Paper 4 The Soul of Urban Neighborhoods and its Relation to the City Soul

MARIE H. WAHLSTRÖM

Not yet submitted 1. Introduction

There seems to be a growing interest among planners in identifying and reinforcing features that differentiate their city from others (see for example Lalli, 1982; Zenker et al., 2013). On a local level, it seems increasingly common that neighborhoods are described and branded based on their distinguishing features. This may well be motivated by assumed effects such as increased real estate prices, enhanced residential commitment, and a more prosperous cultural/commercial life. It thereby appears important for planners and developers to understand what makes neighborhoods attractive and what distinguishes them from other neighborhoods in the city.

Researchers within various disciplines study distinctive features of places, from diverse perspectives. While urban planners/designers, geographers, and economists tend to study the physical and functional qualities of places, environmental psychologists and urban sociologists appear to focus on the so called people-place relation. Within marketing, place distinctiveness is primarily studied from a branding perspective, focusing on physical or relational qualities that build up and can help sell the place brand.

The research also varies with regards to geographical scope. Researchers within urban and regional planning, geography, and economics commonly study metropolitan regions, cities or municipalities. The reason for this might be their interest in migration and what makes people move, rather than what makes them stay (Mellander et al., 2011). The community or city scope is quite often used in literature on city branding and urban design, while research within psychology and sociology commonly focuses on neighborhoods, probably due to the assumption that the people-place relation is strongest at this scale of place (Lewicka, 2010; Hernandéz et al., 2007; Devine-Wright, 2013). Researchers within psychology and sociology have also studied the effect of place scale on the people-place bond (Hidalgo and Hernandez, 2001; Hernandez et al., 2007; Lewicka, 2010; Jorgensen and Stedman, 2011; Casakin et al., 2015), suggesting that the relation between the residents and their city is stronger than the attachment with their neighborhood. Still, this literature is limited both regarding the number of studies and their empirical materials. Furthermore, few studies allow for comparisons between different neighborhoods within the same city. They also tend to focus on the people-place relation and do not take notable consideration of the city’s physical and functional characteristics. As urban planners (practitioners) usually are able to influence the physical and functional characteristics, rather than for example the residents’ relation, these characteristics are considered central in this paper.

Wahlström (2017) uses the concept ‘city soul’, as expressed by the residents, to analyze cities’ distinctive features and defines it as a set of physical, functional and emotional characteristics that influence the residents’ relation to their city as well as their perceptions of its distinctiveness. ‘Public services’ and ‘accessible nature’ are examples of characteristics, and sense of ‘belonging’ and feelings of ‘pride’ examples of relational aspects. Applying this definition, the overall aim of this study is to compare the residents’ perception of their city’s soul with the soul of their neighborhood, to identify potential similarities and differences between these two scales of place.

This study utilizes data from a quantitative survey of 6,600 residents (2,573 respondents) in four Swedish cities. Specifically, 1,300 respondents living in Stockholm (capital of Sweden) will be used to analyze the strength of the city soul vis-à-vis the soul of neighborhoods. The results show that both are significantly related to the corresponding people-place relations and the physical/functional characteristics. Regardless of spatial scale, the residents’ assessments of the soul seem to be stronger influenced by the people-place relation than by their assessments of the characteristics (though the characteristics’ influence is both direct and indirect, via the relation). In 3 particular, the relative importance of the relation is observed for residents that were born in the city or moved there a long time ago.

The neighborhood-specific analyses show that the Stockholm districts differ in several respects – differences that are not equalized by controlling for the district-specific socio-demographic mix – which should be especially interesting from a practical planning perspective. If a district ranks substantially lower than others regarding, for example, ‘Nice routes for cycling/walking’ or ‘Visible art in open places’ it provides the planners with a strong signal to look closer into the corresponding issues.

One contribution of this paper is the suggested conceptual model, which adds to the mainly psychological-sociological literature, in which few studies seem to be theory-driven. Furthermore, the paper pays considerable attention to the physical nature of places, which is another shortage in this literature (Lewicka, 2011b; Zenker & Petersen, 2014). Finally, the paper suggests a methodology that aims at explaining and analyzing neighborhood soul in a systematic, quantitative manner and in comparison with the soul of the city, which together might make the methodology more useful to research within economics, geography as well as urban and regional planning.

After presenting a summary of relevant literature, the methodology is described together with the research questions. The subsequent sections present the empirical analysis, the results, and finally a discussion and some concluding remarks.

2. Literature Review

Studies of city and place distinctiveness are found within different disciplines resulting in a wide variety of concepts, models, and methods. Some of the most frequently used concepts are 'place identity', 'city image', 'sense-of-place' and (more to describe attractiveness in general) 'quality-of- life', which are summarized below (see Wahlström, 2017, for a comprehensive summary).

Place identity has been heavily discussed within environmental psychology, sociology, urban planning, urban design, geography, and branding, often aiming for a better understanding of the relation between people and place. It has also been interpreted and defined in numerous ways; e.g., in terms of identification (as part of individual or group identity, or of a group with a certain territory), the identity of the place itself (distinctive features) or as sense-of-place (the emotional experience of a place). Based on these interpretations place identity is understood as comprised of physical features, functions, meanings and symbols, and as created from both external image and internal culture. (Relph, 1976; Proshansky, 1983; Southworth and Ruggeri, 2011: Kalandides, 2011; Zenker and Pedersen, 2014) According to Wahlström (2017), these interpretations appear to rely on physical, functional as well as emotional factors.

City image is a concept found mainly in marketing literature, though it was an urban planner/architect that introduced the concept over 50 years ago. Marketing studies often aim to understand how cities are perceived, in order to improve their branding. Image, which is seen as external, is sometimes believed to create identity together with the internal culture. However, from a planning perspective city image has been related to perceptions, emotions, culture, history, as well as to the natural and built environment. (Lynch, 1960; Kavaratzis and Hatch, 2013; Anholt, 2010). According to Wahlström (2017), these interpretations mostly rely on physical and emotional factors.

Sense-of-place often appears in literature within anthropology, environmental psychology, architecture, geography, and urban sociology. As with place identity, sense-of-place studies commonly aim for a better understanding of the human-place relationship. Sense-of-place is 4 interpreted in various ways, e.g. as one aspect of place identity, as a summary evaluation of place identity, place attachment, and place dependence, as the spirit of place or as the unique/truly authentic features of a place (Relph, 1976; Bell et al, 2001; Jivén and Larkham, 2003; Jorgensen and Stedman, 2006; Kalandides, 2011). According to Wahlström (2017), these interpretations predominantly rely on emotional factors.

Quality-of-life is mostly found in literature within economics and geography, with the main aim to study city/region attractiveness and migration patterns. It is often interpreted as a set of functional factors, such as job outlook, educational opportunities, health care, and crime rates or as place- based amenities, such as physical setting, consumer services, and life style. A related concept is ‘place surplus’ that links city attractiveness to the economic, human, cultural and social capital. (Power, 2005; Whisler, 2008; Florida, 2008; Florida et al., 2011, 2013; Mellander et al., 2011; Westlund, 2006). According to Wahlström (2017), both physical and functional aspects are central in these understandings of quality-of-life.

City soul (or the soul of the city) is a concept less frequently found in literature related to city distinctiveness, though it occurs in some pieces of work within economics and geography. In line with the quality-of-life literature, these studies often aim to achieve a better understanding of city attractiveness and migration patterns, but also of city branding. These studies interpret city soul as a part of the city’s cultural capital, as a description of community belonging, as the city atmosphere/buzz or as the truly authentic features of a place. (Andersson, 1998; Florida, 2009, 2012; Zukin, 2011). According to Wahlström (2017), city soul - though being a rare concept within the literature connected to distinct features of cities and places - has the advantage of taking physical, functional as well as emotional aspects into account. As this is also observed for place identity, the combined concept ‘city identity/soul’ is used by Wahlström (2017) to further explore the distinct features of cities.

Except for Wahlström (2017), few (if any) previous studies link relational-emotional with functional-physical factors, to explore an intermediary concept like the ‘soul’ of a place (with the interpretation made here), see figure 1. Especially when it comes to comparing different scales of place, such as comparing the city and the neighborhood.

Figure 1. Illustration of this paper’s approach; linking factors used within different strands of literature to explore the distinctiveness of cities and neighborhoods, through the concept ‘soul’ of the city and its neighborhoods.

There are some examples of quantitative studies linking places’ physical setting to the residents’ social and psychological processes. Lalli (1982) compares the neighborhood and city scale concerning the residents’ perceived relation to and evaluation of these places in terms of their perceived uniqueness and special character. One conclusion is that each spatial scale level needs its 5 own theoretical account and specific empirical operationalization. In a more recent study, Stedman (2002) presents and test a model for sense-of-place, based on both perceptions regarding the place’s physical characteristics and the relation to the place. This study suggests that the relation to place (attachment and satisfaction), as well as the place’s symbolic meanings, influence the residents’ willingness to engage in maintaining or improving the area, which is a lake area, i.e., neither a city nor a neighborhood. Another example is Permentier et al. (2011) who uses both stated and revealed assessments of neighborhood attributes together with social and socio-economic variables, to analyze neighborhood satisfaction and reputation. Brown and Raymond (2007) combine regression and spatial cross-correlation to show that aesthetic, recreational, economical, spiritual and therapeutic values are likely to contribute to place attachment. They surveyed 1,400 residents and 500 visitors in the Ottawa region, i.e., neither a city nor a neighborhood. Westford (2010) study neighborhood design and travel in four Swedish sub-urban neighborhoods. The survey covers approximately 450 parents and 750 children, and by means of logistic regression she analyzes the connections between neighborhood design, parents’ satisfaction with urban qualities and children’s behavior.

In line with the aim of my study, the remainder of this review mainly covers literature on how residents’ relation to a place, or their opinions regarding its characteristics, differ with regards to the scale of that place. In this strand of literature, which is mainly found within environmental psychology and sociology, the central concepts for the people-place relation seem to be ‘place identity’ and ‘place attachment’ (see Lalli, 1982; Casakin, 2015; Hernandez et al., 2007; Westin, 2016; Knez, 2005; Anton and Lawrence, 2014; Lewicka, 2008; Bernardo and Palma-Oliviera, 2016). It is important to note that the concept ‘place identity’ in these studies, with almost no exceptions, refers to the residents’ identification with a certain territory, e.g., as a SOHO-resident or a Londoner, a Brooklynite or a New Yorker. Sometimes it refers to a broader range of aspects of the people-place relation, but there seem to be few examples of place identity in the meaning of place distinctiveness, i.e., the identity of the place itself.

Place attachment is another concept commonly used to analyze the relation to place. It is usually assessed through multiple items, e.g., belonging, loyalty, and satisfaction. See for example Casakin (2015); Hernandez et al. (2007); Hidalgo and Hernandez (2001); Zenker and Rütter (2014); Westin (2016); Bonauito et al. (1999, 2003, 2006); Stedman (2002); Knez (2005); Anton and Lawrence (2014); Brown et al. (2003); Lewicka (2011a). Sometimes place identity, in terms of identification with place, is part of place attachment (e.g., Brown and Raymond, 2007).

Quite a few studies suggest that the relation differs with regards to the scale of place (Hidalgo & Hernandez, 2001; Hernandez et al., 2007; Lewicka, 2010; Jorgensen & Stedman, 2011; Casakin et al., 2015). These studies indicate that the relation between the residents and their city is stronger than that with their neighborhood. A variety of statistical analytical methods are used, e.g., ANOVA, correlation, reliability, structural equation modeling, path analysis and regression.

A recent, relatively large study (Westin, 2016) analyzes over 3,000 residents’ attachment to their neighborhood, municipality, and region. Using logistic regression, the study found a relationship between the residents’ mobility and attachment to the neighborhood and municipality, but not with the city region. Social life seems to have a significant influence on attachment to the neighborhood only, while residential history (time living in the city) influences the attachment to all three scales of place. Many studies (see Casakin, 2015; Hernandez et al., 2007; Anton and Lawrence, 2014; Brown et al., 2003; Bonaiuto et al., 1999, Knez, 2005) also support the link between residential time and the relation to place. A couple of studies indicate a link between relation to place and citizen behavior, e.g., between place attachment and citizen involvement (Zenker and Rütter, 2014; Stedman, 2002). 6

Several studies consider the place’s physical characteristics. Bonauito et al. (1999, 2003) show that perceived residential environment qualities influence residential satisfaction. Brown and Raymond (2007) investigate the relationships between perceived qualities and place attachment, Permentier et al. (2010) study the link between urban amenities and residential satisfaction as well as reputation, Stedman (2002) focuses on the sense-of-place and place attachment while Jorgensen and Stedman (2006) focus on sense-of-place.

3. Methodology and Research Questions

This section presents the paper’s central concepts ‘city soul’ and ‘neighborhood soul’. It further explains these concepts’ relationship to urban characteristics and aspects of the relation that residents have with their location of residence. Finally, the research questions are presented, and the questionnaire design described.

3.1 Central concepts

When defining the soul of the neighborhood, this paper make use of the concept city soul, as defined by Wahlström (2017), i.e. as “a set of physical, functional and/or emotional characteristics that influence the residents’ relation to their city as well as their perceptions of its distinctiveness” (p 43).1

Analogously, this paper will assume that the soul of an urban neighborhood refers to a set of characteristics that influences the residents’ relation to their neighborhood as well as their perceptions of its distinctiveness. It is further assumed that the soul of the neighborhood can be described in terms of content as well as assessed in terms of strength.

It is important to study the soul of the neighborhood because it might, for example, be assumed that residents generally have more information and awareness regarding the characteristics that define the vicinity (e.g., neighborhood) than the characteristics of a broader area (e.g., city). Additionally, the residents’ social capital, mobility and residential history could influence their perception of the neighborhood soul more. The idea of an urban amenity premium, first presented by Glaeser et al. (2001) and later discussed by Florida and Mellander (2010), who imply that neighborhoods have different identity and characteristics depending on their social mix. According to them, artists are for example attracted to authenticity and aesthetics and will, therefore, be more likely to reside in areas with such amenities. At the same time, artists are producers of culture and, as a consequence, their neighborhood will probably have more cultural amenities than other neighborhoods. Thus, if the population mix varies between different parts of the city, the urban amenity premium will probably differ as well. Assessments of this premium will thereby not be the same for a specific neighborhood as for the entire city. From this line of reasoning, it is reasonable to assume that the perceptions of the place identity/soul will not be the same for the city as for the neighborhoods within that city.

1 The criterion that it should take physical, functional as well as emotional aspects into account is assumed to be important for urban planners as their role increasingly involves understanding how people perceive and use places. As both ‘place identity’ and ‘city soul’ meet this criterion, as respondents seemed to have a strong willingness to describe the ‘soul of their city’ but tended to mix up ‘identity’ with ‘image’, and as identity seems to be more quantifiable than soul, a combination of the two concepts is used. (Wahlström, 2017) 7

The research presented in this paper is based on a survey of residents in four Swedish cities. The respondents were asked to evaluate both their ‘city’ and their ‘city district’.

‘City’ was not explicitly defined in the questionnaire, but it seems plausible that the most common frame of reference is the city, rather that the inner city or the Stockholm region. The use of ‘city district’ was explained as ‘the area where you live’ and was chosen instead of ‘neighborhood’ for a number of practical reasons.2

3.2 Operationalizing the concept

3.2.1 Conceptual Model and Research Questions

Based on the definition of city soul, Wahlström (2017) suggests a conceptual model in which both the resident’s relation to their city and their perceptions of its characteristics influence their perceptions of the city soul. The influence of city characteristics can be direct and/or indirect, as the characteristics also influence the relation.3 Figure 2 outlines the model adjusted to fit the neighborhood perspective applied in this paper. By way of example, residents’ socio-demographics are assumed to influence both the residents’ perceptions of neighborhood characteristics and their relation to the neighborhood.

Through the following research questions, this paper investigates the concept neighborhood soul and its relation to the soul of the city:

1 Is the conceptual city soul model applicable on neighborhoods? In what way is it different from the city soul model?

2 Which characteristics and/or aspects of the relation to the neighborhood influence the soul of the neighborhood? In what way are they different from those influencing the city soul?

3 How do the Stockholm city districts differ with respect to neighborhood relations, characteristics, and soul?

3.2.2 Design of the Survey and Questionnaire

The empirical material used in this paper is based on a quantitative self-report survey among residents in Swedish cities. As illustrated by figure 2, the residents’ perceptions of the neighborhood soul are assessed through two different types of questions. First, through an open-ended question,

2 There are several explanations behind using the term ‘city district’ in the questionnaire. The questions were written in Swedish, which does not seem to have a feasible/operational translation of the English word ‘neighborhood’. Additionally, as discussed by Lewicka (2010), the concept of neighborhood is very ambiguous, and it is difficult to know what respondents have in mind when they are asked to value different qualities of their ‘neighborhood’ – it can be a street or a block or a larger residential area. Therefore, it might be easier to interpret the results if the examined space is pre-defined, such as a town district. In this study, at least one of the cities (Stockholm) has city districts with borders that probably are more or less familiar to most city residents. Finally, when a quantitative survey, it is preferred to use sufficiently large areas to enable comparisons between results regarding the city and results regarding parts of that city. Using too small parts can be associated with data collection becoming too time and budget consuming. 3 A third component is covered by the survey – so called ‘associations’ based on short descriptions of features that can be associated with a city/neighborhood (e.g., open, green, authentic). However, as initial analysis shows that their relationship with city/neighborhood soul is rather weak, the associations have not been analyzed in this paper. 8 where the respondents are asked to describe the soul of the neighborhood in their own words. Here, the wording ‘neighborhood soul’ is used instead of ‘neighborhood identity/soul’. Secondly, the strength of the neighborhood soul is assessed as the respondents are asked to evaluate the strength on a Likert scale. Here, the term ‘identity/soul’ is used in order to facilitate quantification.

The assessments of neighborhood characteristics are derived from statements relating to physical and functional features, but emotional aspects are included as well. The types of characteristics included are illustrated by figure 2. Only positive statements regarding characteristics are included4, and the respondents answered using a Likert scale from 1 (‘I don’t agree at all’) to 9 (‘I totally agree’). They could also choose the response alternative ‘I don’t know/not relevant’. Aspects of the respondents’ relation to their neighborhood were assessed through seven statements, e.g., “I am overall satisfied with living in this neighborhood” using the same type of Likert scale. The types of relational aspects that these statements refer to are illustrated by figure 2. All statements regarding the neighborhood are presented in table 2.

The same types of questions are asked about their perceptions of the city soul, characteristics, and relation. Detailed information regarding the survey can be retrieved from the author upon request.

4 It was a strategic decision to include only positive statements, in order to keep the questionnaire short and still be able to include questions on both the neighborhood and the city. 9

Figure 2. Illustration of how the concept neighborhood soul is operationalized through the questionnaire. Aspects of the relation are represented by seven items (statements), while the characteristics are represented by twelve groups, each reflecting up to five specific questions (see table 2 for the full list of all 33 questions).

The questionnaire further contains questions about the respondent. As illustrated by figure 2 these questions aim to describe the ‘socio-demographics’ of the residents regarding gender, age, income, etc. As socio-demographics are considered difficult to influence for an urban planner these variables are mainly used as control variables in the analyses.

After initially reviewing relevant literature, a pilot questionnaire was constructed and tested through a pre-study of 1,600 residents in four Swedish cities. Open-ended questions were used to understand how the soul of the city as well as of the neighborhood is perceived and described. The main study was developed based on learnings from the pilot. 10

The main survey, which is reported in this paper, was conducted during fall 2014. 6,600 postal paper questionnaires were distributed to a stratified random sample of residents in the Swedish cities Stockholm, Göteborg, Malmö, and Umeå. The cities are of different size, located in different parts of the country and they differ regarding history, culture, and character.5 A total of 2,573 questionnaires were completed, corresponding to a response rate of 40 %. The city district analyses are based on the Stockholm sample, where there are 1,300 respondents to this survey and a sufficient number of responses per city district. All respondents live in one of the five city districts defined as the inner city; together they account for 36% of the total city population. The districts are (Swedish name in brackets): Downtown (Norrmalm), South (Södermalm), East (Östermalm), Northwest (Vasastan), and West (Kungsholmen). Of these, Stockholm South has the largest residential population, while Stockholm Downtown is the most office-dense district with the smallest population.

Being inner city districts, the districts are relatively similar regarding the natural surroundings, type of development and real estate prices. Still, there are differences in lifestyle, commercial services, recreational space/facilities, etc., which makes them interesting to compare. As they are part of the same city, the potential differences between them cannot easily be explained by for example different job outlooks and housing markets.

4. Results

The survey results are presented in three sections. After an overview of descriptive statistics for key results, the research questions are dealt with through different types of analyses.

4.1 An Overview

The open ended question “Please, describe the soul of your neighborhood in your own words” shows that the respondents’ perceptions about their neighborhood’s soul are more or less in line with the definition used here. Thus, the descriptions are based on both their relation to the neighborhood and their perceptions of its characteristics. Some examples are given to illustrate this (the quotes have been translated into English by the author):

“Many beautiful houses…lovely parks…. In short, I feel comfortable and at home here.”

“Has everything I need within reach, there is a pride in living here, creating a sense of community and safety.”

“A peaceful, green island with thriving center and traffic. Biking and walking routes alongside the shore line.”

The residents’ descriptions of their neighborhood’s soul differ from their descriptions of the city soul in some aspects. It is, for example, twice as common to mention ‘people’ as well as ‘trendy’ when

5 Stockholm (912,000 inhabitants) is located on the east coast and Göteborg (541,000 inhabitants) on the west coast, Malmö (318,000 inhabitants) in the far south, and Umeå (120,000 inhabitants) in the far north of Sweden. Stockholm is the capital and often described as the knowledge center, while Malmö is part of a larger region that includes Copenhagen, Denmark. Both Malmö and Göteborg are old industrial cities. Umeå is the smallest of the four and today a popular student city. Malmö has the youngest citizens and the highest share of immigrants. 11 describing the soul of the neighborhood and it is far more common to mention ‘water’, ‘beautiful’, and ‘buildings’ when describing the soul of the city.

Overall, the surveyed residents seem to perceive both their neighborhood and their city to have a rather strong soul. According to table 1, the average scores vary between 6.4 and 7.7 on the 9-point Likert scale. The average scores for the total data (all four cities) show that the city soul is perceived as stronger than the soul of the neighborhood, even though the differences are small (especially in Stockholm).

The districts show smaller variation between the scores for the city than for the neighborhood soul. The difference between the scores shows no clear pattern. Those living in South (Södermalm) perceive the neighborhood soul to be stronger than the city soul, but the opposite holds for those living in West (Kungsholmen). The difference is not significant for the other three districts.

Table 1. Average scores on a) “My neighborhood has a strong identity/soul” and b) “My city has a strong identity/soul”. Likert scale 1-9, where 1=I don’t agree at all and 9=I totally agree. “I don’t know / not relevant” is treated as missing values. N=unweighted no. of respondents. Weighted results are reported for the total (all four cities), unweighted results for Stockholm and Stockholm’s city districts (Swedish names in brackets). a) Strength of b) Strength of N Significant difference neighborhood soul city soul between a) and b) according to t-test (95%) Total (weighted) 6.4 6.9 2573 Yes Stockholm 7.2 7.3 1311 No Stockholm’s city districts: East (Östermalm) 7.4 7.4 238 No South (Södermalm) 7.7 7.3 481 Yes Downtown (Norrmalm) 6.7 7.3 113 No West (Kungsholmen) 6.7 7.4 273 Yes Northwest (Vasastan) 7.1 7.3 204 No

The smaller variation between the city scores, than between the neighborhood scores, is in line with what can be expected since the districts belong to the same city. It is more difficult to explain the difference between neighborhood scores. However, it should be said that the high neighborhood score for South (Södermalm) is supported by anecdotal evidence about the strong feelings those living there have for their district and for ‘their’ football team (Hammarby).

According to the conceptual model outlined by figure 2, the perceived strength of the neighborhood soul is influenced by perceptions of neighborhood relations and characteristics. Table 2 presents the average scores for all relational aspects as well as all physical and functional characteristics - per city district, for Stockholm city, and for the total dataset. The upper part of the table shows that the relation to the neighborhood does not differ significantly between Stockholm’s city districts. In essence, ‘Stockholmers’ seem to feel more or less the same level of loyalty, pride, attachment, etc. to their neighborhood, regardless of which neighborhood they live in. However, supporting the anecdotal evidence just mentioned, it is notable that South (Södermalm) receives the highest scores on six out of seven relational aspects. The residents in this district seem to have the strongest relations to their neighborhood as well as the strongest perceptions of its soul. 12

Table 2. Mean scores for the neighborhood soul, aspects of the neighborhood relation, and neighborhood characteristics (all measured on a Likert scale 1-9). “I don’t know / not relevant” is treated as missing values. Weighted results are reported for the total (all four cities), unweighted results for Stockholm and Stockholm’s city districts (Swedish names in brackets). Down- North- East South town West west Total Stock- (Öster- (Söder- (Norr- (Kungs- (Vasa- (weighted) holm malm) malm) malm) holmen) stan) Neighborhood Soul: My neighborhood has a strong identity/soul 6.4 7.24 7.4 7.67 6.73 6.67 7.1 Aspects of the Neighborhood Relation: Attachment: I feel attached to this neighborhood, it is part of me 6.66 7.22 7.14 7.35 6.9 7.15 7.31 I feel I belong in my neighborhood 7.01 7.51 7.49 7.61 7.21 7.43 7.57 I am proud of living in my neighborhood 6.89 7.51 7.42 7.61 7.2 7.44 7.68 I'd rather not move from this neighborhood 6.82 7.32 7.34 7.47 7.15 7.13 7.25 I identify with the people living in my neighborhood 5.89 6.16 6.13 6.36 5.86 6.13 5.93 In total, I am very satisfied with my neighborhood 7.06 7.55 7.51 7.66 7.3 7.44 7.59 I would definitely recommend a friend/relative to move to my neighborhood 7.31 7.71 7.69 7.82 7.33 7.7 7.68 Neighborhood Characteristics: Public services Good child care and schools 6.57 6.86 7.13 6.61 7.05 6.89 7.04 Good health care 6.18 6.65 6.75 6.43 6.52 6.98 6.74 Good supply of facilities for sports etc. 6.21 6.33 6.67 6.45 5.93 6.31 5.91 Safety and security 6.52 6.85 7.18 6.7 6.75 6.76 6.98 Good reputation among people not living here 6.81 7.45 7.39 7.51 7.55 7.29 7.56 Commercial services: Good supply of shopping 5.58 6.87 6.69 7.29 7.09 6.29 6.78 Good supply of culture 4.39 5.97 6.48 6.73 6.78 3.87 5.99 Good supply of restaurants/cafés/entertainment 5.24 7.39 6.83 7.82 7.99 6.64 7.78 Offers exciting experiences 4.85 6.21 6.31 6.57 6.65 5.41 6.04 Full of life, a 24-7 neighborhood 3.76 5.62 5.08 6.17 6.25 4.9 5.62 Good universities nearby 5.92 6.55 7.73 5.96 7.41 5.35 7.36 Meeting places: Access to nice places for meeting people 6.21 7.36 7.56 7.46 7.3 7.07 7.3 Access to cozy parks 7.05 7.69 8.18 7.67 6.92 7.66 7.62 Image and stories: A story I like to tell others 5.7 6.74 6.79 7.01 7.02 6.6 6.08 A distinct image 5.71 6.9 7.15 7.61 6.74 5.75 6.43 Diversity: Citizens from many different cultures and countries 6.02 5.58 5.45 5.71 5.74 5.84 4.99 A good supply of religious institutions 5.68 6.78 7.08 6.84 6.84 6.41 6.63 Getting around: Easy to get to/around - by car 6.66 5.4 5.47 5.36 4.49 5.88 5.24 Easy to get to/around - by public transit 6.79 7.48 7.26 7.47 7.41 7.52 7.75 Easy to get to/around - by bicycle 7.88 7.43 7.33 7.47 7.22 7.64 7.26 13

Total Stock- Down- North- Table A1 (continued) (weighted) holm East South town West west Walkability: Nice to walk around in 7.62 8.12 8.15 8.25 7.88 8.12 7.88 Well integrated part of the city 6.62 7.81 7.94 7.7 8.18 7.6 7.97 Nice routes for walking/cycling 7.41 7.91 8.07 8.12 7.3 8.14 7.2 Housing: Good housing at reasonable price 3.55 2.46 2.53 2.43 2.24 2.69 2.25 Sustainable housing, e.g. "green" buildings 3.27 3.28 3.81 3.7 2.69 3.06 2.31 Beauty: Beautiful neighborhood 6.97 7.89 8.16 8 7.67 7.73 7.65 Beautiful houses, buildings 6.07 7.55 7.83 7.53 7.72 7.11 7.78 Attractions, places with a view, etc. I gladly recommend 6.21 7.44 7.51 7.95 7.2 7.27 6.54 Natural environment: Accessible larger green parks or green belt 7.11 7.07 8.44 6.99 6.13 6.49 6.92 Accessible, nice places close to the water 6.81 7.95 7.98 8.37 7.04 8.48 6.76 Symbolic features: At least one building that is strongly associated with the neighborhood 6.19 7.61 7.76 7.36 7.8 7.99 7.44 Statues, signs, symbols that are associated with the neighborhood 4.72 6.25 6.57 6.29 6.74 5.85 6.06 Visible arts in open places, streets, on buildings, etc. 4.56 6.06 6.33 6.47 6.36 5.58 5.29

The variation across neighborhoods is, in general, larger for the assessments of characteristics, for example, the score for ‘Good supply of culture’ is 3.9 for West (Kungsholmen) and 6.8 for Downtown (Norrmalm) and ‘Accessible, nice places close to the water’ ranges between 6.8 for Northwest (Vasastan) and 8.4 for South (Södermalm). As should be expected, the variation is lower for different kinds of public services. It is also interesting to note the low level of scores concerning ‘Good housing at reasonable price’ and ‘Sustainable housing.’

Figure 3 provides a simple overview of the average scores for groups of characteristics, both for the neighborhood and the city level. According to figure 3a, there are differences between the perceptions of the city and neighborhood characteristics, primarily regarding the categories ‘commercial services’, ‘diversity’, and ‘symbolic features’. For all of these, the city scores are higher than the neighborhood scores. Perceptions of the neighborhood’s characteristics also differ between the districts of Stockholm, see fig. 3b. 14

Figure 3a-b. Average scores on groups of city and neighborhood characteristics. Likert scale 1-9, where 1=I don’t agree at all and 9=I totally agree. “I don’t know / not relevant” is treated as missing values. Results for a) Stockholm and b) Stockholm’s city districts. *) N.B. ‘Aesthetics’ includes 'cozy neighborhoods' for the city, not for the neighborhood and ‘Walkability’ includes 'well-integrated part of the city' for the neighborhood, not for the city.

One reason for the districts of Stockholm to differ regarding perceptions of their neighborhood characteristics could be that they differ regarding the socio-demographic mix of residents. According to the appendix, table A1, the districts seem to differ slightly from each other regarding their socio-demographic profiles. By way of example, the residents in East (Östermalm) appear to be somewhat older, richer and more educated while the residents in West (Kungsholmen) to a greater extent are younger and work full-time.

4.2 Is the city soul model applicable on neighborhoods?

In this section, the assumed relationships between the neighborhood soul and the two components - characteristics and relation - are analyzed through bivariate correlations and linear regressions. The analyses are done for the total dataset, for Stockholm city, and for Stockholm’s city districts. A ‘relation index’ is used, based on the average score of the seven relational aspects. Similarly, the average score of all 33 neighborhood characteristics is used. The strength of these relationships is initially analyzed through bivariate correlations. The correlations are significant, and the coefficients are about the same for the city and the neighborhood, which imply that the model is equally applicable on both scales. The strongest link in the model is the one between the soul and the relation index (0.72 and 0.77 for city and neighborhood respectively). The analysis further indicates a substantial indirect influence of the characteristics, implying that the more the residents appreciate their neighborhood’s characteristics, the stronger is their relation to the neighborhood and the stronger do they perceive its soul.

According to table 3, the correlation results are supported by regression analysis. Using the total dataset, a linear regression with neighborhood soul as dependent variable and the relation index as well as the mean score on neighborhood characteristics as independents, indicates a much stronger quantitative influence of the relation index than of the characteristics. As reported by table 3, column 2, the results are similar when running the regression with city soul as the dependent variable. 15

Table 3. Results from OLS regression on the total sample, using 1) “My neighborhood has a strong identity/soul” respectively 2) “My city has a strong identity/soul’ as the dependent variable. Standardized beta coefficients, adjusted R-square, and N= total number of respondents (unweighted). 1) Dependent: 2) Dependent: Neighborhood soul City soul Relation Index 0.607 0.586 Mean score on characteristics 0.245 0.202 R-square adjusted 0.602 0.533 N 2,573 2,573 p≤0.05, if not otherwise indicated

Table 4.Results from OLS regression on Stockholm city (column 1), as well as Stockholm’s city districts (column 2-6), using “My neighborhood has a strong identity/soul” as the dependent variable. Standardized beta coefficients, adjusted R-square, and N= total number of respondents (unweighted). (1) Stock- (2) (3) (4) (5) (6) holm city East South Downtown West Northwest Relation Index 0.587 0.637 0.550 0.667 0.623 0.665 Mean score on characteristics 0.267 0.157 0.297 0.189 0.272 0.202 R-square adjusted 0.589 0.533 0.575 0.606 0.612 0.667 N 1,300 238 481 113 273 204 p≤0.05, if not otherwise indicated

According to table 4, column 1, the same pattern is observed for the Stockholm sample: the influence of the relation index is more than twice as strong as of the characteristics. However, there seem to be some differences between the districts of Stockholm (see table 4, column 2-6). South (Södermalm) shows a relatively stronger quantitative impact from the characteristics and a relatively weaker impact from the relation index, compared to the other districts. Does this imply that to the residents in South (Södermalm), the neighborhood characteristics are more important to their perceptions of the strength of the neighborhood’s soul, as compared to the residents in the other districts? And is the opposite true for East (Östermalm), in which the influence of the characteristics is the weakest? As this result might also be explained by the districts having diverse socio-demographic mix of residents, the next step is to run the regressions while at the same time controlling for the residents’ age, income, education, occupation and time in the city (natives/non- natives). The results, reported in the appendix, table A2, indicate that South (Södermalm) have a stronger relative influence from characteristics, compared to the other districts, even when controlling for residents’ socio-demographics. Overall, the control variables have small or no significant effect on the perceived strength of the neighborhood soul, except for age in Northwest (Vasastan) and education in West (Kungsholmen).

The same regression is also run for sub-sets of native, young, high-income, high-educated respectively working residents. The results from these regressions are reported in the appendix, table A3. These analyses imply that the relation has an even stronger influence on the neighborhood soul among natives than among non-natives, while it seems to have a weaker influence on young residents, residents working full-time or part-time, and high-income residents.

All in all, the correlation and regression analyses imply that the conceptual model should be applicable at the neighborhood as well as the city scale.

4.3 Which characteristics and aspects of the relation influence the soul of the neighborhood?

This section will deepen the analysis presented in table 3 by relating the scores for the city and neighborhood soul to individual characteristics and relational aspects. However, since the 33 characteristics and the seven relational aspects are likely to be inter-correlated, it is necessary to 16 first sort away the variables that are too highly correlated with each other to be included in the regression. The selection is mainly based on Principle Component Analysis (PCA). The total dataset is used, i.e., respondents from all four cities are included in the analyses.

The result from a PCA on all 33 neighborhood characteristics is reported in table 5 and, to enable comparisons, the result from a PCA on all 35 city characteristics is shown in the appendix, table A4. Table 5 shows that the 33 neighborhood factors can be grouped into seven uncorrelated factors. By way of example, factor four, named ‘Public services’, includes the three variables: ‘Good child care and schools’, ‘Good health care’, and ‘Good supply of facilities for sports’. A comparison with table A4 shows that the city level PCA yields similar but not identical results, i.e., the residents’ perceptions regarding the neighborhood characteristics follow a similar but not identical latent structure, as their perceptions of city characteristics. There are seven PCA factors in both tables, and two of the factors are more or less the same; factor 3) ‘Commercial services and Meeting places’, and factor 5) ‘Getting around’. However, there are also differences. By way of example, the image of the neighborhood seems to be closer connected to its physical characteristics, such as beautiful buildings and symbolic features (see factor 1, table 5) while the city image appears to be more ‘stand-alone’ (see factor 2, table A4), perhaps associated with a slogan rather than physical characteristics. The reputation of the neighborhood is more associated with, e.g., safety and security, while the city’s reputation appears closer connected to public services, such as health care and schools. In the neighborhood PCA, cultural diversity (‘citizens from many different cultures and countries’) is grouped with safety and reputation and has a negative factor loading, implying that a low score on cultural diversity is associated with a high score on safety and reputation. For the city, diversity is a factor based on both cultural diversity and religious institutions.

Twelve variables are selected to be included in the regression analysis, as reported by table 5. The selection is based on the variables’ factor loadings and inter-correlations, but also on theoretical reasoning (e.g., that a more generally phrased variable sometimes has been preferred to a more specific). Table 5. Resulting factors from principal component analysis using all 33 neighborhood characteristics (varimax rotation). The number and type of characteristics related to each resulting factor and the variables chosen for regression analysis. Factor Groups of characteristics Variables chosen as independent (number of variables in brackets): variables in the regression: 1)Aesthetics, Symbols and Image Aesthetics (3) ‘Beautiful buildings’ Ten variables related to:: Symbolic features (4) ‘Statues and symbols’ Image/stories (2) ‘A story I’d like to tell others’ Well integrated (1) * ‘Well integrated’ 2) Safety and reputation Safe and secure (1) ‘A safe and secure neighborhood’ Three variables related to: A good reputation (1) ‘Cultural diversity’ Cultural diversity (1) 3) Commercial services and Meeting places Commercial services (5) ‘Exciting experiences’ Eight variables related to: Meeting places (2) ‘Nice places to meet’ 4) Public services Three variables related to: Public services (3) ‘Good health care’ 5) Getting around Getting around (3) ‘Easy access by bicycle’ Four variables related to: Walkability (1) 6) Housing Two variables related to: Housing (2) ‘Housing at reasonable prices’ 7) Nature Nature, water (2) ‘Accessible nature, green belt’ Four variables related to: Biking/walking routes (1) Parks & green spaces (1)

The corresponding PCA on the seven aspects of relation results in one factor only, both regarding the city and the neighborhood. The strong underlying correlations make it impossible to include more than one individual aspect in the regression analysis. As an alternative, the average score on all seven aspects of the relation will be used. 17

Using the variables selected for the neighborhood model (as reported by table 5), table 6a reports four regressions. The dependent variable is the perceived strength of the neighborhood soul in column 1, 3, and 4 and the neighborhood relation index (the average rating of the seven aspects) in column 2. The results reported by column 3-4 are based on subsets of residents that are ‘natives’ (born in the city or lived there since 1987) as well as ‘young’ (under the age of 40). Table 6a. OLS regression results for the total sample (column 1-2), ‘native’ residents (column 3), and ‘young’ residents (column 4). Selected variables from the PCA on neighborhood characteristics (as reported by table 5) are used as independents. Dependent variables: ‘My neighborhood has a strong identity/soul’ (columns 1, 3, and 4), and the neighborhood Relation Index (column 2). Standardized beta coefficients, adjusted R2. Dependent: 1) Neighbor-hood 2) Neighborhood 3) Neighborhood 4) Neighborhood Soul Relation Index Soul (Natives) Soul (Young) Constant Beautiful buildings 0.115** 0.165** 0.073** 0.111* Statues and symbols 0.104** 0.02 0.102** 0.111** A story I'd like to tell 0.094** 0.224** 0.072** 0.118** Integrated with city 0.012 0.086** 0.03 0.077 Safety and security -0.067** 0.22** -0.061** -0.067 Cultural diversity 0.005 -0.038 0.006 0.039 Exciting experiences 0.023 0.053* 0.043 -0.089* Nice places to meet 0.015 0.021 0.007 0.076 Good health care 0.025 0.07** 0.011 -0.04 Easy access - bicycle -0.003 0.093** -0.003 -0.038 Housing at reasonable prices -0.004 -0.02 -0.017 0.003 Accessible nature -0.019 0.124** -0.014 -0.032 Relation index 0.626** 0.699** 0.591** R2 adjusted 0.623 0.424 0.688 0.584 Indication of statistical significance: ** p≤ 0.01, * p≤0.05

It should first be noted that the variable ‘Relation index’ has a positive and significant impact on the perceived strength of the neighborhood soul in all regressions. The same holds for the first three neighborhood characteristics: ‘Beautiful buildings’, Statues and symbols’, and ‘A story I’d like to tell’ The variable ‘Safety and security’ has a negative impact but is not significant when rated by the young. Columns 2 shows that seven characteristics have a significant influence on the Relation index, which supports the model assumption that the characteristics both play a direct and an indirect role for the ratings of the neighborhood soul.

Table 6b reports regression results for the city soul model, using the selected characteristics, as reported by table A4 in the appendix.6 The influence of the Relation index is positive and significant for the rating of the city soul in all three regressions, but rather few characteristics have a significant impact. The most obvious exceptions are the variables ‘Visible art’ and ‘A story I’d like to tell’ – they also have a positive and significant impact in all three regressions.

6 Differences from the results reported by Wahlström (2017) are due to the fact that Wahlström’s paper reports unstandardized coefficients, while this paper reports standardized coefficients. 18

Table 6b. OLS regression results for the total sample (column 1-2), native residents (column 3), and young residents (column 4) when using the selected variables by Wahlström (2017) as independents. Dependent variables: ‘My city has a strong identity/soul’ (columns 1, 3 and 4), and the city Relation Index (column 2). Standardized coefficients, adjusted R 2. Dependent: 1) City Soul 2) City Relation 3) City Soul 4) City Soul Index (Natives) (Young) Constant A beautiful city 0.03 0.203** 0.025 0.018 Visible art 0.136** 0.008 0.098** 0.154** A story I'd like to tell 0.095** 0.191** 0.122** 0.091* Listening to citizens -0.014 0.074** -0.042 -0.025 A good reputation -0.037* 0.178** -0.022 -0.049 Cultural diversity 0.01 -0.004 -0.014 0.03 Exciting experiences 0.036 0.086** 0.052* 0.054 Nice places to meet 0.034 0.052* 0.033 0.079 Easy access - public transport 0.025 0.032 -0.004 0.033 Housing at reasonable prices 0.005 0.063** 0.022 0.009 Nice walking/cycling routes -0.014 0.152** -0.017 -0.037 Relation index 0.59** 0.64** 0.545** R2 adjusted 0.547 0.458 0.594 0.494 Indication of statistical significance: ** p≤ 0.01, *p≤0.05

There are both similarities and differences between the neighborhood soul and the city soul regressions. The most obvious similarity is the strong influence of the relation index on both the neighborhood and city soul. Furthermore, both tables indicate a similarly significant effect of ‘A story I’d like to tell’. It can also be noted that ‘Statues and symbols’ show the same direct effect on neighborhood soul as ‘Visible Art’ does on city soul. And that ‘Safety and security’ show the same negative direct and positive indirect effect on the neighborhood soul as ‘A good reputation’ does on the city soul.

A striking difference between table 6a and 6b is that ‘Beautiful buildings’ have both a direct and indirect effect on the neighborhood soul, while the effect of ‘Beautiful city’ on the city soul is only indirect via the relation index.

The research question regarding which aspects of the relation that influence the neighborhood soul is, as already mentioned, difficult to answer as the aspects are highly correlated with each other. However, attempts to test the aspects one by one as independents in the neighborhood soul regression indicate that ‘Pride’, ‘Attachment’ and ‘Belonging’ seem to be the relational aspects with the strongest influence on neighborhood soul. These are also the relational aspects with the strongest influence on the city soul (Wahlström, 2017). 19

5. Discussion and Concluding Remarks

The analyses reported in this paper indicate that the conceptual city soul model should be applicable at the neighborhood scale. The neighborhood model might differ from the city soul model for a number of reasons, but the results presented here indicate small differences overall. However, this study does not include information on, for example, the residents’ social relations or mobility. Furthermore, these results indicate variations between Stockholm’s districts – both with regards to the perceived strength of the neighborhood soul and with regards to the relative influence on the soul of the two components; perceptions of the neighborhood’s characteristics and the relation to the neighborhood.

The regression results show that the characteristics ‘Beautiful buildings’, ‘Statues and symbols’, and ‘A story I’d like to tell’ have the strongest direct influence on the perceived strength of the neighborhood soul. Additionally, the neighborhood’s safety and security, accessibility to nature, accessibility by bicycle, health care, and degree of integration with the city have at least an indirect influence on the neighborhood’s soul, via the relation. It is difficult to compare the neighborhood and city analyses, as the underlying/latent structures of perceptions regarding the characteristics differ between them (in line with, e.g., Lalli, 1982). The relation index has a strong significant effect on both the neighborhood and the city soul. This effect is somewhat stronger on the neighborhood soul, and it is stronger among ‘native’ residents. The relational aspects with the strongest influence on the neighborhood soul are the same as for the city soul; Pride, Attachment, and Belonging.

One of the main challenges with the research presented here is the strong interdependencies among the assessed perceptions of characteristics as well as among the aspects of the relation. Further research is needed to improve the reliability and validity of the survey questionnaire. As another suggestion for future research, analyses of the neighborhood soul would probably benefit from collecting more information about the residents regarding lifestyle, social relations, etc. These factors appear to be important to the perception of places in general, and of neighborhoods in particular (see, e.g., Westford, 2010). This type of information was suppressed in the survey presented here, mainly based on an aim to keep the questionnaire short.

It is difficult to say if the residents perceive the soul of their neighborhood as stronger, or weaker, than the soul of the city. By way of example, the neighborhood soul is perceived as stronger than the city soul among the residents in South (Södermalm), but the opposite holds for West (Kungsholmen). These results hold even after controlling for the socio-demographic mix in the city districts. The variation between Stockholm’s districts can probably be explained by their different perceptions of neighborhood characteristics since the districts have very similar scores on the seven relational aspects. The residents’ overall relation to the city does not differ significantly from their relation to the neighborhood. Thus, this study does not support the results from previous research showing that the residents’ relation to their city is stronger than the relation to their neighborhood (see, e.g., Hernandez, 2007; Lewicka, 2010; Casakin, 2015).

In spite of all caveats, the presented results should be useful in practical planning. The extensive comparison between city districts, shown in table 2, is an obvious example. The table provides the residents’ views about the most important strengths and weaknesses of their districts. It can hence be used by the planners to decide which issues they should address when trying to improve the districts. The neighborhood-specific analyses show that Stockholm’s districts differ in several respects – differences that are not equalized by controlling for the district-specific socio- demographic mix – which should be especially interesting from a practical planning perspective. If a district ranks substantially lower than others regarding, for example, ‘Nice routes for cycling/walking’ or ‘Visible art in open places’ it provides the planners with a strong signal to look closer into the corresponding issues. 20

The methodology and results of this study should also be useful in other ways. It could, for example, be used as a basis for citizen dialogue when planning a new or existing urban area. The empirical material has indicated the type of characteristics that potentially influence the residents’ relation to the neighborhood and that are important components of a strong neighborhood soul. Due to the difference between a city’s natives and newcomers, it may be even more important to consider relational aspects when refurbishing existing areas. The relation’s importance to native residents is probably part of the explanation for the so called NIMBY phenomenon. As the residents’ relation to their neighborhood and city appears to play such an important role, it may be valuable for planners to collaborate more with urban sociologists and environmental psychologists.

According to, e.g., Stedman (2002) the residents’ relation to their neighborhood also influences their behavior in terms of for example commitment to area improvements/changes. Thus, it seems fair to assume that improved understanding of how residents’ perceptions of different neighborhood characteristics influence their relation to the neighborhood, might also improve the chances for successful implementation of local plans.

Obviously, the presented information in this paper might also become a valuable complement to the concept/branding reports (commonly provided by marketing or media agencies) that typically merely contain qualitative and subjective information. 21

6. References

Andersson, R. (1998) Attractive cities. An Economic Analysis (in Swedish), Byggforskningsrådet.

Anholt, S. (2010) Places. Identity, Image and Reputation. Palgrave Macmillian UK.

Anton, C.E., Lawrence, C. (2014) Home is where the heart is: The effect of place of residence on place attachment and community participation. Journal of Environmental Psychology 40, 451-461. Doi: 10.1016/j.envp.2014.10.007.

Bell, P.A., Green, T.C., Fisher, J.D., Baum A. (2001) Environmental Psychology. (5th ed.), Psychology Press. Taylor and Francis Group. New York.

Bernardo, F., Palma-Oliveira, J-M. (2016) Urban neighbourhoods and intergroup relations: The importance of place identity. Journal of Environmental Psychology 45, 239-251. Doi: 10.1016/jenvp.2016.01.010.

Bonaiuto, M., Aiello, A., Perugini, M., Bonnes, M., Ercolani, A.P. (1999) Multidimensional perception of residential environment quality and neighbourhood attachment in the urban environment. Journal of Environmental Psychology 19, 331- 352.

Bonaiuto, M., Fornara, F., Bonnes, M. (2003). Indexes of perceived residential environment quality and neighborhood attachment in urban environments: a confirmation study on the city of Rome. Landscape and Urban Planning 65, 41-52.

Bonaiuto, M., Fornara, F., Bonnes, M. (2006). Perceived residential environment quality in middle- and low-extension Italian cities. Revue Européenne de psychologie appliquée 56, 23-34.

Brown, B., Douglas, D.P., Brown, G. (2003) Place attachment in a revitalizing neighborhood: Individual and block levels of analysis. Journal of Environmental Psychology 23, 259-271.

Brown, G., Raymond, C. (2007) The relationship between place attachment and landscape values: Toward mapping place attachment. Applied Geography 27, 89-111.

Casakin, H., Hernández, B., Ruiz, C. (2015) Place attachment and place identity in Israeli cities: The influence of city size. Cities 42, 224-230.

Devine-Wright, P. (2013) Think global, act local? The relevance of place attachments and place identities in a climate changed world. Global Environmental Change 23, 61–69. Doi: 10.1016/j.gloenvcha.2012.08.003.

Florida, R. (2008) Who’s your city? Basic Books, New York.

Florida, R. (2009) The soul of the city. The Atlantic, October 1. Retrieved from http://www.theatlantic.com/national/archive/2009/10/soul-of-the-city/27526.

Florida, R., Mellander, C. (2010) There goes the metro: how and why bohemians, artists and gays affect regional housing values. Journal of Economic Geography 10 (2010) pp. 167–188. Doi: 10.1093/jeg/lbp022.

Florida, R. (2012) The Rise of the Creative Class, Revisited. New York, Basic Books.

Florida, R., Mellander, C., Rentfrow, P.J. (2013) The happiness of cities. Regional Studies, 47:4, pp 613-627. Doi:10.1080/00343404.2011.589830.

Florida, R., Mellander, C., Stolarick, K. (2011) Beautiful Places: The Role of Perceived Aesthetic Beauty in Community Satisfaction. Regional Studies, 45:1, pp 33-48. Doi.org/10.1080/00343404.2010.486784.

Glaeser, E.L., Kolko J., Saiz, A. (2001) Consumer City? Journal of Economic Geography 1, 27-50. Doi: 10.1093/jeg/1.1.27.

Hernández, B., Hidalgo, M.C., Salazar-Laplace, M.E., Hess, S. (2007). Place Attachment and Place Identity in Natives and Non-natives. Journal of Environmental Psychology 27, 310-319.

Hidalgo, M.C., Hernández, B. (2001) Place attachment: Conceptual and empirical questions. Journal of Environmental Psychology 21, 273-281. 22

Jivén, G., Larkham, P.J. (2003) “Sense of Place, Authenticity and Character: A Commentary.” Journal of Urban Design, Vol. 8, No. 1, pp 67-81. Doi: 10.1080/1357480032000064773.

Jorgensen, B.S., Stedman, R.C. (2006) A comparative analysis of predictors of sense of place dimensions: Attachment to, dependence on, and identification with lakeshore properties. Journal of Environmental Management 79, 316-327. Doi: 10.1016/j.envman.2005.08.03.

Jorgensen, B.S., Stedman, R.C. (2011) Measuring the spatial component of sense of place: a methodology for research on the spatial dynamics of psychological experiences of places. Environment and Planning B, Vol. 38, 795-813.

Kalandides, A. (2011) The problem with spatial identity: revisiting the “sense of place”. Journal of Place Management and Development, Vol.4, No 1, pp 28-39. Doi: 10.1108/17538331111117142.

Kavaratzis, M., Hatch, M.J. (2013) The dynamics of place brands: An identity-based approach to place branding theory. Journal of Environmental Psychology 21, pp 273-281. Doi: 10.1006/jevp.2001.0221.

Knez, I. (2005) Attachment and identity as related to a place and its perceived climate. Journal of Environmental Psychology 25, 207-218. Doi: 10.1016/j.jenvp.2005.03.003.

Lalli, M. (1982) Urban-related Identity: Theory, measurement, and empirical findings. Journal of Environmental Psychology 12, 285-303.

Lewicka, M. (2008). Place attachment, place identity, and place memory: Restoring the forgotten city past. Journal of Environmental Psychology 28, 209-231.

Lewicka, M. (2010). What makes neighborhood different from home and city? Effects of place scale on place attachment. . Journal of Environmental Psychology 30, 35-51.

Lewicka, M. (2011a) On the varieties of people’s relationships with places: Hummon’s typology revisited. Environment and Behavior 43(5), 676-709.

Lewicka, M. (2011b) Place attachment: How far have we come in the last 40 years? Journal of Environmental Psychology 31, 207-230.

Lynch, K. (1960) The image of the city. MIT Press, USA.

Mellander, C., Florida, R., Stolarick, K. (2013) Here to stay – The effects of community satisfaction on the decision to stay. Spatial Economic Analysis, 6:1, 5-24. Doi: 10.1080/17421772.2010.540031.

Permentier, M., Bolt, G., van Harn, M. (2010) Determinant of neighbourhood satisfaction and perception of neighbourhood reputation. Urban Studies 48(5), 977-996. Doi: 10.1177/0042098010367860.

Power, M.P. (2005) The supply and demand for natural amenities: an overview of theory and concepts. in: Green, G.P., Deller, S.C., Marcouiller, D.W. (eds.) Amenities and Rural Development. Theory, Methods and Public Policy. Edward Elgar, Cheltenham, UK; Northampton, US.

Proshansky, H.M., Kaminoff, F., Kaminoff, R. (1983) Place-identity: physical world socialization of the self. Journal of Environmental psychology, 3, pp 57-83. Doi:10.1016/S0272-4944(83)80021-8.

Relph, E. (1976) On the Identity of Places in Relph, E., Place and Placelessness. Pion, London, pp 44–62.

Rogerson, R.J. (1999) Quality of Life and City Competitiveness, Urban Studies, Vol. 36, Nos 5-6, 969-985. Doi: 10.1080/0042098993303.

Southworth, M., Ruggeri, D. (2011) Beyond placelessness. Place identity and the global city, in: Banerjee, T. and Loukaitousideris, B. (eds.) Companion to Urban Design. Routledge, New York

Pfiffer, D. & Cloutier, S. (2016). Planning for Happy Neighborhoods. Journal of American Planning Association, Vol. 82, No. 3.

Stedman, R.C. (2002) Toward a social psychology of place. Predicting behavior from place-based cognitions, attitude, and identity. Environment and behavior, Vol. 34 No. 5, 561-581. 23

Wahlström, M. (2017) What about the soul of the city? Applying a resident perspective in four Swedish cities. Romanian Journal of Regional Science, Vol. 11, No.1, Summer 2017.

Westford, P. (2010) Neighborhood Design and Travel. A Study of Residential Quality, Child Leisure Activity and Trips to School. Doctoral dissertation. Stockholm: Royal Institute of Technology.

Westin, K. (2016) Place Attachment and Mobility in City Regions. Population, Space and Place 22, 722-735.

Westlund, H. (2006) Social Capital in the Knowledge Economy. Theory and Empirics. Springer-Verlag Berlin Heidelberg.

Whisler, R.L., Waldorf, B.S., Mulligan, G.F., Plane, D.A. (2008) Quality of Life and the Migration of the College-Educated: A Life-Course Approach. Growth and Change, Volume 39, No 1, 58-94. Doi: 10.1111/j.1468-2257.2007.00405.

Zenker, S., Eggers, F., Farsky, M. (2013) Putting a price tag on cities: Insights into the competitive environment of places. Cities 30, 133–139. Doi:10.1016/j.cities.2012.02.002.

Zenker, S., Petersen, S. (2014) An integrative theoretical model for improving resident-city identification. Environment and Planning A, volume 46, 715-729.

Zenker, S., Rütter, N. (2014). Is satisfaction the key? The role of citizen satisfaction, place attachment and place brand attitude on positive citizenship behavior. Cities 38, 11-17.

Zukin, S. (2011) Naked city Death and life of authentic urban places (in Swedish), Daidalos AB, Göteborg. 24

7. Appendix

Table A1 Socio-demographic distribution of residents in Stockholm’s city districts. East South Downtown West Northwest Number of respondents (unweighted): 238 481 113 273 204 Natives Born here or moved here 58 % 59 % 59 % 52 % 59 % before 1987 Moved here 1987 or later 42 % 41 % 41 % 48 % 41 % Newcomers Old-timers 77 % 79 % 82 % 74 % 81 % Moved here 2005 or later 23 % 21 % 18 % 26 % 19 % Income category Less than 500 tkr/year 37 % 49 % 51 % 44 % 40 % 500 000 SEK/year or more 63 % 51 % 49 % 56 % 60 % Age category up to 39 years 22 % 30 % 18 % 35 % 32 % 40-64 years 40 % 41 % 52 % 40 % 39 % 65 years or older 39 % 29 % 30 % 24 % 29 % Occupation Not working 52 % 47 % 44 % 41 % 47 % Working full or part time 48 % 53 % 56 % 59 % 53 % Education level Less than 3 years of 42 % 45 % 53 % 44 % 43 % university University 3 years or more 58 % 55 % 47 % 56 % 57 %

Table A2. OLS regressions results using “My neighborhood has a strong identity/soul” as dependent variable, for Stockholm, as well as Stockholm’s city districts: standardized coefficients, R2 adjusted. East South Downtown West Northwest Number of respondents (unweighted): 238 481 113 273 204 Relation Index 0.62 0.542 0.614 0.611 0.641 Mean score on all the characteristics 0.128 0.314 0.231 0.262 0.198 Higher education- at least 3 years university studies (1/0) Not sign. Not sign. Not sign. -0.122 Not sign. Higher income – annual income above 500 KSEK (1/0) Not sign. Not sign. Not sign. Not sign. Not sign. Young – less than 40 years old (1/0) -0.13 0.074 Not sign. Not sign. -0.138 Working – full-time or part-time (1/0) Not sign. Not sign. Not sign. Not sign. Not sign. Non-natives – moved to the city 1987 or later (1/0) Not sign. Not sign. Not sign. Not sign. Not sign. Adjusted R Square 0.544 0.578 0.611 0.631 0.685 p≤0.05, if not otherwise indicated 25

Table A3. Results from OLS regression, using “My neighborhood has a strong identity/soul” as dependent variable: standardized coefficients and R2 adjusted for different subgroups. Stockholm city as well as Stockholm’s city districts. N= number of respondents (unweighted).

Stockholm East South Downtown West Northwest ‘Native residents’ (born in the city or lived there for at least 30 years) Relation Index 0.655 0.740 0.584 0.719 0.666 0.734 Mean score all characteristics 0.212 Not sign 0.289 Not sign. 0.234 Not sign. R2 adjusted 0.618 0.568 0.617 0.682 0.571 0.677 N 754 137 285 67 143 121 ‘Young residents’ (under 40 years old) Relation Index 0.527 0.682 0.548 0.591 0.460 0.685 Mean score all characteristics 0.307 Not sign. 0.242 Not sign. 0.412 Not sign. R2 adjusted 0.553 0.616 0.507 0.336 0.585 0.682 N 378 51 144 20 96 95 ‘High educated residents’ (at least 3 years in university) Relation Index 0.554 0.681 0.505 0.733 0.586 0.550 Mean score all characteristics 0.305 0.180 0.340 Not sign. 0.296 0.306 R2 adjusted 0.586 0.620521 0.553656 0.613948 0.595863 0.636732 N 728 139 266 53 152 117 ‘Employed residents’ (working full time or part time) Relation Index 0.530 0.582 0.549 0.702 0.530 0.572 Mean score all characteristics 0.325 Not sign. 0.310 Not sign. 0.377 0.315 R2 adjusted 0.591 0.424 0.596 0.588 0.645 0.686 N 703 114 254 83 161 109

‘High-income residents’ (at least 500 000 SEK per year) Relation Index 0.543 0.623 0.493 0.712 0.507 0.582 Mean score all characteristics 0.302 Not sign 0.357 Not sign. 0.370 0.277 R2 adjusted 0.576 0.481 0.587 0.510 0.599 0.630 N 643 135 216 49 134 109 p≤0.05, if not otherwise indicated

Table A4. Resulting factors from principal component analysis on all 33 neighborhood characteristics (using varimax rotation). The number and type of characteristics related to each resulting factor and the variables chosen for regression analysis (as chosen by Wahlström, 2017). Factor Groups of characteristics Variables chosen as independent (number of variables in brackets): variables in the regression: 1) Aesthetics, Symbols and Nature Aesthetics (4) * ‘A beautiful city’ Ten variables related to: Symbolic features (3) ‘Visible (street) art’ Natural environment (1) ‘Nice routes for walking/biking’ Walkability (2) 2) Image and Stories Two variables related to: Image and stories (2) ‘A story I’d like to tell others’ 3) Commercial services and Meeting places Commercial services (5) ‘Exciting experiences’ Eight variables related to: Meeting places (3) ‘Nice places to meet’ 4) Public services Public services (3) ‘A good reputation’ Five variables related to: Reputation (1) Safety (1) 5) Getting around Getting around (3) ‘Easy access - public transport’ Four variables related to: Walkability (1) 6) Housing and Governance Housing (2) ‘Housing at reasonable prices’ Four variables related to: Governance (2) ** ‘City commitment’ 7) Diversity Two variables related to:: Diversity (2) ‘Cultural diversity’ * One of the aesthetic items (cozy neighborhoods) was included among the city characteristics only. ** Only included among the city characteristics.