statistic is non-significant (Chi-Sq.=7.860; p.=0.726), which indicates that the departure of the data from the model is not significant. The Goodness of
Fit Index (GFI) and Adjusted Goodness of Fit Index (AGFI) are 0.998 and 0.995 respectively and the Root Mean Square Residual (RMR) is 0.005, where 1.0 for GFI and AGFI are ideal and an
RMR=O is the best possible fit for the model. In addition, modification indexes indicated that no change to the model would be required and that no error terms were correlated.
Table 3.2 presents details on the Full Factor Analysis, including standardized regression coefficients, significance tests and squared multiple correlations. As indicated in Table 3.2, standardized regression coefficients are 0.5 or above, regression coefficients operate in the expected positive direction, and all regression coefficients are significant at the 0.01 level. As noted above, Industries Pay, Air Pollution, and Money were used to create metrics for each factor, thus their regression weights are 1.0. Overall, the results are consistent with expectations and, as required for Second Order Factor Analysis, a total of three factors with at least two indicators per factor were identified. In short, this First Order Factor Analysis was performed as the necessary first step in creating a Second Order Factor that reflects an individual's level of environmental concern consistent with the concept of cultural capacity.
Subsequent to the First Order Factor Analysis, Second Order Factor Analysis was performed by creating an additional unobserved variable consistent with recommendations made by Kline (1998) and Maruyama (1998). Using the original model in AMOS, the correlation curves were replaced by linear relationships between each factor and the Second Order Factor. The variance of the Second Order Factor also was set to 1.0, as per requirements. Error terms were then assigned to each factor, as they were treated as outcome variables in analysis. The Second
Order Factor, Environmental Concern, is presumed to be a function of Behaviors,
Knowledge/Awareness and Attitudes on Protection and Industry.
73 .54
.38
Subscribe Contribute Aware Poll. AwarePoll. Habitat Industries Protect Publications Money ater Regio Air Region Protection Pay Poll. Wilderness
Figure 3.1--First Order CfA from Polychoric Correlation Matrix
Chi-Square=7.860, df=11, p.=.726 GFl=.998, AGFl=.995, RMR=.005
Table 3.2: First Order Factors, Attitudes, Awareness, and Behaviors
Factors and Reg Std. Critical Std. Reg. Sq. Mult. Indicators Estimate Error Ratio Wt. Correlation
Attitudes Industry Pay Poll. 1.000* 0.605 0.261 Habitat Protection 1.051* 0.098 10.737 0.511 0.366 Protect Wilderness 1.160* 0.101 11.484 0.618 0.382
Awareness Aware Air Poll. 1.000* 0.707 0.421 Aware Water Poll. 0.748* 0.086 8.662 0.649 0.500
Behavior Contribute Money 1.000* 0.850 0.723 Subscribe Pubs. 0.837* 0.056 14.981 0.775 0.600
*Indicates statistically significant at the 0.001 level
74 The Second Order Factor analysis model is presented in Figure 3.2. As with the previous model, fit measures for this model indicate that the model fits the data well. The Chi-Square statistic is non-significant (Chi-Sq.=7.860; p=0.726), which indicates that the departure of the data from the model is not significant. The Goodness of Fit Index (GFI) and Adjusted Goodness of Fit
Index (AGFI) are 0.998 and 0.995 respectively and the RMR (RMR) is 0.005, where 1.0 for GFI and AGFI are ideal and an RMR=O is the best possible fit for the model. In short, as above, the model fits well.
Table 3.3 presents details on the Second Order Factor, including regression weights, standardized regression coefficients, and factor weights. As indicated in Table 3.3, standardized regression coefficients are 0.5 or above, regression coefficients operate in the expected positive direction, and all regression coefficients are significant at the 0.01 level. The most important coefficients in this table are those dealing with the Second Order Analysis (top three lines) and the factor weights for each item. Also indicated is the importance of behavioral and attitudinal indicators in environmental concern relative to awareness. Squared Multiple Correlations indicate that 52.5%, 30.8%, and 55.8% of the variance in the Behaviors, Awareness and Attitudes respectively. Finally, the factor weights used to calculate the environmental concern variable for this analysis are included, where factor weights were multiplied by each observed value in the data set and summed to create environmental concern in terms of factor scores (see Kim and
Mueller 1978a, 1978b; Tarrant and Cordell 1997).
Independent Variables at Level 1
The remaining variables used in analyses at Level 1 include social class, time lived in residence, migration status and age. To create a composite measure of social class, the variables income and education were transformed in the following ways:
1. Income categories were recoded into mid-point values reflecting the mid-point in the income range for that category (e.g., "O to $15,000"; recoded this would be $7,500).
2. Education categories were recoded to fit with the appropriate number of years of education that reflects completion of the category (e.g., "College Degree"; recoded this would be 16 years). 75 .50 .38
Aware Poll. Aware Poll. Habitat Industries Protect ater Regio Air Region Protection Pay Poll. Wilderness
Figure 3.2--Second Order CFA from Polychoric Correlation Matrix
Chi-Square =7.860, df=11, p.=.726 GFl=.998, AGFl=.995, RMR=.005
Table 3.3: Second Order Factor Analysis on Attitudes, Awareness, and Behaviors
Factors and Reg Std. Reg. Sq. Mult. Factor Indicators Estimate Weights Correlation Weight Environmental Concern Attitudes 0.407* 0.747 0.558 Awareness 0.143* 0.555 0.308 Behaviors 0.302* 0.727 0.525 Attitudes Industry Pay Poll. 1.000* 0.605 0.366 0.189 Habitat Protection 1.051* 0.511 0.261 0.110 Protect Wilderness 1.160* 0.618 0.382 0.175 Awareness Aware Air Poll. 1.000* 0.649 0.421 0.286 Aware Water Poll. 1.337* 0.707 0.500 0.295 Behaviors Contribute Money 1.000* 0.850 0.723 0.617 Subscribe Pubs. 0.837* 0.775 0.600 0.424 *indicates statistically significant at the 0.001 level
76 Based on these transformations, a composite measure of social class was created by using the first principal components score for these items (based on the approach adopted by Sampson,
Raudenbush, and Earls (1997). The social class variable was used as the primary independent variable, as the literature and hypotheses focus on the relationship between individual economic position and environmental concern within economic contexts. The variables migrant status and
Time Lived in Region were used as given. As suggested in the literature, time spent in region and migrant status may influence the degree of environmental concern such that migrants may bring with them higher levels of concern. The literature suggests that both age and social class may be curvilinear in their relationship to environmental concern, thus both were subjected to curve-fit analysis. Each of these Level-1 independent variables was included in initial models prior to multilevel analysis.
It is important to note that variables such as race, gender and political ideology are important to the literature on environmental concern but were excluded from analyses in this research. The primary focus of this research is on the relationship between economic contexts and level.s of environmental concern between contexts. Although indicators like race, gender and political ideology are related to social class and possibly influenced by economic context, use of these variables in multilevel models may result in misleading results. For example, quota sampling was performed to achieve a representative sample of urban and rural residences across sub-regions in Southern Appalachia; that is, the sample is stratified by area of residence and not race or gender. Once the quota of households for a sub-region was filled, no more calls were made, thus affecting the degree to which one may generalize in terms of race and gender. The race distribution for the sample also was overwhelmingly white (92%) and some counties were exclusively white, which prevented cross-county comparison of the race and environmental concern relationship in multi-level models. The political ideology variable was coded in such a way to prevent accurate assessment of ideological position beyond Liberal, Conservative, and It
Depends. While this form of measurement is well suited to public opinion polling, difficulties arose in that the variable suffered from high levels of missing data and a high number of individuals selecting the "It Depends" category (a category that offers little information on political 77 ideology). Despite these problems, the variables reviewed in Chapter Two (age, social class, place of residence characteristics) and more heavily theorized were adequate for analyses and thus used to test hypotheses.
Context Variables at Level 2 Multi-level studies must use accurate measures of aggregate characteristics in a manner such that context effects will be present. Contexts were measured in a manner as proximate to the individual as is possible for each of the counties. Measures also were drawn from existing literature on economic, social, and natural environmental impact specific to the region and discussed in the literature previously. The indicators used to identify context effects for this study may be presented as follows (note that for all indexes the subscript "c" refers to the county value and the subscript "u" refers to the national value, both in year "j" (1995 for this study) and the subscript "idx" refers to the variable as an index):
1. Economic Distress (index recommended by the Appalachian Regional Council 2005; Wood 2005)--This index measures overall county productivity, unemployment, labor force dependency, and spending on non-productive populations all as an indicator of economic distress or poor economic health in the county overall:
ECONidxci = [PCMlidxci + URT1dxci + LFPOPidxci + TFPidxcil / 4
Where:
PCMl1dxcJ = Per Capita Market lncome--Defined as per capita income earned from wages, dividends, and rent, adjusted for residence and excluding transfer payments (i.e., market earnings) relative to national county average.
= PCMlidxcJ 100 / [1 - [(PCMluJ - PCMlcJ) / PCMlui ]]
URTidxcj = Unemployment Rate-Defined as proportion of working aged persons not employed relative to national county average.
= URTidxci [100 * URTcil / URTuJ
LFPOPidxcj = Labor Force Dependency--Defined as relative number of non working persons to working labor force; that is, a measure of dependency relative to national county average.
LFPOPidxcJ = 100/ [1 - [(LFPOPui - LFPOPcJ) / LFPOPuJ ]]
78 TFPidxcj = Per Capita Transfer Payments--Defined as payments made to individuals or entities that do not generate self-sustaining levels of income relative to national county average.
c TFPidx j = 100 [PCTPcj / PCMI cil / [PCTPuj / PCMluj ]
Higher scores on this index indicate poorer economic health (or higher distress) in counties, particularly those scoring above 100, which is the national average in calculations (Reliability: Chronbach's Alpha=0.84).
2. Social Distress (index based on items recommended in calculating county-level quality of life indexes (see Gardner and Raudenbush 1991; Sampson et al. 1997; Hanham, Hanham, and Banasick 2000)--This index measures income inequality, vulnerable populations, and relative consumptive practice all as an indication of social stressors that may lead to differences among social groups, as well as those stressors that will be experienced more directly by individuals:
SOCcj = [EDATTidxcj + FHPOVidxcj + KPOVidxcj + POVidxcj + GINlidxcj] / 5
Where:
EDATTidxci = Proportion of individuals 25 years or older without a Bachelor's or higher (i.e., higher education rate) compared to national county average. This measure indicates the proportion of persons with low educational attainment, which may restrict job opportunities or prevent industry location in an area.
EDATTidxcj = (100 * EDATTd) / EDATTuj
FHPOVEidxci = Proportion of female headed households living in poverty compared to national county average. This measure reflects the proportion of persons more susceptible to certain forms of environmental risk
FHPOV1dxcj = (100*FHPOVcj)/ FHPOVuj
KPOVidxci = Proportion of children living in poverty compared to national county average. This measure reflects the proportion of children more susceptible to certain forms of environmental risk.
KPOVidxcj = (1OO*KPOV cj)/ KPOVuj
POVidxcj = Proportion of individuals living at or below poverty level (i.e. poverty rate) compared to national county average. This measure is used to estimate the relative vulnerability of
POVidxci = (100 * POVci) / POVui
79 GINlidxci = GINI Coefficient of income inequality compared to national average.
GINl1dxci = (100 * GI Nici) / GINlui
Higher scores on this index indicate poorer quality of life ( or higher social distress) in counties, particularly those scoring above 100, which is the national average in calculations. Note that measures from the most proximate year were used from figures provided by the U.S. Census (U.S. Census USA Counties 2005). In the instance that information was not available for 1995 the proportions were used to estimate the relevant population based on 1995 figures. Information on education was estimated from 1993 data and information for female headed households was estimated from 1990 data, as this information was the most proximate available at the time of analysis. Poverty rates and gini coefficients were identified for the year 1995 and used to create the indexes. The overall reliability of this index is reported at: Cronbach's Alpha=0.74.
3. Pollution Context (based on items reported by SCORECARD (2005) and U. S. Department of Agriculture (2005))-This index measures degree of toxin release, land loss, and development all as indications of environmental harm.
NATidxcj = [ NCNR1dxcj + CNR1dxcj + DEV1dxcj ] / 3
Where:
NCNRidxcj = Noncancer risk score as calculated from environment releases by type of pollution as compared to national county average.
NCNR1dxcj = 100 / [1 - [(NCNRuj - NCNRci) / NCNRuj ]]
CNRidxcj = Cancer risk score as calculated from environment releases by type of pollution as compared to national county average.
CNR1dxcj = 100 / [1 - [(CNRuj - CNRcj) / CNRuj ]]
DEVidxcj = Proportion of developed land (land not used for farming or other uses) relative to total land area in county as compared to national county average.
DEVidxcj = (100 * DEVcj) / DEVuj
Higher scores on this index indicate poorer environmental quality in counties, particularly those scoring above 100, which is the national average in calculations (Reliability: Chronbach's Alpha=O. 78).
Each of the variables used in analysis for comparison comes from a governmental or private industry data source. Information on economic patterns was obtained either through the Bureau of Labor Statistics based on recommendations from the Appalachian Regional Council.
80 Population density and demographics come from the U.S. Census. Information on pollution characteristics of certain counties, as well as increased risk, comes from Scorecard
Environmental Assessment, Environmental Protection Agency, and U.S. Department of
Agriculture.
Statistical Models Used in Analysis HLM techniques began with an examination of the variation in the Level-1 dependent variable, followed by specification of the multi-level model. The initial model is known as the random effects model, which is essentially a one-way random effects Analysis of Variance
(ANOVA) model. Such models are sometimes referred to as variance components models
(VCM) and are used in HLM to generate components used in intraclass correlation coefficients (ICC); thus representing the proportion of total variability accounted for by differences in higher level units ( see Raudenbush and Bryk 2002; Raudenbush et al. 2002). Subsequent to initial models, multi-level models were run on the variables discussed above. To illustrate, the hypotheses for this project are as follows (with sample formal models based on Raudenbush and
Bryk 2002; Raudenbush et al. 2002; Singer 1998; Snijders and Bosker 2002):
H1. Individuals will have lower environmental concerns in counties with higher economic distress (i.e., the intercepts of environmental concern will vary across counties and the contextual variables will explain some of the cross-county variation). LEVEL 1: Yg = �o; + rg LEVEL 2: �o; = Yoo + Yo1 � + Uoj
Where: Yu = Values for individual level Environmental Concerns �o; = County means for Environmental Concerns rg = Individual deviation from county mean Yoo = Grand mean across counties Yo1 = Influence of Economic Distress � = Values for county level Economic Distress u0i = Unique increment in intercept associated with county j
Note that models of this nature are referred to as means as outcomes, where the grand mean and influence of LEVEL 2 variables are treated as fixed, but the deviations at both levels are treated as random, such that: fixed random
Yu = Yoo + Yo1 � + Uoj + rg 81 H2. Social class and economic distress will interact across levels such that the relationship between social class and environmental concern scores will vary across counties, where economic distress explains some of this cross-county variation (i.e., slope coefficients between social class and environmental concern scores will vary across counties).
Basic model form with LEVEL 1 and LEVEL 2 predictors:
= LEVEL 1: Yg Poi + P,i X1j + rg LEVEL 2a: Poi = Yoo + Y o1 � + Uoj (Predicting Intercepts) LEVEL 2b: P1i = Y10 + Y11 � + U1j (Predicting Slopes)
Based on the basic model form, a reduced form of the equation, substituting LEVEL 2a and LEVEL 2b into LEVEL 1, may be presented to identify the cross level interaction for this test (placing random terms at the end of the equation and the cross-level interaction in bold):
random
Where: yq = Values for individual level Environmental Concerns rg = Individual deviation from county mean Yoo = Grand mean across counties Yo1 = Effectof economic distress y,0 = Average Social Class-Environmental Concern slope across counties � = Values for county level Economic Distress Xij = Values for Individual Social Class Uoj = Unique increment to intercept associated with county j u11 = Unique increment to slope associated with county j
Note that this sample model also is applicable to the final hypothesis dealing with time in residence. Discussion and Conclusion
All of the statistical analyses, variables, and additional techniques outlined above are well grounded in sociological literature, but rarely used in combination to test multi-level models. The rarity of use is a function of the newness of such approaches in sociological analysis, and not based on any specific problems associated with any of the afore-mentioned techniques. Indeed, many across the sub-disciplines of sociology are now calling for approaches similar to the type outlined above, but a lag in methodological sophistication and a lack of clarity in conceptualization 82 have been barriers to these approaches. In fact, the required statistical procedures are now available in most statistical packages (SPSS, SAS, HLMS, LISREL), so opportunities for these types of analyses are increasing. The advent of these techniques represents something of a paradigm shift in theorizing and testing relationships across levels. In this period of shift, researchers must conduct analyses using extant theories and tested forms of measurement to extend the knowledge of the discipline, especially in guiding future theory construction, conceptualization, and measurement. This project was specifically designed to take many such first steps by incorporating aggregate and individual level measures into models to identify potential context effects using a dataset that was originally created for this purpose. Thus, this project fulfills many needs in the overall discipline of sociology and specifically in environmental sociology by taking some of the necessary preliminary steps toward understanding the role of context, especially in terms of Environmental Concern and Cultural Capacity Building.
As will be discussed in the next chapter, adopting this approach does provide useful information on the relationship between context and levels of environmental concern. By following the recommendations of Dunlap and Jones (2001) and carefully measuring environmental concern multi-level models may be run that adequately assess the influence of context on individual level environmental concern. Environmental concern is a necessary component to building the cultural capacity necessary for ecological modernization.
Understanding the effects at the local level may provide a fruitful common ground for dialogue between stake-holders in various issues at the local level. Information on how environmental concern may vary across contexts may also provide natural resource professionals and environmental groups useful information on channeling environmental concerns through political mobilization. All of these potential outcomes are contingent upon careful measurement and model development across levels of analysis. More specifically, economic contexts play a role in environmental concerns and cultural capacity, which affect the degree of institutional changes consistent with ecological modernization. By contextualizing understandings of environmental attitudes, consumptive behaviors, and political behaviors as environmental concern, clarity may be gained on how economic context provides opportunities or constrains collective action within 83 local areas. Thus, by using the measures and methods outlined above, a great deal of information is gained and such information is indispensable to fomenting environmental concern.
84 CHAPTER FOUR
TESTING HYPOTHESES AND EXPLORING MULTILEVEL MODELS
Introduction and Overview
Much of what sociologists study is inherently multilevel, and the multilevel character of sociological data is highly salient to environmental issues. Humans are embedded in environmental and economic contexts. These contexts influence how humans collectively respond to environmental pressures and how humans individually make sense of environmental issues. To collect information only on aggregates or individuals may lead to ecological or atomistic fallacies whereby inferences and explanations are flawed. In fact, as Luke (2004) notes, much of the difficulty in keeping contexts and individuals separate is more a problem of inference rather than of measurement. Statistically, sociologists have typically followed the path of least resistance by focusing on individual level phenomena even though they often are implicitly claiming that context influences individual level phenomena. The consequence of this research strategy is that contexts are not modeled and thus left as residuals in individual level error terms (Luke, 2004). Practical information derived from multilevel models may be used to inform natural resource professionals or environmental activists on how to approach communities that may vary in the degree to which they express environmental concerns. In regard to the environment, information on the effect of context is highly important to understanding the degree to which individuals may express or act on environmental concerns, thus placing pressures on social institutions for environmental reform. As will be discussed in the next section, many natural resource professionals in the Southern Appalachian region acknowledge both the theoretical and practical importance of contextualizing information on the Southern Appalachian region for natural resource management. Thus, for these reasons, this project began with the assumption that multilevel modeling would inform both the literature on environmental concern and inform natural resource professionals in the region.
This chapter presents results of tests for the contextual influence of economic distress, social distress and pollution risks on environmental concerns within and between counties in the
85 Southern Appalachian Region (SAR). The first section of this chapter provides a thumbnail sketch of the history and environmental issues of the SAR. In this section, values of the contextual variables were collapsed into high, moderate and low categories and mapped using
Arcview GIS, to illustrate the distribution of various types of distress or risk residents of counties face. The relationships between the contextual variables are discussed and implications for models are presented. The next major section of this chapter provides information on Level-1 variables that will be used to explain within county variation and cross-level interaction effects. In particular, the variables social class, time in residence, migration status and age are explored in this section. For each of these variables, the potential for different effects at Level-1 are discussed and curvilinear relationships are tested. In the following section, multilevel models are presented. This section begins with a recap of hypotheses and an exploration of unconditional random effects for counties in the region, to identify the amount of variation between counties that may be explained. Subsequent analyses in this section focus on the influence of Level-2 on
Level-1 variables, with additional tests for fixed contextual effects on variables important to Level-
1 relationships. This section also includes explorations beyond initial hypotheses to offer clarity on the effect of context on environmental concern. Finally, a summary of findings and a discussion and conclusion are presented.
The Southern Appalachian Region
A Brief History of the Region and Importance of Study
The SAR includes approximately 34. 7 million acres of cropland, pastures and forested areas across Alabama, Georgia, North Carolina, South Carolina, Tennessee, Virginia and West
Virginia (Cordell et al. 1996). Originally occupied by various Native American tribes and explored
th by the Spanish, land in the region was first sold and settled by European colonists in the mid-18 century {about 1748) (Guffey 2005). For the most part, early communities formed around primary water sources and residents engaged in hunting or subsistence farming until the beginning of the
th 20 century (Cordell et al. 1996). As industry grew in many other parts of the nation, particularly the East, subsequent to the Civil War, the SAR was considered somewhat unique in that its residents were poorer, less educated, and governed by kinship ties (Cordell et al. 1996). 86 Eventually private and public organizations began to focus on development in the region, with an emphasis on improving living conditions of residents in the areas. Examples of such organizations include the Appalachian Regional Commission, the Economic Development
Administration, and the Tennessee Valley Authority, which continue all their respective work today (Cordell et al. 1996). As a result, the SAR has experienced a great deal of growth and increased urbanization. For example, over the course of the 20th century, the population in the region nearly tripled from just under one million to over three million (English and Huss 2005).
Further, population density in the region has increased from 40 persons per square mile to 110 persons per square mile over the same time period (English and Huss 2005). This kind of growth has focused a great deal of attention on the effects of change on natural resources and pollution in the region.
As noted in the previous chapter, some of the most pressing issues in this region are those dealing with economic health, poorly planned growth and pollution effects on the region.
Indicative of this growing concern is the trend toward providing recycling centers and public landfills, especially in Tennessee counties rep'resenting large populations in the region (Bradley and Nolt 2005; English and Huss 2005). Yet, as noted by Bradley and Nolt (2005), the SAR is still deficient in its efforts to curb illegal dumping or disposing of highly toxic wastes. For example,
Tennessee ranks fourth among the highest contributors to toxic wastes in the United States, with much of this toxic waste generated in the region (Bradley and Nolt 2005). Likewise, the SAR is characterized by mining practices that continue to deposit mine slurry or runoff into water sources, despite regulations and efforts to protect these water sources (Bradley and Nolt 2005).
Further, farmland converted to other uses such as housing or other forms of development is a problem in the region. For example, using National Resources Defense Council data, English and Huss (2005} estimate that between 1992 and 1997 nearly 2.1 million acres of land in the region was converted to development, with nearly 28% of that amount classified as prime farmland. Still further, between 1990 and 1999 air pollutants increased 19% in the region and
59% of this increase may be attributed to area sources (English and Huss 2005). In short, the
87 environment of the SAR is experiencing increases in negative effects of due to development in the area.
Given these conditions and trends, residents in the Southern Appalachian region have expressed a greater need for understanding, and local input, into regulation of natural resources, which makes this data particularly pertinent to the research questions posed. As noted by Cordell et al. (1996), residents in the area want more input into the decision making process and want more information about decision making processes. Those community residents included in focus groups indicate that they support resource management but fear outside interests and their influence on management practice (Cordell et al. 1996). In particular, there is a general concern that outside interests may harm local communities and there exists a desire for more information on ecosystem management (Cordell et al. 1996). The central problem is that many local economies are tied to natural resources (Cordell et al. 1996). This condition has created the need for greater attention to local circumstance, particularly in terms of the economic health of an area and its residents (Cordell et al. 1996). Thus, as Cordell et al. note (1996:42):
Planners and policy staff of public agencies will be challenged to account for these differences and be aware of the potentials for inequitable effects of resource management choices, particularly negative ones, among local communities. Of particular concern is the need to balance local interests among retirement, logging and tourism communities.
Based on the diversity of communities and the desire for more information on the relationship between economic health and concerns of residents, this data set is particularly well-suited for the multilevel analyses conducted below, as well as contributing to the literature on cultural capacity within context.
Contextual Indicators and Regional Maps
To better characterize this region in terms of the measures used, analyses began with a general look at the distribution of economic distress, social distress and pollution risks in the area.
Using Arc-View GIS, an overall map of the counties was generated and the contextual variable scores for each county were mapped based on position relative to national averages. Figure 4.1 presents a general map of the region with counties included in subsequent analyses.
88 Figure 4.1. Southern Appalachian Counties
89 As noted in Chapter Three, certain counties were excluded from analyses because of either excessive missing data among cases within the county or single respondents in the county. The final data set included 97 (out of an original 135) counties with 1,044 respondents.
To better characterize those counties included in analyses in terms of their relative economic distress, social distress, and pollution scores, the following maps were generated.
Counties were color coded relative to national average on specific indicator. The darker the color of the county the higher the level of distress as identified in terms of standard deviations from the mean, where:
• White=2 Standard Deviations Below National Mean;
• Light Gray=1 Standard Deviation Below the National Mean.
• Dark Gray=1 Standard Deviation Above the National Mean
• Black=2 Standard Deviations Above the National Mean
As discussed in Chapter Three, raw scores on each index indicate the percentage difference from national average (e.g., a score of 150=50% above national average), and higher scores indicate worse conditions.
The first map, Figure 4.2, illustrates the variation in levels of cancer and non-cancer pollution risk distributed throughout the region. As noted in the graphic, counties in white represent much lower than nation average risks (10.3% of all counties included in analysis), counties in light gray indicate moderately lower risks than national average (49.5%), counties in dark gray indicate moderately higher risk than national average (26.8%), and counties identified in black indicate a much higher than national average risks (13.4%). Overall, most counties in the region are relatively low on pollution risks when compared to national averages, with exceptions including more industrialized counties like Calhoun, AL and Knox, TN that are slightly above national averages (ranging from about 1 % to 5% above).
90 Black=2 Std. Dev. Higher than National Average Dk. Gray=1 Std. Dev Higher th an NationalAverage Lt. Gray= 1 Std. Dev. Lower than National Average White=2 Std . Dev.Lower than National Average
Figure 4.2. Southern Appalachian Region Pollution Index Scores
91 Figures 4.3 and 4.4 illustrate the variation in levels of economic and social distress distributed throughout the region by county. As with the pollution scores, the color codes in this map represent the score of the county relative to the national average scores. Of note in Figure
4.3 is that roughly 51 % of all counties in the region are moderately worse or much worse than national averages on measures like per capita market income, transfer payments, unemployment, and transfer payments, thus indicating the relative economic distress of the region. Such counties represent the more commonly referenced poorer areas in the SAR. Also of note is the fact that nearly half of the severely distressed counties are found in Tennessee, followed by
Georgia, North Carolina, Alabama, and Virginia. In addition, the moderately and severely distressed counties appear to be distributed throughout the region, instead of concentrated as with the pollution risk scores.
Finally, Figure 4.4 illustrates the relative social distress experienced by residents of each county, as determined by proportion of female headed households in poverty, proportion of individuals with lower educational attainment, proportion of children in poverty, overall poverty rate, and inequality index scores. Overall, when compared to economic distress, the region has fewer counties in severe social distress relative to national averages. Those counties experiencing the lowest social distress represent 14.4% of counties in the region. Moderately low social distressed counties represent 30.9% of all counties and moderately high social distress make up 41.2% of all counties. Those counties experiencing the highest levels of inequality make up 13.4% of counties in the region. In many instances, there is some overlap between economic distress and social distress such that higher economic distress corresponds to higher social distress.
92 Black=2 Std. Dev. Higher than National Average Dk. Gray=1 Std. Dev. Higher than National Av erage Lt. Gray=1 Std. Dev.Lower than National Av erage White=2 Std. Dev. Lower than National Average
Figure 4.3. Southern Appalachian R.egion Economic Distress Index Scores ·
93 Black=2 Std.Dev. Higher than National Average Dk. Gray=1 Std. Dev. Higher th an National Average Lt . Gray,=1 Std. Dev. Lower than National Average White=2 Std. Dev. Lower than National Average
Figure 4.4. Southern Appalachian Region Social Distress Index Scores
94 To illustrate the relationship between these contextual variables, a correlation was performed among the Level-2 indicators. Table 4.1 presents the Pearsons' Correlation
Coefficients for these measures. As indicated in the table, the correlation between economic distress and social distress is 0.725 and statistically significant, thus indicating that as values of economic distress increase so too do values of social distress. The correlation between economic distress and social distress is strong, which suggests that both measures are tapping into forms of distress but may be collinear in multilevel models. The correlation between economic distress and pollution is inverse and statistically significant, suggesting that as values of economic distress increase pollution scores decrease. This correlation also operates in the expected direction such that areas with higher pollution have more industry as a basis for economic growth. The correlation between social distress and pollution scores is essentially O and non-significant, indicating that there is no relationship between levels of social distress and pollution risk. The lack of a relationship between social distress and pollution risk is unexpected, since from the literature one might conclude that vulnerable populations or poverty would be related to pollution risks. One possible explanation is that the relative risks of pollution are concentrated in specific, highly populated areas in the county, while social distress is distributed across the region. This concentration coupled with low pollution risk, on average, across the region may explain this lack of a relationship In addition, the lack of a relationship between social distress and pollution context may be an artifact of measurement, which is fully discussed in the next chapter on recommendations for future research.
Table 4.1 : Correlations Between Contextual Variables
Economic Social Pollution Distress Distress Risk Economic Distress 1 Social Distress 0.725** 1 Pollution Risk -0.435** -0.004 1 97 97 97 ** Correlation is significant at the 0.01 level (2-tailed)
95 Characterizing Contexts in SouthernAppalachia
While most of the region is relatively low in terms of overall pollution risk when compared to national averages, it is apparent that some parts of the region experience not only greater pollution risks but also higher degrees of economic distress and social distress. In some instances, there is a great deal of overlap between these characteristics such that counties experiencing economic distress also experience higher degrees of social distress. When comparing across indicators, only 14.4% of all counties included in analyses reported lower than national average scores across all measures. Those counties with at least one score on one indicator falling above the national average represented 35.1% of all counties in the region.
Counties scoring high on two context measures represented 41.2% and counties scoring high on all three context measures consisted of 9.3% of all counties in the Southern Appalachian region.
Although a rough guideline, these scoring positions may be thought of as a range from better to worse contextual conditions in this region.
To briefly characterize these patterns, Table 4.2 presents a list of these counties by number of contextual scores above the national average. In effect, the counties do vary in the extent to which they experience distress or risk, but in many instances counties that experience one form of distress or risk also experience others. As noted above, policy makers, natural resource professionals, and local citizens wish more dialogue on the state of the region and wish greater efficiency in planning for natural resource management. By taking into account these conditions, a better understanding of the region and its residents may provide opportunities for better planning in dealing with these conditions. Thus, in the models that follow, contextual variables will be used to explain environmental concerns within and between counties.
Level 1 Model of Environmental Concern
Expectations from the Literature, Distributions� and Correlations
Prior to performing multilevel analyses, the Level 1 data was subjected to correlation and regression analysis to determine the extent' to which social class, time in residence, migration status, and age may be used to explain environmental concern. As noted in Chapter Two, the relationship between social class and environmental concern is complicated. Often the 96 Table 4.2: Overall Summaryof County Scores for Southern Appalachian Region
No High Scores on Any Single High Score Two High Scores High Scores on All Measure Habersham Ga CleburneAL Cherokee AL Calhoun AL Murray GA DeKalb AL Randolf AL Talladega AL White GA Bartow Ga Chattooga GA Haralson GA Surry NC Catoosa GA Dawson GA Polk GA Wilkes NC Gordon GA Fannn GA Stephens GA Loudon TN Hall Ga Floyd GA Transylvania NC Amherst VA Lumpkin GA Franklin GA Hamblen TN Augusta VA Paulding GA Glilmer GA Unicoi TN Botetourt VA Pickens GA Union GA Pulaski VA Franklin VA Rabun GA Walker GA Greene VA Whitfield GA Alleghany NC Madison VA Buncombe NC Ashe NC Rockbridge VA Burke NC Jackson NC Shenendoah VA Caldwell NC Macon NC Henderson NC Madison NC McDowell Mitchell NC Greenville SC Yancey NC Oconee SC Anderson TN Pickens SC Bradley TN Blount TN Carter TN Sevier TN Claiborne TN Albemarle VA Cumberland TN Bath VA Greene TN Bedford VA Hamilton TN Floyd Hawkins TN Frederick VA Jefferson TN GIies Knox TN Montgomery VA Marion TN Nelson VA McMinn TN Page VA Monroe TN Patrick VA Roane TN Warren VA Sullivan TN Hardy WV Union TN Pendleton WV Washington TN Carrol VA Grayson VA Smyth VA Washington VA Wythe VA Hampshire WV
97 relationship between social class and environmental concern is a function of how one measures environmental concern. The measure used for environmental concern as an element of cultural capacity for this project includes items on financial contribution, subscriptions to environmental publications, knowledge and protection of land. Given this measurement of environmental concern, the expectation is that the relationship between social class and environmental concern will be positive. Proximity to environmental harms and attachment to an area may also increase awareness of problems and, therefore, environmental concern, although length of time in residence may reduce the importance of the environment as an issue. Such higher concerns also may be a function of the migration status of individuals, where migrants may have higher concerns and longer time in residence may be inversely related to environmental concern. For example, as noted by Jones et al (2003), the influx of affluent migrants to rural areas may awaken or enhance existing concerns in the population, because migrants bring with them higher levels of concern. Finally, age and environmental concern are expected to be inversely related such that younger individuals will express higher levels of concern, and older cohorts may be expected to express lower concerns than younger cohorts.
To assess these relationships at Level 1, correlations and simple regression were performed. As migrant is dichotomous and all other variables are interval, the relationships between migrant and all other variables are point biserial coefficients instead of Pearson's r coefficients. Table 4.3 presents the correlations among variables to be used in subsequent regression analyses. As indicated in the table, social class is positively correlated with environmental concern, although the degree of association is moderately low. Time in residence and age are inversely related to environmental concern, although the correlation with time in residence is low and age is very low with environmental concern. Migrant is dummy coded for
1 =Migrant and O=Lifelong Resident. The correlation between migrant and environmental concern is positive and low, which indicates a slight tendency for migrants to express greater concern.
These findings are consistent with what may be expected in terms of Level 1 relationships discussed in Chapter Two.
98 Table 4.3: Correlations Between Environmental Concern, Social Class, Time in Residence, Age, and Migrant Status
Environmental Social Time in Migrant Concern Class Residence Age Status Environmental Concern 1.00 Social Class 0.33** 1.00 Time in Res. -0.21** -0.27** 1.00 Age -0.09** -0.11** 0.55** 1.00 Migrant -0.19** 0.23** -0.65** 0.05 1.00 --- N 1044 1044 1044 1044 ** Correlation is significa
Level egressions
The variables social class, time in residence, migration status, and age were separately included in simple regression models on environmental concern. The purpose of these regressions is to better understand the potential effect of each variable on environmental concern without controlling for the effect of county context. Identifying relationships at Level 1 provides a way to elaborate on how and why context effects are identified and important.
Table 4.4 presents the simple regression for social class. All coefficients and the overall model are significant, and the model explains 10.9% of the variance. The class coefficient indicates that an increase in social class results in an increase in environmental concern. As the literature suggests that the relationship between social class and environmental concern may be curvilinear, a curve-fit analysis was performed to asses the degree to which a quadratic term increases the explanatory power of social class. Table 4.5 presents the results of this curve-fit analysis. As indicated in the table, including the quadratic term in the model increases the explained variance to 11.4%. The quadratic model indicates the following:
2 Y = 2. 593 + (social class * 0. 253) - (social class * 0. 045)
In effect, the negative value for the quadratic term suggests that when social class scores are above 2.8, environmental concern is expected to decrease (2.8 may be identified by the following equation: [0.253 / (2*0.045)]-see Norusis, 2003).
To illustrate this relationship, Figure 4.5 is presented. When considering the figure, there does appear to be a tendency for individuals of lower social class and higher social class 99 Table 4.4: Regression of Environmental Concern on Social Class Parameter B Std. Error Beta t Sig. Intercept 2.547 0.021 123.635 0.000 Social Class 0.234 0.021 0.331 11.31 5 0.000 R-Sq.=0.109; F=128.03; df=1, 1043; p.<0.001
Table 4.5: Curve-fit Analysis for Regression of Environmental Concern on Social Class Equation R-Sq F df1 df2 Sig Intercept b1 b2 Linear 0.107 128.04 1 1042 0.00 2.548 0.231 Quadratic 0.1 14 69.25 . 2 1041 0.00 2.593 0.253 -0.045
0 Observed ---Linear 00 0 00 -· ·-0ua dratic 400
0 00 $ I C a• • I
3DO '
• § 0 2DO 0 • .. • i
100 : . 0
-200000 -1.00000 0.00000 1.00000 200000 3.00000 Social Class
Figure 4.5--Cmve-frtfor Social Class and Environmentil1 Concem
100 to have lower environmental concern scores. This nonlinearity is more apparent at the upper end of social class. Subsequent investigation into the cases that influence the relationship between social class and environmental concern in this way suggest that while this curvilinear relationship may be substantively interesting and statistically significant, little may be gained by including the quadratic in multilevel linear models. More specifically, when controlling for county in multilevel models the social class scores of these cases are an attribute of county distributions, where they are not as influential on the relationship between social class and environmental concern. Thus, social class will be treated as linear in subsequent models.
In the next model, time in residence is added to social class as a predictor of environmental concern. Table 4.6 presents the results of the regression model for class and time in residence. Both social class and time in residence are statistically significant in this model.
The social class coefficient is positive and indicates that an increase in social class results in an increase in environmental concern when holding time in residence constant. Time in residence is negative and the coefficient indicates that an increase in time in residence results in a decrease in environmental concern, which is a small but detectable effect. The model explains 12.5% of the variance in environmental concern. In this model, social class appears to be the stronger predictor of environmental concerns, although the addition of time in residence does increase the explained variance (incremental F-test when compared to class alone: F=19.02; df=1, 1043; p<0.001 ). Both social class and time in residence are central to hypotheses tested in the next section and both variables are included in the multi-level models.
Table 4.6: Regression of Environmental Concern on Social Class and Time in Residence Parameter B Std. Error Beta t Sig. Intercept 2.694 0.039 68.843 0.000 Social Class 0.208 0.021 0.295 9.775 0.000 Time in Residence -0.005 0.001 -0.133 -4.400 0.000 R-Sq.=0.125; f=74.828; df=2, 1041 ; p.<0.001
101 Given the significance of time in residence and the importance of migrant status in the literature, migrant status was included in the following regression model. As indicated in the correlations in Table 4.3 above, migrant status and time in residence are highly correlated, thus time in residence was eliminated from the model, because the literature suggests migrant status may be the more theoretically sound predictor. This model will allow for the identification of differences between migrants and non-migrants in terms of the relationship between social class and environmental concern (Migrant= 1, Native=O). In addition to its theoretical importance, migrant status may be a more conceptually sound way of identifying the effect of time in residence, which is discussed in the explorations at the end of the next section.
Table 4.7 presents the regression model for migrant status and social class. As indicated in the previous table, the model and all coefficients for this model are statistically significant, with
12.3% of the variance in environmental concerns explained. The coefficient for migrant status provides the difference in intercepts for regression lines between migrant and native residents.
As the lines of each group are forced to be parallel in this model, this coefficient also represents the constant separation between migrants and natives. The class coefficient represents the common slope for migrants and natives, which may be interpreted as a 0.215 increase in environmental concern for every one unit change in social class for migrants and natives. In effect, the significance of the migrant coefficient indicates that migrants are, on average, slightly more concerned than natives and this difference is statistically significant (incremental F-test when compared to social class alone: F=16.65; df=1, 1043; p<0.001).
To aid in seeing these differences, Figure 4.6 illustrates the regression lines of social class and environmental concern between migrants and native residents in the counties of the region. In effect, migrants have, on average, higher environmental concerns than life long residents in this sample. As some counties may have higher rates of migration than others in the region, the migrant effect will be incorporated into multilevel models near the end of this chapter, to assess its importance in explaining differences in environmental concern when controlling for county. 102 'f
Table 4.7: Regression of Environmental Concern on Migrant Status and Social Class
Parameter B Std. Error Beta t Sig. Intercept 2.475 0.027 91 .12 0.000 Migrant Status 0.169 0.042 0.1 19 3.989 0.000 Social Class 0.215 0.021 0.304 10.203 0.000 R-sq.=0.123; F=72.489; df=2,1041; p.<0.001
Environmental Concern ------
2 - -Migrant - Non-Migrant 1. 5
0.5
-2 -1.5 -1 -0.5 0 0.5 1.5 2 2. Social Class 5
Figure 4.6-Migrant Effect on Social Class and Environmental Concern
103 In addition to time in residence and migrant status, the relationship between age and environmental concern was assessed. The primary importance of identifying this relationship is not to directly test hypotheses but to inspect the degree to which age may have an effect on environmental concern. As indicated in Chapter Two, age is a consistently important predictor of environmental concern and age cohort is considered theoretically important in explaining environmental concern. This relationship may be important in understanding environmental concern within economic context, thus the following models were run. As the literature also suggests that the age-environmental concern relationship also may be curvilinear, a simple regression and curve-fit analysis are presented. Table 4.8 presents the results of the simple regression of age and environmental concern. Of note in Table 4.8 is the expected inverse relationship and the low percentage of variance explained. In this model, age explains less than
1 % of the variance in environmental concerns. As noted in the literature review above, research on the relationship between age and environmental concern over time suggest that from about
1980 forward, age has become less important in explaining environmental concerns, although the variable is important conceptually to understanding environmental concern (see Jones and
Dunlap 1992). One explanation for the low proportion of variance explained by age in this model is, as with social class, that the relationship between age and environmental concern is curvilinear.
Table 4.9 and Figure 4. 7 present the results of curve-fit analysis on the relationship between age and environmental concern. As indicated in Table 4.9, including the quadratic does increase the amount of variance explained. But, the addition of the quadratic term appears to have little effect on the overall model. More specifically, the quadratic model indicates the following:
2 Y = 2. 593 + (age * 0.031) - (age * 0. 0004)
In effect, the negative value for the quadratic term suggests that when age increases beyond about age 39, environmental concern is expected to decrease (this point may be identified by the following equation: [0.031 / (2*0.0004)]-see Norusis, 2003). 104 Table 4.8: Regression of Environmental Concern on Age
Parameter B Std. Error Beta t Sig. Intercept 2.726 0.066 41.447 0.000 Age -0.004 0.001 -0.087 -2.810 0.000 R-Sq.=0.008; F=7.796; df=1, 1043; p.<0.001
Table 4.9--Curve-fit Analysis for Regression of Environmental Concern on Age Equation R-Sq F df1 df2 Sig Intercept b1 b2 Linear 0.008 7.896 1 1042 0.005 2.726 -0.004 Quadratic 0.029 15.309 2 1041 0.000 1.979 0.031 -0.0004
0 Observed ---Linear 0 0 " 00 4DO -- · Quadratic
3DO
2DO
0 .. 1DO
20 60 80 Age Figure4.7 --Curve-fitAnatys:is for Age and Em"ironmental Concern
105 Of note, the curved line at the lower end of age is misleading in this graphic, as no respondent was included in analysis that was younger than 18 years old. The most apparent divergences from linear are for persons aged 18 to about 30 and about 65 and above. Although not completely apparent in the graphic above, the results indicate that those born roughly between
1942 (age 53) and 1964 (age 31) have higher environmental concern scores than other respondents. Unlike the curvilinear relationship between social class and environmental concern, this relationship deserves greater attention. As noted in the literature, the curvilinear relationship between age and environmental concern may, in fact, be a cohort effect.
To understand the nature of this potential cohort effect, a cohort variable was calculated by subtracting age from the year of the survey (1995) and cohorts were then recoded into specific age groups: 1) Born Prior to 1942; 2) Born Between 1942 to 1964; and 3) Born 1965 and After.
Figure 4.8 presents the mean environmental concern score for each of the cohort groups. Figure
4.8 demonstrates that individuals born between 1942 and 1964 have higher average environmental concern scores as a group, when compared to the other age cohort groupings.
Likewise, individuals born 1965 and after have higher average environmental concern scores than individuals born prior to 1942. To determine if these differences were significant, Analysis of
Variance (ANOVA) and all pairwise comparisons were run on the cohort variable. Results from pairwise comparisions are presented in Table 4. 10 below.
As noted in Table 4. 10, the null hypothesis of no difference in group means is rejected
(F=12.872, p<0.001). The variances of the groups are assumed equal (Levene=0.686, p>0.05} and differences in the groups explain 2.4% of the variance in environmental concern (Eta
Sq.=0.024). Contrasts indicate that cohorts born prior to 1942 and those born in 1965 or later are not significantly different, although very close (p=0.053). The cohorts born between 1942 and
1964 are statistically different from other groups. To control the familywise significance level for the three comparisons, Bonferroni post hoc comparisons were also performed. Results indicate that the cohorts born 1965 and after are not significantly different from the cohorts born prior to
1942 and between 1942 and 1964. In short, cohorts born 1965 and after are similar to both the cohorts born prior to 1942 and cohorts born 1942 to 1964. 106 2.00
2.eo
� (.> ; 2.55 u
- 2.00 ·� ...UJ ¢ 2.�
2.4J
2.35
B or n Prior to 1942 Born B elwee n 1942an d 1934 Bo rn 1005 and Petter Birth CohortGrou p
Figure 4.,8 • .:JA.ean EnvironmentalConcern Scores by Group
Table 4.10--ANOVA, Contrasts, and Post Hoc Tests for Mean Environmental Concern of Cohorts Sum of ANOVA Squares df Mean Sq. F Sig. Between Groups 12.502 2 6.251 12.872 0.000 Within Groups 505.535 1041 0.486 Total 518.037 1043
Contrasts Compared to Diff Std. Error Sig. Lower Upper Prior to 1942 1942 to 1964 -0.253 0.050 0.000 -0.352 -0.154 1965 and After -0.123 0.630 0.053 -0.247 0.002 1942 to 1964 1965 and After 0.130 0.057 0.023 0.018 0.242
Bonferronl Compared to Diff Std. Error Sig. Lower Upper Prior to 1942 1942 to 1964 -0.253 0.050 0.000 -0.374 -0.132 1965 and After -0.123 0.630 0.159 -0.275 0.029 1942 to 1964 1965 and After 0.130 0.057 0.069 -0.007 0.267
107 Although difficult to untangle, results from age analyses thus far suggest that there may be a cohort effect and that the cohorts born between 1942 and 1964 are different from the birth cohorts prior to 1942 and possibly different from birth cohorts born after 1965. Birth cohorts born prior to 1942 and those born 1965 or after are not different. The problem at hand is to determine how the birth cohorts from 1942 to 1964 are different from the remaining two. Two suggestions from the literature may be used to disentangle this relationship. On the one hand, studies indicate that cohorts born in the mid-1960s and later {a time of economic downturn} tend to express lower support for environmental issues {Bakvis and Nevitte 1992; George and Southwell
1986; Mohai and Twight 1987}. This finding corresponds to those in the study conducted by
Jones and Dunlap (1992} suggesting that, by the 1980s, age became less important in explaining environmental concern. The implication of these findings, coupled with the indications from analyses above, is that there may be a relationship between current economic position, economic context historically, and age such that persons raised in periods of economic downturn are more sensitive to economic context, especially when economically vulnerable. Unfortunately, these assertions cannot be tested with the data set at hand; it would require measures within subjects over time. But, it does appear that cohorts born after 1942 have, on average, higher concerns and social class may play less of a role for these cohorts in explaining environmental concern than for those born prior to 1942. These relationships will be further explored in multilevel models below after hypothesis testing.
Multi-Level Models
Hypotheses and Unconditional Random Effects Model
With the Level 1 relationships identified, the next step in analyses was to begin testing for differences in levels of environmental concern between counties, so that hypotheses tests were 1 clear. This first step allowed for the identification of the amo_unt of variance in environmental concern that may be explained by Level-2 variables. This amount of variance then became the
1 This type of model is often referred to as a "'random effects" or "empty'' model because it has no predictor variables at Level 1 or Level 2. The model does not explain differences between groups in terms of fixed group differences, thus contains only random groups and random variation. It is the most basic form of Hierarchical Linear Model and is appropriate when testing for Level 2 effects and when group sizes have a lower limit of 2 per group It is used to determine the basic variability partition between Level 1 and Level 2. (See Snijders and Bosker 2002.) 108 basis for partitioning variance between Level 1 and Level 2. Before proceeding to analyses, it is necessary to restate the hypotheses, so that the steps of analyses are clear. The hypotheses for this project are as follows:
1. Individuals will have lower environmental concerns in counties with higher economic distress (i.e., the intercepts of environmental concern will vary across counties and the contextual variable will explain some of the cross-county variation).
2. Social class and economic distress will interact such that the relationship between social class and environmental concerns will vary across counties, where economic distress explains some of this cross-county variation (i.e., slope coefficients between social class and environmental concernswill vary across counties).
3. Time of residence and economic distress will interact such that the relationship between time of residence and environmental concerns will vary across counties, where economic distress explains some of this cross-county variation (i.e., slope coefficients between time of residence and environmental concerns will vary across counties).
Based on these hypotheses, one may expect first and foremost that some of the variation in environmental concern may be explained by the higher level unit of analysis; that is, county. Put more specifically, a model with a fixed intercept and random county effect allows for the identification of variance in environmental concern scores between counties and the variance of environmental concern scores within counties. The variation of environmental concern scores within counties may be referred to as residual variance and the variation in mean environmental scores between counties may be referred to as county variance. The proportion of county variance to total variance (residual variance + county variance) indicates the amount of variation that may be attributed to difference between counties. In short, analyses began with the identification of the amount of variance in environmental concern that may be attributed to differences between counties.
To identify the amount of variation in environmental concern that may be explained by county, an unconditional random effects model was specified with environmental concern (Level
1) as the outcome/dependent variable and county as a factor, where the variance components of environmental concern between counties were treated as random; that is, the county variance and error or residual variance were estimated. As with the Level 1 models presented above,
Table 4. 11 includes estimates from across multiple imputations and presents information on fixed
109 Table 4.1 1: Estimates for Unconditional Random EffectModel, County 95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.456 0.038 95.997 64.687 0.000 2.381 2.532
95% COVARIANCE Estimate Std. Error Wald Z Sig. Lower Upper Residual 0.412 0.019 21.760 0.000 0.377 0.451 County 0.102 0.020 5.045 0.000 0.069 0. 151 Total 0.514 95% ICC Estimate Lower Upper County 19.877 15.546 25.058
2 3 4 effects parameters , covariance parameters , and the intraclass correlation coefficient (ICC) • As this model includes a single fixed effect, the intercept, the F-statistic tests the null hypothesis that the mean environmental concern score for all counties is 0, which is rejected (p<0.001). The estimate for the overall mean environmental concern score for all counties is 2.456, with a lower limit of 2.381 and an upper limit of 2.532. The Covariance Estimates section of Table 4. 11 presents the variance estimates for county (0. 102) and residual (0.412). In effect, the county variance indicates how much average environmental concern scores vary between counties and the residual variance indicates the degree to which environmental concern scores vary within counties. The sum of the two variances, 0.514, indicates the total variance of environmental concern scores. The ICC indicates the amount of total variance that may be attributed to differences between counties: 19.88% of the variance in environmental concern scores may be explained by differences between counties (with a lower limit of 15.55% and an upper limit of
2 Fixed effects refer to tests where the dependent variable is presumed to be a linear combination of all explanatory variables; that is, a regression. In multilevel analysis, such models presume that Level 1 and Level 2 differences are fully explained by the grouping variable. In this instance, no variable is assigned as a predictor and the fixed effect is of no substantive interest beyond identifyingthe mean for the population of groups, although it is typically presented as a component in hierarchical modeling. (See Norusis 2003 or Snijders and Bosker 2002 for more detail.) 3 Covariance values represent the partitioned variance for the random components of the model. In this analysis, the random components are county and residual. When the group variance is a small proportion of the total variance, then most of the variability may be attributed to units within groups, as opposed to differences between groups. These values allow for the calculation of the Intraclass Correlation Coefficient. (Norusis 2003 or Snijders and Bosker 2002.) 4 The Intraclass Correlation Coefficient(IC C) is the proportion of the variance accounted fordifferences between Level 2 units divided by the total amount of variance. It is equal to the correlation between values of 2 randomly drawn Level 1 units in the same randomly drawn Level 2 unit. (See Norusis 2003 or Snijders and Bosker 2002 for more detail.) 110 25.06%). The 19.88% of the variance attributed to county (Covariance Value=0.102) will be the variance partitioned when introducing contextual variables in the next section. Level 2 Random Coefficients
Given that 19.88% of the variation in mean environmental concern scores may be explained by county, it becomes important to identify which contextual effects may explain some of this variation. As noted in Chapter Three, contextual variables tapping economic distress, social distress, and pollution risk were created for this purpose. To assess the degree to which each context may explain variation in environmental concern scores, each contextual variable was included in models separately to identify which contexts were or were not important in explaining this variation. As with Level 1 model investigations, estimates were calculated from each imputation and then averaged. Simple regression models were used in these initial tests to determine the variance of environmental concern each explains. Further, all variables were centered, as recommended for the technique (see Raudenbush et al 2001). More formally, the following models were run:
5 1. Environmental Concem=lntercept + 81 (Economic Distress) + County Random Effect + error 2. Environmental Concern=lntercept + 81 (Social Distress) + County Random Effect + error 3. Environmental Concem=lntercept + 81 (Pollution) + County Random Effect + error
Table 4.12 presents the results of the economic distress model, with Fixed Effects, Covariance
Parameters, and Estimates of Explained Variance. The regression coefficients for economic distress and Intercept are Fixed Effects, with county and residual treated as random effects. The intercept for this model is very close to the intercept for the unconditional random effects model above and statistically significant. The slope for economic distress (-0.011) is statistically significant and indicates that environmental concern decreases by 0.011 for every one unit increase in economic distress. The county random effects indicate the degree to which mean environmental concern scores vary when adjusting for economic distress. The residual random effects indicate how much individuals within counties vary in terms of environmental concern scores. More importantly, the variance estimate for the county random effect is 0.071 when
5 The County Random Effectterm represents fitting a separate intercept for each county, thus providing the amount of variance in environmental concern that may be explained by the Level-2 variable. This differsfrom the error term, which is unexplained variance in this model. 111 Table 4.12: Estimates for Hierarchical Linear Model, Economic Distress Context
95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.560 0.039 95.256 66.216 0.000 2.484 2.637 Econ. Index -0.011 0.002 95.203 -5.362 0.000 -0.015 -0.007
95% COVARIANCE Estimate Std. Error Wald Z Sig. Lower Upper Residual 0.412 0.019 21.762 0.000 0.377 0.451 County 0.071 0.016 4.480 0.000 0.046 0.110
adjusted for economic distress. This estimate also is statistically significant (p<0.001}, thus indicating that the null hypothesis of no significant variation in county environmental concern means when controlling for economic distress may be rejected. Put more directly, economic distress explains 30.4% of the variance in environmental concern between counties (the formula
[0.102-0.071J/0. 102 provides the percentage of variance explained}.
Table 4. 13 presents the results of the social distress model, with fixed effects, covariance parameters, and estimates of explained variance. As with the previous model, the regression coefficient for social distress and intercept are fixed effects, with county and residual as random effects. The intercept for this model is 2.515 and statistically significant. The slope for social distress (-0.01O} is statistically significant and indicates that environmental concern decreases by
0.01O for every one unit increase in social distress. The county random effects indicate the degree to which mean environmental concern scores vary when adjusting for social distress. The residual random effects indicate how much individuals within counties vary in terms of environmental concern scores. More importantly, the variance estimate for the county random effect is 0.083 when adjusted for social distress. This estimate also is statistically significant
(p<0.001} and social distress explains 18.6% of the variance in environmental concern between counties.
Finally, Table 4. 14 presents the results of the pollution risk model with fixed effects, covariance parameters, and estimates of explained variance. The fixed effects intercept for this model is 2.519 and statistically significant. The slope for pollution risk (0. 789} is statistically
112 Table 4.13: Estimates for Hierarchical Linear Model, Social Distress Context 95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.515 0.038 95.360 65.790 0.000 2.439 2.591 Social Index -0.010 0.002 94.956 -4.012 0.000 -0.014 -0.005
95% COVARIANCE Estimate Std. Error Wald Z Sig. Lower Upper Residual 0.412 0.019 21 .763 0.000 0.377 0.451 County 0.083 0.018 4.732 0.000 0.055 0.126
Table 4.14: Estimates for Hierarchical Linear Model, Pollution Risk Context 95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.519 0.043 95.069 59.191 0.000 2.434 2.603 Pollution Index 0.789 0.274 95.103 2.876 0.005 0.244 1.334
95% COVARIANCE Estimate Std. Error Wald Z Sig. Lower Upper Residual 0.412 0.019 21.760 0.000 0.377 0.451 County 0.092 0.019 4.877 0.000 0.062 0.138
113 significant and indicates that environmental concern increases by 0.789 for every one unit increase in pollution risk. The variance estimate for the county random effect is 0.092 when adjusted for pollution risk. This estimate is statistically significant (p<0.001) and pollution risk explains 9.8% of the variance in environmental concern between counties.
Across these models, it is evident that context explains some variance in environmental concerns between the counties. Based on these findings, Hypothesis One focusing on the importance of contextual variables in explaining some of the variance in environmental concerns between counties is supported. As indicated above, economic distress does reduce, on average, the county environmental concern score. Likewise, social distress reduces, on average, county environmental concern scores. Pollution risk increases environmental concern scores, on average, between counties. Overall, these findings suggest that all of the included contextual variables do have some influence on individual level environmental scores. When comparing the proportion of the variance explained in each instance, the economic distress variable explains more of the variance in environmental concerns between counties (30.4%) than do social distress
(18.6%) and pollution risk (9.6%). Hence, in models thus far it appears that economic distress is the most important of these contextual factors in explaining variation in environmental concern scores. This finding has some practical consequence in that an understanding of economic distress in an area prior to efforts of building support for environmental initiatives must be addressed.
Although not directly related to the hypotheses for this project, possible interactions between various contexts were explored because of their potential practical significance. Based on the literature, one may expect that economic distress will interact with both social distress and pollution risk such that effects of social distress and pollution risk may on environmental concern may increase or decrease in magnitude under certain levels of economic distress. A full model including economic distress, social distress, and pollution risk with interactions was fit first, where non-significant terms were removed to find the best model fit. The interaction between pollution risk and economic distress remained significant; social distress and the interactions between social distress and economic distress and social distress and pollution risk were not significant. 114 Table 4.15 presents the results of this analysis, with fixed effects, covariance parameters, and estimates of explained variance. As indicated in the table above, economic distress and the interaction between economic distress and pollution risk are statistically significant (p<0.001 ).
Although the constituent estimate for pollution risk is non-significant, this variable is retained in the model per the requirements of interaction models (see Fox 1997). Of note, the interaction coefficient is positive and indicates that for every one unit increase in pollution risk the slope of economic distress on environmental concern increases by 0.0375. Further, by adding the interaction, the model results in a percentage increase in explained variance of 2.2%. The influence of pollution risk is, however, not as great as one might expect when considering the literature on pollution and environmental concern discussed in Chapter Two, with only a slight increase in average environmental concerns in the presence of greater pollution risk. While this is an interesting finding, the relative increase in explained variance at Level 2 is not great enough to warrant inclusion of interactions in subsequent models (based on -2 Log Likelihood change:
Chi. Sq.=1.953, df=2, p=0.38). In addition, given findings from contextual discussions above, the interaction may only be applicable in highly urban settings.
Random Coefficient Models and Cross-Level Interactions
Given that 80% of the variance in environmental concern scores is unexplained in the previous models, it became important next to explore the relationship between Level-1 variables central to hypotheses and environmental concerns within contexts. As noted previously, the most theoretically important Level 1 variable to multi-level analyses in this project is social class. The purpose of these analyses is to determine how much of the variability within each county may be accounted for by differences in social class among respondents. In effect, a random coefficient model was run under the assumption · that each county has a different slope and intercept in
6 relationships between social class and environmental concern.
6 A random coefficient model essentially fits a single regression model for each of the Level 2 units. In these regression models, slopes and intercepts are allowed to vary between Level 2 units. The fixed effects represent average slope and intercept, and the covariance parameters indicate variation between Level 2 units. Such models provide more detail on partitioned variance and, therefore, will differ in appearance from previous tables. (See Norusis, 2003 or Snijders and Bosker, 2002 for more detail.) 115 Table 4.15: Estimates for Hierarchical Linear Model Economic Distress and Pollution Risk Interaction
95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.599 0.042 93.258 61 .217 0.000 2.514 2.683 Economic Index -0.007 0.003 93.225 -2.590 0.01 1 -0.013 -0.002 Pollution Index 0.097 0.286 93.282 0.338 0.738 -0.471 0.665 Econ. lndex*Poll. Index 0.037 0.017 93.281 2.158 0.034 0.003 0.072
95% COVARIANCE Estimate Std. Error Wald Z Sig. Lower Upper Residual 0.412 0.019 21.762 0.000 0.377 0.451 County 0.068 0.016 4.358 0.000 0.043 0.106
Table 4.16 presents the results of this random coefficient. As indicated in Table 4.16, the estimated county environmental concern mean is 2.492 and statistically significant, which is similar to the previous models. The estimate for social class indicates that a one unit increase in social class results in a 0.239 unit increase in environmental concern. This slope coefficient is statistically significant, thus indicating that social class is related to environmental concern within counties. The covariance section indicates that the residual variance is 18% lower than the original residual variance (0.336 as compared to 0.412). This change indicates that social class explains 18% of the variation in environmental concern within a county, when controlling for county. The UN (1, 1) parameter estimates the variation of county intercepts (0.123), the UN (2, 1) parameter estimates the covariance of the slope and intercept (0.042), and the UN (2,2) parameter estimates the variance of county slopes (0.034). All of these estimates are statistically significant and indicate that the slope and intercept values do vary among counties and that there is a relationship between slopes and intercept estimates. In short, counties with higher average environmental concern scores have higher slopes; that is, the effect of social class on environmental concern is greater in counties with higher average environmental concerns.
Thus far contextual variables are important in explaining variation in environmental concerns between counties and social class is important in explaining within county variance. As
116 Table 4.16: Estimates for Random Coefficient Hierarchical Linear Model Social Class
95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.492 0.040 83.568 61.662 0.000 2.412 2.573 Social Class 0.239 0.029 65.994 8.243 0.000 0.181 0.297
95% COVARIANCE Estimate Std. Error Wald Z Sig. Lower Upper Residual 0.336 0.017 20.206 0.000 0.305 0.371 Intercept UN(1, 1) 0.1 23 0.024 5.035 0.000 0.083 0.1 81 UN(2,1) 0.042 0.014 2.911 0.004 0.014 0.070 UN(2,2) 0.034 0.012 2.727 0.006 0.016 0.069
Random Structure Intercept Social ClasslCounty lnterceptlCounty 0.123 0.042 Social ClasslCounty 0.042 0.034
indicated in the tables above, economic distress, social distress, and pollution risk explain 30.4%,
18.6%, and 9.6% respectively of the variation in mean environmental concerns between the counties. Of these contexts, economic distress explains a greater proportion of the variation and appears to have the greatest influence on environmental concerns. Hence, the literature pointing to the importance of economic contexts in affecting environmental concerns is supported. But, the effect of economic distress on environmental concerns requires specification through a cross level interaction test. As noted above in discussions on the Level 1 model, social class is the best predictor of environmental concerns, as measured to include attitudes, knowledge, and behaviors that reflect cultural capacity. Social class also explains roughly 18% of the variance in environmental concerns when adjusting for county, and models suggest that counties with higher average environmental concern scores tend to have higher slopes as a function of social class.
To test the hypothesis that social class and economic distress will interact, models with cross level interaction effects were fit.
Table 4. 17 presents the results of the hierarchical linear model to investigate cross-level interactions between social class and economic distress.
117 Table 4.17: Estimates for Hierarchical Linear Model, Economic Distress and Social Class 95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.573 0.043 80.080 59.864 0.000 2.488 2.659 Social Class 0.265 0.032 65.897 8.365 0.000 0.201 0.328 Economic Distress -0.009 0.002 80.921 -3.917 0.000 -0.014 -0.005 Econ. Distress*Social Class -0.004 0.002 60.466 -2.090 0.041 -0.007 0.000
95% COVARIANCE Estimate Std. Error Wald Z Sig. Lower Upper Residual 0.337 0.017 20.157 0.000 0.306 0.372 Intercept UN( 1,1) 0.100 0.021 4.719 0.000 0.066 0.152 UN(2, 1) 0.032 0.013 2.508 0.012 0.007 0.057 UN(2,2) 0.030 0.012 2.578 0.010 0.014 0.063
All coefficients in this model are statistically significant at the 0.05 level and the mean level of environmental concern is reported at 2.573 (consistent with previous models). The social class coefficient represents the average slope within counties and the economic distress coefficient represents the estimated county slope. The interaction effect between social class and economic distress indicates that the slopes for counties are significantly different and, as the coefficient is negative, that the 1slopes relating environmental concern to social class are smaller in counties with higher economic distress. When considering the covariance patterns, the variability of the intercept is 0. 10 as compared to 0. 123 in the random coefficients model (Table 4.16), which indicates that the introduction of economic distress has reduced the variability of county mean estimation by roughly 23%. Yet, all covariance estimates are statistically significant and the model indicates that most of the variability between intercepts and slopes of counties is explained by economic distress.
Using recommendations by Aiken and West (1991), three hierarchical models were run with the following specifications: 1) Low Economic Distress (Economic Distress + 1 Standard
Deviation of Economic Distress); 2) Moderate Economic Distress (Economic Distress as measured in data set); and 3) High Economic Distress (Economic Distress - 1 Standard
Deviation of Economic Distress). The regression coefficientsfor each of these models were then used to calculate plausible environmental concern values based on rough ranges forsocial class.
These values were then plotted to illustrate the interaction effect (values for social class go 118 beyond ranges in previous models to aid in viewing the interaction effect when using this technique).
Figure 4.9 presents the graphed regression equations for each of these models. To clarify before interpretation:
• The Low Economic Distress line is used to illustrate relationship between social class and environmental concern in counties with better economic conditions;
• The Moderate Economic Distress line is used to illustrate the relationship between social class and environmental concerns as observed in the dataset; and,
• The High Economic Distress lines is used to illustrate the relationship between social class and environmental concerns in counties with very poor economic conditions.
As indicated in Figure 4.9, the interaction is located among individuals with lower social class scores and, as social class scores increase, differences in the models become apparent. For example the dotted line representing low economic distress intersects the vertical axis at 2.73, which means that the average environmental concern score for counties with low economic distress is 2.73. By comparison, high economic distress (the dashed line) has a lower mean than both the moderate economic distress (2.57) and low economic distress (2. 73). Likewise, the coefficients for the slopes within counties (the social class coefficient) vary such that counties with high economic distress have smaller slopes. This graph indicates that the effect of social class on environmental concern is more pronounced in counties with lower economic distress. In addition, for all levels of economic distress, the higher the social class the greater the environmental concern.
To test the final hypothesis, the relationship between time in residence and environmental concern was assessed in a random coefficient model. Table 4.18 presents the results of this random coefficient model with fixed effects, covariance parameters, and estimates of explained variance. The estimated county environmental concern mean is 2.581 and statistically significant, which is higher than previous models. The fixed estimate for time in residence indicates that a one unit increase in years spent in the area results in a 0.005 decrease in environmental concern. While this slope coefficient is statistically significant, the covariance
119 4.00
3.50 Environmental Concern
3.00
• • • Low Economic Distress - Moderate Economic Distress 2.00 -- High Economic Distress
1.50
1.00
0.50
-5 -4 -3 -2 -1 0 1 Social Class Figure 4.9-·Economic Distress and Social Class Interaction (Low, Moderate, and High Economic Distress)
Table 4.18: Estimates for Random Coefficient Hierarchical Linear Model, Time in Residence 95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.581 0.089 96.000 28.890 0.000 2.575 2.776 Time in Residence -0.005 0.001 95.103 -3.120 0.003 -0.009 -0.005
95% COVARIANCE Estimate Std. Error Wald Z Sig. Lower Upper Residual 0.495 0.022 22.825 0.000 0.454 0.540
120 section of Table 4.18 indicates that the residual variance has increased when compared to the baseline model (Table 4.11), thus indicating that time in residence does not explain variation in environmental concern, when controlling for county. Put more succinctly, as with the Level 1 models, time in residence has little effect on environmental concern within counties and no real effect on environmental concern between counties. Thus, the third hypothesis is rejected; that is, time in residence does not explain variation in environmental concern between counties, nor would an interaction between economic context and time in residence.
Exp/orations of Age and Migration in Fixed Effe cts Models
Three additional models were run to determine the extent to which migrant status and age cohort may explain some of the variance in environmental concerns between counties, as well as to check for potential spuriousness in previous findings. These models were run for the purposes of gaining clarity on the relative importance of migrant status for the region (Jones et al
2003) and under the assumption that economic experiences may influence postmaterial concerns
(lnglehardt 1990). Literature reviewed in Chapter Two and analyses above indicate that there may be a cohort effect such that individuals experiencing greater economic hardship during formative years may have lower concerns about issues like the environment (lnglehard 1990).
Likewise, literature in Chapter Two points to the growing importance of migrant status in explaining increases in environmental concerns in rural regions. Information on both types of relationships may help to inform on the types of cultural capacity available in counties in this region. For example, if migrants moving to the region express higher levels of concern, they may serve as a basis for mobilizing greater pressure on institutions for reform. As noted above from the SAA technical report, many communities in the region at the time of the survey were becoming tourism and retirement oriented. In this sense, differences in age cohort and migrant status may influence environmental concerns within contexts. While a detailed analysis on these relationships is beyond the scope of this dissertation, as age and migrant status are considered important in the literature, it is necessary to identify these relationships for further clarification of context effectsin future research.
121 Given the significance of the interaction effects in explorations of the Level 1 model, the
Pre-1942 cohort variable and the migrant status variable were used as fixed factors in multilevel analyses. The goal of these analyses was to identify possible additional Level-1 variables that may be associated with environmental concern. To accurately asses the influence of both age cohort and migration status, each of the independent variables will be treated as fixed effects and county will be treated as random. By adopting this approach, generalizations on the relationship between cohort and migration may be made to counties the region. Fixed effects models, therefore, differ from the previous random coefficients models in that estimates are conditional upon the factor values in the model. The first model presented assesses the degree to which being born after 1942 has an effect on environmental concerns. As noted above in Level 1 models, the relationship between social class and environmental concern is moderated by membership in the 1942 and later cohorts, which will be assessed next. Thus, the purpose of this analysis was to determine if the cohort effect may be generalized to all counties and if the moderating effectmay be generalized to all counties.
Table 4.19 presents the results of the fixed effects model for Pre-1942 cohort. As indicated in the table above, the fixed coefficient for the pre-1942 cohorts is statistically significant, thus confirming that there is a relationship between cohort group and environmental concern. Individuals born prior to 1942 (-0.162) appear to have lower environmental concern scores than those born 1942 or after. This relationship may be generalized to counties in the region, as county is treated as random. It appears that across counties the relationship between cohort grouping and environmental concern is consistent such that persons born prior to 1942 have lower environmental concerns. The residual estimate indicates the amount of unexplained variance of environmental concern scores after the cohort effects are fit. When compared to the original unconditional county model (Table 4.11), fitting cohort as fixed explains roughly 7% of the variance in environmental concern ([0.412-0.384]/0.412). The county coefficient also has reduced from the original model and the interaction term is statistically significant, which indicates
122 4.19: Estimates for Fixed Effects for Mixed Model, County and Cohorts 95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.507 0.043 134.137 58.603 0.000 2.422 2.591 Pre-1942 Cohort -0.162 0.06 1 86.881 -2.651 0.010 -0.283 -0.041
95% COVARIANCE Estimate Std. Error Wald Z Sig Lower Upper Residual 0.384 0,018 20.955 0.000 0.350 0.422 County 0.053 0.025 2.085 0.037 0.021 0.136 County*Pre-1942 Cohort 0.071 0.026 2.770 0.006 0.035 0.144
that the relationship between cohort membership and environmental concern is not the same across all counties. This finding suggests that the composition of some counties may affect the degree to which this relationship holds.
Table 4.20 presents the results of the fixed effects model for non-migrants to the region.
The coefficient for non-migrants is statistically significant, thus confirming that there is a relationship between migrant status and environmental concern. Individuals living in the region their entire lives appear to have lower environmental concern scores (-0.275) than those moving to the region. When compared to the original unconditional county model (Table 4.11), fitting migrant status as fixed explains roughly 9% of the variance in environmental concern ([0.412-
0.375)/0.412). More importantly, the county coefficient has reduced from the original model and the interaction term is statistically significant, which indicates that the relationship between migrant status and environmental concern is not the same across all counties. As with the general cohort model above, this finding suggests that the composition of some counties may affect the degree to which this relationship holds.
Given the potential for these effects and the possibility that they may confound the relationship between economic distress and environmental concern, one additional set of analyses was performed. These analyses focused on determining whether or not the environmental concern is better explained by social class, migrant composition, and cohort composition than by economic context; that is, a check was performed to see if the effect of the contextual variable identified above was spurious. To be certain that the findings on economic
123 4.20: Estimates for Fixed Effects for Mixed Model, County and Migrants 95% FIXED Estimate Std. Error df t Sig Lower Upper Intercept 2.629 0.051 191 .407 51.811 0.000 2.529 2.729 Non-Migrant -0.275 0.062 84.474 -4.413 0.000 -0.398 -0.151
95% COVARIANCE Estimate Std. Error Wald Z Sig Lower Upper Residual 0.375 0.018 20.881 0.000 0.341 0.412 County 0.031 0.025 1.232 0.218 0.006 0.150 County*Non-Migrant 0.088 0.026 3.142 0.002 0.047 0.163
context above are not spurious, three models were run. First, all Level-1 variables were included in a full model, where all intercepts and slopes were treated as random and cross-county variation was taken into account. This model provided information on the degree to which class, migrant status, and cohort were explaining the bulk of the variation in environmental concern across counties. Second, economic context was introduced as a predictor for the intercepts and slopes of each Level-1 variable and environmental concern. This model provided information on whether or not economic context could be used to predict the slopes and intercepts of each regression. Finally, migrant and cohort were set to fixed, consistent with the models presented above. Social class was treated as random in slope and intercept and economic context was used to account for slope and intercept differences across counties in the social class environmental concern relationship.
Table 4.21 presents the results of the full fixed and random effects hierarchical linear. model. Of note, the economic context coefficient remains significant in predicting social class and environmental concern slopes and intercepts, even when controlling for potential composition effects. The variance components also remain statistically significant for this model. Both of these findings indicate that the variation in environmental concern across counties is a function of economic context through social class; that is, the relationship is not spurious. The only major difference betweenthis model and previous models is the higher p-value for the cohort variable
124 Table 4.21 : Estimates for Full Fixed and Random Effects Model Social Class, Pre-1 942 Cohort, Non-Migrant, and Interaction FIXED Estimate Std. Error df t Sig. Intercept 2.721 0.032 95.000 7.932 0.000 Social Class -0.275 0.062 84.474 -4.413 0.000 Economic Distress -0.100 0.003 95.000 -3.147 0.001 Economic * Social Class -0.005 0.002 95.000 -3.204 0.000 Pre-1942 -0.109 0.063 1038.000 -1.726 0.084 Non-Migrant -0.191 0.062 1038.000 -3.085 0.003
COVARIANCE Estimate Std. Error df Chi. Sq. Sig. Residual 0.1 65 0.406 0.000 County 0.031 0.025 94.000 372.015 0.000 Class Slope 0.088 0.028 94.000 345.803 0.000
(p=0.084), which implies non-significance. Given that no hypotheses directly address the cohort variable, this change will not be fully explored in this dissertation. But note that the tests of significance for these types of models are very conservative and two-tailed, which suggests that a p-value of less than 0.10 may be interpreted as statistically significant in this instance.
Overall, it appears that the fixed effects of both migrant status and cohort are important in explaining environmental concern in multilevel models. Of note is the indication in both models that the composition of a county in terms of numbers of migrants or age structure may affect the level of environmental concern. This finding is consistent with the literature reviewed in previous chapters, especially the work of Jones et al (2003). They argue that new residents migrating to counties in the region may be bringing with them higher levels of environmental concern.
Likewise, the age structure of an area may play a role in average levels of environmental concern at the county level, as well as individual levels of environmental concern. For example, if the age structure of a population in a given area is skewed toward persons born prior to 1942, then the relative concern of citizens in that area may be lower, on average, than in areas where the age structure differs. Unfortunately, additional analyses required to uncover such relationships are beyond the scope of this dissertation and would require a data set specifically designed for multilevel models. These limited findings coupled with information from the Level 1 models
125 discussed above do suggest that a future avenue of research focus specifically on age structure and migrant proportions in estimating context effects on environmental concern. Discussion and Conclusions from All Analyses
As noted earlier in this chapter and earlier in this dissertation, the Southern Appalachian region is one experiencing a great deal of growth and the negative consequences of that growth.
Historically, residents of the Southern Appalachian region have been poorer, less educated, and the region developed more slowly than the remainder of the country. But recent trends in growth of population density and urbanization highlight the need to better understand how local areas are affected differently within the region and how differential effects (or context effects) may shape individual attitudes and opinions regarding the natural environment. For example, greater urbanization and greater population growth lead to greater conversion of farmland to other uses, higher rates of pollution and greater susceptibility of individuals to economic, environmental and social harms. These consequences, however, are not uniformly experienced by individuals across all counties in the region, despite the fact that the negative consequences impact the entire region. As noted above in a discussion on the SAA technical report, residents of the region recognize these trends as problematic and wish greater input into decisions made regarding natural resource management and more information about how those decisions are made.
Information on potential contextual influences may be used t� highlight the differential impact of the growing problems faced by residents and highlight how economic tradeoffs may take shape in some communities and not others. In short, the Southern Appalachian region serves as an excellent laboratory for identifying the potential for economic, social and pollution influences on individual attitudes.
In the earlier sections of this chapter, maps were used to identify the distribution of different types of harm in the region. While more descriptive in character, it appears that the more developed, urban counties typically experience higher levels of pollution than the remainder of the region. Counties adjacent to the more developed areas also typically experience similar effects, although to a slightly lesser degree. In effect, pollution effects do seem to be concentrated in some areas but not limited to the counties experiencing the most direct 126 consequences in terms of pollution risks. These higher risk areas do appear to have lower economic distress when compared to national averages, possibly because of higher industry and higher numbers of jobs. Yet, the most severe instances of economic distress appear to be distributed more broadly throughout the region than specific pollution risks. While not particularly revealing given the history of the region, the social consequences of this distribution are apparent in that areas with higher levels of economic distress tend to experience higher levels of social distress in the form of disproportionately poorer, less educated individuals, particularly among vulnerable groups like female headed households, children and persons over 25 without college degrees. More importantly, many counties in the region experience multiple forms of economic, social and pollution risks when compared to national averages, with just under 10% of all counties experiencing high scores on all contextual measures of distress and roughly half scoring higher than national averages on two or more of the indicators used. In short, as often noted in the literature on the region, Southern Appalachia is a distressed region in need of considered natural resource planning, thus information on the cultural capacity of the region is needed.
Although the analyses above are detailed and, at times, complicated, the findings presented above suggest a great deal of information. First, to aid in understanding context effects on individual environmental concerns as a form of cultural capacity, several Level 1 models were run using social class, time in residence, age, and migration status as predictors.
Overall, models were consistent with the literature on environmental concern and suggested that social class is the most important predictor of environmental concern. While recognizing that this finding may, in part, be a function of the measurement of environmental concern, the most apparent effects and the better model fits were those that included social class. Likewise, time in residence was significant in the Level 1 model, but its effect was smaller and appears less important than social class. More importantly, migrant status appears to be a better indicator in terms of time in residence, suggesting that migrants do, in fact, bring with them higher environmental concerns to the region. Further, the distinction in the literature between cohort groupings appears to be important to understanding environmental concern at the individual level such that cohorts born prior to 1942 do appear to have lower concerns and that the effect of 127 social class may be dampened in later cohorts. In short, social class, migrant status, and cohort appear to be important in explaining environmental concern, which is consistent with the literature.
Second, based on Level 1 analysis, multilevel models were developed to identify the differences within and between counties. Multilevel analyses began with identification of an unconditional random effects model by estimating county and residual variance. Results from this initial �odel indicate that 19.877% of the total variance in environmental concern may be attributed to differences between counties. Using this variance figure as a baseline, the economic, social and pollution contexts were each included in separate hierarchical linear models to determine the amount of variance in county average environmental concern. Findings confirm that not only did each of the variables explain variance in county environmental concern means but that all contextual variables used in analysis are statistically significant. These results may be summarized as follows:
• Economic distress explains 30.4% of the variance in environmental concern means between counties for the Southern Appalachian region.
• Social distress explains 18.6% of the variance in environmental concern means between counties
• Pollution risk explains 9.8% of the variance in environmental concern means between counties
Of the contextual variables, economic distress appears to be the most important predictor of differences in environmental concern means between counties. Based on these initial models and the importance of economic distress, Hypothesis One was confirmed; intercepts (or means) of environmental concern vary across counties and economic distress explains 30.4% of this variance. Further, in explorations of potential interactions, pollution risk and economic distress appear to interact such that higher pollution risk may increase environmental concern, even in economically distressed counties. Although this effect was statistically significant, the relative effect was quite small and not important enough to incorporate into addition models. Yet, this type of relationship may be highly important in more urban areas. Thus, economic distress,
128 social distress and pollution risk all explain variance in environmental concern means between counties and thus context does have an effect on individual level environmental concern.
To elaborate on these initial findings, social class was added hierarchical models as a
Level 1 predictor of the variability within counties. The purpose of this analysis was to identify the degree to which the social class relationship identified at Level 1 was consistent across counties.
Next, models were run to determine the extent to which economic context may have an effect on the relationship between social class and environmental concern; that is, a cross-level interaction effect. Findings from these analyses may be summarized as follows:
• When controlling for county, social class explains 18% of the variance within counties in the region.
• The slope and intercept values for the social class and environmental concern relationship vary among counties such that counties with higher average environmental concern scores have higher increases in environmental concern as a function of that individual's social class.
• Counties with higher economic distress have smaller increases in environmental concern as a function of social class than do counties with lower economic distress; that is, social class may be more important in explaining environmental concerns when economic distress is lower.
In effect, economic context does matter when thinking about the relationship between social class and environmental concern. Economic distress may make environmental concerns less important to individuals of lower social class, particularly those considered very economically vulnerable. This context effect explains lower county average environmental concerns and smaller increases in environmental concern as �ocial class increases. Social class is still important to explaining environmental concern, but economic context has the effect of reducing concerns across class when economic conditions worsen. Thus, Hypothesis Two was supported:
The relationship betweens social class and environmental concern does vary across counties when economic distress is higher.
Finally, time in residence, migrant status and age cohort were assessed in terms of their effect on environmental concern in multilevel models. As noted above, the relationship between time in residence and environmental concern was statistically significant but minimal. Time in residence did not explain ariy of the variance within county when controlling for county. This
129 suggests that the relationship between time in residence and environmental concern does not vary among counties in the region. As this relationship did not vary, additional tests for interactions were not necessary and Hypothesis Three, therefore, was not supported. Given that time in residence was significant as a fixed effect but not as a random effect in explaining environmental concern, the variables age and migrant status were included in models as fixed effects. These explorations were performed to determine the extent to which the effects identified in level 1 analyses may be generalized to all counties in the region. In these models, county was treated as random, with migrant status and age cohort grouping treated as fixed factors. These relationships were statistically significant and important in explaining some of the within county variance. Findings from these analyses may be summarized as follows:
• The relationship between age cohort membership and environmental concern does differ between counties and age cohort membership explains roughly 7% of the variance in environmental concern scores.
• The relationship between migrant status and environmental concern does differ between counties and migrant status explains roughly 9% of the variance in environmental concern scores.
Although not directly related to hypothesis testing, these analyses suggest that the composition of a county in terms of numbers of migrants or age distributions may have an effect on environmental concern scores. This effect may manifest as lower environmental concerns for older residents living in counties for longer periods of time when compared to younger, possibly retiring, in-migrants. Such migrants may be attracted to an area for reasons of quality of life, thus bringing with them attitudes and behaviors that are pro-environmental and consistent with other, similar attitudes. In short, although time in residence was not important in multilevel models, migrant status and age cohort were important and imply that the social composition of an area may influence levels of environmental concern. More research on this relationship using multilevel models is warranted.
To conclude this chapter, it is important to note that much of the research in this area is in its infancy, but it appears that multilevel analyses do help clarify environmental concern and cultural capacity. For example, analyses above indicate that economic context does play a role in shaping relationships between individual level economic factors and environmental concern. 130 More specifically, economic distress alters the relationship between social class and environmental concern such that higher levels of economic distress appear to diminish the social class and environmental concern relationship This finding has several practical implications including the suggestion that individuals living in areas of higher economic distress may be more resistant to an environmental initiative than individuals living in areas of lower economic distress
(or better economic health). The cultural capacity for ecologically modern transformations may be diminished if the economic context is poor. This implication is consistent with the literature on ecological modernization and consistent with the literature on social class and environmental concern. Further, social distress and pollution contexts appear to be important, with pollution context possibly reducing the influence of economic distress· on environmental concerns.
Unfortunately, the data set used in these analyses cannot be used to test these specific assertions on pollution and social distress as they relate to the experience of environmental or social injustice. But, the importance of these variables in explaining environmental concerns should not be ignored; they are important, although not as strong as economic context in models above. In the end, the cultural capacity of an area may be impacted by any number of contextual conditions and models above suggest that context, particularly economic distress, is important in explaining environmental concerns.
131 CHAPTER FIVE
SUMMARY, CONCLUSIONS, AND FUTURE RESEARCH
Introduction and Overview
By way of a conclusion, this final chapter serves two purposes. First, this chapter is used to summarize much of what has been presented in earlier chapters. In an effort to avoid too much redundancy, the first major section of this chapter recapitulates the central arguments and findings in a highly condensed form. In some instances, main arguments or findings are bulleted to aid in summarizing what is a great deal of complex material. Next is presented a discussion on the caveats and limitations of this study, which point to both weaknesses and opportunities for future research. These problems are presented in the spirit of guiding future research in a manner not available during analyses for this dissertation. Second, this chapter is used to elaborate on the types of theoretical development needed to better understand the influence of context. Focusing on the broader issues, this section provides a thumbnail sketch of what kinds of information and what kinds of linkages are necessary to more fully understand contextual influence on environmental concerns. These issues and avenues for future research are discussed relative to the level of analysis at which data must be collected, as well as ways to refine measurement at each level. · Finally, a discussion and conclusion section is used to reiterate the importance of theory and context in understanding the issues important to creating environmental messages the public may accept.
Theories, Empirical Findings and Caveats
Overview of Th eories
Context is a central but often untested feature of sociological theories; the importance of context may be illustrated in �he environmental concerns of individuals and environmental policy.
From the sociological perspective, social and economic contexts are presumed to influence the norms and values of individuals. Social institutions like the economy are, therefore, a product of contextual influence and replicated in the behaviors and beliefs of individuals. Identifying variation in contexts and their potential effects on beliefs or actions of individuals have several consequences for understanding public opinions and policy on the natural environment. Among 132 these consequences is the potential that environmental policies created at the national level and applied homogenously to all contexts may result in Jow public support or resistance to such policies. Further, standardized application of environmental policies when contexts vary may result in strategies that fail to meet local environmental needs, thus eroding public faith in governmental legitimacy. One way to create a better fit between environmental policy and environmental needs of an area is to identify social, economic or specific environmental context effects on public environmental concerns. By identifying contextual effects across local areas, the complexity of public environmental concerns and ways to tailor environmental policies may be understood and theorized.
Ecological Modernization theory and economic contingency research suggest that economic growth may lead to higher levels of concern in the general public. From these perspectives, theory and research should focus on how environmental reform arises from the relationship between economic growth and public opinion. Public pressures for institutional reform may be viewed as the cultural capacity of the public to push for and accept pro environmental reforms. As economic growth and environmental harm increases, public awareness of the environment and public environmental concerns will increase pressures for change in policy and industrial practice. Changes in policy will result as the inefficiency of the state becomes more apparent, and market changes will occur as a result of greater public demand for corporate environmental responsibility and environmentally friendly products. The work of many researchers supports the proposed relationship between economic conditions and public opinion by suggesting that environmental attitudes vary over time, where economic downturns correspond to lower environmental concerns. Research on the social bases of concern and research on perceptions of place support claims on the importance of context in understanding local environmental concern. If, as theorists and research suggest, public opinion varies in relation to economic context at the national level, then so too may attitudes and behaviors on the environment vary relative to localized economic conditions. Hence, economic conditions affect levels of environmental concerns (or cultural capacity), thus economic distress, social distress, and pollution contexts are tested in this dissertation. 133 Overview of Hypotheses and Findings
Much of the emergent literature on how individuals perceive the natural environment focuses on the context within which individuals experience the natural environment. If the possible influence of context is not taken into account when researching or theorizing on environmental concerns, then theories may fail to draw out important links and generalizations from research may not be valid. To account for such influences, relevant contexts must be identified and incorporated into models at appropriate levels of analysis. Recent sociological theorizing suggests that, as a feature of American culture, environmental concerns vary relative to historical, social, economic, and environmental harm contexts at the national and local levels.
At the individual level, characteristics such as social class position shape how environmental concerns may manifest in attitude or behavior. Across levels of analysis, social class positions vary, where some positions are more vulnerable than others in certain contexts, thus environmental concerns may also vary. Likewise, additional factors such at time in residence and age may vary in their influence within contexts. These individual characteristics link individuals to potential economic conditions such that in-migrants may be less economically vulnerable and persons experiencing the Great Depression may be more concerned about broader economic conditions. Thus, these theories and the empirical literature on environmental concern serve as the basis for empirical tests in this dissertation.
Consistent with the theories and empirical literature presented, hypotheses were derived to focus specifically on economic context, although social distress and pollution levels were incorporated into some models. As economic context was the focus of this research, the following hypotheses guided empirical testing:
• Individuals will have lower environmental concerns in counties with higher economic distress.
• Social class and economic distress will interact across levels such that the relationship between social class and environmental concerns will vary across counties, where economic distress explains some of this cross-county variation. • Time of residence and economic distress will interact across levels such that the relationship between time of residence and environmental concerns will vary across counties, where economic distress explains some of this cross-county variation.
134 Confirmatory factor analysis was used to identify and create a composite measure of environmental concern. This dependent variable was then used in hierarchical linear models for empirical testing. Contextual variables were also identified in the extant literature focusing on county level characteristics in the Southern Appalachian Region. Based on calculations presented in this literature, county level variables were generated and set relative to national averages, as recommended in the literature. These variables were then used as Level-2 contextual variables in hierarchical linear models. Further, a social class variable was generated using principal components scores and age and time in residence were used as given. Finally, analysis proceeded from map generation to demonstrate distribution of contextual characteristics to Level 1 analyses and, subsequently, to hierarchical analyses.
Overview of Specific Analyses Conducted and Findings
Analyses began with a focus on the distribution of contextual characteristics reflecting different types of harm in the region and Level 1 simple regression models focusing on the effects of social class, time in residence, and migrant status. As demonstrated in maps generated:
• Pollution effects do seem to be concentrated in some areas but not limited to the counties experiencing the most direct consequences in terms of pollution risks.
• The most severe instances of economic distress appear to be distributed broadly throughout the region with higher risk areas experiencing lower economic distress
• Social distress appears to be distributed across the region but areas with higher levels of economic distress tend to experience higher levels of social distress
• Many counties in the region experience multiple forms of economic, social and pollution risks, with just under 10% of all counties experiencing high scores on all contextual measures of distress and roughly half scoring higher than national averages on two or more of the indicators used.
Following presentation of the risks in the region, several Level 1 models were run using social class, time in residence, age, and migration status as predictors. Key findings include the following relationships:
• Social class is the more important predictor of environmental concern, with, the most apparent effectsand the better model fits.
• Time in residence was significant in the Level 1 model, but its effect was smaller and appears less important than social class.
135 • Migrant status appears to be a better indicator than time in residence, suggesting that migrants do, in fact, bring with them higher environmental concerns to the region.
• Further, the distinction in the literature between cohort groupings appears to be important to understanding environmental concern at the individual level such that cohorts born prior to 1942 do appear to have lower concerns and that the effect of social class may be dampened in later cohorts.
In short, social class, migrant status, and cohort appear to be important in explaining environmental concern, which is consistent with the literature, although the relationship between social class and environmental concern may be less important for later cohorts.
Multilevel models were then used to identify the differences within and between counties.
As noted previously, multilevel models were run on each of the imputed data sets and models included unconditional random effects, specific contextual effects, and cross-level interactions.
Key findings from these tests include the following:
• Unconditional Random Effects: 19.877% of the total variance in environmental concern may be attributed to differences between counties.
• Separate Contextual Models: 1) Economic distress explains 30.4% of the variance in environmental concern means between counties for the Southern Appalachian region; 2) Social distress explains 18.6% of the variance in environmental concern means between counties; and 3) Pollution risk explains 9.8% of the variance in environmental concern means between counties
• Explorations of potential interactions, pollution risk and economic distress appear to interact such that higher pollution risk may increase environmental concern, even in economically distressed counties. Although this effect was statistically significant, the relative effect was quite small and not important enough to incorporate into addition models. Yet, this type of relationship may be highly important in more urban areas.
Social class and time in residence were then added hierarchical models as a Level-1 predictors of the variability within counties. Findings from these analyses may be summarized as follows:
• When controlling for county, social class explains 18% of the variance within counties in the region.
• The slope and intercept values for the . social class and environmental concern relationship vary among counties such that counties with higher average environmental concern scores have higher increases in environmental concern as a function of that individual's social class.
• Counties with higher economic distress have smaller increases in environmental concern as a function of social class than do counties with lower economic distress; that
136 is, social class may be more important in explaining environmental concerns when economic distress is lower.
• Time in residence did not explain any of the variance within county when controlling for county, which suggests that the relationship between time in residence and environmental concern does not vary among counties in the region. As this relationship did not vary, additional tests for interactions were not performed.
Thus, findings provide the basis for the following conclusions regarding the hypotheses tested:
• Hypothesis One: The intercepts· of environmental concern vary across counties and economic distress explains 30.4% of this variance. Th erefore, hypothesis one is supported.
• Hypothesis Two: Economic context has the effect of reducing concerns across class when economic conditions worsen. That is, the relationship betweens social class and environmental concern does vary across counties when economic distress is higher. Therefore, hypothesis two is supported.
• Hypothesis Three: Time in residence was significant in fixed effect but not in random coefficients models and does not appear to be a function of context. Th erefore, hypothesis three is not supported. Hence, economic context does matter when thinking about the relationship between social class and environmental concern. Time in residence is still salient to understanding individual level environmental concern as a fixed effect, but its effect in context is negligible. In the end, economic distress may serve to reduce the environmental concerns of lower social class individuals, thus leading to lower county environmental concerns and smaller increases in environmental concern as social class increases.
Based on findings at Level 1 and the importance of age and migration to the environmental concern literature, the variables age and migrant status were included in models as fixed effects. Although not related to hypotheses, the tests were performed to determine the extent to which the effects identified in Level 1 analyses may be generalized to all counties in the region. In these models, county was treated as random, with migrant status and age cohort grouping treated as fixed factors. These relationships were statistically significant and important in explaining some of the within county variance. Findings from these analyses may be summarized as follows:
• The relationship between age cohort membership and environmental concern does differ between counties and age cohort membership explains roughly 7% of the variance in environmental concern scores.
137 • The relationship between migrant status and environmental concern does differ between counties and migrant status explains roughly 9% of the variance in environmental concern scores.
Analyses suggest that the composition of a county in terms of numbers of migrants or age distributions may have an effect on environmental concern scores. For example, older residents living in counties for longer periods of time may be less environmentally concerned than younger or retiring migrants. Counties that attract migrants or retirees may have higher environmental concerns for the very reasons that such counties may be attractive. In addition, the effect of context on the social class and environmental concern relationship remains when controlling for migrant status and cohort. Given the importance of migrant status, cohort, and social class, more research on these relationships using multilevel models is warranted.
Caveats and Recommendations for Future Research
Although this dissertation drew on newer statistical techniques in an effort to identify context effects on environmental concerns, it is important to view the findings of this study as early efforts that require refinement. Indeed, this study should be viewed as one of a very few studies that form a "beachhead" in understanding the effects of context on environmental concerns. Multilevel models are growing in their popularity in the social and behavioral sciences.
Uncovering the relationship between context and public opinion does offer some hope in understanding the complexity of environmental concern. Yet, , the use of newer techniques does not equate to substantive understanding, and much more research of this sort is needed. Much like this dissertation, the few multilevel articles reviewed drew on empirical literature in the hopes of better informing what is a more heavily empirical and under theorized field. Such studies do offer insight, but researchers should not rely on method to dictate substance or the consequence may be a continuing difficulty in understanding the effects of context on environmental concerns.
More importantly, there is a need for incorporating more and more sophisticated statistical techniques into the environmental social sciences, especially if findings from such studies may be applied to real world circumstances. It is in this spirit that the following caveats are presented and in the hopes of · theoretical development and practical application that the following recommendations are made. 138 Caveats and Methodological Recommendations
No study is perfect and no study stands on it own; this study is no different. Several issues arose in theorizing, conceptualization, measurement and analysis that, while limitations, may be more positively viewed as opportunities to learn from experience. First and foremost among these limitations was the data set used in analyses. The data set was selected because of its historical and ecological importance, because the data set has been subjected to a great deal of scrutiny in other research efforts, and because data were collected with an eye toward hierarchical relationships. Indeed, of the possible data sets evaluated for use in this dissertation, the data set selected was clearly the best among all options available. Yet, the method for data collection on specific environmental concern scales presented numerous problems. For example, the survey instrument included multiple environmental concern scales but only one set of attitudinal questions was asked of all respondents. In effect, the most commonly asked scales in environmental social science could not be used in hierarchical analyses because of case number requirements within county. This type of randomization is an efficient method for conducting survey analysis, but it limits the types of analyses that may be performed or the types of composite measures that may be created. In the end, the Environmental Attitudes for Southern
Appalachia scale proved useful in identifying core attitudes consistent with the literature on environmental concern, but the more commonly asked scales were not available in analyses.
Note that the data limitations in no way suggest that the measures or methods employed in this study are inadequate; the measures could have been better with specific types of data collection techniques. The real issue with this study was that the more commonly used scales in the environmental social sciences were not available. This type of limitation is important for two reasons. First, it is fairly common that public opinion polling requires asking the fewest possible questions to obtain higher response rates and reduce costs. A great deal of cross-sectional data suffers from limited information at Level 1, which limits hierarchical analyses of the type performed in this dissertation. Second, the type of variables that may be used in creating composite measures also limits the understanding that may be gained from hierarchical analysis.
Recognizing this limitation, it becomes important for researchers to pay greater attention the 139 possibility of measuring the effects of context in research design. Taking context into account requires that researchers refine traditional sampling techniques to maximize higher and lower level units in administration. Researchers may then tailor survey instruments to include the more commonly used environmental concern scales and enhance understandings of contextual effects on public opinion. In addition, although hierarchical models will work with as few as two cases per higher level unit, far more variation within contexts provides better comparisons across contexts, thus increasing the reliability and validity of findings.
Next, identifying relevant and appropriate contextual measures became particularly challenging. The measures used in this dissertation were selected because they are commonly used by national, state and local governments in identifying county-level problems. In fact, the most accurately measured variable also serves as the most important in models; that is, economic context. Measures of social distress and pollution were based on criteria set forth by governmental institutions, but may not be as direct a measure of context as the economic context variable. For example, social distress often is a matter of definition, where decreased infrastructural investment, higher crime rates or even abandoned buildings might have been better. The use of these types of measures reflects a different and, perhaps, more environmental operational definition of social distress that may be more important in explaining concerns.
Measures of pollution impact also may be too indirect to adequately reflect the degree to which individuals and communities experience pollution. Many of the more objective measures of pollution suffer from a similar problem. For example, pollution in the form of emissions in tons may be less direct in measurement and context effect than an individual's perceived level of smog in the area. Further, context measures at the county level may not be proximate enough for pollution or social distress, especially when considering that neighborhoods experience these types of problems differently within a local area. In short, the contextual measures used in this dissertation are one approach and better measures for some contexts may exist.
Realistically, the difficulties in identifying appropriate contexts and measuring contexts emerge from a lack of theory and a lack of available data. Government sources do provide a great deal of information at the county and community levels on demographic characteristics and 140 economic patterns, but the available data are typically not annual. For example, some census information is estimated between collection years and a great deal of information on natural resource use is typically every four years. Further, when dealing specifically with natural resource data, estimates below the level of the state lose their accuracy because of sampling procedures; that is, sub-state level resource information may not be estimable in certain circumstances. Adding to these issues is the lack of a clear theoretical framework suggesting which contexts affect individuals directly and how to identify the contexts most salient to individuals in perception. Some theory does exist on the potential effects of context but few of these theories point to specific contexts or how one might define and measure these contexts.
For example, the theories used as a framework for this investigation suggest that economic context may influence environmental concerns. None of these theories specifically indicated how economic context may be defined or what measures are more appropriate for measuring economic context. In the end, better theory on contextual effects and greater demand for contextual measures may lead to consistency in measurement and greater data availability, thus offering more opportunities for multilevel studies.
One of the better solutions to contextual data problems is for researchers to take contexts into account when measuring at the individual level and then aggregate specific indicators for multilevel studies. For the most part, data availability and theoretical issues arise only when dealing with sub-national studies. National and international studies that use multilevel approaches have greater flexibility in that national data is typically available for cross-national comparison. But, these multilevel studies and the theories that underpin them point to the lack of sub-national research in understanding how and why environmental concern may vary. Given the importance of sub-national studies to the literature, issues of data availability and theories on contextual influence must be addressed. To address such problems, researchers may work to either collect contextual and individual data as part of the same project or develop individual measures that may be aggregated to reflect an overall context. For example, asking individuals to rate the environment where they live using well defined criteria may then be used to create an aggregate measure of perceived environmental quality at a higher level. The same may also be 141 asked about economic conditions or any other contextual variable relevant to the study at hand.
The advantage of this approach is that researchers may derive more proximate measures as additional levels of analysis in hierarchical models. While such solutions may not be ideal if adequate objective measures are possible, such approaches provide a basis for theorizing, measuring and testing relationships not possible in most public opinion data sets.
Additional problems to take into account when reviewing this study include limited generalization and too few cognitive measures included in models. At the time of data collection, the United States was experiencing an increased level of prosperity. Although this increased prosperity was not experienced uniformly across the nation, the possible influence of this time period must be taken into account because the relationships identified in this study may not hold from one time period to another. Likewise, the data used was only applicable to the Southern
Appalachian Region, thus generalizations beyond the region may be untenable. Further, this study focused on a limited number of Level-1 variables suggested as important in explaining environmental concern. Additional cognitive measures on the dimensions of environmental concern coupled with economic perceptions may illuminate the effect of context. For example, the ways that individuals see the economy and environmental tradeoffs may influence environmental concerns, as might the degree to which individuals feel they can influence policy.
In this sense, lower social class may lead to lower political efficacy and subsequently lower pro environmental behavior. This relationship may then vary relative to social, political and economic contexts. Of course, further speculation on potential relationships requires better theory, but the main issue at hand is that additional cognitive measures and samples better suited to generalization across time and place may refine understandings of how contexts affect environmental concerns. Recommendations for Future Th eoryand Research
Any discussion on the types of future research addressing context effects also should include an emphasis on the overarching ideologies that guide much of environmental thinking both in policy and individual behavior. For example, much of the thinking on the natural environment is economic. As a consequence, the solutions to environmental problems that are 142 considered to have merit also are economic. An over-reliance on economic solutions results in marginalizing arguments on the intrinsic value of nature and discourse on how humans are embedded in natural contexts (Barry 2000). Such solutions also have a tendency to marginalize commonly accepted scientific findings like current conclusions on Global Warming as a growing problem. In public debate, these positions have a tendency to distill down into "wedge" issues based on false dichotomies like "jobs vs. environment." These types of debates also lend themselves to inflammatory rhetoric and symbolic debate, instead of substantive understanding or reform. From the perspective of theory, these types of issues and debates reflect an increasing disconnect from nature, and this disconnect should be the focus of theorizing at collective and individual levels. For example, modern industrial or post-industrial societies often operate as if they exist as separate from nature and treat natural resources as commodities distinct from nature. As this type of disconnect grows, so too does the lack of a recognition that the natural environment is a context, thus reducing cultural capacity for environmental reform.
Hence, theoretical development on how and why macro conditions may lead to perpetuating an economic ideology becomes the key to further understanding environmental concerns and decisions of individuals.
The real importance of this kind of theorizing is found in how the disconnect between nature and societies becomes disconnects between individuals, between communities, and between generations. As Habermas (1984) and Giddens (1990) all note, recognition of environmental problems is, in part, a recognition that the lived experiences of humans in modern societies is increasingly becoming alienated, bureaucratic, and consumption oriented. As alienated and consumption oriented beings, individuals see themselves as self-interested, short term economic actors that do not take into account collective or future interests in their behaviors.
The rise of environmental politics and environmental concerns is, therefore, an expression of a broader set of issues challenging traditional institutions that lead to exploitation of nature and humans. In other words, the lived experiences of humans have increasingly become influenced by economic institutions and environmental politics have become a way to address moral questions of identity and lifestyle (Habermas 1984; Giddens 1991 ). In effect, environmental 143 politics is an expression of a growing recognition that modern societies create contradictions that cannot be resolved under market logic. Hence, environmental politics arises from more than environmental crises; it arises from the more fundamental problem that modern societies dominated by economic thinking do not address issues of social justice, morality, and political efficacy (see Giddens 1991 }.
These types of issues are under theorized in Ecological Modernization theory, which rests heavily on the importance of modernization and market mechanisms. As noted earlier, a central claim in Ecological Modernization theory is that environmental problems are a type of institutional failure in adequately responding to environmental issues. Under Ecological
Modernization, the twin forces of economic growth and increased public awareness of environmental problems may lead to institutional transformation. For example, pressures from the public in the form of environmental concern may manifest as demand for environmentally responsible products. Based on this higher demand, industries will be required to engage in a process of super-industrialization, where industries must incorporate more efficient and environmentally responsible production. In essence, the preferences of consumers become a driving force to initiate change. Higher levels of industrial pollution and ineffective efforts or policies in dealing with environmental issues results in a greater concern from the public, which may not only forces changes in production processes but also results in greater pressures on the government to act to protect the natural environment. In short, public concern over the environment and demand for environmental responsibility result in the types of institutional reform predicted by Ecological Modernization theorists, under the economics as ideology model.
Unfortunately, such transformations have yet to have reached fruition to the extent predicted by the theory and, more importantly, ecological crises must occur for environmental concern to rise. Recognizing the reliance on an economic ideology and environmental crises, revisions to the theory include economic context as a potential barrier to cultural capacity and institutional reform. As noted in previous chapters, the United States is one of the nations expected to have begun the shift toward becoming ecologically modern, yet has not made substantive movement toward that end. As to why this shift has not occurred in the expected 144 direction for the United States, the lack of an active, progressive political leadership and the character of public concern for the environment are suggested as the primary causes. Of these two factors, the latter is theorized as the most crucial, especially in modern democracies, in that pressures from the general public to hold leaders accountable for environmental protection become the key to transformation. Yet, public opinion in the United States remains relatively high on protecting the natural environment, despite a lack of institutional reform. This discrepancy has led many Ecological Modernization theorists to contend that additional factors such as the economic conditions nationally and economic conditions locally may affect the degree of cultural capacity, thus reducing pressures for institutional reform. Put more simply, opportunities for institutional transformation may only manifest during periods of economic growth. If economic growth plays such a central role in this process, then economic growth may well increase the likelihood of institutional transformation, but only when a critical mass is reached in terms of the cultural capacity of a nation or local area. The implication of this relationship is that reaching the public with environmental messages is the crucial factor in restructuring institutional relationships and policy, particularly during periods of economic growth. More importantly, by understanding the influence of economic context on environmental concerns, messages may be tailored to specific areas to either enhance or create environmental awareness. In short, economic context may be understood as either a potential barrier to or potential opportunity for growth in environmental awareness and civic action.
Given the problems of relying on market mechanisms and the importance of public action on behalf of the environment, future theory and research should focus specifically on expanding understanding of cultural capacity in Ecological Modernization. This research and theory should take into account three interrelated themes. First, theory and research should focus on how, in some contexts, attachment to an economic ideology varies in public opinion within contexts. This will rely on developing specific cognitive measures that tap the degree to which an individual privileges economic growth as an overarching goal for society, as opposed to social justice goals like environmental protection or decreasing poverty. Second, theory and research should take into account how social justice may be defined and manifest in specific contexts. Often, the 145 definitions of social justice are broad and poorly defined, thus making it difficult to tap values that reflect social justice concerns. For example, research might focus on how to specifically identify salient issues in specific contexts and to rank the salience of these values accordingly. With patterns identified within context, natural resource professionals and environmental social scientists may tailor messages or address the more salient concerns prior to introducing the environment as a basis for political action. Third, research and theory should work to identify the dimensions for perceiving environmental problems and pro-environmental behavioral options.
For example, information on how an individual defines environmental problems and at what level
(global, national, local) such problems are defined may be used to explain the salience of the issue to the individual. Likewise, information on how an individual makes environmental choices in purchasing would illuminate the degree to which societies may rely on market mechanisms, instead of pro-active effort, to provide "green" options. Hence, research and theory should focus on how the life-worlds of individuals may be affected by economic concerns, how these concerns are related to broader social justice values, and how these elements translate into environmental practice within specific contexts ( see Brand 1997).
The goal for this type of research should be to link broader social, political, and cultural patterns to the lived experience of individuals, with an eye toward application to real world circumstances. One of the problems in effectively communicating an environmental message is the difficulty in delineating how individual interests are met through responsible environmental behavior. If individuals orient themselves toward personal economic concerns during periods of economic downturn, an understanding of realistic behavioral choices available to economically stable and economically vulnerable individuals or families must be acquired. For example, purchasing a hybrid vehicle may be beyond the financial limits of some families, but behaviors like recycling or choosing green products when available and affordable may be possible.
Likewise, carpooling or reducing energy consumption may be demonstrated to reduce both personal financial costs and collective environmental costs. In application, a practical environmentalism would require greater efforts to educate the public on the economic and environmental costs of choices for individuals, families and future generations. Put another way, 146 an environmental message that emphasizes the need to protect the environment more broadly for intrinsic or aesthetic reasons may miss the mark in a culture that promotes consumption or a context of economic downturn. The key is to delineate how consumption costs individuals, why it is economically sound for individuals to be pro-environmental, and how it is collectively beneficial for societies to become environmentally responsible. While the availability and costs of "green" products are a function of market forces, a growth in practical options and values encouraging pro-environmental behaviors may result in greater availability and affordability. In short, greater environmental awareness in behavior must be built on practical options viable in good or bad economic conditions to increase cultural capacity and the likelihood of market reform.
Research and theory also should focus on how contexts may provide opportunities to create higher cultural capacity for greater attention to local conditions and greater levels of civic participation in environmental decision-making. Any practical environmentalism at the local level must first be built on a strong foundation of overall civic engagement. This type of civic engagement would involve greater participation in decision-making on any issue of local interest, with the environment as one of many issues addressed. Issues of political and personal efficacy may play a prominent role in how individuals understand their relationship to the natural environment. When citizens see themselves as self-interested individuals, belief in the possibility for political change and personal responsibility decreases, thus affecting the likelihood of accepting an environmental message. By getting people vested in collective concerns at the local level, the links to broader concerns like the natural environment become easier to understand. This type of civic engagement may increase the need for greater environmental knowledge on how the environmental decisions of individuals and communities impact others. In effect, by localizing issues like the environment and increasing public interest in decision-making, not only may environmental awareness/concern increase but also the desire for governmental accountability to citizens. In its most ideal form, these elements are the foundation for a democracy that lends itself to grassroots movements. More importantly in terms of the natural environment, a greater level of civic engagement on issues like the natural environment act as a check against tendencies to set homogenous environmental policies across contexts and act as a 147 check against hierarchical approaches to environmental management. Thus, research and theory along these lines would also require understanding mechanisms leading to an informed civic engagement that increases awareness of environmental issues locally and their connections to environmental issues beyond the local level.
Discussion and Conclusion
Researchers must acknowledge that solutions to such environmental issues are not clear-cut, but may be better understood when contextualized. By paying attention to the economic contexts of environmental problems, it becomes possible to understand such issues analytically, especially in terms of the dynamic between public opinion and action on the natural environment. For example, by looking backward only a few years one may see that in the United
States alone efforts were made to restrict the power of environmental enforcement throughout the
th latter half of the 20 Century, despite widespread public support for the natural environment.
Under the Reagan and Bush administrations in the 1980s and early 1990s, environmental enforcement practice was reduced and claims of environmental harm went uninvestigated (see
Zinn 2004). Presently, the same practices are in place under the George W. Bush administration, with the addition of overt censorship of scientific reports indicating that the natural environment is in peril (Devine 2004 ). Such practices come into conflict with not only basic principles of transparency in democratic process but also result in a poorly informed public on the character of environmental problems. Yet, these practices remain consistent with a dominant viewpoint presently coloring debates on the environment-economy relationship. In effect, economic growth as an ideology has become a master historical context within which debates over the natural environment occur and within which solutions to environmental problems are sought.
This master historical context is at the center of the debates over Ecological
Modernization theory, the role of public concerns, and avenues for identifying common ground.
Ecological Modernization theorists suggest that economic expansion promotes greater investment, higher taxation, and higher interest in social responsibility. As a consequence, citizens at the local level experience the benefits of economic growth and become more environmentally conscious. Opponents of the Ecological Modernization acknowledge that some 148 pro-environmental changes have occurred, but few have been sustained for any period of time
(Schnaiberg, Pellow, and Weinberg 2003). The crux of the problem in relying on market mechanisms is that the state, in an economically domin·ated ideological context, will default to economic growth at the expense of the environment. Although these disagreements exist, there is a region of overlap between these perspectives when dealing with the role of the public in fomenting pro-environmental changes. Both perspectives tend to agree that public concern and cultural capacity play an important role in altering environmental practice. By better understanding how economic context may affect the cultural capacity of a local area, environmental social scientists may gain not only greater clarity on how an environmental message is received but also how to tailor environmental messages to fit with concerns of local citizens. Thus, more context specific research is needed to determine how and why economic conditions lead to greater or lesser support for an environmental agenda.
Given the importance of public concerns, researchers must remember that public opinion on most issues tends to moderate and economic context influences shifts in the public mind.
Given the moderate tendencies of the public and the importance of economic context, a practical environmentalism that avoids inflammatory rhetoric in environmental debates is recommended.
Debates over the environment are prone to become conflicts between "tree-huggers" and "neo fascist corporate interests." These types of debates tend to be nonproductive and force a wedge between individuals, groups or even parties that, if one is to believe national polls, agree at some level on natural environmental protection. Unfortunately, this consensus does not translate into action, and, if inflammatory rhetoric is the basis for claims-making, practical action takes a second place to winning the symbolic political argument in discourse. A more practical approach to the natural environment would involve specific claims made based on objective conditions at any level of analysis that are worsening or improving, where improvements become part of the environmental discourse. Further, objective conditions· like pollution must be linked to real human impact, instead of treated as distant or separate from humans. More specifically, environmental degradation is relatively easy to ignore when, for example, the water pollution of one community is experienced in another community entirely. By linking the pollution of one local area to the 149 experiences of pollution in another, the objective conditions of human impact are better understood and the use of an ambiguous apocalyptic rhetoric and the resulting debates is avoided.
Consistent with the work of classical sociological theorists, environmental social scientists must return to context as a way to further understand how a practical environmentalism may be adopted locally to transform environmental practice. The natural environment has become a consensus issue in American politics and discourse. By exploring the effects of contexts, the relationship between economic distress, social distress and pollution in influencing environmental concerns may be identified. This identification allows for a better understanding of how and when certain issues may become important to specific populations, as well as how to tailor environmental messages to suit both environmental and individual needs. Granted, a great deal of research is needed in terms of the influence of context on environmental concerns. But, based on the research conducted in this dissertation, it is apparent that contexts play a role in environmental concern. The implication of this finding for the broader literature is that environmental concern is not monolithic and that context must be considered when theorizing.
Further, it would appear that treating environmental concerns as a consensus issues becomes a way to avoid substantive environmental reform in policy. For substantive reform and an increase in cultural capacity, economic context must be considered and a practical environmentalism may be necessary before ecological change in social institutions is possible. Ultimately, efforts must be made to create conditions and contexts where individuals choose to actively participate in and understand the importance of environmental decisions that affectthem locally and globally.
150 BIBLIOGRAPHY
151 BIBLIOGRAPHY
Adeola, Francis 0. 1998. "Cross-National Environmentalism Differentials: Empirical Evidence from Core and Non-core Nations." Society and Natural Resources 11 :339-364.
Aiken, Leona S. and Stephen West. 1991. Multiple Regression: Testing and Interpreting Interactions. Newbury Park, CA: Sage.
Ajzen, leek. 2001. "Nature and Operation of Attitudes." Annual Review of Psychology 52:27-58.
Allen, Mary J. and Wendy M. Yen. 1979. Introduction to Measurement Theory. Prospect Heights, IL: Waveland Press, Inc.
Allison, Paul D. 2002. Missing Data. Thousand Oaks, CA: Sage.
Ali, S. Harris. 2002. "Disaster and the Political Economy of Recycling: Toxic Fire in an Industrial City." Social Problems 49:129-149.
Althoff, Phillipand William H. Greig. 1977. "Environmental Pollution Control: Two Views from the General Population." Environment and Behavior 9:441-456.
Amato, Joseph A. and John Radzilowski (eds.). 1999. Community of Strangers: Change, Turnover, Turbulence, and the Transformation of a Midwestern Country Town. Marshall, MN: Crossings Press.
Anderson, Terry L. (ed.) 2004. You Have ToAd mit It's Getting Better: From Economic Prosperity to Environmental Quality. Stanford, CA: Hoover Institution Press.
Anderton, Douglas L., Andy B. Anderson, John M. Oakes, and Michael R. Fraser. 1994. "Environmental Equity: The Demographics of Dumping." Demography 31 :221-240.
Environmental Policy Andrews, Richard N. 1994. "Risk-Based Decision-making." Pp. 209-232nd in in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2 ed. Washington, D.C.: Congressional Quarterly.
Arbuckle, James L. and Werner Wothke. 1999. Amos 4.0 User's Guide. Chicago: Small Waters Corporation.
Arbuthnot, Jack and Sandra Ling. 1975. "A Comparison of French and American Environmental Behavior, Knowledge, and Attitudes." InternationalJournal of Psychology 10:275-281.
Arcury, Thomas A. and Eric H. Christianson. 1990. "Environmental Worldview in Response to Environmental Problems: Kentucky 1984 and 1988 Compared." Environment and Behavior 22:387-407.
Arcury, Thomas A. and Eric H. Christianson. 1993. "Rural-Urban Differences in Environmental Knowledge and Actions." Journalof Environmental Education 25:19-25.
Aronowitz, Stanley. 2003. How Class Works: Power and Social Movement. New Haven, CT: Yale University Press.
Ashford, Nicholas A. 2002. "Government and Environmental Innovation in Europe and North America." American Behavioral Scientist 45:1417-1434.
152 Attfield, Robin. 1994. Environmental Philosophy: Principles and Prospects. Brookfield, VT: Ashgate Publishing Company.
Bakan, Joel. 2004. The Corporation: The Pathological Pursuit of Profit and Power. New York: Free Press.
Bakvis, Herman and Neil Nevitte. 1992. ''The Greening of the Canadian Electorate: Environmentalism, Ideology, and Partisanship." Pp. 144-163 in Canadian Environmental Policy: Ecosystems, Politics, and Process, edited by Robert Boardman. Don Mills, ON: Oxford University Press.
Barrett, Brendan. 2005. Ecological Modernization and Japan. New York: Routledge.
Barrow, Chris J. 1997. Environmental and Social Impact Assessment. New York: John Wiley & Sons, Inc.
Barry, John. 2000. Environment and Social Theory. New York: Routledge.
Bartlett, Robert V. 1994. "Evaluating Environmental Policy Success and Failure." Pp. 167-188nd in Environmental Policy in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2 ed. Washington, D.C.: Congressional Quarterly.
Bagguley, Paul. 1992. "Social Change, The Middle Class and the Emergence of "New Social Movements": A Critical Analysis." Th e Sociological Review 40:26-48.
Baugh, Joyce A. 1991. "African-Americans and the Environment: A Review Essay." Policy Studies Journal 19:182-191 .
Baumer, Eric P., Steven F. Messner, and Richard Rosenfield. 2003. "Explaining Spatial Variation in Support for Capital Punishment: A Multilevel Analysis." American Journalof Sociology 108:844-875.
Baumol, William J. and Wallace E. Oates. 1975. The Theory of Environmental Policy. Englewood Cliffs, NJ: Prentice-Hall.
Beale, Calvin L. 1977. "The Recent Shift of United States Population to Non-Metropolitan Areas, 1970-1975." International Regional Science Review 2: 113-122.
Beale, Calvin L. and Kenneth M. Johnson. 1998. "The Identification of Recreational Counties in Non-Metropolitan Areas of the United States." Population Research and Policy Review 17:37-53.
Beckert, Jens. 1996. "What Is Sociological about Economic Sociology? Uncertainty and the Embeddedness of Economic Action." Theory and Society 25:803-840.
Been, Vicki. 1994. "Locally Undesirable Land Uses in Minority Neighborhoods: Disproportionate Siting or Market Dynamics?" The Yale Law Journal 103:1383-1421.
Beierle, Thomas C. and Jerry Cayford. 2002. Democracy in Practice: Public Participation in Environmental Decisions. Washington, D.C.: Resources for the Future.
Bell, Michael M. 1998. An Invitation to En vironmental Sociology. Thousand Oaks, CA: Pine Forge Press.
153 Benford, Robert D. and David A. Snow. 2000. "Framing Processes and Social Movements: An Overview and Assessment." Annual Review of Sociology 2�:611-639.
Bennett, Keith and Mark K. McBeth. 1998. "Contemporary Western Rural USA Economic Composition: Potential Implications for Environmental Policy and Research." Environmental Management 22:371-381.
Benton, Ted. 2002. "Social Theory and Ecological Politics: Reflexive Modernization or Green Socialism?" Pp. 252-273 in Sociological Theory and the Environment, edited by Frederick H. Buttel, Peter Dickens, Riley E. Dunlap, and August Gijswijt. Lanham, MD: Rowman and Littlefield.
Berry, Edna H. 2000. "Review Essay: Rural Sociology, Migration, and Community Change." Rural Sociology 65:658-667.
Berger, Ida E. 1997. "The Demographics of Recycling and the Structure of Environmental Behavior." Environment and Behavior 29:515-531.
Beutel, Ann M. and Margaret M. Marini. 1995. "Gender and Values." American Sociological Review 60:436-448.
Biel, Anders, Bengt Hansson, and Mona Martensson (eds.). 2003. Individual and Structural Determinants of Environmental Practice. Burlington, VT: Ashgate Publishing Company.
Blaikie, Norman. 1992. "The Nature and Origins of Ecological Worldviews: An Australian Study." Social Science Quarterly 7 3: 144-165.
Blake, Donald E. 2001. "Contextual Effectson Environmental Attitudes and Behavior." Environment and Behavior 33:708-725.
Block, Fred. 1990. Postindustrial Possibilities. Los Angeles, CA: University of California Press.
Blocker, T. Jean and Douglas L. Eckberg. 1989. "Environmental Issues as Women's Issues: General Concerns and Local Hazards." Social Science Quarterly 70:586-593.
1997. "Gender and Environmentalism: Results from the 1993 General Social Survey." Social Science Quarterly78:84 1-858.
Bogart, W.T. 1998. The Economics of Cities and Suburbs. Upper Saddle River, NJ: Prentice Hall.
Books, John and Charles Prysby. 1999. "Contextual Effects on Retrospective Economic Evaluations the Impact of the State and Local Economy." Political Behavior 21 :1-16.
Booth, Douglas E. 2004. Hooked on Growth: Economic Addictions and the Environment. Lanham, MD: Rowman & Littlefield Publishers.
Bord, Richard J. and Robert E. O'Connor. 1997. "The Gender Gap in Environmental Attitudes: The Case of Perceived Vulnerability to Risk." Social Science Quarterly 78:830-839.
Bordieu, Pierre. 2002. Outline of a Theoryof Practice. New York: Cambridge University Press.
Bosso, Christopher J. 1994. "After the Movement: Environmental Activism in the 1990s." Pp. 31- E. 50nd in Environmental Policy in the 1990s, edited by Michael Kraft and Norman J. Vig, 2 ed. Washington, D.C.: Congressional Quarterly. 154 Brace, Paul, Kellie Sims-Butler, Kevin Arceneaux, and Martin Johnson. 2002. "Public Opinion in the American States: New Perspectives Using National SurveyData." American Journal of Political Science 46:173-189.
Bradley, Athena L. and John Nolt. 2005. "Consumption and Waste." Pp. 261-282 in A Land Imperile d, edited by John Nolt. Knoxville, TN: University of Tennessee Press.
Brand, Karl. W. 1997. "Environmental Consciousness and Behavior: The Greening of Lifestyles." Pp. 204-217 in The InternationalHandbook of En vironmental Sociology, edited by Micheal Redclift and Graham Woodgate. Cheltingham, UK: Edward Elgar.
Bramwell, Anna. 1989. Ecology in the Twentieth Century: A History. New Haven. CT: Yale University Press.
Brechin, Steven R. 1999. "Objective Problems, Subjective Values, and Global Environmentalism: Evaluating the Postmaterialist Argument and Challenging a New Explanation." Social Science Quarterly 80:793-809.
Brechin, Steven R. and Willet Kempton. 2003. "Global Environmentalism: A Challenge to the Postmaterialism Thesis?" Pp. 326-344 in Environment, Energy, and Society: Exemplary Works, edited by Craig R. Humphrey, Tammy L. Lewis, and Frederick H. Buttel. Belmont, CA: Wadsworth.
Brody, Charles. 1984. "Differences by Sex in Support for Nuclear Power." Social Forces 63:209- 228.
Brown, Phil, Stephen Zavestoski, Brian Mayer, Sabrina McCormick and Pamela S. Webster. 2002. "Policy Issues in Environmental Health Disputes." The Annals of the American Academy of Political and Social Science 584:175-202.
Brulle, Robert J. 1996. "Environmental Discourse and Social Movement Organizations: A Historical and Rhetorical Perspective on the Development of U. S. Environmental Organizations." Sociological Inquiry 66:58-83.
Bryant, Bunyan (ed.). 1995. Environmental Justice, Issues, Policies, and Solutions. Washington, D.C.: Island Press.
Bullard,· Robert D. 1990. Dumping in Dixie: Race, Class, and Environmental Quality. Boulder, CO: Westview.
1993. Confronting Environmental Racism: Voices from the Grassroots. Boston, MA: South End Press.
1994a. "Environmental Racism and 'Invisible Communities." We st Virginia Law Review 96:1037-1050.
1994b. "Legacy of American Apartheid and Environmental Racism." St. John's Journalof Legal Commentary9:4 45-4 72.
1996. "Environmental Justice: It's More than Waste Facility Siting." Social Science Quarterly 77:493-499.
2002. "Confronting Environmental Racism in the Twenty-First Century." Global Dialogue 4:34-48. 155 Bullard, Robert D. and Glenn S. Johnson. 2000. "Environmental Justice: Grassroots Activism and Its Impact on Public Policy Decision Making." Journalof Social Issues 56:555-578.
Bullard, Robert D. and Beverly H. Wright. 1986."The Politics of Pollution: Implications for the Black Community." Phy/on 47:71-78.
1987. "Environmentalism and the Politics of Equity: Emergent Trends in the Black Community." Mid-American Review of Sociology 12:21-38.
1990. "The Quest For Environmental Equity: Mobilizing the African-American Community For Social Change." Society and Natural Resources 3:301-311.
Bullard, Robert D., Glenn S. Johnson and Beverly H. Wright. 1997. "Confronting Environmental Injustice: It's the Right Thing to Do." Race, Gender and Class 5(1): 63-79.
Burbank, Matthew J. 1995. "How Do Contextual Effects Work? Developing a Theoretical Model." Pp. 165-178 in Spatial and Contextual Models in Political Research, edited by Munroe Eagles. Londo�, UK: Taylor and Francis.
1997. "Explaining Contextual Effectson Vote Choice." Political Behavior 19:113-132.
Burn, Shawn. 1991. "Social Psychology and the Stimulation of Recycling Behaviors: The Block Lead er Approach." Journalof Applied Social Psychology 21 :611-619.
Bush, Judith, Suzanne Moffatt and Christine Dunn. 2001.. "'Even the Birds Round Here Cough': Stigma, Air Pollution and Health in Teesside." Health and Place 7:47-56.
Buttel, Frederick H. 1975. "The Environmental Movement: Consensus, Conflict, and Change." Journal of Environmental Education 7:53-63.
1978a. "Economic Growth and the Welfare State: Implications for the Future of Environmentalism." Social Science Quarterly 58:693-699.
1978b. "Environmental Sociology: A New Paradigm?" The American Sociologist 13:252- 256.
1979. "Age and Environmental Concern: A Multivariate Analysis." Youth and Society 10:237-256.
1987. "New Directions in Environmental Sociology." Annual Review of Sociology 13:465- 488.
1992. "Environmentalization: Origins Processes, and Implications for Rural Social Change." Rural Sociology 57:1-27.
1993. "Environmentalization and Greening: Origins, Processes, and Implications." Pp. 252-256 in The Greening of Rural Policy: International Perspectives, edited by Sarah Harper. New York: Bellhaven Press.
1997. "Social Institutions and Environmental Change." Pp. 40-46 in The International Handbook of Environmental Sociology,. edited by Michael Redclift and Graham Woodgate. Northhampton, MA: Edward Elgar Publishing.
156 Buttel, Frederick H. and William L. Flinn. 1974. ''The Structure and Support for the Environmental Movement, 1968-1970." Rural Sociology 39:55-69.
1975a. "Methodological Issues in the Sociology of Natural Resources." Humboldt Journal of Social Relations 3:63-72.
1975b. "Sources and Consequences of Agrarian Values in American Society." Rural Sociology 2:134-1 51.
1976a. "Economic Growth Versus the Environment: Survey Evidence." Social Science Quarterly 57:41 1-420.
1976b. "Environmental Politics: The Structuring of Partisan and Ideological Cleavages in Mass Environmental Attitudes." The Sociological Quarterly 17:477-490.
1976c. "Sociopolitical Consequences of Agrarianism." Rural Sociology 41 :473-483.
1977a. ''The Interdependence of Rural and Urban Environmental Problems in Advanced Capitalist Societies: Models of Linkage." Socio/ogia Ruralis 17:255-281.
1977b. "Conceptions of Rural Life and Environmental Concern." Rural Sociology 42:544- 555.
1978a. "The Politics of Environmental Concern: The Impacts of Party Identification and Pol_itical Ideology on Environmental Attitudes." Environment and Behavior 10:17-36.
1978b. "Social Class and Mass Environmental Beliefs: A Reconsideration." Environment and Behavior 10:433-450.
1979. "Sources of Working Class Consciousness." Sociological Focus 12:37-52.
Buttel, Frederick H. and Donald E. Johnson. 1977. "Dimensions of Environmental Concern: Factor Structure, Correlates, and Implications for Research." Journalof Environmental Education 9:49-74.
Buttel, Frederick H. and Peter J. Taylor. 1992. "Environmental Sociology and Global Environmental Change: A Critical Assessment." Society and Natural Resources 40:21 1- 230.
Buttel, Frederick H., Ann P. Hawkins, and Alison G. Power. 1990. "From Limits to Growth to Global Change." Global En vironmental Change 1 :57-66. Buttel, Frederick H., Peter Dickens, Riley E. Dunlap, and August Gijswijt. 2002. "Sociological Theory and the Environment: An Overview and Introduction." Pp. 3-34 in Sociological Theoryand the Environment, edited by Frederick H. Buttel, Peter Dickens, Riley E. Dunlap, and August Gijswijt. Lanham, MD: Rowman and Littlefield.
Cable, Sherry and Michael Benson. 1993. "Acting Locally: Environmental Injustice and the Emergence of Grass-Roots Environmental Organizations." Social Problems 40:464-477.
Cable, Sherry and Charles Cable. 1995. Environmental Problems, Grassroots Solutions: The Politics of Grassroots Environmental Conflict. New York: Saint Martin's Press, Inc.
157 Cable, Sherry, Donald W. Hastings, and Tammy L. Mix. 2002. "Different Voices, Different Venues: Environmental Racism Claims by Activists, Researchers, and Lawyers. Human Ecology Review 9:26-42
Cairncross, Frances. 1995. Green, Inc. Washington, D.C.: Island Press.
Canan, Penelope. 1996. "Bringing Nature Back In: The Challenge of Environmental Sociology." Sociological Inquiry66: 29-36.
Carman, Christopher J. 1988. "Dimensions of Environmental Policy Support in the United States." Social Science Quarterly 79:717-733.
Carruthers, Bruce G. and Brian Uzzi. 2000. "Economic Sociology in the New Millennium." Contemporary Sociology 29:486-494.
Carson, Rachel. 2002. Silent Spring. New York: First Mariner Books.
Catton, William R. Jr. 1982. Overshoot: Th e Ecological Basis of Revolutionary Change. Urbana, IL: University of Illinois Press.
1994. "Foundations of Human Ecology." Sociological Perspectives 37:75-95.
Catton, William R. Jr. and Riley E. Dunlap. 1978. "Environmental Sociology: A New Paradigm." Th e American Sociologist 13:41-49.
1980. "A New Ecological Paradigm for Post-Exuberant Sociology." American Behavioral Scientist 24:14-57.
Champion, Anthony G. 1989. Counterurbanization: Th e Changing Pace and Nature of Population Deconcentration. London, UK: Edward Arnold.
Chase-Dunn, Christopher. 1999. "Globalization: A World-Systems Perspective." Journalof World Systems Research 5:187-215.
Cheng, Anthony S., Linda E. Krueger, and Steven E. Daniels. 2003. '"'Place" As an Integrating Concept in Natural Resource Politics: Propositions for a Social Science Research Agenda." Society and Natural Resources 16:87-104.
Choldin, Harvey M. 1978. "Social Life in the Physical Environment." Pp. 352-383 in Handbook of Contemporary Urban Life, edited by David Street. San Francisco, CA: Jossey-Bass.
Cohen, Maurie J. 1998. "Science and the Environment: Assessing Cultural Capacity for Ecological Modernization." Public Understandings of Science 7:149-167.
2000. "Ecological Modernization, Environmental Knowledge and National Character: A Preliminary Analysis of the Netherlands" Environmental Politics 9:77-106.
Collins, Randall. 2000. "Situational Stratification: A Micro-Macro Theory of Inequality." Sociological Theory 18:17-43.
Commoner, Barry. 1972. The Closing Circle: Nature, Man, and Te chnology. New York: Bantam Press.
158 Cordell, Kenneth H., Gerald Helton, and John Peine. 1996. "Communities and Human Influences in Southern Appalachian Ecosystems." SouthernAppalachian Assessment: Social/Cultural/Economic Technical Report. USDA, Forest Service, Southern Region.
Cotgrove, Stephen. 1982. Catastrophe or Cornucopia: The Environment, Politics, and the Future. Chichester, UK: John Wiley and Sons.
Cotgrove, Stephen and Andrew Duff. 1981. "Environmentalism, Values, and Social Change." The British Journal of Sociology 32:92-110.
Crossley, Nick. 2002. Making Sense of Social Movements. Philadelphia, PA: Open University Press.
Cutter, Susan J. 1981. "Community Concern for Pollution: Social and Environmental Influences." Environment and Behavior 13:105-1 24.
Cutter, Susan L., Bryan J. Boruff, and W. Lynn Shirley. 2003. "Social Vulnerability to Environmental Hazards." Social Science Quarterly 84:242-261.
Daniels, Thomas L. and Mark B. Lapping. 1996. "The Two Rural Americas Need More, Not Less Planning." Journalof the American Planning Association 62:285-288.
Davidson, Debra J. and William R. Freudenberg. 1996. "Gender and Environmental Risk Concerns: A Review and Analysis of Available Research." Environment and Behavior 28:302-339.
Davies, Paul and Kenneth Newton. 1972. "The Social Pattern of Immigrant Areas." Race 14:43- 57.
De Young, Raymond. 2000. "Expanding and Evaluating Motives for Environmentally Responsible Behavior." Journalof Social Issues 56:509-526.
Decker, Jill and John L. Crompton. 1990. "Business Location Decisions: The Relative Importance of Quality of Life and Recreation, Park, and Cultural Opportunities." Journal of Park and Recreation Administration 8:79-93.
Derr, Thomas S. 1975. "Religion's Responsibility for the Ecological Crisis: An Argument Run Amok." A Journalof Religion and International Affairs 18:39-45.
Dersken, Linda and John Gartrell. 1993. "The Social Context of Recycling." American Sociological Review 58:434-442.
Des Jardins, Joseph R. 1997. Environmental Ethics. Belmont, CA: Wadsworth Publishing Company.
Devine, Robert S. 2004. Bush versus the Environment. New York: Random House.
Dickens, Peter. 2002. "A Green Marxism? Labor Processes, Alienation, and the Division of Labor." Pp. 51-72 in Sociological Th eory and the Environment, edited by Frederick H. Buttel, Peter Dickens, Riley E. Dunlap, and August Gijswijt. Lantham, MD: Rowman and Littlefield.
159 Diekmann, Andreas and Axel Franzen. 1999. "The Wealth of Nations and Environmental Concern." Environment and Behavior 31 :540-549.
Dietz, Thomas, Paul C. Stern, and Robert W. Rycroft. 1989. "Definitions of Conflict and the Legitimation of Resources: The Case of Environmental Risk." Sociological Forum 4:47- 70.
Douglas, Mary. 1992. Risk and Blame: Essays in Cultural Theory. London, UK: Routledge.
Douglas Mary and Aaron Wildavsky. 1982. Risk and Culture: An Essay on the Selection of Technological and Environmental Dangers. Berkeley, CA: University of California Press.
Dewie, Mark. 1995. Losing Ground: American Environmentalism at the Close of the Twe ntieth Century. Cambridge, MA: The MIT Press.
Downs, Anthony. 1972. "Up and Down with Ecology: The Issue-Attention Cycle." Public Interest 28:38-50.
Draper, David. 1995. "Inference and Hierarchical Modeling in the Social Sciences." Journalof Educational and Behavioral Statistics 20:115-147.
Duffy, Robert J. 2003. The Green Agenda in American Politics: New Strategies for the Twenty First Century. Lawrence, KS: University Press of Kansas.
Dunlap, Riley E. 1975. ''The Impact of Political Orientation on Environmental Attitudes and Actions." Environment and Behavior 7:429-454.
1980. "Paradigmatic Change in Social Science: From Human Exemptionalism to an Ecological Paradigm." American Behavioral Scientist 24:5-14.
1987. "Polls, Pollution, and Politics Revisited: Public Opinion on the Environment in the Reagan Era." Environment 29:6-11, 32-37.
1989. "Public Opinion and Environmental Policy." Environmental Politics and Policy, edited by James R. Lester. Raleigh, NC: Duke University Press.
1991. "Public Opinion in the 1980s: Clear Consensus, Ambiguous Commitment." Environment 33:10-15, 32-37.
1992. ''Trends in Public Opinion toward Environmental Issues: 1965-1990." Pp. 89-115 in American Environmentalism: The US Environmental Movement, 1970-1990, edited by Riley E. Dunlap and Angela G. Mertig. New York: Taylor and Francis.
1997. "The Evolution of Environmental Sociology: A Brief History and Assessment of the American Experience." Pp. 21-39 in The InternationalHandbook of Environmental Sociology, edited by Michael Redclift and Graham Woodgate. Northamption, MA: Edward Elgar.
1998. "Lay Perceptions of Global Risk: Public Views of Global Warming in a Cross National Context." International Sociology 13:473-498.
2002. "Paradigms, Theories, and Environmental Sociology." Pp. 329-350 in Sociological Theory and the Environment, edited by Frederick H. Buttel, Peter Dickens, Riley E. Dunlap, and August Gijswijt. Lanham, MD: Rowman and Littlefield. 160 Dunlap, Riley E. and Matthew P. Allen. 1976. "Partisan Differenceson Environmental Issues: A Congressional Roll-Call Analysis." We stern Political Quarterly 29:384-397.
Dunlap, Riley E. and William R. Catton, Jr. 1979. "Environmental Sociology." Annual Review of Sociology 3:243-273.
1983. "What Environmental Sociologists Have In Common (Whether Concerned with "Built" or "Natural" Environments)." Sociological Inquiry 53:113-135.
2003. "Struggling with Human Exceptionalism: The Rise, Decline and Revitalization of Environmental Sociology." Pp. 97-1 19 in Environment, Energy, and Society: Exemplary Works, edited by Craig R. Humphrey, Tammy L. Lewis, and Frederick H. Buttel. Belmont, CA: Wadsworth.
Dunlap, Riley E. and Richard P. Gale. 1974. "Party Membership and Environmental Politics: A Legislative Roll-Call Analysis." Social Science Quarterly 55:670-690.
Dunlap, Riley E. and Robert E. Jones. 2001 . "Environmental Concern: Conceptual and Measurement Issues." Pp. 482-524 in Handbook of Environmental Sociology, edited by Riley E. Dunlap and William Michelson. Westport, CT: Greenwood University Press.
Dunlap, Riley E. and Angela G. Mertig (eds.). 1992a. American Environmentalism: The U. S. Environmental Movement, 1970- 1990. New York: Taylor and Francis.
1992b. "The Evolution of the US Environmental Movement From 1970-1990: An Overview." Pp. 1-10 in American Environmentalism: The US Environmental Movement, edited by Riley E. Dunlap and Angela G. Mertig. New York: Taylor and Francis.
1995. "Global Concern for the Environment: Is Affluencea Prerequisite?" Journalof Socia/ Issues 51:122-1 37.
1997. "Global Environmental Concern: An Anomaly for Postmaterialism." Social Science Quarterly 78:24-29.
Dunlap, Riley E. and Rik Scarce. 1991 . "Environmental Problems and Protection." Public Opinion Quarterly 55:651-672.
Dunlap, Riley E. and Kent D. Van Liere. 1978a. "Commitment to the Dominant Social Paradigm and Concern for Environmental Quality." Social Science Quarterly 10:13-28.
1978b. "The "New Environmental Paradigm." Journalof Environmental Education 9:10- 19.
1984. "Commitment to the Dominant Social Paradigm and Concern for Environmental Quality." Social Science Quarterly 65:101 3-1028. Dunlap, Riley E. and Marvin E. Olsen. 1984. "Hard-Path versus Soft-Path Advocates: A Study of Energy Advocates." Policy Studies Journal 13:328-413.
Dunlap, Riley E., George H. Gallup Jr., and Alex M. Gallup. 1993. "Of Global Concern: Results of the Health of the Planet Survey." Environment 35:7-39.
Dunlap, Riley E., Frederick Buttel, Paul Dickens, and August Gijswijt (eds.). 2002. Sociological Theory and the Environment. New York: Rowman & Littlefield Publishers, Inc. 161 Dunlap, Riley E., Kent D. Van Liere, Angela G. Mertig, and Robert E. Jones. 2000. "Measuring Endorsement of the New Ecological Paradigm: A Revised NEP Scale." Journalof Social Issues 56:425-442.
Durant, Will. 1968. The Lessons of History. New York: Simon & Schuster.
Durkheim, Emile. 1984. The Division of Labor in Society. Translated by W .D. Halls. New York: The Free Press.
Durr, Robert H. 1993. "What Moves Policy Sentiment?" The American Political Science Review 87: 158-170.
Eagles, M. (ed.). 1995. Spatial and Contextual Models in Political Research. London: Taylor and Francis.
Eckberg, Douglas L. and T. Jean Blocker. 1989. "Varieties of Religious Involvement and Environmental Concern: Testing the Lynn White Thesis." Journalfor the Scientific Study of Religion 28:509-517.
Ehrenreich, Barbara. 2001. Nickel and Dimed. New York: Henry Holt and Company.
Elliot, Euel, James L. Regens, and Barry J. Seldon. 1995. "Exploring Variation in Public Support for Environmental Protection." Social Science Quarterly 76:41-52.
Elliot, Fratkin. 1997. "Pastoralism: Governance and Development Issues." Annual Review of Anthropology 26:235-261.
Engel, Uwe and Manuela Potschke. 1998. 'Willingness to Pay for the Environment: Social Structure, Value Orientations and Environmental Behavior in a Multilevel Perspective." Innovation 11 :315-332.
English, Mary R. and Sean T. Huss. 2005. "Population and Urbanization." Pp. 155-172 in A Land Imperiled, edited by John Nolt. Knoxville, TN: University of Tennessee Press.
Esty, Daniel. 1999. "Toward Optimal Environmental Governance." New York University Law Revie w 74: 1495-1574.
Fisher, Dana R. and William R. Freudenburg. 2001. "Ecological Modernization and Its Critics: Assessing the Past and Looking Towards the Future." Society and Natural Resources 14:701-709.
2004. "Postindustrialization and Environmental Quality: An Empirical Analysis of the Environmental State." Social Forces 83: 157-188.
Forsyth, Timothy. 2003. Critical Political Ecology. New York: Routledge.
Fortmann, Louise and Jonathan Kusel. 1990. "New Voices, Old Beliefs: Forest Environmentalism Among New and Long-Standing Rural Residents." Rural Sociology 55:214-232.
Foster, John B. 1992. "The Absolute General Law of Environmental Degradation Under Capitalism." Capitalism, Nature, Socialism 3:77-82
------. 1994. The Vulnerable Planet: A Short Economic History of the Environment. New York: Monthly Press Review. 162 ------. 1999. "Marx's Theory of Metabolic Rift: Classical Foundations for Environmental Sociology." American Journalof Sociology 105:366-405.
------. 2003. "The Crisis of the Earth." Pp. 120-136 in En vironment, Energy, and Society: Exemplary Works, edited by Craig R. Humphrey, Tammy L. Lewis, and Frederick H. Buttel. Belmont, CA: Wadsworth.
Frank, David, John Hironaka, and Evan Schaefer. 2000. "The Nation-State and the Natural Environment over the Twentieth Century." American Sociological Review 65:96-1 16.
Freeman 111, Myrick A. 1994. "Economics, Incentives, and Environmental Regulation." Pp. 189- 208 in Environmental Policy in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2nd ed. Washington, D.C.: Congressional Quarterly.
Freudenburg, William R. 1991. "Rural-Urban Differences in Environmental Concern: A Closer Look." Sociological Inquiry 61 :186-1 98.
Freudenburg, William R. and Robert Gramling. 1989. "The Emergence of Environmental Sociology: Contributions of Riley E. Dunlap and William R. Catton, Jr." Sociological Inquiry 59:439-452.
Fuchs, Stephan. 2001 . "Beyond Agency." Sociological Theory19 :24-40.
Fukuyama, Francis. 2000. The Great Disruption. New York: Simon and Schuster.
2002. The End of History and the Last Man. New York: Perennial.
Garner, Craig L. and Stephen W. Raudenbush. 1991. "Neighborhood Effectson Educational Attainment: A Multi-level Analysis." Sociology of Education 64:251-262
Gates, Joseph. 2001 . Democracy at Risk: Rescuing Main Street from Wall Street. Cambridge, MA: Perseus Publishing.
Gaventa, John. 1982. Power and Powerlessness . Chicago, IL: University of Illinois Press.
Gawande, Kishore and Hank Jenkins-Smith. 2001. "Nuclear Waste Transport and Residential Property Values: Estimating the Effectsof Perceived Risks." Journalof Environmental Economics and Management 42:207-233. George, David and Pricilla L. Southwell. 1986. "Opinion on the Diablo Canyon Nuclear Power Plant: The Effectsof Situation and Socialization." Social Science Quarterly 39:722-735. Geteris, Joseph. 1998. "Political Alignment and the American Middle Class, 1974-1994." Sociological Forum 13:639-666. Gibbs, David. 2000. "Ecological Modernization, Regional Economic Development, and Regional Development Agencies." Geoforum 31 :9-19.
2003. "Ecological Modernization and Local Economic Development: The Growth of Eco industrial Development Initiatives." InternationalJournal of Environment and Sustainable Development 2:1-17.
Giddens, Anthony. 1986. The Constitution of Society. Los Angeles, CA: University of California Press. 163 1990. The Consequences of Modernity. Stanford, CA: Stanford University Press.
1991. Modernity and Self-Identity. Stanford, CA: Stanford University Press.
Gieryn, Thomas F. 2000. "A Space for Place in Sociology." Annual Review of Sociology 26:463- 496.
Gill, James D., Lawrence A. Crosby, and James R. Taylor. 1986. "Ecological Concern, Attitudes, and Social Norms in Voting Behavior." Public Opinion Quarterly 50:537-554.
Glenn, Norval D. 1977. Cohort Analysis. Beverly Hills, CA: Sage Publications.
Gobster, Paul H. 1995. "Perception and Use of a Metropolitan Greenway System for Recreation." Landscape and Urban Planning 33:401 -41 3.
1998. "Nearby Neighborhood Residents' Images and Perceptions of the River." Pp. 5-48 in People and the River: Perception and Use of Chicago Waterways for Recreation, edited by P.H. Gobster and L.M. Westphal. Milwaukee, WI: U.S. Department of Interior, National Park Service Rivers, Trails, and Conservation Assistance Program.
2002. "Managing Urban Parks for a Racially and Ethnically Diverse Clientele." Leisure Sciences 24:1 43-1 59.
Goldman, Michael and Rachel A. Schurman. 2000. "Closing the "Great Divide": New Social Theory on Society and Nature." Annual Review of Sociology 26:563-584.
Gould, Kenneth A. 1993. "Pollution and Perception: Social Visibility and Local Environmental Mobilization." Qualitative Sociology 16:1 57-1 78.
Gould, Kenneth A., Allan Schnaiberg, and Adam S. Weinberg. 1996. Local Environmental Struggles: Citizen Activism in the Tre admillof Production. Cambridge, UK: Cambridge University Press.
Gray, David. 1985. Ecological Beliefs and Behaviors. Westport, CT: Greenwood Press.
Greenbaum, Allan. 1996. "Taking Stock of Two Decades of Research on the Social Bases of Environmental Concern." Pp. 125-1 52 in Environmental Sociology: Theory and Practice, edited by Michael D. Mehtat and Eric E. Oulette. North York, ON: Captus Press.
Guagnano, Gregory A., Paul C. Stern, and Thomas Dietz. 1995. "Influences on Attitude Behavior Relationships: A Natural Experiment with Curb Side Recycling." Environment and Behavior, 26:699-71 8.
Guber, Deborah L. 1996. "Environmental Concern and the Dimensionality Problem: A New Approach to an Old Predicament." Social Science Quarterly 77:644-662.
2003. The Grassroots of a Green Revolution. Cambridge, MA: The MIT Press.
Guerrin, Daniel, Jean Crete, and Jean Mercier. 2001 . "A Multilevel Analysis of the Determinants of Recycling Behavior in the European Countries." Social Science Research 30:195-21 8.
Guffey, Stan. 2005. "History." Pp. 15-48 in A Land Imperiled, edited by John Nolt. Knoxville, TN: University of Tennessee Press.
164 Habermas, Jurgen. 1984. Th e Theory of Communicative Action: Reason and the Rationalization of Society. Translated by Thomas McCarthy. Boston, MA: Beacon Press.
2000. The Structural Tra nsformation of the Public Sphere. Translated by Thomas Burger. Cambridge, MA: The MIT Press.
Hackett, Paul. 1995. Conservation and the Consumer: Understanding Environmental Concern. New York: Routledge.
Hajer, Maarten A. 1995. Th e Politics of Environmental Discourse: Ecological Modernization and the Policy Process. New York: Oxford University Press.
Hambleton, Ronald K., Hariharan Swaminathan, and H. Jane Rogers. 1991. Fundamentals of Item Response Theory. London: Sage.
Hamilton, Lawrence D. 1985a. "Concern about Toxic Wastes: Three Demographic Predictors." Sociological Perspectives 28:463-486.
--. 1985b. "Who Cares about Water Pollution? Opinions in a Small-Town Crisis." Sociological Inquiry 55:170-181.
Hand, Carl and Kent Van Liere. 1984. "Religion, Mastery-Over-Nature, and Environmental Concern." Social Forces 63:555-570.
Hanham, Alison, Robert Hanham, and Shawn Banasick. 2000. "A Human Development Index for Pennsylvania Counties: An Applied Regional Geography." The Pennsylvania Geographer 38:91-105.
Hannigan, John A. 1995. Environmental Sociology: A Social Constructionist Perspective. New York: Routledge Press_.
Hannon, Bruce M. 1994. "Sense of Place: Geographic Discounting by People, Animals, and Plants." Ecological Economics 10:157-174.
Hannon, Bruce M. and Bryan G. Norton. 1997. "Environmental Values: A Place-Based Theory." Environmental Ethics 19:227-245.
Hardin, Garret. 1968. "The Tragedy of the Commons." Science 162:1243-1248.
Hardy, Melissa A. 1993. Regression with Dummy Variables. Newbury Park, CA: Sage Publications.
Hawken, Paul. 1993. Th e Ecology of Commerce. New York: Harper Collins.
Hawken, Paul, Amory Lovins, and L. Hunter Lovins. 1999. Natural Capitalism: Creating the Next Industrial Revolution. Boston, MA: Little, Brown and Company.
Heberlein, Thomas A. and J. Stanley Black. 1976. "Attitudinal Specificity and the Prediction of Behavior in a Field Setting." Journalof Personality and Social Psychology 33:474-479.
Hertz, Noreena. 2003. The Silent Ta keover: Global Capitalism and the Death of Democracy. New York: HarperCollins.
165 Honadle, George. 1999. How Context Matters: Linking Environmental Policy to People and Place. West Hartford, CT: Kumarian Press.
Honnold, Julie A. 1984. "Age and Environmental Concern: Some Specification of Effects." Journal of Environmental Education 16:4-9.
Hoover, Kenneth R. 2003. Economics as Ideology. Lanham, MD: Rowman & Littlefield Publishers, Inc.
Hornborg, Alf. 1998. "Ecological Embeddedness and Personhood: Have We Always Been Capitalists?" Anthropology Today 14:3-5.
2003. "Cornucopia or Zero-Sum Game: The Epistemology of Sustainability." Journalof World-Systems Research 9:205-216.
House, James S., 1981. "Social Structure and Personality." Pp. 525-561 in Social Psychology: Sociological Perspectives, edited by Morris Rosenberg and Ralph H. Turner. New York: Basic Books.
1977. "The Three Faces of Social Psychology." Sociometry 40:161-177.
Howell, Susan E. and Shirley B. Laska. 1992. "The Changing Face of the Environmental Coalition: A Research Note." Environment and Behavior 24:134-144.
Huber, Terri and Paul Pederson, 1997. "Meteorological Knowledge and Environmental Ideas in Traditional and Modern Societies: The Case of Tibet." Journalof the Royal Anthropology Institute 3:577-598.
Hughes, Donna E. 1995. "Environmental Sociology: A Distinct Field of Inquiry?" Pp. 61-82 in Environmental Sociology: Th eory and Practice, edited by Michael D. Mehta and Eric Ouellet. ON: Captus Press.
Humphrey, Craig R., Tammy L. Lewis, and Frederick H. Buttel, (eds.). 2003. Environment, Energy, and Society: Exemplary Works. Belmont, CA: Wadsworth Publishing Company . lnglehart, Ronald. 1977. "Long Term Trends in Mass Support for European Unification." Governmentand Opposition 12:150-177.
------. 1990. Culture Shiftin Advanced Industrial Society. Princeton, NJ: Princeton University Press.
------.1995. "Changing Values, Economic Development and Political Change." International Social Science Journal 4 7:379-403.
------. 1997. Modernization and Postmodernization: Cultural, Economic, and Political Change in 43 Societies. Princeton, NJ: Princeton University Pres.
Jaccard, James, Robert Turrisi, and Choi K. Wan. 1991. Interaction Effects in Multiple Regression. Newbury Park, CA: Sage Publications.
Jaeger, William K. 2005. Environmental Economics for Tree Huggers and Other Skeptics. Washington, DC: Island Press.
166 Johnson, Dallas E. 1998. AppliedMultivariate Methods for Data Analysis. Pacific Grove, CA: Brooks/Cole Publishing Company.
Jones, Robert E. 1998. "Black Concern for the Environment: Myth versus Reality." Society and Natural Resources 11 :209-228.
2002. "Blacks Just Don't Care: Unmasking Popular Stereotypes about Concern for the Environment among African-Americans." InternationalJournal of Public Administration 25:221-251.
Jones, Robert E. and Lewis F. Carter. 1994. "Concern for the Environment Among Black Americans: An Assessment of Common Assumptions." Social Science Quarterly 75:560- 579.
Jones, Robert E. and Riley E. Dunlap. 1992. "The Social Bases of Environmental Concern: Have They Changed Over Time?" Rural Sociology 57:28-47.
Jones, Robert E. and Shirley A. Rainey. 2006. "Examining Linkages between Race, Environmental Concern, Health, and Justice in a Highly Polluted Community of Color." Journalof Black Studies 36:4 73-496.
Jones, Robert E. and Aaron Routhe. 2003. "Predicting and Understanding Public Support for Water Supply Proposal in Cumberland County, Tennessee." A Report Prepared for the Waste Management Research and Education Institute, a State Supported Center of Excellence at the University of Tennessee.
Jones, Robert E., J. Mark Fly, and H. Ken Cordell. 1999. "How Green Is My Valley? Tracking Rural and Urban Environmentalism in the Southern Appalachian Ecoregion." Rural Sociology 64:482-499.
Jones, Robert E., J. Mark Fly, James Talley, and H. Ken Cordell. 2003. "Green Migration into Rural America: The New Frontier of ·Environmentalism?" Society and Natural Resources 16:1-18.
Jorgenson, Andrew K. and Edward L. Kick. 2003. "Globalization and the Environment." Journalof World-Systems Research 9:1 95-203.
Kahn, James R. 1998. The Economic Approach to Environmental and Natural Resources. Fort Worth, TX: Dryden Press.
Kanagy, Conrad L. and Fern Willits. 1993. "A "Greening" of Religion? Some Evidence from a Pennsylvania Sample." Social Science Quarterly 74:676-683.
Kanagy, Conrad L., Craig R. Humphrey, and Glenn Firebaugh. 1994. "Surging Environmentalism: Changing Public Opinion or Changing Publics?" Social Science Quarterly 75:804-819.
Kaplan, Steven. 2003. "Human Nature and Environmentally Responsible Behavior." Journalof Social Issues 56:491-508.
Katz, Cindi. 2004. Growing Up Global: Economic Restructuring and Children's Everyday Lives. Minneapolis, MN: University of Minnesota Press.
Kelley, Johnathan and Mariah D. Rupert-Evans. 1995. "Class and Class Conflict in Six Western Nations." American Sociological Review 60:157-178. 167 Kempton, Willett, James S. Boster and Jennifer A. Hartley. 1995. Environmental Values in American Culture. Cambridge, MA: MIT Press.
Kim, Jae 0. and Charles W. Mueller. 1978a. Factor Analysis. Newbury Park, CA: Sage Publications.
1978b. Introduction to Factor Analysis. Newbury Park, CA: Sage Publications.
Klein, Joe. 2002. Th e Natural: The Misunderstood Pre sidency of Bill Clinton. New York: Doubleday.
Kline, Rex B. 1998. Principles and Practice of Structural Equation Modeling. New York: The Guildford Press.
Klineberg, Stephen L., Matthew McKeever, and Bert Rothenbach. 1998. "Demographic Predictors of Environmental Concern: It Does Make a Difference How It's Measured." Social Science Quarterly 79:734-753.
Kluegel, James R. and Eliot R. Smith. 1981. "Beliefs about Stratification." Annual Review of Sociology 7:29-56.
Kowalewski, David and Karen L. Porter. 1992. "Eco-protest Alienation, Deprivation, or Resources?" Social Science Quarterly 73:523-534.
Kraft, Michael E. 1994. "Environmental Gridlock: Searching for Consensus in Congress." Pp. 97-
120nd in En vironmental Policy in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2 ed. Washington D.C.: Congressional Quarterly.
Kraft, Michael E. and Norman J. Vig. 1994. "Environmental Policy From the 1970s to the 1990s:
Continuity and Change." Pp. 3-30 in Envirnd onmental Policy in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2 ed. Washington, DC: Congressional Quarterly Inc.
Kreft, lta and Jan De Leeuw. 2002. Introducing Multilevel Modeling. Thousand Oaks, CA: Sage Publications.
Krieg, Eric J. 1998. ''The Two Faces of Toxic Waste: Trends in the Spread of Environmental Hazards." Sociological Forum 13:3-18.
Lester, James P. 1994. "A New Federalism? Environmental Policy in the States." Pp. 51-68nd in En vironmental Policy in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2 ed. Washington, DC : Congressional Quarterly.
Llewelyn-Davies, Richard. 1972. "The American City through English Eyes." Daedalus 101 :185- 193.
List, Peter C. 1993. Radical Environmentalism: Philosophy and Tactics. Belmont, CA: Wadsworth Publishing Company.
Logan John R. and Harvey L. Molotch. 1987. Urban Fortunes: The Political Economy of Place. Los Angeles, CA: University of California Press.
Lomborg, Bjorn (ed.). 2004. Global Crises, Global Solutions. New York: Cambridge University Press. 168 Long, J. Scott. 1983. ConfirmatoryFactor Analysis. Newbury Park, CA: Sage Publications.
Lorey, David E. 2003. Global Environmental Challenges of the Twenty First Century. Wilmington, DE: SR Books.
Lowe, Philip D. and Wolfgang Rudig. 1986. "Political Ecology and the Social Sciences - The State of the Art." British Journalof Political Science 16:513-550.
Lowe, George D., Thomas K. Pinhey and Michael D. Grimes. 1980. "Public Support for Environmental Protection: New Evidence from National Surveys." Pacific Sociological Review 23:423-445.
Luke, Douglas A. 2004. Multilevel Modeling. Thousand Oaks, CA: Sage Publications.
MacDermid, Robert and H. Michael Stevenson. 1991. Identification with New Social Movements: The Structure of Public Opinion on Environmental Issues. North York, ON: York University Institute for Social Research.
Macnaghten, Phil. 2003. "Embodying the Environment in Everyday Life Practices." The Sociological Review 51:6 3-84.
Macnagten, Phil and John Urry. 1999. "Towards a Sociology of Nature." Sociology 29:203-220.
Manza, Jeff, Michael Hout, and Clem Brooks. 1995. "Class Voting in Capitalist Democracies Since World War II: Dealignment, Realignment, or Trendless Fluctuation?" Annual Review of Sociology 21 :137-162.
Margai, Florence M. 1997. "Analyzing Changes in Waste Reduction Behavior in a Low-Income Urban Community Following a Public Outreach Program." Environment and Behavior 29:769-792.
I Marger, Martin N. 1987. Elites and Masses: An Introduction to Political Sociology. Belmont, CA: Wadsworth Publishing Company.
Marsden, Terry. 2004. "The Quest for Ecological Modernization: Re-Spacing Rural Development Sociologia Ruralis and Agri.-Food Studies." 44:129-146. Marshall, Brent K. 2001. Bridges and Barriers to Ecosystem-Based Approaches: The Case of the Tennessee ValleyAu thority's Adoption of the Wa tershed Approach. Unpublished Dissertation. Knoxville, TN: University of Tennessee.
Maruyama, Geoffrey M. 1998. Basics of Structural Equation Modeling. Thousand Oaks, CA: Sage Publications.
Maslow, Abraham H. 1954. Motivation and Personality. New York: Viking Press.
1962. Toward a Psychology of Being. Chicago, IL: Rand McNalley.
Mazur, Allan and Jinling Lee. 1993. "Sounding the Alarm: Environmental Issues in the U.S. National News." Social Studies in Science 23:681-720.
169 Mazur, Allan. 2003. "Global Environmental Change in the News: 1987-90 Versus 1992-96." Pp. 262-272 in Environment, Energy, and Society: Exemplary Works, edited by Craig R. Humphrey, Tammy L. Lewis, and Frederick H. Buttel. Belmont, CA: Wadsworth.
McCluney, William R. 2004. Humanity's Environmental Future: Making Sense in a Troubled World. Cape Canaveral, FL: Sun Pine Press.
McKechnie, Rosemary B. and Ian Welsh. 2002. "When the Global Meets the Local: Critical Reflections on Reflexive Modernization." Pp. 286-310 in Sociological Th eory and the Environment, edited by Frederick H. Buttel, Peter Dickens, Riley E. Dunlap, and August Gijswijt. Lantham, MD: Rowman and Littlefield.
Meadows, Donella H., Jorgen Randers, and Dennis Meadows. 2004. The Limits to Growth: The 30-Year Up date. White River Junction, VT: Chelsea Green Publishing.
Michelson, Wesley. 1976. Man and His Urban Environment. Reading, MA: Addison-Wesley.
Milbrath, Lester W. 1975. "Environmental Beliefs, Perceptions, and Actions." Th e CornellJournal of Social Relations 10:139-149.
1984. Environmentalists: Vanguard for a New Society. Albany, NY: State University of New York Press.
Mohai, Paul. 1985. "Public Concern and Elite Involvement in Environmental-Conservation Issues." Social Science Quarterly 66:820-838.
1990. "Black Environmentalism." Social Science Quarterly 71 :744-765.
1992. "Men, Women, and the Environment: An Examination of the Gender Gap in Environmental Concern and Activism." Society and Natural Resources 5:1-1 9.
Mohai, Paul and Bunyan Bryant. 1998. "ls There a "Race" Effect on Concern for Environmental Quality?" Public Opinion Quarterly 62:475-505.
Mohai, Paul. and Ben W. Twight. 1987. "Age and Environmentalism: An Elaboration of the Buttel Model Using National Survey Evidence." Social Science Quarterly 68:798-815.
Mol, Arthur P. 1995. The Refinement of Production: Ecological Modernization Th eory and the Chemical Industry. Utrecht, The Netherlands: Van Arkel.
1997. "Ecological Modernization: Industrial Transformations and Environmental Reform." Pp. 138-49 in The InternationalHandbook of Environmental Sociology, edited by Michael Redclift and Graham Woodgate. Northampton, MA: Edward Elgar.
2001 . Globalization and Environmental Reform: The Ecological Modernization. Cambridge, MA: MIT Press.
2003. "Ecological Modernization: Industrial Transformations and Environmental Reform." Pp. 401-411 in Environment, Energy, and Society: Exemplary Works, edited by Craig R. Humphrey, Tammy L. Lewis, and Frederick H. Buttel. Belmont, CA: Wadsworth.
Mol, Arthur P. and David A. Sonnelfield. 2000a. "Ecological Modernization around the World: An Introduction." Environmental Politics 9:3-1 6.
170 {eds.). 2000b. Ecological Modernization: Perspectives and Critical Debates. London, UK: Routledge.
Moore, Keith M. (ed.). 2005. Conflict, Social Capital, and Managing Social Resources. Cambridge, MA: Wallingford.
Murphy, Raymond. 1994. Rationality & Nature: A Sociological Inquiry into a Changing Relationship. Boulder, CO: Westview Press.
Myers, Norman and Jennifer Kent. 2004. The New Consumers: The Influence of Affluence on the Environment. Washington, DC: Island Press.
Namboodiri, Krishnan. 1988. "Ecological Demography: Its Place in Sociology." American Sociological Review 53:619-633.
Newton, Lisa H. 2005. Business Ethics and the Natural En vironment. Malden, MA: Blackwell Publishing.
Nibert, David. 1994. "Animal Rights and Human Social Issues." Society and Animals 2:115-124.
Nord, Mark, Albert E. Luloff and JeffreyC. Bridger. 1998. "The Association of Forest Recreation with Environmentalism." Environment and Behavior 30:235-246.
Nolt, John. 2005. A Land Imperiled: The Declining Health of the SouthernAppalachian Bioregion. Knoxville, TN: University of Tennessee Press.
Norgaard, Richard B. 1997. "A Coevolutionary Environmental Sociology." Pp. 158-168 in The International Handbook of Environmental Sociology, edited by Michael Redclift and Graham Woodgate. Northampton, MA: Edward Elgar.
Norusis, Marija J. 2003. SPSS 12.0 Statistical Procedures Companion. Upper Saddle River, NJ: Prentice Hall.
O'Brien, Robert and Marc Williams. 2004. Global Political Economy: Evolution and Dynamics. New York: Palgrave Macmillan.
O'Connor, James. 1991. "On the Two Contradictions of Capitalism." Capitalism, Nature, Socialism. 2:107-109.
O'Connor, Robert E., Richard J. Bord, and Ann Fisher. 1999. "Risk Perceptions, General Environmental Beliefs, and Willingness to Address Climate Change." Risk Analysis 19:461-471.
Offe, Claus. 1985."New Social Movements: Challenging the Boundaries of Institutional Politics." Social Research 52:817-868.
Ohiorhenuan, John F.E. and Stephen M. Wunker. 1995. Capacity Building Requirements for Global Environmental Protection. Washington DC: Global Environmental Facility.
Olsen, Marvin E. Dora G. Lodwick and Riley E. Dunlap. 1992. Viewing the World Ecologically. Boulder, CO: Westview.
171 Opschoor, Johanes B. 1997. "Industrial Metabolism, Economic Growth, and Institutional Change." Pp. 274-286 in The International Handbook of Environmental Sociology, edited by Michael Redclift and Graham Woodgate. Northampton, MA: Edward Elgar.
Organization for Economic Cooperation and Development. 2004. Environmentally Harmful Subsidies: Challenges for Reform. Paris, FR: OECD.
2005. Environment and the OECD Guidelines for Multinational Enterprises: Corporate Tools and Approaches. Paris, FR: OECD.
Oskamp, Stuart. 2000. "Psychological Contributions to Achieving an Ecologically Sustainable Future for Humanity." Journalof Social Issues 56:373-390.
Outhwaite, William. 1994. Habermas: A Critical Introduction. Oxford, UK: Polity Press.
Pae hike, Robert C. 1994. "Environmental Values and Public Policy." Pp. 349-68nd in Environmental Policy in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2 ed. Washington D.C.: Congressional Quarterly.
Park, Robert E. 1936. "Succession, an Ecological Concept." American Sociological Review 1 :171-179.
Park, Robert E. and Ernest W. Burgess. 1921. Introduction to the Science of Sociology. Chicago, IL: University of Chicago Press.
Partridge, Ernest. (ed). 1981. Responsibilities to Future Generations: Environmental Ethics. Buffalo, NY: Prometheus Books.
Peine, John D. 2005. "Future Prospects." Pp. 303-340 in A Land Imperiled, edited by John Nolt. Knoxville, TN: University of Tennessee Press.
Peine, John D., Robert E. Jones, Mary R. · English, and Samuel E. Wallace. 1999. "Contributions of Sociology to Ecosystems Management." Pp. 61-72 in Integrating Social Sciences with Ecosystems Management, edited by H. Ken Cordell and John C. Bergstrom. Champagne, IL: Sagamore Publishing.
Pellow, David N., Allan Schnaiberg, and Adam S. Weinberg. 2006. "Putting the Ecological Modernization Thesis to the Test: The Promises and Performances of Urban Recycling." Governing Environmental Flows: Global Challenges to Social Theory, edited by Gert Spaargaren, Arthur P. Mol, and Frederick H. Buttel. Cambridge, MA: MIT Press.
Perkin, Harold. 1980. The Rise of Professional Society. New York: Routledge.
Picou, J. Steven. 1999. "Theoretical Trends in Environmental Sociology: Implications for Resource Management in the Modern World." Unpublished paper presented at the Social and Economic Planning Conference, Minerals Management Service. Park City, Utah.
Ponting, Clive. 1991. A Green History of the World. New York: Penguin Books.
Pestone, Moishe. 1999. "Contemporary Historical Transformations: Beyond Postindustrial Theory and Nee-Marxism." Current Perspectives in Social Theory 19:3-53.
Raudenbush, Stephen W. and Anthony S. Bryk. 2002. Hierarchical Linear Models. Thousand Oaks, CA: Sage Publications.
172 Raudenbush, Stephen W., Anthony S. Bryk, Y.F. Cheong, and R. Congdon. 2001. HLM 5: Hierarchical Linear and Nonlinear Modeling. Lincolnwood, IL: Scientific Software International, L.L.C.
Raudsepp, Maaris. 2001. "Some Socio-Demographic and Socio-Psychological Predictors of Environmentalism." Trames 5:355-367.
Redclift, Michael and Graham Woodgate (eds.). 1997. The International Handbook of Environmental Sociology. Northampton, MA: Edward Elgar Publishing Limited.
Renner, Michael and Molly 0. Sheehan (eds.). 2003. Vital Signs 2003. New York: W.W. Norton & Company.
Reser, Joseph P. and Joan M. Bentrupperbaumer. 2000. "Unpackaging Nature and Management Implications of 'Environmental Concern."' Presented at the Eighth International Symposium on Society and Resource Management, Bellingham, WA: Western Washington University.
Riazanov, David. 1973. Karl Marx and Friedrich Engels: An Introduction to TheirLives and Work. New York: Monthly Press Review.
Ridley, Matt and Bobbi S. Low. 1993. "Can Selfishness Save the Environment?" Atlantic Monthly 3:76-86.
Ringquist, Ryan J. 1997. "Equity and the Distribution of Environmental Risk: The Case of TRI Facilities." Social Science Quarterly 78:812-829.
Roberts, J.Timmons. 2003. "Global Inequality and Climate Change." Pp. 254-261 in Environment, Energy, and Society: Exemplary Works, edited by Craig R. Humphrey, Tammy L. Lewis, and Frederick H. Buttel. Belmont, CA: Wadsworth.
Roberts, J.Timmons and Peter E. Grimes. 2002. 'World-System Theory and the Environment: Toward a New Synthesis." Pp. 167-198 in Sociological Theory and the Environment, edited by Frederick H. Buttel, Peter Dickens, Riley E. Dunlap, and August Gijswijt. Lantham, MD: Rowman and Littlefield.
Rohrschneider, Robert. 1988. "Citizens' Attitudes toward Environmental Issues: Selfish or Selfless?" Comparative Political Studies 21 :347�367.
1990. ''The Roots of Public Opinion toward New Social Movements: An Empirical Test of Competing Explanations." American Journalof Political Science 34:1-30.
Rosenbaum, Walter A. 1994. "The Clenched Fist and the Open Hand: Into the 1990s at EPA." Pp. 121-44 in Environmental Policy in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2nd ed. Washington D.C.: Congressional Quarterly.
Ross, Catherine E. and John Mirowsky. 2001 . "Neighborhood Disadvantage, Disorder, and Health." Journalof Health and Social Behavior42:258-276.
Ross, Catherine E., John Mirowsky and Shana Pribesh. 2001. "Powerlessness and the Amplification of Threat: Neighborhood Disadvantage, Disorder, and Mistrust." American Sociological Review 66:568-591 .
Rostow, W.W. 1960. The Stages of Economic Growth. New York: Cambridge University Press. 173 Routhe, Aaron. 2003. Applying Attitude Theory to Environmental Concern and Public Support for Environmental Policy. An Unpublished Master's Thesis. Knoxville, TN: University of Tennessee.
Rubin, David M. 1987. "How the News Media Reported on Three Mile Island and Chernobyl." Journalof Communication 37:42-57.
Ryan, Robert L., Rachel Kaplan, and Robert E. Grese. 2001. "Predicting Volunteer Commitment in Environmental Stewardship Programs." Journalof Environmental Planning and Management 44:629-648.
Sagoff, Mark. 1986. "Values and Preferences." Ethics 96:301-316.
Saha, Robin and Paul Mahal. 2005. "Historical Context and Hazardous Waste Facility Siting: Understanding Temporal Patterns in Michigan." Social Problems 52:618-648.
Samdahl, Diane M. and Robert Robertson. 1989. "Social Determinants of Environmental Concern." Environment and Behavior 21:57 -81.
Sampson, Robert J., Stephen W. Raudenbush, and F. Earls. 1997. "Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy." Science 277:918-924.
Schahn, Joachim and Erwin Holzer. 1990. "Studies of Individual Environmental Concern: The Role of Knowledge, Gender, and Background Variables." Environment and Behavior 22:767-786.
Schnaiberg, Allan. 1975. "Social Syntheses of the Societal-Environmental Dialectic: The Role of Distributional Impacts." Social Science Quarterly 56:5-20.
1980. The Environment: From Surplus to Scarcity. New York: Oxford University Press.
1986. "Reflections on Resistance To Rural Industrialization: Newcomers' Culture of Environmentalism." Pp. 229-258 in Differential Impacts of Rural Resource Development, edited by Phyllis D. Elkind-Savatsky. Boulder, CO: Westview Press.
Schnaiberg, Allan. and K.A. Gould. 1994. Environment and Society: The Enduring Conflict. New York: St. Martin's Press.
Schnaiberg, Allan, David N. Pellow, and A. Weinberg. 2003. "The Treadmill of Production and the Environmental State." Pp. 412-426 in Environment, Energy, and Society: Exemplary Works, edited by Craig R. Humphrey, Tammy L. Lewis, and Frederick H. Buttel. Belmont, CA: Wadsworth.
Schulman, Bruce J. 2002. The Seventies: Th e Great Shiftin American Culture, Society, and Politics. Cambridge, MA: Da Capo Books.
Schultz, P. Wesley. 2000. "Empathizing with Nature: The Effects of Perspective Taking on Concern for Environmental Issues." Journalof Social Issues 56:391-406.
Schultz, P. Wesley and William F. Stone. 1994. "Authoritarianism and Attitudes toward the Environment." Environment and Behavior 26:25-37.
Schumaker, Randall E. and Richard G. Lomax. 1996. A Beginner's Guide to Structural Equation Modeling. Malwah, NJ: Lawrence Erlbaum Associates. 174 Scorecard. 2005. Environmental Justice Data for U. S. States and Counties. Available at: www.scorecard.org. (accessed June 2005).
Scott, David and Fern K. Willits. 1994. "Environmental Attitudes and Behavior: A Pennsylvania Survey." Environment and Behavior 26:239-260.
Seippel, Ornulf. 2002. "Modernity, Politics, and the Environment: A Theoretical Perspective." Pp. 197-229 in Sociological Theory and the Environment, edited by Frederick H. Buttel, Peter Dickens, Riley E. Dunlap, and August Gijswijt. Lantham, MD: Rowman and Littlefield.
Shabecoff, Phillip. 1993. A Fierce Green Fire: The American EnvironmentalMovement. New York: Hill & Wang.
1996. "Greens vs. Congress: A Play-by-Play." The Amicus Journal 18:24-29.
2001. Earth Rising: American Environmentalism in the 21st Century. Washington, D.C.: Island Press.
Shove, Elizabeth. 1997. "Revealing the Invisible: Sociology, Energy, and the Environment." Pp. 261-273 in The International Handbook of En vironmental Sociology, edited by Michael Redclift and Graham Woodgate. Northampton, MA: Edward Elgar.
Shove, Elizabeth and Alan Warde. 2002. "Inconspicuous Consumption: The Study of Consumption, Lifestyles, and the Environment." Pp. 230-251 in Sociological Theory and the En vironment, edited by Frederick H. Buttel, Peter Dickens, Riley E. Dunlap, and August Gijswijt. Lanham, MD: Rowman and Littlefield.
Simmons, James and Nancy Stark. 1993. "Backyard Protest: Emergence, Expansion, and Persistence of a Local Hazardous Waste Controversy." Policy Studies Journal21 :470- 491.
Singer, Judith D. 1998. "Using SAS PROC MIXED to Fit Multilevel Models, Hierarchical Models, and Individual Growth Models." Journalof Educational and Behavioral Statistics 23:323- 355.
Singer, Thomas 0. and Robert Stumberg. 1999. "A Multilateral Agreement on Investment: Would It Undermine Subnational Environmental Protection?" The Journalof Environment and Development. 8:5-23
Snijders, Tom and Roel Bosker. 2002. Multilevel Analysis. Thousand Oaks, CA: Sage Publications.
Snow, David A. 1999. "1998 PSA Presidential Address: The Value of Sociology." Sociological Perspectives 42:1-22.
Soros, George. 2000. Open Society: Reforming Global Capitalism. New York : Public Affairs.
Spaargaren, Gert. 2003. "Sustainable Consumption: A Theoretical and Environmental Policy Perspective." Society and Natural Resources 16:687-701.
Spaargaren, Gert and Bas Van Vliet. 2000. "Lifestyles, Consumption and the Environment: The Ecological Modernization of Domestic Consumption." Environmental Politics 9:50-76.
175 Spaargaren, Gert, Arthur Mol, and Frederick H. Buttel (eds.}. 2006. Governing Environmental Flows: Global Challenges to Social Theory. Cambridge, MA: MIT Press.
Stearns, Peter N., Michael Adas, and Stuart B. Schwartz. 1996. World Civilizations: The Global Experience. New York: Longman Publishers.
Steger, Mary A., John C. Pierce, Brent S. Steel, and Nicholas P. Lovrich. 1989. "Political Culture, Postmaterial Values, and the New Environmental Paradigm: A Comparative Analysis of Canada and the United States." Political Behavior 11 :233-254.
Stern, Paul C. 1992. "Psychological Dimensions of Global Environmental Change." Annual Review of Psychology 43:269-302.
2000. "Toward A Coherent Theory of Environmentally Significant Behavior." Journalof Social Issues 56:407-424.
Stern, Paul C. and Thomas Dietz. 1994. "The Value Basis of Environmental Concern." Journalof Social Issues 50:65-84.
Stern, Paul C., Thomas Dietz, and Gregory A. Guagnano. 1995. "The New Ecological Paradigm in Social-Psychological Context." Environment and Behavior 27:723-743.
1998. "A Brief Inventory of Values." Educational and Psychological Measurement 58:984- 1001.
Stern, Paul C., Thomas Dietz, and Linda Kalaf. 1993. ''Value Orientations, Gender, and Environmental Concern." Environment and Behavior 25:322-348.
Stern, Paul C., Thomas Dietz, Thomas Abel, Gregory A. Guagnano, and Linda Kalaf. 1999. "Value-Belief-Norm Theory of Support for Social Movements: The Case of Environmentalism." Human Ecology Review 6:81-97. ·
Stern, Paul C., Thomas Dietz, Linda Kalof, and Gregory A. Guagnano. 1995. ''Values, Beliefs, and Pro-Environmental Action: Attitude Formation toward Emergent Attitude Objects." Journal of Applied Social Psychology 25: 1611-1636.
Stevenson, Randolph T. 2001. "The Economy and Policy Mood: A Fundamental Dynamic of Democratic Politics?" American Journalof Political Science 45:620-633.
Stimson, James A. 1999. Public Opinion in America. Boulder, CO: Westview Press.
Summers, Gene F. and Kristi Branch. 1984. "Economic Development and Community Social Change." Annual Review of Sociology 10:141-166.
Swidler, Ann. 1986. "Culture in Action: Symbols and Strategies." American Sociological Review 51 :273-286.
Switzer, Jacqueline V. and Gary Bryner. 1998. Environmental Politics: Domestic and Global Dimensions. New York: St. Martin's Press.
Szasz, Andrew. 1986. "The Process and Significance of Political Scandals: A Comparison of Watergate and the "Sewergate" Episode at the Environmental Protection Agency." Social Problems 33:202-217.
176 2003. "A New Mass Issue Is Born." Pp. 280-310 in Environment, Energy, and Society: Exemplary Works, edited by Craig R. Humphrey, Tammy L. Lewis, and Frederick H. Buttel. Belmont, CA: Wadsworth.
Szasz, Andrew and Michael Meuser. 2003. "Environmental Inequalities: Literature Review and Proposals for New Directions in Research and Theory." Pp. 23-38 in Environment, Energy, and Society: Exemplary Works, edited by Craig R. Humphrey, Tammy L. Lewis, and Frederick H. Buttel. Belmont, CA: Wadsworth.
Talley, James. 2001. Demographic Shiftsand Green Values in Rural America: A Southern Appalachian Case. An Unpublishe� Dissertation. Knoxville, TN: University of Tennessee.
Tarrant, Michael A. and H.Ken Cordell. 1997. "The Effect of Respondent Characteristics on General Environmental Attitude-Behavior Correspondence." Environment and Behavior 29:618-637.
Taylor, Dorceta E. 1989. "Blacks and the Environment: Toward an Explanation of the Concern Gap Between Blacks and Whites." Environment and Behavior21 :175-205.
Thomas, Keith. 1983. Man and the Natural World. New York: Pantheon Books.
Tindall, David B. 2004. "Social Movement Participation over Time: An Ego-Network Approach to Micro-Mobilization." Sociological Focus 37:163-184.
Tindall, David B., Scott Davies and Celine Mauboules. 2003. "Activism and Conservation Behavior in an Environmental Movement: The Contradictory Effects of Gender." Society and Natural Resources 16:909-932.
Tremblay, Kenneth R. Jr. and Riley E. Dunlap. 1978. "Rural-Urban Residence and Concern with Environmental Quality: A Replication and Extension." Rural Sociology 43:474-491.
Tucker, Robert C. (ed.) 1978. The Marx-Engels Reader. New York: W.W. Norton & Company, Inc.
United Church of Christ .Commission on Racial Justice. 1987. Toxic Wastes and Race in the United States. New York: United Church of Christ.
Urry, John. 2001. "Sociology of Space and Place." Pp. 3-15 in The Blackwell Companion to Sociology, edited by Judith R. Blau. Malden, MA: Blackwell Publishers.
U.S. Bureau of the Census. 2005. USA Counties. Available at: http://censtats.census.gov/usa/usa.shtml. (accessed June 2005).
U.S. Bureau of Economic Analysis. 2005. State and Local Personal Income Characteristics. Available at: http://www.bea.gov/bea/regional/data.htm. (accessed June 2005).
U.S. Department of Agriculture. 2005. Major Land Uses Data. Available at: http://www.ers.usda.gov/Data/MajorLandUses/. (accessed June 2005).
U.S. General Accounting Office (GAO). 1983. Siting of Hazardous Waste, Landfills, and Their Correlation With Racial and Economic Status of Surrounding Communities. Washington, DC: U.S. General Accounting Office.
177 Van Liere, Kent D. and Riley E. Dunlap. 1980. "The Social Bases of Environmental Concern: A Review of Hypotheses, Explanations and Empirical Evidence." Public Opinion Quarterly 44: 181-197.
1981. "Environmental Concern: Does It Make a Difference How It's Measured?" Environment and Behavior 12:651-676.
Vig, Norman J. 1994. "Presidential Leadership and the Environment: From Reagan and Bush to
Clinton." Pp. 71-96nd in En vironmental Policy in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2 ed. Washington D.C.: Congressional Quarterly.
Vig, Norman J. and Michael E. Kraft (eds.). 1994. Environmental Policy in the 1990s. Washington D.C.: Congressional Quarterly.
Weatherford, Stephen M. 1983. "Economic Voting and the "Symbolic Politics" Argument: A Reinterpretation and Synthesis." Th e American Political Science Review 77:158-174.
1983. "Evaluating Economic Policy: A Contextual Model of the Opinion Formation Process." Th e Journalof Politics 45:866-888.
Weatherford, Stephen M. and Boris Sergeyev. 2000. ''Thinking about Economic Interests: Class and Recession in the New Deal." Political Behavior 22:311-339.
Weaver, Alicia A. 2002. "Determinants of Environmental Attitudes." InternationalJournal of Sociology 32: 77-108.
Weber, Max. 2002. Th e Protestant Ethic and the "Spirit" of Capitalism and Other Writings. New York: Penguin Books.
Wehling, Peter. 2002. "Dynamic Constellations of the Individual, Society, and Nature: Critical Theory and Environmental Sociology." Pp. 1. 44-165 in Sociological Th eory andthe Environment, edited by Frederick H. Buttel, Peter Dickens, Riley E. Dunlap, and August Gijswijt. Lanham, MD: Rowman and Littlefield.
Weinberg, Adam S., David N. Pellow, and Allan Schnaiberg. 2000. Urban Recycling and the Search for Sustainable Community Development. Princeton, NJ: Princeton University Press.
Wenner, Lettie M. 1994. "Environmental Policy in the Courts." Pp. 145-164 innd En vironmental Policy in the 1990s, edited by Michael E. Kraft and Norman J. Vig, 2 ed. Washington D.C.: Congressional Quarterly.
White, Damian F. 2004. "Environmental Sociology and Its Future(s)." Sociology 38:389-397.
White, Lynn. 1967. ''The Historical Roots of Our Environmental Crisis." Science 155:1203-1207.
Wiedner, Helmut. 2002. "Capacity Building for Ecological Modernization." American Behavioral Scientist 45:1340-1368.
Wood, Lawrence E. 2005. Trends in National and Regional Economic Distress: 1960-2000. Appalachian Regional Council Technical Report, available at http://www.arc.gov/index.do?nodeld=57 (accessed June 2005).
178 Woodrum, Eric and Thomas Hoban. 1994. "Theology and Religiosity Effectson Environmentalism." Review of Religious Research 35: 193-206.
Worsley, Peter. 1990. Marx and Marxism. New York: Routledge.
Wright, Eric Olin. 1985. Class. London, UK: Verso.
1997. Class Counts: Comparative Studies in Class Analysis. New York: Cambridge University Press.
Xu, Zhi and David N. Bengston. 1997. 'Trends in National Forest Values Among Forestry Professionals, Environmentalists, an·d the News Media, 1982-1993." Society and Natural Resources 10:43-59. York, Richard and Eugene A. Rosa. 2003. "Key Challenges to Ecological Modernization Theory." Organization and Environment 16:277-288. Zelenzy, Lynette C., Poh-Pheng Chua, and Christina Aldrich. 2000. "Elaborating on Gender Differencesin Environmentalism." Journalof Social Issues 56:443-457.
Zelezny, Lynette C. and P.Wesley Schultz. 2000. "Promoting Environmentalism." Journalof Social Issues 56:365-371.
Zinn, Howard. 2005. A People's History of the United States. New York: HarperCollins Publishers.
Zinn, Howard and Anthony Arnove (eds.). 2004. Voices of a People's History of the United States. New York: Seven Stories Press.
179 VITA
Sean Huss currently holds a Bachelor of Arts degree in s.ociology and Anthropology from the University of Arkansas, Little Rock and a Master of Arts degree in Sociology from the
University of Tennessee. He currently is completing his Ph.D. at the University of Tennessee, with specializations in research methods, social psychology, and statistics. Sean has studied statistical methods at the University of Michigan's Institute for Social Research and International
Consortium for Political and Social Research, as well as at the University of North Carolina. He has worked as a consultant on numerous projects, ranging from qualitative investigations on the dynamics of fraud to quantitative analysis of demographic data. In his consulting work, Sean has worked for the State of Tennessee, the Energy, Environment and Resources Center, the Human
Dimensions Laboratory, and the Centers for Toxicological and Environmental Health. Sean also has conducted research on computer hackers, where he was a subject in a documentary by Tiger
Television in New York. This research on hackers also has served as the basis for subsequent research on computer deviance conducted internationally. More recently, Sean has focused on environmental public opinion and community breakdown leading to drug use. This dissertation is the culmination of his most recent investigation on environmental public opinion, based on initial research conducted at the Energy, Environment and Resources Center and the Human
Dimensions Laboratory. His research on drug use, focusing on the production, distribution and consumption of crystal methamphetamine, has resulted in national recognition. Sean, along with a team of researchers, recently won a Distinguished Service Award from the Office of the
President for his work in finding patterns of community disengagement and increased drug availability and use. Sean also is one of the primary subjects of a documentary (still in production) entitled Temporary Sparrows: Making a Difference in the River Valley. He presently works as an assistant professor at Arkansas Tech University, where he sits on the .academic standards review board. Sean also serves as a member of the River Valley Meth project, where he organized a symposium on the consequences of methamphetamine, has organized several large community events, and sits on the Federal Drug Court review board.
180 58748!rl2. � 11 s:1 M/84/G7 UR it