DENSE BUT NOT CROWDED: MAINTAINING A SENSE OF NEIGHBORHOOD COMMUNITY IN A WORLD OF INCREASING URBAN DENSITY by Eric Douglas B.A. Architecture, University of California, Berkeley, 1999 MArch, University of California, Los Angeles 2010 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Planning) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2021 © Eric Douglas, 2021 ii The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled: Dense but not crowded: Maintaining a sense of neighborhood community in a world of increasing urban density submitted by Eric Douglas in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Planning Examining Committee: Dr. Maged Senbel, Associate Professor, School of Community and Regional Planning, UBC Supervisor Dr. Penelope Gurstein, Professor, School of Community and Regional Planning, UBC Supervisory Committee Member Dr. Nathaniel Lauster, Associate Professor, Department of Sociology, UBC Supervisory Committee Member Dr. Ute Lehrer, Professor, Faculty of Environmental and Urban Change, York University External Examiner Dr. Patrick Condon, Professor, School of Architecture and Landscape Architecture, UBC University Examiner Dr. Michael Hooper, Associate Professor, School of Community and Regional Planning, UBC University Examiner iii Abstract This study investigates the relationship between urban density and sense of community. In particular, it tries to establish whether residents’ sense of community diminishes as density increases, and, if so, what can be done to moderate this relationship. It used an explanatory sequential mixed-methods approach that included an online survey and semi-structured interviews. The findings suggest that there is a negligible relationship between density and sense of community at all density levels except for very dense environments, in which case the relationship is negative. Several types of public space may moderate this relationship, however. Both the survey and the interviews suggest that high-quality parks, walkways, and community centers may increase residents’ sense of community. iv Lay Summary With more and more people moving to cities, and with cities becoming more and more dense, are we squeezing the life out of our neighborhoods? Or, are there ways that our cities can accommodate growing populations and also provide excellent neighborhoods? Can we offer ever more people a higher quality of life in our urban environments? If so, how? The purpose of this study is to investigate the relationship between urban density and sense of community. While sense of community is only one aspect of a person’s quality of life, it serves as a fairly good proxy for quality of life in general. When we feel at home in our neighbourhood and comfortable around our neighbours, it makes the good times better and the tough times easier to handle. In a world of increasing density, sense of community is a great thing to have. v Preface The entirety of this research was designed, performed, analyzed, and described by the author alone. No part of this thesis has been previously published except for citations that have been properly referenced. This research was approved by the UBC Behavioral Ethics Research Board (id# H17-02209). vi Table of Contents Abstract ......................................................................................................................................... iii Lay Summary ............................................................................................................................... iv Preface .............................................................................................................................................v Table of Contents ......................................................................................................................... vi List of Tables ..................................................................................................................................x List of Figures ............................................................................................................................... xi Acknowledgements ..................................................................................................................... xii Chapter 1: Introduction ................................................................................................................1 Problem statement ....................................................................................................................... 2 State of the field .......................................................................................................................... 3 Research objective ...................................................................................................................... 3 Key research questions ............................................................................................................... 4 Personal Interest .......................................................................................................................... 4 Chapter 2: Literature Review .......................................................................................................6 Architectural affordance as a theoretical framework .................................................................. 6 What is sense of community? ................................................................................................... 30 How do we react to urban density? ........................................................................................... 35 How is density an issue for city planners? ............................................................................ 36 Definitions of density ........................................................................................................ 37 Issues associated with low and high urban density ........................................................... 39 Density and high-rise development .................................................................................. 41 What are the effects of increasing density? .......................................................................... 46 How density affects marketability .................................................................................... 46 How density affects sustainability .................................................................................... 47 How density affects livability ........................................................................................... 48 How does urban density influence sense of community? ......................................................... 51 The role of public space ............................................................................................................ 58 Measuring public space......................................................................................................... 60 vii Chapter 3: Methodology..............................................................................................................64 Survey ....................................................................................................................................... 65 Creating the questions ........................................................................................................... 66 Selecting the target areas ...................................................................................................... 68 Advertising the survey .......................................................................................................... 69 Processing the data ................................................................................................................ 70 Analyzing the data ................................................................................................................ 70 Limitations ............................................................................................................................ 71 Lessons learned ..................................................................................................................... 72 Interviews .................................................................................................................................. 72 Creating the questions ........................................................................................................... 72 Selecting interviewees .......................................................................................................... 73 Conducting the interviews .................................................................................................... 73 Processing the data ................................................................................................................ 74 Analyzing the data ................................................................................................................ 74 Limitations ............................................................................................................................ 74 Lessons learned ..................................................................................................................... 75 Site Observations ...................................................................................................................... 75 Selecting sites........................................................................................................................ 76 Selecting metrics ................................................................................................................... 76 Conducting observations ....................................................................................................... 76 Processing data...................................................................................................................... 77 Limitations ............................................................................................................................ 77 Lessons learned ..................................................................................................................... 77 Chapter 4: A quantitative inquiry into the relationship between urban density and sense of community in the Greater Vancouver Regional District .........................................................78 The relationship between urban density and sense of community ........................................... 78 Potential confounding variables ................................................................................................ 88 Potential moderating variables .................................................................................................. 95 Related considerations .............................................................................................................. 96 Relationship between sense of community and past experience .......................................... 96 viii Relationship between density and crowding......................................................................... 97 Relationship between density and safety ............................................................................ 100 Effectiveness of test items ...................................................................................................... 102 Chapter 5: Gaining a deeper understanding of residents’ sense of community through semi-structured interviews ........................................................................................................109 Perceptions of terms ................................................................................................................ 109 What do you consider to be your neighbourhood? ............................................................. 109 What do you think it means to have a sense of community? .............................................. 110 What are the public/common spaces in your neighbourhood? ........................................... 111 Perceptions of neighborhoods ................................................................................................. 114 How would you describe your sense of community in your neighbourhood? ................... 115 What do you like about your neighbourhood? .................................................................... 117 If you could change anything about your neighbourhood, what would it be? .................... 118 Do you wish you spent more time or less time speaking with your neighbours? ............... 119 Do you consider your neighbourhood to be very dense? .................................................... 119 Is it crowded? ...................................................................................................................... 119 Would you rather live in a less dense neighbourhood? ...................................................... 119 Tell me about how safe your neighbourhood is. ................................................................. 120 What would make it safer? ................................................................................................. 120 Perceptions of connections ..................................................................................................... 121 Recommendations for public space ........................................................................................ 126 Chapter 6: The role of public space in maintaining a sense of community in high-density neighborhoods ............................................................................................................................133 Appropriating automobile space to facilitate social interaction ............................................. 133 How interaction in public space has led to friendships ........................................................... 137 Interviewees’ objections to density and terms for accepting it ............................................... 144 Applying architectural affordance to public space recommendations .................................... 149 Chapter 7: Conclusion ...............................................................................................................153 Applications ............................................................................................................................ 157 Practice ................................................................................................................................ 157 Academia ............................................................................................................................ 158 ix Limitations .............................................................................................................................. 159 Future research ........................................................................................................................ 159 References ...................................................................................................................................161 Appendix A: Summary of previously used sense of community measures……………….173 Appendix B: Site selection……………………………………………………………………186 Appendix C: Advertising the survey………………………………………………………...197 Appendix D: Survey questions……………………………………………………………….209 Appendix E: Survey variables……………………………………………………………….218 Appendix F: Interview questions…………………………………………………………….229 Appendix G: Interview and survey summaries for interviewees………………………….230 Appendix H: Interview response summaries by question………………………………….245 Appendix I: Site observation template………………………………………………………271 Appendix J: Site observation results and images…………………………………………..273 Appendix K: Density quintile examples along Dunbar transect…………………………..289 Appendix L: A discussion of case study research…………………………………………..294 x List of Tables Table 1 - Chain of influence from environment to user ................................................................ 26 Table 2 - Survey results ................................................................................................................ 83 Table 3 - Effectiveness of various test items by Spearman test. ................................................. 104 Table 4 - Results of regression model of SOC test items showing predictive power of successive items ............................................................................................................................................ 106 Table 5 - Places that Interviewees considered 'public space' in their neighborhoods ................. 111 Table 6 - How interviewees describe their sense of community ................................................ 116 Table 7 - Whether interviewees consider their neighborhoods to be dense/crowded................. 120 Table 8 - How interviewees would improve the safety of their neighborhoods ......................... 121 Table 9 – Culture with which interviewee identifies and how interviewee thinks that culture values sense of community ......................................................................................................... 123 Table 10 – Communities to which interviewee feels connected an how interviewee ranks the importance of these connections. ................................................................................................ 126 Table 11 - Survey respondents’ use of public space and interaction in public space ................. 128 xi List of Figures Figure 1 - Postal code outlines (blue areas) with dots showing postal code centroids ................. 67 Figure 2 - Postal code centroids shown within dissemination block outlines .............................. 67 Figure 3 – Relationship between sense of community and urban density .................................... 80 Figure 4 - Relationship between sense of community and urban density at very low density ..... 84 Figure 5 - Relationship between sense of community and urban density at low density ............. 84 Figure 6 - Relationship between sense of community and urban density at medium density ...... 84 Figure 7 - Relationship between sense of community and urban density at high density ............ 84 Figure 8 - Relationship between sense of community and urban density at very high density .... 84 Figure 9 - Relationship between sense of community and urban density for parents with young children ......................................................................................................................................... 84 Figure 10 - Map of density by postal codes in Kitsalano and downtown Vancouver .................. 85 Figure 11 - Relationship between sense of community and urban density in Kitsalano and downtown Vancouver ................................................................................................................... 85 Figure 12 - Map of density by postal codes in and near the Klahanie area .................................. 86 Figure 13 - Relationship between sense of community and urban density in and near the Klahanie area ................................................................................................................................ 86 Figure 14- Map of density by postal codes at the University of British Columbia ...................... 87 Figure 15- Relationship between sense of community and urban density at the University of British Columbia ........................................................................................................................... 87 Figure 16 - Potential influences on sense of community and on the relationship between urban density and sense of community ................................................................................................... 88 Figure 17 - Relationship between sense of community and use of public spaces ........................ 90 Figure 18 - Relationship between sense of community and interaction in various public spaces 91 Figure 19 - Relationship between sense of community and feeling of crowding ......................... 93 Figure 20 - Relationship between sense of community and feeling of safety .............................. 94 Figure 21 - Relationship between density and feeling of crowding ............................................. 99 Figure 22 - Relationship between feeling of crowding and type of housing ................................ 99 Figure 23 - Relationship between density and feeling of safety ................................................. 101 Figure 24 - Relationship between feeling of safety and housing type ........................................ 101 Figure 25 - Effectiveness of sense of community test items ...................................................... 105 Figure 26 - Effectiveness of feeling of crowding test items ....................................................... 107 Figure 27 - Effectiveness of feeling of safety test items ............................................................. 107 xii Acknowledgements Thanks to my wife, Teresa, for suggesting I get a PhD (because she recognized I needed something to do), and for then supporting me through the many years it’s taken to get it. Thanks to my advisor, Dr. Maged Senbel, for his guidance and patience. Thanks to Dr. Penny Gurstein and Dr. Nathan Lauster for agreeing to serve on my committee. Thanks to Dr. Scott McIntyre for guidance regarding my quantitative analysis. And, thanks to my colleagues in the School of Community and Regional Planning and my neighbors in Acadia Park for the many lively and enjoyable discussions, debates, and conversations. The ride’s been bumpy but the scenery not without merit. 1 Chapter 1: Introduction Many contemporary urban designers advocate “compact city” principles for North American urban areas (Farr, 2007; Hester, 2006; Duany et al, 2010; Calthorpe, 2011). These principles include walkability, dense clusters around transit nodes, and high-density mixed-use development (Talen, 1999). Benefits of compact city design may include increased social interaction (Raman 2010), lower carbon emissions (Calthorpe, 2011), and improved access to amenities (Hester, 2006). However, the increased density of such compact city design has also been associated with several negative effects, such as increased aggression (Evans 2000), decreased privacy (Evans et al. 1989), and decreased neighborhood satisfaction (McCarthy & Saegert 1978). While researchers have studied residential density’s relationship with many conditions, one association that has not been sufficiently explored is the relationship between high-density residential areas and residents’ sense of community. Putnam (2000) showed a relationship between low-density suburban neighborhoods and reduced sense of community (although he also showed it was difficult to theorize a specific causal relationship from the data he used). Other researchers have considered the association between higher-density areas and sense of community, but these were just higher-density sections of low-density areas (Wilson & Baldassare 1996; French et al. 2014). Also, their results were inconclusive. Researchers have investigated factors that may influence sense of community, such as community empowerment (Amad et al. 2016), sense of place (Wise 2015), diversity (Neal & Neal 2014), neighborhood associations (Kingston et al. 1999), and social capital (Long & Perkins, 2007). Other studies have investigated how the built environment in general may influence sense of community (Jung et al., 2015; Ebrahim, 2015; Kaźmierczak, 2013; Francis et al., 2012; Schwaller, 2012; Talen, 1999). While much has been written about urban density and sense of community separately, few studies have sought to link these topics empirically. Interest in the concept of sense of community among sociologists, community psychologists, and city planners has grown since Sarason (1974) popularized the term. Chavis et al. (1986) provided further theoretical structure to the definition of sense of community by claiming that the construct required four elements, namely, membership, influence, shared values, and shared emotional connection (also McMillan & Chavis 1986). Researchers have 2 associated sense of community with several personal and societal benefits, such as improved mental health (Hall 2017; Davidson & Cotter 1991; Pretty 2006), reduced crime (Jacobs 2011), and resiliency after disasters (Wickes et al. 2015). Others have noted potential hazards, such as exclusion (Halamova 2016), rigid conformity (McMillan & Chavis 1986), compromise of personal safety or values (Sense of Community Partners 2004), lack of diversity (Walker & Ravel 2017), and compromise of solidarity to other groups (Pretty et al. 2006). While planners tend to assume that building a sense of community is a beneficial endeavor and somehow within their purview (www.planning.org), it is unclear when this effort is appropriate (some people would rather be left alone (Brower 2011)), whether the effort is specially challenged in high-density areas (and why), and, if it is, what may and should be done about it and by whom. Given the interest among North American city planning departments in both compact design and community building (Brower 2011), it is important to fill the gap in understanding regarding the relationship between high-density development and neighborhood sense of community. This study considers this relationship, as well as built environment strategies that may affect sense of community in high-density neighborhoods. My intent is not to question whether cities should become more compact, but rather to understand possible negative ramifications of this process and what might be done to ameliorate or alleviate them. As cities continue to develop high-density neighborhoods, understanding the associated effects on sense of community, and strategies to address them, will continue to be relevant. Problem statement Some urban theorists have contended that compact neighborhoods generally create a stronger sense of community among residents than do low-density suburban neighborhoods (Talen, 1999). Sense of community refers to an individual’s perception that she feels connected to her community, feels invested in it, and feels a shared set of goals with her neighbors (Ebrahim, 2015). Sense of community has been shown to have several societal benefits, such as civic involvement (Sense of Community Partners, 2004), emotional wellbeing (Hall, 2017; Lardier et al. , 2017; Francis et al., 2012), perception of safety (Sense of Community Partners, 2004), and community resilience (Walton 2016). While New Urbanism theorists have argued that very low residential densities negatively correlate with sense of community (Audirac, 1999), they do not discuss the potential negative effects of very high densities on sense of community. High urban 3 residential density brings potential benefits (such as increased return on investment for developers, increased housing options for residents, and reduced environmental impacts), but it may also diminish a neighborhood’s sense of community if the urban form it requires reduces, rather than increases, opportunities for residents to interact (Francis et al., 2012). For example, highly dense environments, such as apartment towers, may afford little opportunity for neighbors, even adjacent neighbors, to interact. Although sense of community can form in virtual (online) environments, neighborhood sense of community tends to require some form of direct, in-person interaction (Francis et al., 2012). If it is the case that high residential density leads to a lack of semi-public space in which residents can interact, and if it is true that the resulting lack of interaction leads to a loss of sense of community, then sense of community would clearly be a casualty of increased density. It is, therefore, critical to examine the relationship between sense of community and density, especially with a view to understanding public space strategies intended to encourage sense of neighborhood community in high-density environments, if we wish to ensure that density and sense of community are compatible. State of the field Researchers have discussed built environment design strategies specific to both high-density environments (Lehman 2016; Moroni 2016) and to sense of community (Walton 2016; Ebrahim 2015), but there has been almost no discussion in the literature attempting to bridge these two concerns (Francis et al. 2012; Talen 1999). I have found no studies that have examined the relationship between sense of community and increasing density in high-density environments, nor studies that discuss the role of various types of public space in enhancing residents’ sense of neighborhood community in high-density environments. While no published studies have sought to establish this relationship directly, many studies have shown an association between high-density environments and outcomes, such as aggression and withdrawal, that researchers commonly consider antithetical to sense of community (Cramer et al. 2004, Evans 2003, Boyko & Cooper 2011, Burton 2000, Audirac 1999). Research objective The primary objective of this study was to investigate how residents’ sense of community relates to increased density in high-density urban environments. A secondary objective was to 4 investigate how public space can increase residents’ sense of community in high-density urban environments. Results of this study may inform stakeholders interested in providing, maintaining, using, or understanding high-density urban environments in which residents experience a high sense of community. Such stakeholders may include design professionals (such as architects), regulators (such as city planners), housing suppliers (such as developers), and researchers (such as environmental psychologists and community psychologists). Key research questions The primary question of this study is, How does population density relate to residents’ sense of neighborhood community? Derivative questions include the following: • Does sense of community tend to diminish in high-density neighborhoods? • If so, is this tendency due to a lack of opportunities for residents to experience informal meetings? • If so, can developers and planners increase sense of community with the thoughtful addition of public open space? • Finally, what other factors might mitigate any potential negative effects of high density on sense of community? Note that these derivative question are predicated upon the relationship I expected to find. Personal Interest During my architectural training, I took special interest in how the built environment could provide venues for people to interact. As part of my planning education, I also learned about the value of public space and the importance of providing pedestrian-centric infrastructure. So, it was with particular delight that I moved into my current neighborhood, Acadia Park, the student family housing section of the University of British Columbia. Here, for the first time, I was able to experience a pedestrian-oriented master-planned community. It was as if the ideas on which I had been academically raised but could never find architecturally expressed had finally been put to use. I was also delighted to find that the area seemed to work just as the architectural and planning theorists imagined. My neighbors seemed to have something. What was it? Yes, they enjoyed being here, but there was something more. Over time, I came to understand this ‘something more’ as ‘sense of community’ and learned that it was a real thing, a thing that 5 people studied. I wanted to study it, too. I wanted to know if the ideas that I had learned in school—that good urban design could create better living experiences—had a real basis. But, I also wanted to know about density. The university has a long waiting list of students who would like to live on campus, but have no place available to them. The university is currently investigating how many new units it can build on its remaining land. It is even considering demolishing Acadia Park to make way for denser housing. This brings up the question, if there are aspects of Acadia Park that make it a “high quality”1 area, what are those aspects, and can they be applied to a neighborhood that accommodates more people? Or, how dense can one make Acadia Park without losing what Acadia Park is? And, more generally, how can we build better neighborhoods that accommodate more people? How can we provide more and better housing? These were some of the thoughts that got me started on the topic of this thesis. What follows is what I found out. 1 Ideally, aspects that we can identify with the concept of “quality” without going mad (Pirsig 1974). 6 Chapter 2: Literature Review Architectural affordance as a theoretical framework This investigation into the relationship between sense of neighborhood community and urban density exists within the more general inquiry into the relationship between quality of life and the built environment. I find this investigation compelling because I am interested in understanding how to improve quality of life through changes to the built environment. Further, I believe this research is topical because more and more people are living in increasingly dense environments. These dense environments create conditions that some research has suggested may prove challenging to residents’ sense of community (Cramer et al. 2004, Evans 2003, Boyko & Cooper 2011, Burton 2000, Audirac 1999, Baldassare 1982, Nguyen 2010). Yet, the explicit relationship between density and sense of community remains poorly researched and poorly theorized. This study seeks to fill this gap in knowledge by investigating this relationship. The findings of this research may include applications in theory, practice, and industry. An application to industry might be the question, ‘How much non-rentable/non-salable space should a developer set aside in a housing project for amenities?’ An application to (planning) practice might be, ‘What concessions should a city require of a developer who wishes to increase the density of a residential tower above that allowed by typical zoning law?’ An application to theory might be an argument as to whether physical design decisions (such as a provision of public space) are able to influence human values (such as sense of community). A commonality across this range of inquiries is that we must assume the ability of the built environment to shape people’s perception and experience. While commonly taken for granted, this assumption is unproven and, even if true, may represent a chaotic rather than a mechanistic relationship. Given our intended objectives, we need to find a theoretical framework that allows for a causal relationship between the built environment and human perception. We can begin this search by trying to understand the nature of causal relationships generally. If we can gain this understanding, we may be able to then see which causal framework is most suitable to our research question. Thus, we begin with a discussion of causality. *** 7 What is causality? How does one thing make another thing happen? A review of the relevant literature suggests that no one seems to know. We can imagine that there are different kinds of causality. For example, we might distinguish physical causation (one billiard ball strikes another and ‘causes’ it to move) from psychological causation (an advertisement ‘causes’ someone to make a purchase) and from social causation (a prominent event ‘causes’ a population to vote for a particular candidate). But, in all of these variants, we never see the causal mechanism. We don’t observe how the electro-magnetic force transfers energy from one billiard ball to another, which aspect of an advertisement tips a person to enter a store he otherwise would not have considered entering, or how an event alters an election. We can theorize about how these mechanisms work2, but we can never falsify our theories because we can never test them. As 18th-century philosopher David Hume noted, we never observe laws or causes, but from the manifestations and results of them we merely assume causality (Durant 1926). So, perhaps the best we can do is to rationalize that some theories of causality make more sense than others. An academic search for books and articles on causality returns few fruitful entries. Much of the current literature on causality has little to say about the nature of causality, but rather bypasses this discussion and moves straight to discussions of how to model relationships (for example, see Halpern 2016, Berzuini et al. 2012, Pearl 2009, and Morton & Williams 2010). Yet these causal models are constructed with pre-conceived assumptions about the causality of the relationships (Kleinberg 2012). Research scientists tend to eschew reference to causality, preferring to talk about association and correlation instead (Illari et al. 2011). They may point out that randomization is critical to the identification of causation, but the mere introduction of randomization tells us nothing about the mechanics of causation (Berzuini et al. 2012). In fact, randomized controlled trials, considered in most research fields as the best indicators of causal associations, still give no indication as to the fundamental nature of these associations (Kleinberg 2012). The paucity of attempts to define causality may reflect a fundamental lack of understanding among scientists as to its very nature. 2 Little (2011) describes “causal realists” as those who maintain that “we can only assert that there is a causal relationship between X and Y if we can offer a credible hypothesis about the sort of underlying mechanism that might connect X to the occurrence of Y.” 8 Some writers, however, have struggled with the problem of what causality is. The most prominent writer on the topic of causality is Hume. He argued that causality had three essential elements, namely, contiguity (cause and effect must be proximate in time), temporal priority (the cause must precede the effect in time), and necessary connection (the effect requires the cause) (Kleinberg 2012). He claimed that causes may be objects, events, or processes (Kleinberg 2012). More recently, John Leslie Mackie has argued that a cause may be an Insufficient but Non-redundant part of an Unnecessary but Sufficient condition (INUS3) (Kleinberg 2012). This consideration of sufficient non-necessity brings up the messy potential of “overdetermination,” which Kleinberg (2012) describes as a case in which “there are two or more possible causes for an effect and all are present (such that) all causes will turn out to be spurious aside from the earliest.4” Little (2011) defines a cause as “a condition that either necessitates or renders more probable its effect.” Kleinberg (2012) also asserts that a causal relationship may be either deterministic or probabilistic, and notes that probabilistic theories of causality may consider the lack of determinacy in a system to be inherent (ontological) or based in limits of observation (epistemic). He also places a distinct emphasis on temporality, noting that a fundamental aspect of causality, though one that is often overlooked, is the time range within which causality may happen (Kleinberg 2012). Still, even these refinements and digressions may be no more than circular tautologies if all they do is tell us that, in one way or another, causes are things that make other things happen (or more likely to happen). But, this was all I could find. So, I still don’t know what causality is. And, I don’t think anyone else does, either. Yet, we have to deal with it. We have to assume that causality exists. Most research relies upon this polite assumption. So, how do researchers work around the necessity of causality? Pearl (an often-cited authority on the topic) sees it this way: 3 “Unpacking this, we have that: 1. C ∧ X is sufficient for E. 2. C ∧ X is not necessary since Y could also cause E. 3. C alone may be insufficient for E. 4. C is a non-redundant part of C ∧ X.” (Kleinberg 2012) 4 Kleinberg (2012) suggests that one way to resolve this problem of causal attribution when multiple causes are involved is to say that a single necessary and sufficient cause would have a significance of unity, but shared causes would have a significance between 0 and 1. 9 “We view the task of causal discovery as an induction game that scientists play against Nature. Nature possesses stable causal mechanisms that, on a detailed level of descriptions, are deterministic functional relationships between variables, some of which are unobservable. These mechanisms are organized in the form of an acyclic structure, which the scientist attempts to identify from the available observations.” Pearl (2009, p. 43) I would challenge most of these statements, though. How can induction inform the causality involved in a singular event?5 Does nature really posses stable causal mechanisms, or could the mechanisms be chaotic (or absent, leaving mere chimeras of chance)? Are the relationships really deterministic, or could they be probabilistic6? Are only some of the relationships unobservable, or are all of them unobservable? Are the mechanisms really acyclic, or could they be recursive, involving variables that are both influencing and being influenced by other variables? Despite these criticisms, I think Pearl’s viewpoint reflects the tacit assumptions of most natural scientists. But, what can be done for those in the social sciences? They seem to be the most challenged by causality. Theoretical fields, such as mathematics, may perhaps legitimately claim causality (in that the addition of two numbers can be shown to result in the creation of a new one), and physical sciences may be able to make a strong argument of causality based upon the consistencies of associations that they observe among inanimate matter, but social sciences must account for a virtually infinite number of potential variables. A social science researcher, trying to explain the entire string of causality involved when someone reacts to a given stimuli in a certain way (say a child was stung by a bee, started crying, then tried to look brave in front of his friends), might need to account for the causal factors associated with fields ranging from botany to biology to entomology to anatomy to psychology to sociology, along with subfields within each. Where was the real cause of the reaction? What is one to do when none of the causes is observable? 5 Causal relationships may be typical, applicable universally and generally used for prediction, or token, applicable to only a specific instance of a relationship and generally used to explain a past event (Kleinberg 2012). 6 Pearl (2009) also notes that most causality models express causality in terms of probability. 10 How much more fraught than a single reaction is the search of causality in a group or chain of reactions7? Yet policy research must attempt to do just this, if it is to have relevance (Illari et al. 2011). Unfortunately, political scientists have not invented a methodology to assign causality. Instead, they borrow from the methodologies of the statisticians and economists (Morton & Williams 2010). According to Morton & Williams (2010), the standard experimental approach in political science involves four principles, namely, 1-designate a target population 2-apply an intervention 3-account for confounding variables 4-randomly assign control and treatment to sample group. This approach should look familiar to any laboratory technician. It may even produce the best data that can be produced. What it can not do, however, is show causality. This may not be an impediment, depending upon the level of scrutiny the research receives. Still, at some point, every policy position will face criticism. Every policy based on research should have someone asking the researcher, ‘but how do you know that x causes y?’ The likely response will be a reference to methodology, but, as Illari et al. (2011) note, “causality is at the crux of metaphysical, epistemological and methodological issues in the sciences (and) giving a methodological answer to someone concerned about the metaphysics of this question, or vice versa, will not help them.” (italics theirs) Further complications for the social sciences include the uncomfortable fact that, unlike epidemiology, causal relationships in sociology may not even have an obvious statistical association (Little 2011), thus knocking the wind out of this primary touchstone of causality in the natural sciences. Also, it is impossible to perform true experiments in real-world sociological 7 Tolstoy (2017, p571) confronts this dilemma with respect to the folly of assigning causality to major historical events: “When the apple is ripe and falls—why does it fall? Is it because it is drawn by gravitation to the earth, because its stalk is withered, because it is dried by the sun, because it grows heavier, because the wind shakes it, or because the boy standing under the tree wants to eat it? Not one of those is the cause. All that simply makes up the conjunction of conditions under which every living, organic, elemental event takes place. And the botanist who says that the apple has fallen because the cells are decomposing, and so on, will be just as right as the boy standing under the tree who says the apple has fallen because he wanted to eat it and prayed for it to fall. The historian, who says that Napoleon went to Moscow because he wanted to, and was ruined because Alexander desired his ruin, will be just as right and as wrong as the man who says that the mountain of millions of tons, tottering and undermined, has been felled by the last stroke of the last workingman's pick-axe. In historical events great men—so called—are but the labels that serve to give a name to an event, and like labels, they have the least possible connection with the event itself.” 11 settings, as one can not observe the same sample both with and without an intervention (Berzuini et al. 2012). The world is not a laboratory. Finally, we should note that some laws that appear universal may instead be localized, like an orderly set of numbers within a larger random string (Svozil 2018). What applies at one level of government or scale of population might be just opposite of the application in another. What to do? Is there any way forward? The lack of causal observation is even more damaging to the social sciences than to the natural ones because there are more variables and less control. We have to live in the experiment whilst it plays out and wonder what might have turned out if things had been different8. We have to study it while it is happening and hope that our guesses are no worse than random. Yes, we can still run regression models and look for correlations, but we can not escape the trite axiom that ‘correlation does not imply causation’9. This taunting reminder lures us with the promise of a metaphysical escape from responsibility, but offers no clue as to what does imply—or, better yet, show—causality. Unfortunately, this metaphysical pardon doesn’t excuse us entirely. We may not understand how causality works, how to observe or demonstrate it, or even what it is, but still we must answer for it. If we look for associations between independent and dependent variables, even variables as complex as density and sense of community, we do so because we expect these relationships to be causal, not random. Otherwise, what is the point of the research? Does or does not the independent variable determine (at least to some extent) the outcome of the dependent one? (If not, why call it dependent?) In this sense, causality is inextricably linked to determinism. In another sense, they are, in fact, the very same thing. And, while “causality” has no precedent as a theoretical framework, “determinism” most certainly has. Therefore, is determinism the most appropriate theoretical framework under which to consider our particular research question? *** This study is premised on an assumed relationship between the built environment and human perception/behavior, a relationship that is sometimes obvious and uncontested (as when a locked door impedes entry) and sometimes obscure and subject to debate (as whether a park 8 One test for causality involves an evaluation of counterfactuals, or, what would have happened if the proposed causal agents had been different, absent, or abetted or supplanted by other factors (Berzuini 2012). 9 Of course correlation implies causality. That’s way we use it. It just doesn’t prove causality. But, then again, neither does anything else. 12 bench facilitates social interaction). It builds on a rich history of interest in the reciprocal relationship between how societies shape, and are shaped by, their environments. Discussions regarding the influence of the environment on perception and behavior usually fall within the theoretical framework of environmental determinism. This is unfortunate for two primary reasons. First, while environmental determinism has had a long and extensive discussion and produced substantial bodies of literature and thought, the more traditional scope of environmental determinism concerns the primal effects of the natural environment rather than the mediated effects of the built environment. Second, environmental determinism has become tainted with unpalatable political applications (such as justification of colonialism or even ethnic purges) that have left many academics dismissive of the entire corpus of the theory out of hand. Despite its current unpopularity, the framework under which this study would fall by default is environmental determinism. In the spirit of due diligence, however, we should review whether some other option may be even more appropriate. With our discussion of causality in mind, let us turn to the field of determinism generally, along with several descendent subcategories (including environmental determinism), to see which linkage between the built environment and perception might be most appropriate for our current needs. Determinism involves one phenomenon being determined, in whole or in part, by another. A strict deterministic doctrine would hold that one condition (or set of conditions) is both necessary and sufficient to determine another10 (Faubion 2008) and that “all events, without exception, are just effects” (Honderich 2005). While determinism provides a basis for understanding causal effects of environmental conditions, it connotes a fatality that has little application to the intent of the inquiry of this study. 10 A philosophical debate concerns whether human free will can be completely subsumed by external factors that predetermine human actions. Philosophers have argued (though not necessarily using the word ‘determinism’) that human actions result primarily, or exclusively, as a result of an omniscient deity (Luis de Molina (1535–1600), Baruch Spinoza (1632–1677), Gottfried Wilhelm Leibniz’s (1646–1716)), of a relentlessly mechanistic universe (Galileo Galilei (1564–1642), Rene´ Descartes (1596–1650), David Hume (1711–1776), Immanuel Kant (1724–1804), Pierre-Simon de Laplace (1749–1827)), or some combination thereof (Chene, 2004). While there could be many potential types of determinism, only a few (such as theological, physical, environmental, cultural, sociological, technological) have received serious discussion (Faubion 2008) (though Ballinger (2008) would counter that most of these ‘determinisms’ are false, in that they are themselves influenced by outside sources, and, thus, not deterministic because they are not first causes). Advocates of strict determinism may consider the human mind to be indistinguishable from the human brain, a complex machine that produces a series of electro-chemical responses that would be predictable if all of the inputs could be known (Osborne 2005). 13 Physical determinism, a sub-category of determinism, builds on the principle of universal laws of motion and extrapolates these, to varying degrees, to the idea that all actions and consequences can be derived from the positions and motions of physical entities (Chene 2004). For example, according to Laplace, “Given for one instant an intelligence which could comprehend all the forces by which nature is animated and the respective situation of the beings who compose it...for it, nothing would be uncertain” (Laplace 1814/1951, p. 4 as quoted in Moxley 1999 p 100). Leibniz held that the universe, given a sufficient knowledge, was as predictable as the motions of billiard balls. (James Clerk Maxwell (1882-1969), on the other hand, countered that the predictability of simple, stable systems did not necessarily extend to complex unstable ones (Moxley 1999)). While the question of whether we live in a predetermined universe receives little attention among contemporary theorists, it has received much attention in the past. Philosophers who address physical determinsim have generally fallen into one of three camps: determinism (human actions result directly from universal laws), libertarianism (human actions result from uninfluenced free will), or compatibilism (universal laws are compatible with free will) (Honderich 2005, Ernste & Philo 2009). Like determinism, physical determinism carries a fatalistic implication and takes little interest in the social dimension, making it poorly suited to this study. Environmental determinism more directly focuses on the balance of influence between the environment and individual free will. For those who assume that the environment is subject to predictable forces, environmental determinism may be seen as a subcategory of physical determinism, (itself, as noted, a subcategory of determinism). As with physical determinism (and determinism generally), environmental determinism can be viewed from a deterministic, libertarian, or compatibilistic perspective (Ernste & Philo 2009). Although environmental determinism engages and draws from many fields, it has fallen most directly within the field of geography, “the field that has the longest sustained record of engagement with questions of human-environment relations” (Meyer & Guss 2017). Definitions of environmental determinism include arguments that social and cultural features such as creativity, productivity, and diversity result only from environmental factors (Thomas 2008, Johnston 2009), that “human existence and society, arguably including everything from settlement to language, (can) be determined by prior and external natural environmental conditions” (Ernste & Philo 2009, p 102), and that “human activity, culture, and physical and mental characteristics are, at once, informed and 14 inhibited by the geographical and climatic conditions of the physical environment” (Keighren 2015, p 720). However, according to Meyer and Guss (2017, p 5), environmental determinism need not be fatalistic, but can be defined as merely “treating the environment as a factor influencing human affairs independently and from the outside.” Others have provided similar non-fatalistic definitions, proclaiming that environmental determinism “treats the environment as a separate, simple cause or ‘factor’ not mediated by culture: something external to culture and influencing it from the outside” (Blaut 1993, 69 as quoted in Meyer and Guss 2017, p 6), or that it sees the natural environment as “an active factor exerting simple and direct causal influence on human life” (Platt 1948, 351 as quoted in Meyer and Guss 2017, p 6). Although ideas and mythologies relating the physical environment to social and cultural development date from antiquity, environmental determinism, as it developed in Western thought, traces its ancestry to such thinkers as Hippocrates of Cos (c.460–377 BC), Aristotle (384–322 BC), and Strabo of Amaseia (c.63 BC–AD 23) (Keighren 2015). Aristotle, seeking to explain the superiority of Greek civilization, suggested relationships among climate, race, and intelligence (Keighren 2015)—a suggestion that would later influence, and then taint, environmental determinism in the 19th and 20th centuries. More recently, the idea that climate and availability of natural resources strongly influence the evolution and capabilities of living organisms developed with the writings of Jean-Baptiste de Lamarck (1790–1869), Thomas Malthus (1766–1834), Charles Darwin (1809–1882), and Alfred Russell Wallace (1823–1913) (Thomas 2008). These writers strongly influenced early advocates of environmental determinism including Friedrich Ratzel (1844–1904), Ellen Churchill Semple (1863–1932), Ellesworth Huntington (1876–1947), and Griffith Taylor (1880–1963) (Fellman et al. 2009). While highly influential in the 19th and early 20th centuries, and with some notable recent exceptions (such as Diamond 2017), geographers and other social scientists have generally rejected environmental determinism since the mid-20th century (Keighren 2015, Meyer & Guss 2017). The primary point of contention for many has been the application of the theory, rather than newly-discovered flaws in its internal logic (though there have been some). For example, the idea that a scientific rationale explained the disparities in levels of civilization, prosperity, and sophistication of societies based on geography provided a justification in the minds of some for the exploitation and subjugation of ‘primitive’ societies (Thomas 2008). This rationalization resulted in an academic backlash against the theory behind it. 15 In addition to the exploitative applications of environmental determinism in the early 20th century, changes in dominant perceptions of the physical universe also eroded its perceived legitimacy. The rigidly causal form of determinism that borrowed legitimacy from a mechanistic, Newtonian view of the universe suffered from the indeterminate nature of quantum physics as it ascended to prominence in the mid-20th century (Ballinger 2008). It would no longer do to compare the determinacy of human actions to the determinacy of billiard balls if the billiard balls might take unpredictable routes. The theory has also been attacked for a lack of tangible evidence. Ewing et al. (2016), in their critique of studies claiming that the built environment affects travel behavior, note that most of these studies are cross-sectional and thus lack a theoretical basis for claiming causation. They further note that non-built environment factors (demographic, social, economic, etc.) may also be influential, or even exclusive, behavioral determinants. Of course, this has always been the primary counterargument, or, rather, counter position, to environmental determinism. Other counter positions include humanism, which argues that human ingenuity can overcome the natural environment, and materialism, which argues that societies and environments co-produce one another (Johnston 2009). A further challenge to environmental determinism is the varied nature of human response at both the group and individual level. Theories such as post-colonialism, feminism, and intersectionality (Crenshaw 1991, Crenshaw 1989) strongly challenge the concept that people react in a universal way to a given stimulus. Criticism of physical determinism may be mild (accepting the premise but claiming its effects are minimal), limited (for example, accepting that built environment effects are substantial but challenging the interpretation of the effects), or severe (claiming that built environment effects are insignificant or non-existent) (Jabareen & Zilberman 2017, Gans 2017). Franck (1984) points out four specific areas in which she considers physical determinism to be vulnerable to criticism, namely, 1) an exaggerated claim of influence of the built environment, 2) an assumption of only direct effects, 3) ignoring people’s capacity to exercise discretion, and 4) ignoring people’s ability to modify their environment. She suggests a remedy to such totalizing claims is to consider the influence of mitigating factors when investigating built environment influences. While a strict, fatalistic, version of environmental (or any other type of) determinism is unlikely to enjoy a renaissance, the basic concept of environmental influence is unlikely to 16 disappear entirely, either. Environmental determinism can still refer to environment/social relationships that are merely influential, or even mutually influential (Meyer & Guss 2017). Also, concerns about the effects of climate change on humanity have brought a renewed interest in environmental determinism, even if this specific phrase is not used and if the locus of concern has shifted from the field of human geography to the field of paleoanthropology (Livingstone 2012). Still, the emphasis on first cause and universal reaction leave environmental determinism more in need of qualification than I would prefer. Before we abandon it entirely, though, we should consider whether modified versions, such as possibilism and probabilism, reconcile its deficiencies sufficiently to adopt it for our purposes. Possibilism represented a counterpoint to environmental determinism. Advocated by writers such as Lucien Febvre (1878–1956), Paul Vidal de la Blache (1845–1918), Jean Brunhes (1869–1930), Isaiah Bowman (1878–1950) and Carl Sauer (1889–1975), possibilism emphasized free will over fatalism, yet retained an assumption of environmental influence (Fellman et al. 2009). It promoted the idea that, while environments offer a range of possibilities and opportunities from which people may choose, it is primarily human decisions and actions, rather than influences of the natural environment, that shape culture (Herbert 2014, Sullivan 2009, Johnston 2009). Possibilism was introduced by Vidal de la Blache in the late 1800’s as a framework for the field of geography that did not rely on strict environmental deterministic explanations of human development (Berdoulay 2009). It was further popularized by Febvre, who claimed that “there are no necessities, but everywhere possibilities; and man, as master of the possibilities, is the judge of their use” (Febvre 1932 p 27 as quoted in Johnston 2009, p 560). Despite its being a response to the increasingly unpopular theory of environmental determinism, possibilism failed to gain nearly as much attention as its rival, perhaps because geographers and other social scientists had abandoned the discussion entirely, and perhaps because “possibilism seemed to threaten the very raison d’etre of geographical study...by reducing it to...sociology with some locational reference” (Spate 1958). However, this study, like geography, is inextricably linked to location and the relationship of location and society, thus making possibilism (in addition to its lack of theorization) poorly suited as an underlying theory. Others who found possibilism lacking responded with the theory of probabilism, in a sense, the sysnthesis of possibilism (as antithesis) and environmental determinism (as thesis). 17 Probabilism is “a thesis about the relationship between culture and nature, which proposes that while the physical environment does not determine how human societies will react to its influence, it renders some responses more likely or probable than others.” (Johnston 2009) The concept of probabilism was introduced by O.H.K. Spate (1911-2000) in 1952 as a middle ground between determinism and possibilism. While he made a clear distinction between environmental determinism and the free-will-acknowledging probabilism, his distinction between probabilism and possibilism was less defined, the main argument being that not all possibilities are equal—the environment renders some options more probable than others (Flowerdew 2009). Spate criticized possibilist geographers of “writing sociology, without sociological techniques” (Flowerdew 2009 p 449) Although simplistic in summary, probabilism has room for nuance and application in several fields11. A probabilistic view of the world may be either subjective (epistemic) or objective (ontological) in that it may assume either a randomness based upon incomplete knowledge or one that is inherent in the workings of our universe (Duus-Otterstrom 2009). Empirically, both conditions look the same, but, theoretically, the difference is fundamental. For example, a researcher might observe a bus stop for several hours and notice that half of the passengers sat and half stood. However, she would have no way of knowing from her observation whether the presence of a bench predestined exactly half of all passengers to sit, whether it predestined a different percentage to sit that she would have discovered had she looked longer, whether it predestined a range of sitting percentages that included the percentage observed, whether the percentage were a necessary product of all factors involved at that particular time and location, or whether the observation were a simple fluke. The answer to what she would have actually observed could only be theoretical, not empirical. Despite reconciling many of the conflicts of both determinism and possibilism, probabilism suffered even greater disregard than its predecessors. There are many potential reasons that probabilism did not gain a wide acceptance at the time: It suffered some of the drawbacks of both previous theories without fully reconciling the failings of either; it was a 11 A strain of probabilism applies to the field of ethics, wherein one may be ethically compelled to act contrary to one’s conscience if the preponderance of one’s expert peers holds a belief different from one’s own (Schwartz 2014). Environmental risk managers may follow a deterministic rationale, evaluating all possible hazards solely upon their potential outcomes, or a probabilistic rationale, weighing hazards as products of both their potential outcome and their probability of occurrence (Basta 2014). 18 response to a response to an issue that had generally died in the minds of its primary guardians and thus spoke to an issue that no one still found compelling; it could be interpreted as simply a clarification of possibilism, rather than a competing theory; and, human effects on the environment were becoming a more immediate concern than the obverse (Flowerdew 2009). Also, it was not necessarily a repudiation of determinism since probability may, in fact, simply be determinism viewed at a larger scale. Duus-Otterstrom (2009) notes that “even if we settle for explaining patterns of outcomes, it might be that what on aggregate adds up to probabilism is the effect of complete determinism on the level of the individual case.” Finally, the move from determinism to probabilism does not necessarily make room for free will, since the odds of someone taking a certain action may be just as fixed as the certainty of the person doing so. Contrariwise, the probability of a number of people in a group holding a given opinion may not necessitate the probability of any member of the group believing it (Duus-Otterstrom 2009). Probabilism has the potential to be a useful tool in theorizing the relationship between society and environment. It can tap into the legacy of discussions surrounding its antecedents without evoking the viscerally antagonistic response associated with environmental determinism. However, it may also suffer from a lack of depth in its own right, saying nothing of real importance (such as that ‘some things make other things likely to happen’). Further, it still focuses on the unidirectional relationship of environmental effects on society, tending to ignore reciprocity. More useful to many studies involving the built environment would be an accounting of how cultures and their environments influence each other. For example, Alexander von Humbolt (1769-1859), in his book Kosmos, discussed the reciprocal and inter-related nature of society and the environment. He viewed nature as influencing, rather than determining, human actions and considered how human actions, in turn, might influence environmental systems (Keighren 2015). Unfortunately, this thinking did not spawn a following and, to date, the socio-spatial inter-relationship does not have a substantial body of theoretical literature, despite the tacit assumption of its existence in many fields, such as architecture. While probabilism reconciles some aspects of environmental determinism, it fails to address the reciprocal nature of society and the environment, and gives only marginal differentiation between the built and the natural environment, making it only marginally useful for this study. One field that attempts to address the relationship between humans and the built environment is “environment-behavior studies.” Rappoport (2008) discusses environment- 19 behavior studies and notes that it has gained little traction. He traces the origins of environmental-behavioral studies to the 1960’s, growing out of “dissatisfaction with the lack of knowledge about how people and environments interact,” and cites three primary questions that it seeks to address, namely, “(1) What bio-social, psychological, and cultural characteristics of human beings...influence characteristics of the built environment? (2) What effects do which aspects of which environments have on which groups of what sets of conditions, and why? (3) What are the mechanisms of these two-way interactions between people and environments?” (p 277) While there is considerable interest in environmental design research (as evidenced by such bodies as The Environmental Design Research Association (www.edra.org), the International Association of People-Environment Studies (iaps-association.org), and the Man-Environment Research Association (www.ebs-net.info) ), the field lacks a substantive body of theoretical literature (Rappoport 2008). This may be due, in part, to the highly interdisciplinary nature of the fields and topics involved in understanding the reciprocal relationship between humans and their built environment (Demsky & Mack 2008). It would be helpful to have a robust body of theory discussing the dialectic between society and its use of space. Soja (1980) discusses what he calls the Socio-Spatial Dialectic, but his interest focuses on Marxist spatial analysis rather than a more general discussion of the reciprocal influences on each other of society and urban space. So, if we can not find a body of theory that discusses a reciprocal relationship between society and the built environment, can we at least locate a discussion that emphasizes the built, rather than the natural, environment, even if it means a return to determinism? In fact, we can, if we turn to architectural determinism. Architectural determinism, a sub-category of environmental determinism, argues that the built environment influences behavior, either directly, through constraint or opportunity, or indirectly, through subliminal pedagogy or mnemonic devices—the later enjoying far less consensus than the former (Pop 2014). There is a robust history of architects and planners who attempted to influence society by means of the built environment. Jabareen & Zilberman (2017) trace modern interest in architectural determinism to architects and urban designers such as Clarence Perry (1872-1944), Le Corbusier (1887-1965), Walter Gropius (1883-1969), Frank 20 Lloyd Wright (1867-1959), and Ludwig Mies van der Rohe (1886-1969). Other notable figures in this realm include Frederick Law Olmsted (1822-1903), who sought to uplift the lower-class masses of New York City by providing them a venue (Central Park) in which to view their more refined urban counterparts, Daniel Burnham (1846-1912), whose City Beautiful movement was intended to purify society through the construction of elegant structures, and Ebenezer Howard (1850-1928), whose Garden Cities were to cure social ills and make society more productive (Riggs 2014). Although usually well-intended, architectural determinism has also had questionable applications, such as the attempt of Jesuit priests in 19th-century Montana to change the culture of the local American Indian population through the use of architectural and spatial interventions (Van west 1987). Despite widespread acceptance among practitioners of the behavioral and perceptional effects of design12, the theory of architectural determinism is backed by little empirical data, largely because post-occupancy evaluations are rarely commissioned and thus purported design benefits are rarely verified (Marmot 2002). While some studies, such as Newman’s (1973) investigation of tenement buildings’ negative effects on residents, make a strong argument for the validity of environmental influence, others, such as Atlas’s (1982) study of the relationship between spatial and architectural factors and rates of violence in prisons, find that cultural variation seems to have greater explanatory power than do living conditions. Thus, not only does the degree of influence of environmental design on behavior remain unresolved, so does the question of whether this influence even exists. Finally, there is the ethical imperative to consider--architects and planners often proceed under the general assumption that thoughtful and skillful adjustments to the built environment can make the world and the people in it better off (Gans 2017, Lang & Moleski 2010), but this requires the dual assumptions that people’s lives need improvement and that it falls to the architect or planner to effect this improvement (Broady 1966, Simon 2016). Its failings (a lack of theoretical articulation, a paternalistic legacy (and perhaps a paternalistic nature), a lack of serious discussion and debate, a lack of empirical findings, and a lack of successful application) notwithstanding, architectural determinism comes closer than competing theories to describing the relationship between the built environment and society that 12 According to Broady (1966, p173), architectural determinism is “more often found implicit in architects’ thinking than in any clearly argued form: and it is probably the more dangerous for that.” 21 I am seeking to explore. It might be useful to combine the ideas of architectural determinism and probabilism, but architectural probabilism might not afford any insights that could not be accommodated by architectural determinism (which carries no significant historical burden of fatalism). It would also be more helpful to have a theory that considers the wider built environment (architecture, to me, connotes a limitation to buildings) and that addresses the type of reciprocity between people and the environment that Rapporport considers. But, again, such a theoretical discussion seems to be lacking in the literature. Therefore, within the context of the above critiques, considerations, and disclaimers, it seems that architectural determinism may be an appropriate existing theoretical framework to use to consider the effects of urban density on people’s perception of their neighborhood and of their quality of life. Is it also appropriate as a framework for considering sense of community? While this study will involve several variables (sense of community, perception of density, fear of crime, etc.) each with its own body of related theories, the primary issue it will seek to address involves the relationship between the built environment and people’s individual and collective responses to it. Several studies suggest connections between built environment features and social interaction (Talen 2000). A review by Talen (2000) of planning documents of twenty major U.S. cities found that such documentation showed a general acceptance that the built environment could increase sense of community through facilitated social interaction. However, the mechanisms by which the built environment may influence sense of community remain poorly theorized (Moustafa 2009) and researchers have found little empirical evidence linking any specific feature of the built environment to any specific component of sense of community (French et al. 2014). As Kingston et al. (1999) note, it may be that sense of community is determined solely by individual characteristics, such as a personal desire for interaction, or by socioeconomic status, rather than directly by any environmental factor. Parsing which factors contribute to sense of community is difficult, and attributing the portion for which the built environment is responsible is even harder. Researchers have produced inconclusive and contradictory assessments as to whether the physical environment can affect sense of community at all (Jung et al. 2015; Ebrahim 2015). Despite a growing body of popular and academic literature linking New Urbanist design principles to sense of community, there remains a paucity of empirical evidence to support this link (Talen 1999). Even if urban design can influence social interaction, it is unclear how much 22 social interaction alone influences sense of community (Talen 1999). Talen (2000) highlights three practical limitations to the link between physical design and aspects of community: 1) most research examines effects on social interaction as a proxy for sense of community rather than on sense of community directly, 2) most research has focused on the scale of sites rather than of neighborhoods, and 3) most research considers only indirect effects of the built environment rather than aspects of the built environment directly. Supporting this last point, a study by French et al. (2014) showed that residents’ perception of their neighborhoods were more closely associated with their sense of community than were objective measures of environmental characteristics (see also Francis et al. 2012). Also, neighborhood design elements that do increase sense of community may do so indirectly by encouraging a homogeneous population rather than directly by facilitating interaction (Talen 1999). While some studies seeking to understand psychological effects of physical typologies fail to account for non-physical factors (Jabareen & Zilberman 2017), other studies emphasize them, supporting the notion that neighborhood residents have been ‘liberated’ from the need to make social connections within their neighborhood (Talen 1999). Given the multiple proposed components of sense of community, including shared emotional connection, neighborhood attachment, membership, influence, reinforcement, and sense of place, it is unclear to what extent these all might be affected by simply facilitating random encounters among residents with strategically placed public space (Talen 1999). A study by Jabareen & Zilberman (2017) found 13 percent of variation in sense of community due to physical typologies (design, compactness and transportation), 13 percent due to a demographic factor (length of residence), and 19 percent due to the socio-cultural perception of trust. Expectations of increasing sense of community by providing nearby social space may be misguided if they fail to predict residents’ preferred methods of finding companionship and associated barriers to doing so (Broady 1966). Notions of spatial determinism that predict an association between sociability and proximity presume that residents put a high “spatial cost” on relationships that are far away, and this may not be the case (Talen 1999). While architectural determinism may provide a venue for such criticisms, it does little to provide a meaningful response to them. *** 23 Another way of conceptualizing the relationship between the environment and users of the environment was proposed by James Gibson in the 1970’s with his theory of affordance, which moved away from determinism by imbuing animals with agency (Withagen et al. 2012). Gibson rejected the behaviorist idea, popular at the time, that animals had little choice in how they reacted to their environments. He suggested, instead, that objects in the environment provide various opportunities to animals that entice them to respondent actions. He extended this theory to humans and noted that, particularly with humans, this relationship could be reciprocal. “Why has man changed the shapes and substances of his environment?” he asked. “To change what it affords him.” (Gibson 2015, p122) But, what, exactly is an affordance? What does the term “affordance” mean? Gibson introduced the term “affordance” in 1979, in his book, “The ecological approach to visual perception.” According to Gibson, “The affordances of the environment are what it offers the animal, what it provides or furnishes, either for good or ill. The verb to afford is found in the dictionary, but the noun affordance is not. I have made it up. I mean by it something that refers to both the environment and the animal in a way that no existing term does. It implies the complementarity of the animal and the environment.” (Gibson 2015, p119, emphasis his) Although researchers universally credit Gibson with coining the term, they have not all accepted his definition without modification. According to Norman (2013), affordances are relationships between physical objects and people. Evans et al. (2017, p. 39) also consider affordances to “belong...to the relationship between individuals and their perceptions of environments.” There seems to be a general agreement that “an affordance indicates the potential for a behavior, but not the actual occurrence of that behavior.” (Maier et al. 2009) There is a lack of consensus, however, as to whether affordances are always helpful. Maier et al. (2009), like Gibson, believe that affordances can be both positive and negative, but Norman (2013) considers objects that prevent or hinder activity to be constraints (though he also refers to them as ‘anti-affordances’). Mehan (2017) explicitly extends the influence of affordances beyond activity by arguing that the environment may offer several types of affordances, including physical, social, emotional, and 24 cognitive13. Still, a commonality among all researchers seems to be an agreement that affordances represent ways that the environment influences users of the environment. Of course, this influence can only take place if the users can perceive the affordances around them, and perception has been intimately linked to the concept of affordance from the outset. In fact, both Gibson and Norman were primarily interested in the issue of perception. As noted above, when Gibson introduced the idea of affordance, it was within the context of perception. He claimed that “the composition and layout of surfaces constitute what they afford...to perceive them is to perceive what they afford” (Gibson 2015, p119), and that “what we perceive when we look at objects are their affordances.” (Gibson 2015, p126) Norman also was keenly interested in the role of perception in the user/environment relationship as well, though he rejected Gibson’s assertion that animals perceive affordances directly. Instead, he believed that “affordancs are not mere opportunities for action, but are perceived action possibilities that suggest actions to an animal.” (Withagen et al. 2012, p. 253, emphasis theirs) Norman (2013) believes that the affordance of some objects can be perceived due to the nature of the object, but that other objects require ‘signifiers’ to make their affordance known. “Affordances determine what actions are possible. Signifiers communicate where the action should take place.” (Norman 2013, p. 14) Do we perceive affordances directly, or do we perceive signifiers and assume that the signifiers represent affordances? Gibson and Norman did not agree on that point. This disagreement regarding perception is indicative of a more profound lack of consensus regarding the very nature, or ontology, of the concept of affordance. As Greeno asks, “Is the affordance that a chair provides for sitting a property of the chair, a property of the person who sits on it or perceives that he or she could sit on it, or something else?” (Greeno 1994, p. 340) This question is fundamental to our conception of how affordances work. In contrast to Gestalt psychologists like Lewin and Koffka, Gibson believed that affordances exist independently of perception (Withagen et al. 2012). He thought that “’values’ and ‘meanings’ of things in the environment can be directly perceived” and that “values and meanings are external 13 Similarly, Montello (2014, p75) posits that “like other physical environments, architecture influences human cognition, experience and behavior by allowing, facilitating, requiring, impeding or preventing various perceptions, thoughts, emotions and acts.” 25 to the perceiver.” (Gibson 2015, p119) He gives us a somewhat ‘quantum’, or, at least, ambiguous description of how he viewed the nature of affordance: “An important fact about the affordances of the environment is that they are in a sense objective, real, and physical, unlike values and meanings, which are often supposed to be subjective, phenomenal, and mental. But, actually, an affordance is neither an objective property nor a subjective property; or it is both if you like. An affordance cuts across the dichotomy of subjective-objective and helps us to understand its inadequacy. It is equally a fact of the environment and a fact of behavior. It is both physical and psychical, yet neither. An affordance points both ways, to the environment and to the observer.” (Gibson 2015, p121) Subsequent researchers have supported Gibson’s contention that affordances are objective realities, detached from the perception of users. Withagen et al. (2012) agree with Gibson (and not Norman) that affordances exist even if they are not perceived. They further agree with Gibson that “affordances do not change as the intentions or needs of the actor change (p. 255). Rietveld and Kiverstein also agree, commenting that “affordances are real... in much the same way as colors are real. Both are there independent of any particular individual’s action.” (Rietveld and Kiverstein 2014, p. 338) The main proponent of the subjective nature of affordances seems to be Norman. Despite the enthusiasm generated by Gibson’s idea, few researchers have done much to advance the theory of affordance or to use it as a theoretical framework in the realm of environmental design. Maier et al. (2009) provide the most comprehensive attempt to apply the concept of affordance to architecture and to explicitly discuss this application as theory. Unfortunately, no one seems to have applied their ideas empirically. Mehan (2017) considers the concept of affordance in his discussion of the public realm and Coolen (2015) suggests using a matrix of affordances as a means of cataloging and evaluating user housing preferences, but neither one tests nor advances the theory. The only example I found of someone using the concept of architectural affordance as a research framework was Bichard (2015), who used it in her thesis exploring publicly accessible toilets. Bichard presents affordance as a more ‘elastic’ concept than determinism, but makes no effort to elucidate architectural affordance as a theory or to reference any related discussion of the topic other than Gibson. 26 Thus, the concept of affordance generally and the application of this concept specifically to the influence of the built environment on human behavior remains largely un-theorized. It certainly lacks consensus. Evans et al. (2017) reviewed 82 communication-oriented scholarly works on the topic of affordances and found little consistency among them on how they applied the term. This leaves open for interpretation (and explication) how, indeed, affordance—if it is real—works. Norman (2013) believes there are six key elements involved in our ability to discover and understand our environment: affordances, signifiers, constraints, mappings, feedback, and the conceptual model of the environmental system. While this may be so, I think we only need a few of these items to describe a system of affordance. I think the flow of influence from environment to environmental user might be represented like this: environment > elements > affordances/ constraints > agency/desires/ limitations > user > feeling/ actions signifiers > perceptions > Table 1 - Chain of influence from environment to user What does this table mean? We can imagine that there is some environment (say, an urban park) with elements in it (say, benches). The elements provide affordances (flat surfaces that afford sitting) and constraints (armrests that make sleeping difficult). These elements also provide signifiers (shape, color, texture, material) that offer clues as to their utility. Within the environment are users who have both agencies (abilities to do things) and desires (such as to sit or sleep). These users also form perceptions about their environment and the elements within it. If their perceptions of the signifiers align with their agency and desires, these signifiers may influence them to take advantage of the affordances (and avoid the constraints) of the elements in the environment. This influence may motivate them to take some action. If it does, the built environment has thus ‘caused’ behavior. It is this series of connections, this application of affordance to the built environment, that I think we can refer to as “architectural affordance.” What is architectural affordance? I have not found an explicit definition (though Maier et al. (2009) provide thoughtful discussion), so I wish to provide one here: Architectural affordance is a theoretical framework that posits that the built environment influences human perception and behavior by providing both affordances (opportunities and encouragement to experience some feeling or perform some action) and constraints 27 (corresponding limitations or discouragement). The environment also contains signifiers14, some of which correspond to its affordances and constraints, and some of which are perceived by environmental users. These users relate their perception of these affordances and constraints (by perceiving the related signifiers) to their own agency (ability and capacity to feel and act) and desires and then respond with a modified feeling and by acting within the limits of the existing constraints. In this way, the built environment (as with the environment generally) influences human perception and behavior. This definition is not based on observation, but it is rather a working definition that I intend to describe my understanding of how architectural affordance works. I think it raises several questions in need of clarity. Those questions (and my answers) follow: • Are affordances always positive? o Yes. I believe affordances are always positive (‘helpful’) to the organism (in this case, the human user) to which they afford a feeling or behavior. I reject the idea that affordances can be negative and I reject the idea of “anti-affordances.” Instead, I believe the opposite of affordance is constraint. I think that constraints can still fall under the theory of architectural affordance, however, because, generally, the intent of the built environment is to afford action rather than prevent it. Some objects (multiple armrests on a bench) may act as both affordances for some activities (resting arms while sitting) and constraints for others (sleeping on the bench). • Are affordances inherent? o No. I reject the idea that affordances are inherent. This is a minority viewpoint, but it is shared by Norman. I believe the idea that affordances are inherent properties of objects (like color or texture) is easily countered. Consider a bench. Is the affordance of ‘sitting’ an inherent property of the bench? I say no. It affords no opportunity for sitting to a giraffe or a whale or a person in a wheelchair. It affords no opportunity for anything if it is on the moon or anywhere that users can not apprehend it. There may be an infinite number of 14 “Affordances determine what actions are possible. Signifiers communicate where the action should take place.... Signifiers can be deliberate and intentional, such as the sign PUSH on a door, but they may also be accidental and unintentional, such as our use of the visible trail made by previous people walking through a field or over a snow-covered terrain to determine the best path.” (Norman 2013, p 14) 28 affordances associated with the bench (or none), depending on the life forms and personalities that encounter it. But, if you throw the bench into the middle of the ocean, its affordances do not sink with it. They simply cease to exist, and new ones arise as it settles into its new home. • Are affordances perceived directly? o No. Here, again, I disagree with Gibson (and Withagen et al. and Rietveld & Kiverstein) and embrace Norman. I contend that affordances are not perceived at all. In order to benefit from the affordance of a bench, one must perceive that it is associated with an agency and a desire (like sitting, or laying down, or performing a skateboard trick). But users do not perceive the affordance before they make use of it. In order to use the affordance, users must first perceive the signifier. They must see that the thing that looks like a bench affords sitting. Of course, the bench may not. It may be wet with water or paint. It may be behind a barrier, such as a fence. You may need to buy a coffee in order to sit in that bench. The bench may be hidden from view and need an explicit signifier such as a sign to point out that sitting is available nearby. Also, there may be a boulder next to the bench that would afford sitting just as well as the bench, but the user would have to interpret a signifier associated with the boulder to mean that it affords this activity. And, the bench may afford many other activities that would only be perceived if suggested by signifiers. This brings us to the next question, which is related. • Do signifiers align with affordances and constraints15? o Sometimes. Signifiers tend to align with affordances and constraints, but there is imperfect overlap. A ramp may be perceived as an affordance for a wheelchair user, but it may still restrict some wheelchair users and it might afford passage for people making deliveries with a dolly. A gate may signify to users that they are not allowed to pass, but it may keep out people who should pass or allow people 15 Although beyond the scope of the current discussion, note that this disconnection between signifier and affordance, which I present as primarily functional, harmonizes with the post-modernist aesthetic disconnection between object and symbol in architecture as popularized by Venturi el at. (1972). 29 who shouldn’t. Thus, the ‘signifiers’ of the ramp and the gate may align with the related affordances and constraints either well or poorly. • Do users consciously perceive signifiers? o Sometimes. When users see a “don’t walk” sign, they likely process this on a fairly low level, changing their behavior (walking) without much thought. Another signifier, such as the smell of smoke, may trigger a much higher level of awareness and thought. A feeling of calm in a natural setting or a feeling of reverence in a large cathedral may not be a conscious reaction at all, or such feelings may be heightened by reflection. • Do affordances and constraints influence perception? o Yes. I agree with Mehan (2017) that affordances may influence many aspects of human perception and behavior, including physical, social, emotional, and cognitive conditions. Affordances and constraints, or the signifiers associated with them, may influence both what we do and how we view the world. • What is the range of influence of affordances and constraints? o The built environment offers a full range of influence, from minimal (as an opportunity for a place to sit) to complete (as a jail cell). By their nature, affordances tend to be optional (things you can do) and constraints tend to be mandatory (things you can’t do). Influence may be either directly by affordances and constraints, or indirectly through related signifiers. • Can we prove a causal pathway? o No. Unfortunately, we have still not resolved the conundrum with which we began our discussion, namely, Can the built environment cause human behavior, and, if so, how? We have, however, suggested a pathway for causality, which puts us, at least, a bit better off than when we started. These may not be the right answers. The true nature of the relationship between the built environment and human behavior might be nothing like this. It might not even exist. But these are the assumptions I adopt as I present my research. To close this discussion, I think it is useful to compare the words “facility” and “facilitate.” Seeing them together makes me think that the purpose of the built environment (echoing Gibson above) is to make it easy for us to accomplish things. The built environment is 30 built purposefully to make it easier to stay dry, to sleep, to eat, to learn, to work, and to do whatever else we need to do. Facilities facilitate. Architecture affords. That’s why we build it. Several researchers have noted the lack of a working theoretical framework in the field of architecture (Gibson 2015, Maier et al. 2009, Coolen 2015, Broady 1966). The nascent concept of architectural affordance may be able to fill this gap. Using this theoretical framework, therefore, I wish to explore the primary research question noted at the outset, namely, “How does population density relate to residents’ sense of neighborhood community?”, along with its derivative questions. I see this question fitting into a hierarchy of human-environment relationships as follow, from general to specific: ➢ How does the built environment influence human behavior and experience? o How does the design of a neighborhood affect residents’ quality of life? ▪ How does urban density relate to residents’ sense of neighborhood community? • How does public space in a neighborhood (among other factors) moderate this relationship? To prepare for this research, I spend the balance of this literature review discussing the relevant topics of sense of community, urban density, and public space. What is sense of community? The concept of community as used by sociologists, community psychologists, and urban planners relates to a group of people (possibly with some other defining characteristics added). What, then, does it mean to have a sense of this thing? Does it mean that one senses that a group exists? This definition would be insufficient to merit the level of attention the phrase ‘sense of community’ has garnered over the last few decades. As a sense, of course, it requires a consciousness to sense it. It may, however, be an ubiquitous sense with a nature that many people can agree upon (like the color blue) such that a group may share a common sense of community and have some expectation that its members are experiencing approximately the same feeling. Or, it may be that this sense is experienced differently (if at all) by each person. When people speak of ‘building’ or ‘strengthening’ a community, they may be referring to modifying the perception individual members have of their community, rather than increasing a community’s numbers or influence (though they could mean these things as well). It is this 31 perception that people have of their community, usually held to be a positive perception, that is usually referred to in related literature as sense of community. However, the definition, and the theory behind it, deserve a more nuanced consideration. Most definitions of sense of community involve some combination of notions of belonging, membership, interdependence, support, connection, commitment, empowerment, sharing, and participation, though they may or may not involve location (Ebrahim 2015). Psychological sense of community generally refers to “how an individual perceives his or her bond to a community and the intensity of these ties to the community” (Halamová 2016). According to Sarason (1974 p157) sense of community involves "the perception of similarity to others, an acknowledged interdependence with others, a willingness to maintain this interdependence by giving to or doing for others what one expects from them, (and) the feeling that one is part of a larger dependable and stable structure." Talen (2000, p174) defines sense of community as “the interrelationship between the individual and the individual’s social structure.” McMillan & Chavis (1986 p9) state that sense of community is “a feeling that members have of belonging, a feeling that members matter to one another and to the group, and a shared faith that members’ needs will be met through their commitment to be together.” In a study of sense of community and New Urbansim, Ebrahim (2015 p26) defines sense of community as “social attachment and togetherness experienced by neighbourhood residents and an attachment to place where the environmental experience of this togetherness happens and people’s needs could be met.” Cochrun (1994 p93) describes sense of community as a psychological construct that refers to “the feeling an individual has about belonging to a group and involves the strength of the attachment people feel for their communities or neighborhoods.” Researchers have associated sense of community with neighboring behaviors, political efficacy, walkability, intended length of residence, neighborhood satisfaction, safety, control over one’s environment, and community bonds (Johnson & Halegoua 2015). Sense of community is related to “neighboring,” which Unger & Wandersman (1985) describe as involving “the social interaction, the symbolic interaction, and the attachment of individuals with the people living around them and the place in which they live.” Sense of community is also similar to attachment to place, but with an emphasis on people rather than on location (Unger & Wandersman 1985). In fact, the importance of neighborhood may be contested in studies that focus on virtual sense of community (Abfalter et al. 2012), multiple senses of community (Bahl et al. 2019), or school 32 sense of community (Prati et al. 2017, Prati et al. 2018, Prati & Cicognani 2019). People with a sense of community feel that they are part of, connected to, and committed to a community whose goals they recognize and are motivated to work together to achieve (Ebrahim 2015). Such motivation is often the focus of studies that consider the related topic of sense of community responsibility (Boyd & Nowell 2020, Yang et al. 2020). Interest in the concept of sense of community has its roots in feelings of dissociation and alienation associated with 19th century industrialization (Halamová 2016). Literature in the field of community psychology tends to present an idealized notion of sense of community that emphasizes positive community involvement and social support structures (Moustafa 2009). In small town or village settings, sense of community may be an expected by-product of residents’ familiarity, shared history, homogeneity, and length of residency, whereas in modern cosmopolitan settings many of these elements may be lacking (Cochrun 1994). Yet, such elements may be more than quaint provincial trappings and may represent actual needs of the human psyche. Baumeister & Leary (1995) reviewed a body of empirical literature to test commonly accepted theories of the human need to belong and found that humans do, indeed, have a fundamental need to belong and that we seek frequent interactions within long-term, caring relationships. Such relationships build social capital. Long & Perkins (2007) propose that sense of community is one of the four elements comprising social capital, along with collective efficacy, neighboring behavior, and formal citizen participation, and that sense of community is the best predictor of these other three elements. Chavis & Wandersman (1990) posit that sense of community can act as a catalyst in community involvement by mobilizing members’ perceptions of their physical environment, of their community relationships, and of their own level of empowerment in the community. McMillan (1996) emphasizes the need for community members to feel safe and rewarded by being in the community. He notes that once a group is confident in their similarities and shared goals, the members may then feel comfortable negotiating resolutions to their differences. Researchers have theorized the structure of sense of community. McMillan & Chavis (1986) propose four elements that form the amalgam of sense of community, namely, “membership,” “influence,” “integration and fulfillment of needs,” and “shared emotional connection” (see also Chavis et al. 1986). McMillan (1996) later reframed the four elements of sense of community using more emotive descriptors categorized as “Spirit, Trust, Trade, and Art.” Chavis et al. (1986) formulated a sense of community index (SCI) as a 33 means to empirically quantify the components of sense of community. Several researchers have used some variant of this index to compare sense of community with other variables. It remains, in one form or another, the most commonly used metric of sense of community found in current literature (but see Appendix ‘A’ for a comprehensive discussion of sense of community measures). Many elements may affect, and be affected by, sense of community.16 Studies have associated sense of community with social engagement (Wells et al. 2019, Tang et al. 2017, Talo et al. 2014, Miranti & Evans 2019, Dinnie & Fischer 2020), life satisfaction (Hombrados-Mendieta et al. 2019, Ditchman et al. 2017), empowerment (Ramos-Vidal et al. 2019), and well-being (Prati et al. 2018, Rollero et al. 2014, Jorgensen et al. 2010, Coulombe & Krzesni 2019, Moustafa 2009). Several studies have shown a strong relationship between quality of life and neighborhood social connections (Talen 2000). Researchers have associated sense of community with benefits at many scales, including the individual (better mental and physical health and higher quality of life), the community (increased pro-social behavior and cooperation), and society in general (greater interest and involvement in civic affairs) (Halamová 2016). Studies have associated sense of community with feelings of safety, self-efficacy, and well being and actions such as volunteering, community participation, voting, and helping others (Sense of community Partners 2004). A study by Davidson & Cotter (1991, referenced in Cochrun (1994)) found that sense of community was associated with more happiness, less anxiety, and greater perceived personal life competency (also Farahani 2016). A potential outcome of sense of community is social support, which can be emotional, functional, or informational (Unger & Wandersman 1985). A study by Forsyth et al. (2015) showed a positive relationship between sense of community and environmental engagement. Davidson & Cotter (1991), found strong positive relationships between sense of community and happiness (r = 0.45, 0.19, 0.34), but weaker relationships with worrying (r = 0.06, 0.11, 0.12) and coping (r = 0.16, .016, 0.17). A study by Gattino et al. (2013) found sense of community to be positively associated with the World Health Organization Quality of Life index. 16 In this section, I discuss relationships in which sense of community tends to be thought of as influencing something else. In the section ‘How does urban design influence sense of community,’ I discuss relationships in which the direction of influence tends to be considered as opposite. Of course, these are just organizational conventions. My point in neither section is to establish causality or suggest uni-directionality. I just want to show what people have researched. 34 Ahmad et al. (2016) found that community projects were more likely to succeed when members felt empowered and had a strong sense of community. In more targeted studies, researchers have associated sense of community among adolescents with such positive outcomes as a more solidified ethnic identity, increased access to positive adult mentors, a reduced tendency to engage in destructive behaviors, increased psychological and social well being, and an increased drive to reduce common problems (Lardier et al. 2017). Mendoza et al. (2016), found sense of community to be the strongest predictor of a college student’s tendency to thrive in the campus environment. Farahani (2016) claims that the advantage of neighborhood sense of community is not an ability to provide the highest levels of intimacy but rather the benefits of access and proximity. An argument in favor of this benefit comes from a study that found that socially isolated residents during a 1995 Chicago heat wave were seven times more likely to die from heat exposure than were those with some social network (Montgomery 2013). People with a strong sense of community tend to have healthy feelings of belonging, control over their environment, shared history with fellow members, personal investment in community success, and conviction that their needs can be met through the collective abilities of their community (Cochrun 1994). While it may be tempting to romanticize sense of community, it would be naïve to imagine that it could never create or exacerbate negative outcomes (Sarason 1974). Communities may be founded, consciously or unconsciously, on constructive ideals such as faith, hope, and tolerance, or on destructive ones, such as fear, hatred, and rigidity (McMillan & Chavis 1986). It is often possible to exploit social cohesion and social capital to nefarious ends (Putnam 2000). The unity of a group is in no way a guarantee of good intentions, harmless actions, or immunity from deception. Even with best intentions, a community member’s sense of community must correspond to the nature and values of the community. Some members may consider identification or association with the community to represent a compromise of their personal values or even a reduction in their safety (Sense of community Partners 2004). The conflicts between members’ values and their understanding of the community’s values may range in severity or may be themselves conflicted (some values may be in harmony, some in minor conflict, and some in fundamental conflict). In a study by Walker & Ravel (2017), the authors interviewed undergraduate students from rural towns about their home communities. They found that the students generally felt a strong sense of community in their home towns, but 35 had felt the communities lacked diversity and access to opportunity. The authors speculated that the students may have felt some obligation to remain in their home towns to help preserve the community. In other cases, residents may develop a negative sense of community in neighborhoods they consider to be more of a threat than a resource (Pretty et al. 2006). In addition to conflicts of values, sense of community may have negative effects if the community is in harmony but built on values that are harmful to society at large (such as racism or drug smuggling). In forming a sense of community, it is important to question whether it is based on exclusion of some members of the community and what types of diversity the community may not tolerate (Halamová 2016). Developing a sense of community may be dangerous if members of the community in question have unsupportive or predatory values (Halamová 2016). Such values may surface more readily in cases of severe heterogeneity and a perceived lack of resources. For example, an influx of immigrants may pose challenges to the sense of community of both immigrants and the established community into which they enter (Pretty et al. 2006). A minority group’s sense of community may be used against its members by outsiders who are antagonistic toward it because of its ethnic, religious, cultural, or political makeup (Pretty et al. 2006). Close-knit, morally homogeneous neighborhoods may prove harshly judgmental of those it perceives as deviants (Unger & Wandersman 1985). While sense of community has many benefits and is generally perceived as benign, there are darker aspects that we should acknowledge. Since the phrase ‘sense of community’ was introduced by Sarason in the 1970’s, it has come into popular use and is often used by planners and developers as a positive aspect associated with a location. While I have attempted to show a comprehensive range of definitions and applications of sense of community in recent literature, the purpose of this larger research project is not to engage in disambiguation of the term or to explore all applications of it. Rather, the focus, as noted earlier, is on the relationship between sense of community, as broadly defined above, and urban density, as broadly defined below. How do we react to urban density? Increased density and compact urban development have become widely accepted goals among city planners and urban design professionals in North American cities. Development, planning, and environmental organizations such as the American Planning Association, the Urban Land 36 Institute, the Congress for New Urbanism, the Natural Defenses Resource Council, and the United States Green Building Council promote compact city and smart growth goals, including increased urban densities. This unanimity of emphasis on increasing density is striking in view of the history of urban planning. From the early to mid twentieth century, urban planners were quite intent on solving the problem of urban density rather than promoting it. It has only been in the last few decades that urban density has become regarded as something of a panacea for many social and environmental ills. In fact, so closely has density become aligned with ecological sustainability that many urban designers consider compact neighborhood design to be a fundamental aspect of sustainable urbanism. Some purported benefits of density include more land for biodiversity and human access to nature (Hester 2006; Farr 2007). Density may also have indirect benefits. Calthorpe (2011) suggests a chain of personal, societal, and environmental benefits stemming from urban density, with dense environments leading to reduced auto use, which leads to reduced pollution and more walking, which lead to better health and stronger communities. While many of these benefits may be achieved with low densities, high density environments often offer economies of scale that allow market forces to align with economic and environmental objectives. Some success stories are available. For example, by using density to encourage transit use and walkable neighborhoods, Portland, Oregon has preserved farmland, increased housing options, and reduced per capita vehicle miles traveled (Calthorpe 2011). Other cities have used design standards in industry guidelines that include minimum residential and commercial density requirements, such as LEED for Neighborhood Development, to shape density policy. While it is clear that density has many advocates, it is useful to consider what it is about density that they find so appealing. To do so, we will discuss the meaning of urban density and the advantages and disadvantages of increasing it. How is density an issue for city planners? Density is not new. While the modern professions of city planning and urban design are only a few generations old, density has been a part of urban structure since antiquity. The most current framework for conceptualizing urban density is the compact city movement (Randolph & Tice 2013). Therefore, it is useful to understand current issues of density through this lens. 37 Density is essential to the compact city. Compact cities are attempts to accommodate more people with less land and fewer resources. This does not necessarily involve a uniform increase in density everywhere. New Urbanist ideals call for concentrated density around transit nodes, but not high density generally (Churchman 1999). According to Ewing et al. (2015) compact development includes medium to high densities with “strong centers,” mixed land uses, and contiguity with existing development. Compact city designs include dense, mixed-use development with an interconnected street network that facilitates mobility by transit, walking, and cycling (Lehmann 2016; Tian et al. 2015). While these features are primarily intended to leverage the resource efficiencies that high density affords, other motives relate to transportation and social interaction. Compact city principles call for urban growth boundaries and higher residential densities as a means to reduce auto use (Churchman 1999). Also, dense urban neighborhoods may offer greater opportunities to share knowledge through face-to-face interactions (Moroni 2016). While some of these ideas may appear novel in the context of modern North American development, Neuman (2005) reminds us that the term “compact city” is a redundancy that only has meaning when contrasted with the term “sprawl.” Indeed, cities have traditionally been defined by, or at least characterized by, density, until personal, affordable automobility enabled the more dispersed settlement patterns known as suburbia. The compact city movement is less a celebration of urbanism than it is a rejection of suburbia, with the critical difference involving density. But, density itself has no intrinsic value. It is not a quality like ‘happiness’ or ‘prosperity’ whose inherent value increases with quantity. It is simply a result of certain market forces and policies that create, or are associated with, certain conditions. (Moroni 2016). Still, to understand its associated values, we should understand what urban density means. Definitions of density The dictionary definitions of the words “compact” and “dense” make their relationship unclear.17 In general use, their meanings are similar (having or made of things that are close 17 The Concise Oxford English Dictionary defines “compact” as “adj. 1 closely and neatly packed together; dense.>having all the necessary components or features neatly fitted into a small space.” and “dense” as “adj. 1 closely compacted in substance. >crowded closely together.” Merriam Webster’s Dictionary of Synonyms contrasts the words “compact” and “dense” this way: “Dense applies to something in which the arrangement of parts or units is exceedingly close....The term commonly implies impenetrability and an extended use may lose the basic notion of close packing of parts. Compact suggests close and firm union or consolidation of parts, especially within a small compass; it often also implies neat or effective arrangement.” 38 together), with “compact” carrying an aura of neatness about it. However, each is a relative term and neither is particularly precise. This base ambiguity is compounded in the application of the term “density” to describe the ratio of human beings to land area. This ambiguity remains at all scales of measurement and has led to confusion and need for interpretation. There is no universally accepted measure for neighborhood density, making comparisons between studies difficult (Regoeczi 2003). Boyko & Cooper (2011) identify 23 working definitions of density and argue that density policy can be very complex. Such complexity may be lost, however, when descriptors of density as high, medium, or low, are used without specifying thresholds. Yet even with thresholds, whether arbitrary, contrived, or based upon some rationale, definitions of density may vary widely according to cultural, political, and geographical regions (Churchman 1999 p399). Three common ways of discussing density as it affects people’s life are as 1) a simple ratio of persons or dwelling units per area, 2) as perceived density (the range of subjective reactions to density), and 3) as crowding, a negative reaction to perceived density (Churchman 1999). Common measures of density include “Net Dwelling Density” (dwelling units per area of residential land), “Gross Residential Density” (persons, households, or dwelling units per residential area, including streets), “Neighborhood Density” (persons, households, or dwelling units per area of land used for residential or community purposes), and “City Density” (using city limits as the denominator) (Alexander 1993). Some factors involved in calculations of density include dwelling form, dwelling size, lot size, block configuration, measurements used and methods used to take the measurements (Alexander 1993). However, measures of density are often ill-defined because it may be unclear which area is included in the denominator (Churchman 1999). Even when the area is clear, the density measure may have little applicability due to variation within the area. For example, while densities within urban boundaries may be instructive, metropolitan area densities may be quite meaningless because they involve both urban areas and rural areas (Demographia 2017). In addition to the complications of deciding how to structure the numerator and denominator of the urban density equation, the resulting ratio may have limited applicability for planning and policy purposes. This is because perceived density, people’s experience of, and reaction to, density, is, ultimately, more important than net or gross measures of density, though it is far more difficult to measure (Hester 2006). Further, it is difficult to translate physical 39 density into a measure of perceived density because the relationship between the two is weak (Alexander 1993, Rapoport 1975). Despite these complications, we can generally understand urban density to refer to the ratio of people per land area. The ambiguity and lack of consensus or consistency in defining urban density create a challenge for those interested in measuring or discussing density. Even focusing on a discussion of high density, as this study does, is challenging. Dave (2011), in his study of neighborhood density and social sustainability, adopts the density thresholds of the Mumbai Metropolitan Authority for low (up to 200 units per hectare), medium (between 201 and 400 units per hectare), and high (between 401 and 600 units per hectare) densities, but these levels would likely have little currency in most other cities (Rapoport 1975). I have found no literature that attempts to define “high density” in either numeric terms or express characteristics in a North American context. While this may be an appropriate response, given that any definition would need to be tied to some geographic context and, even then, would likely be an arbitrary definition (Rapoport 1975), still, it shifts the burden of definition to every author who discusses the term. Issues associated with low and high urban density To understand the kinds of problems compact city designs are intended to address, we have to understand the kinds of problems suburban sprawl causes. Compact city goals can be understood in the context of addressing these problems, usually by containing sprawl, and may be focused on economic benefits, sustainability objectives, or social outcomes (McFarlane 2016). Economic benefits may include increasing opportunities for local retail merchants, concentrating a labor pool, increasing employment opportunities, providing efficient infrastructure, and providing affordable housing (Boyko & Cooper 2011). Higher density may promote sustainability by improving transit efficiency, facilitating walking and biking as mobility options, and reducing auto traffic congestion (Boyko & Cooper 2011). Other environmental benefits include reduced energy use (including options to use district energy systems), reduced auto use (for improved air quality), and preservation of farmland and open space (Churchman 1999; Calthorpe 2011). Compact city principles also include a social component. While early twentieth-century planning focused on reducing density, as cities were generally considered to be crowded, noisy, and dirty (Moroni 2016), compact city advocates believe higher density benefits outweigh such nuisances. Compact cities may increase opportunities for interpersonal interaction by favoring pedestrian mobility and providing public space (Bramley & Power 2009; 40 Talen 1999). A study by Freeman (2001) suggests a strong inverse relationship between automobile use and neighborhood social ties. According to Ewing (1997), “leapfrog” development fails to provide functional open space where authentic communal public life can occur. Other compact city social goals include increasing housing options, bringing vitality to neighborhoods, and improving safety (Churchman 1999; Boyko & Cooper 2011). While these goals will likely appeal to many urban planners, by what mechanisms can density achieve them? Generally, the tools planners have include policies and pricing mechanisms. Pricing mechanisms, especially those with a clear nexus (such as increased infrastructure costs) could, in theory, curb sprawl, but many, such as congestion pricing, are politically difficult to enact (Ewing et al. 2015). Policies may include zoning regulations and growth boundaries. Zoning regulations may allow for denser development but may not be able to sufficiently incentivize it if market conditions are inhospitable. Growth boundaries may be effective at setting physical limits to sprawl, but have not been widely used and so have had few success stories. A study by Anthony (2004) showed that state growth management programs were not effective at limiting sprawl, but a study by Carruthers (2002) suggested that state growth management programs with consistency requirements and enforcement mechanisms might be (Ewing et al. 2015). Portland, Oregon has enacted urban growth boundaries in an effort to contain sprawl and seems to have enjoyed some success (Song & Knaap 2004). However, in order to accept that compact city principles can mitigate the problems associated with sprawl, one has to accept a series of premises. These premises include, in order, that problems exist, that they are caused by sprawl, that aspects of compact city design (including density) can address these problems, and that policy mechanisms can bring about these design solutions. These premises must be both technically viable and theoretically defensible. Are they? Compact city claims have several vulnerabilities. They may be invalid because they are technically infeasible, politically infeasible, over-stated (good, but not as good as claimed), mis-matched (good, but for other problems), misguided (misaligned with the problems they are intended to address), or conceptually flawed. Many of these contestations are considered elsewhere, but this review will only consider the conceptual soundness of compact city claims. While compact cities may offer economic efficiencies of scale and infrastructure and increased access to goods and labor, economic disadvantages are also noteworthy and may include higher construction costs, higher costs of goods and services, and higher costs of housing (Boyko & 41 Cooper 2011). Density is often touted for its ecological superiority, but environmental disadvantages may include loss of urban open space, higher ecological construction costs, and higher pollution due to traffic congestion (Churchman 1999). If we consider extreme examples of density, we see that they are not sustainable by many metrics. For example, Kowloon Walled City in China, with a population density of 1.2 million people per square kilometer (until it was demolished in 1992), was an example of a neighborhood with an unhealthy, unsafe, and unsustainable level of density (Lehmann 2016). While compact cities may offer more opportunities for social encounters, social disadvantages may include increased anxiety, reduced privacy, reduced safety, reduced environmental control, increased competition for resources, increased social segregation, loss of recreational opportunities, difficulty supervising children playing outside, and loss of sense of community (Churchman 1999; Boyko & Cooper 2011; Bramley & Power 2009). Transportation disadvantages may include increased pedestrian and vehicular congestion and a lack of parking (Boyko & Cooper 2011). Even today, many people associate the word “urban” with crime, congestion, poverty, and crowding (Calthorpe 2011). Suburban living still appeals to Americans for several emotional reasons, including feelings of independence, success, privacy, safety, familiarity, luxury, and ownership (Hester 2006). For many such reasons, it would be a mistake to assume that resistance to compact city principles is unfounded (Moroni 2016). After all, it was not that long ago that planners’ prime directive was to alleviate the problems of urban density. Density has gone from a perceived environmental and social liability to a perceived asset in only about a generation (Tonkiss 2013, p37). Density and high-rise development When we consider urban density, it is important to distinguish between area density and building density. Although people may intuitively associate high-rise development with high density, this association is by no means fixed (Churchman 1999). As Lehmann (2016) makes clear, building density and area density are not necessarily associated, as a given area density may be achieved, theoretically, by different housing typologies. But, in practice, high-rise buildings are only financially viable when land costs and housing demand are high, and this combination is usually limited to downtown cores. While different building typologies can, in theory, produce equal area densities, they tend to occur within particular density ranges. Alexander (1993) compares the density ranges of residential buildings by typology (single family detached, row housing, low-rise garden apartments, and high-rise apartments). He finds that 42 single family housing tends to range up to 10 units per acre, row housing and garden apartments tend to range from 20-40 units per acre, and high-rise apartments tend to range from 60-170 units per acre. This provides some rationale for tying density to building typology. However, functional differences in density may not be reducible to simple ratios of only one numerator and one denominator. To provide meaningful comparisons between areas, it may be that several density measures should be considered. Dovey & Pafka (2014) argue that urban densities can only be meaningfully compared when they consider a suite of metrics that include building typology, building density, population density, and open space. They refer to these compilations as ‘density assemblages.’ A genuine understanding of urban density would likely take some such suite of measures into consideration. As noted by Alexander (1993) (and, as seems intuitive), high-rise development is most likely to produce the highest building densities, and, by extension, the highest neighborhood densities. This suggests that literature on the experience of living in high-rise structures would be very helpful in a discussion of urban density. Unfortunately, as several authors note, there seem to be few recent articles that discuss the socio-cultural aspects of the high-rise building typology in the North American context (Nethercote & Horne 2016, Harris 2015, Graham & Hewitt 2012)). This lack is particularly notable when searching for the experience of particular demographics, such as families with children living in high-rise environments (Whitzman & Mizrachi 2012). While a large body of literature from the United States in the 1970’s focussed on issues of social degradation and crowding associated with inner-city high-rise social housing, more recent literature seems to coalesce around theories of “vertical urbanism,” such as by Nethercote & Horne (2016), Harris (2015), Graham & Hewitt (2012), and Harker (2014) (who all seem inspired by the work of Eyal Weizman on power and space in the West Bank in Israel (Harker 2014)), or on more pragmatic issues related to high-rise living in highly-dense Asian cities, such as reported by Randolph & Tice (2013), Karsten (2015), Yeh & Yuen (in Yuen & Yen 2011), and Cho & Lee (2011). The character of residential high-rise (or, “tall”) buildings, as discussed in the literature, allows for some variation. For example, high-rise buildings may contain several uses, but traditionally these uses are only mixed at the ground level; most floors in a high-rise building are mono-functional (Dovey & Pafka 2014). In many locations, residential high-rise buildings are traditionally constructed for the rental market, but not all high-rise residents are renters. 43 Condominium buildings, usually high-rise, have a financial structure that allows residents to own their units. In many cities, this arrangement is very popular. For example condominium units in Toronto increased from about 65,000 in 1981 to around 280,000 in 2011 (Rosen & Walks 2014). Also, the height of a high-rise, or “tall” building, is open to interpretation. Nematollahi et al. (2016), in their study of residents’ attitudes toward density in Perth, defined high-density housing as apartments over four stories tall. Verhaeghe et al. (2016) also use this definition. But, according to the Council of Tall Buildings and Urban Habitat, “tall” buildings are ten stories or more (Yeh & Yuen in Yuen & Yen 2011). Nethercote & Horne (2016), in their case-study investigation of high-rise residents in Melbourne, consider high-rise buildings to be 15 stories or more. Perhaps the height threshold of a high-rise building varies by region. It is difficult to know from the academic literature if this is so, but it critical to the discussion. In the last few years, there seems to be a growing body of theory around the relationship between high-rise buildings and social forces, generally in the field of human geography. Baxter (2017) discusses the origins of high-rise architecture springing from the International Style (popularized by architects such as Le Corbusier and Walter Gropius). He notes that most literature discussing high-rise issues is concerned with the social failure associated with high-rise living (such as Pruitt-Igoe), but he also points to an emerging dialog around vertical urbanism that seeks to understand issues of vertical living that range from power dynamics to ‘ordinary topologies.’ Graham & Hewitt (2012) discuss the relationship between building height and power and money. They point out that in many cities, such as Dubai and Hong Kong, elevation (especially with fast elevators) is a symbol and mechanism of elitism, as wealthy high-rise residents are able to vertically separate themselves from the masses (see also Harker 2014 and Harris 2015). Some researchers have sought to discuss the particular characteristics of high-rise living. Boyko & Cooper (2011, referencing Mitchell 1971 and Bagley 1974) claim that residents of high-density dwellings are more likely to suffer from emotional illness, hostility, and neuroticism. Kitchen et al. (2012), using Statistics Canada data from 2008, found sense of community belonging to be lowest among residents of high-rise apartments. Karacor & Parlar (2017) suggest that an increase in high-rise buildings in a neighborhood in Istanbul has resulted in a reduction in use of public space and thereby a reduction in collective efficacy and place attachment. Other researchers are more nuanced in their conclusions. Van Soomeren et al. 44 (2016), in their study of crime in two neighborhoods with high-rise buildings (one in Amsterdam and one in Barcelona) suggest that the high-rise buildings were less related to crime than was the low-density environments in which they were placed, as this low density led to deserted public spaces, fear of crime, and criminal acts. And Verhaeghe et al. (2016), using data from the 2001 Belgian Census, found that residents in high-rise buildings tended to report having poorer health, but these findings could mostly be accounted for due to socioeconomic and demographic variables (meaning they found no negative health effects associated with high-rise living). Few recommendations seem to be available in academic literature for architects seeking to improve the lot of high-rise dwellers, but the City of Vancouver (1992, p7, 8) has provided a set of guidelines for the construction of high-density (including high-rise) housing for families. Some suggestions include the following: • Provide direct visual and physical access between each unit and at least one common play area (3.4.3) • Strictly segregate children’s play and circulation areas from vehicle traffic (3.5.3) • Design interior corridors to accommodate children’s play and toys (3.6.3) • Provide indoor amenity spaces for play and large gatherings (3.7.2) Such strategies, while directed at family housing, may prove beneficial for any high-rise development. On the other hand, it may be that such concerns are much ado about nothing, or that they can be resolved monetarily. Economist Edward Glaeser is quite sanguine about high-rise development, arguing that “limiting high-rise development...guarantees high prices” (2011, p152). He claims that “canyons of glass and steel and concrete, such as those along New York’s Fifth Avenue, aren’t an urban problem; they are a perfectly reasonable way to fit a large amount of people and commerce on a small amount of land. Only poor policy prevents a long row of fifty-story buildings from lining Mumbai’s seafront....height is the best way to keep prices affordable and living standards high.” (p160). Glaeser suggests replacing poor policies, such as those preventing new construction from blocking light and views, with a fast-track tax system that financially compensates “neighbors who lose light from a new construction project” (p161). While high-rise living has developed a somewhat negative reputation in many Western lands, in Asian cities, such as Hong Kong and Singapore, the common perception is a bit different. The literature on high-rise living favors consideration of Asian cities and dates back 45 several decades. Mitchell (1971), in his study of high-rise residents in Hong Kong in 1967, produced several findings, including the following: • Density within dwelling units had limited effects on occupants • Attitudes toward lack of privacy corresponded with densities within dwelling units • High densities affected worry and unhappiness, but only for the poorest residents • Densities alone did not affect intense emotional strain and hostility • The condition of non-related families sharing a unit caused them stress • Parents living in high-density housing had limited control over children playing outside • High-density housing discouraged interaction among neighbors Other researchers have focussed more on the social aspects of high-rise living in Asian cities. While Dave (2011), studying neighborhoods in Mumbai, found no connection between household density and social interaction, he did find that building form influenced behavior. He found that there was less informal chatting among neighbors who lived in high-rise buildings. This may have been due to a lack of community space. In their study of high-rise residents in Seoul, Cho & Lee (2011) suggest that provision of community spaces and community programs will improve resident satisfaction. Some cities seem to be taking such suggestions to heart. Yuen (in Yuen & Yeh 2011, p136) notes that Singapore is not content to provide minimally acceptable public high-rise housing, but rather “a total living environment” that would support “quality living, recreation and accessibility to facilities and a sense of community spirit and belonging.” Researchers have given special attention to issues of family life in high-rise environments in Asian cities. Rapoport (1975) notes that, in Chinese culture, upper stories of high-rise buildings are far less desirable than lower stories for residents with children. Whitzman & Mizrachi (2012) studied how children living in high-rise buildings in Melbourne used public space as part of their Vertical Living Kids research project. They found that children who lived in public housing tended to have a high level of freedom and a low quality of public space (a ‘wasteland’ condition) but children in private housing tended to have a low level of freedom and a high quality public space (a ‘glasshouse’ condition). Randolph & Tice (2013) studied the demographic data of high-rise occupants in Melbourne and Sydney and found that they are primarily childless renters. They suggest that if planners wish to use high-rise development as a means to produce compact city environments, they should structure these developments so that 46 they will accommodate a wider range of lifestyles. A study by Karsten (2015) of middle-class families with children living in high-rise apartments in Hong Kong found that few interviewees interacted often with their neighbors and most felt that the environment provided poor opportunities for their children to play. On the other hand, considering the culture and lack of housing alternatives, interviewees felt that, overall, high-rise living was compatible with raising children. As Karsten notes, this viewpoint tends to contradict most other findings. What are the effects of increasing density? Researchers have considered density’s relationship with a variety of economic, ecological, and social issues. For example, researchers have recognized the role of density in facilitating agglomeration economies (Boyko & Cooper 2011). Such agglomeration economies lead to an increase in job opportunities within the sector and in supporting sectors (Tonkiss 2013, p39). This, in turn, may increase the desirability of an area for employment, and, in turn, its marketability as a residential area. Yet, some studies have shown cases of a low market demand for high-density neighborhoods (Tian et al. 2015; Bramley & Power 2009). Other studies have shown that residents in high-density areas are often dissatisfied with their neighborhoods, especially in low-income neighborhoods (where residents may have no good options for moving) (Baldassare 1982). Studies associate increased density with reduced automobile and energy use (Hall 1999), but also with decreased affordability (Boyko & Cooper 2011). City planners often seek to enhance a city’s marketability, sustainability, and livability, and may look to density to address all of these goals, yet these goals may be poorly compatible. For example, with respect to energy use, there may be conflicts among the goals of livability (high energy use), sustainability (low energy use), and marketability and affordability (low energy cost), that density cannot resolve. One of the paradoxes of the compact city is that sustainability and livability may be inversely related (Neuman 2005, Howley et al. 2009, Bay & Lehmann 2017). Other relationships among density-affected variables may be similarly complicated. How density affects marketability It is difficult to know how density affects marketability in a given market, since people’s preferences differ. If the question of marketability reduces to maximization of cash value of land, density may offer so much monetary advantage in number of units to sell that any disadvantages may be completely offset. Still, it is useful to consider what advantages and disadvantages dense environments offer on a per-unit basis. One marketing advantage is 47 proximity to employment centers, especially when these tie in to agglomeration economies that may offer robust employment options (Boyko & Cooper 2011, Glaeser 2012). The question is how appealing this proximity is in comparison to other quality of life factors. Several surveys show American preferences both for the high-density advantages of walkable environments with close amenities and short work commutes as well as for the low-density advantages of privacy, space, and free parking (Tian, et al. 2015). Privacy in general, and private outdoor space in particular, is of paramount importance in some cultures (Mulholland Research & Consulting 2003). While compact city and smart growth environments are advocated by many environmentalists, planners, and urbanists, most renters and home buyers in North America have not shown a high demand for them (Tian et al. 2015). In a study of English housing, Bramley & Power (2009) found an inverse relationship between density and neighborhood satisfaction across all demographics they sampled. A survey conducted in Salt Lake City, Utah showed that the highest priority of respondents, when considering where to live, was convenient parking (Tian et al. 2015). However, a balance of several preferences is at play in the marketability, and financial viability, of dense developments. Whatever people’s affinity for, or aversion to, density per se, the North American market has shown that some popular areas, such as downtown San Francisco or midtown Manhattan, continue to maintain high prices for housing irrespective of the densities involved. In the last few decades, the density of Vancouver, B.C. has more than doubled and so has the cost of living in condominiums in the densest parts of the city (City of Vancouver as cited in Montgomery 2013). This is not to suggest that the market is willing to pay more for housing because of high density (more likely it is despite the density), but simply that, in some areas, density and marketability can rise in tandem. Density may also carry a substantial cultural connection which may be positive or negative. For example, in Singapore, today one of the densest cities in the world, residents, who had traditionally lived in low-density villages, had to become accustomed to the high-rise building typology, which they initially viewed as foreign (Lawson in Ng 2009). How density affects sustainability In its EcoDensity charter (City of Vancouver 2008), the City of Vancouver claims that “A denser city uses less energy, provides easier access, promotes public health, and is more affordable than a less dense city.” Several North American cities have made similar claims. While a commonly-cited motive for increasing density is to improve ecological sustainability, 48 there remains a paucity of empirical evidence that the compact city is actually more sustainable than its alternates. Some sectors, such as transportation, offer compelling rationales. The relationship between the built environment and travel demand, according to Ewing et al. (2015), has become the most researched subject in planning literature, with empirical studies being in general agreement that a strong relationship exists. Studies show a positive correlation between density and walking and a negative correlation between density and vehicle miles traveled (VMT), although effect sizes vary greatly. Some studies have shown higher-density cities reduce automobile and energy use, but non-linearly and with decreasing benefits (Hall 1999). Other studies have shown relationships between density and biodiversity (negative), concentration of pollutants (positive), and per capita energy use (negative) (Boyko & Cooper 2011). Buildings in dense environments may have lower energy needs due to the insulative benefits of shared walls, but differences in energy savings in high-density buildings as compared to low-density buildings may be less significant today than in past decades due to overall improved building techniques (Holden & Norland 2005). Phinyawatana (in Schropfer 2016) cites several strategies for enhancing the sustainability of high-density buildings, but all of the strategies Phinyawatana cites would be equally valid for low-density buildings. Public health can be considered an issue of sustainability. Ewing et al. (2015) found that health problems, such as obesity, heart disease, high blood pressure, and diabetes are less common in compact environments than in low density environments. However, given the varied findings in recent literature, it may be argued that compact city principles are neither necessary nor sufficient to achieve urban sustainability (Neuman 2005). How density affects livability Of special interest to this study is the affect of density on quality of life. Studies have associated density with several aspects of quality of life (Macdonald 2007), perception of environment, and social issues. In what the authors consider to be the first study of the effects of density upon quality of life, Cramer et al. (2004) found that quality of life varied inversely with density, even when controlling for factors such as levels of education and income. Mouratidis (2018), though, found the opposite relationship, as did Bardhan et al. (2015). Psychological effects of living in high-density environments may include decreased perception of privacy and increased anxiety (Raman 2010) as well as loneliness and lack of control (Evans 2003). Several studies have shown the deleterious effects of high-density, high-rise, multi-family housing 49 environments on families with young children, especially when compounded by the effects of poverty, restricted play opportunities, and lack of public socializing spaces (Evans 2003, Krysiak 2018, City of Toronto 2017). Studies have shown a negative relationship between density and mental, emotional, and physical health (Boyko & Cooper 2011) though some health agencies claim the opposite (BC Centre for Disease Control 2018). Burton (2000) found that higher density areas tended to be associated with less domestic living space, less affordable housing, higher crime levels, and lower levels of walking and cycling, but higher transit use, less social segregation, and better access to facilities, than lower density areas. Bolleter (2020) claims that adding greenspace to dense environments can relieve many of the psychological stresses common to density. In addition to studies of density, studies have considered perceptions of density. Perceptions of density vary greatly from person to person and may have little relationship to objective measures of density (Raman 2010). A study in New Zealand by Walton et al. (2008) that measured perceived neighborhood quality (to represent residential satisfaction as a component of quality of life) found that their respondents preferred medium-density neighborhoods, but were split as to their lesser preference for low- and high-density neighborhoods. They concluded that resident density preference was based upon trade-offs rather than being linearly associated with density. A national sample of households showed a negative association between density and community satisfaction (Audirac 1999) and a study by Baldassare (1982) suggested that low-income residents in high-density areas showed the most dissatisfaction among the groups sampled. Social effects of density are particularly noteworthy in view of claims that compact cities may positively influence communal interactions. Some studies suggest that residents in high-density neighborhoods form fewer but stronger bonds with neighbors (Boyko & Cooper 2011). Raman (2010), in a study of six UK neighborhoods, found that social interactions in outdoor public spaces were most frequent in medium density areas (80-100 households/hectare) and least frequent at the lowest and highest densities. Studies have suggested that communal spaces are critical for neighborhood social activities, especially in denser neighborhoods (Raman 2010), yet requirements for community and social spaces may hamper efforts to create very high densities while maintaining a highly livable environment (Hall 1999). A study by Nguyen (2010) found that living in a high density area is associated with low social interaction and volunteering, but 50 higher political participation. A study by Morris and Pfeiffer (2017) found no meaningful difference between the amount of time spent socializing by urbanites versus suburbanites. And, studies have shown a negative correlation between density and affordability (Boyko & Cooper 2011), though each may be a product of confounding factors, such as job availability. The effects on livability are thus varied and it is difficult to know whether they are, on balance, more positive than negative. A fundamental livability issue related to high density environments is crowding. Researchers agree that spatial restriction is a prerequisite for crowding, but lack agreement regarding the degree to which it is primarily a physical manifestation or primarily a psychological response. Stokols (1972a, p276) frames crowding as a spatial issue, characterized as a “motivational state directed toward the alleviation of perceived restriction and infringement, through the augmentation of one's supply of space, or the adjustment of social and personal variables, so as to minimize the inconveniences imposed by spatial limitation.” Elsewhere, however, Stokols defines crowding as a “multivariate phenomenon, resulting from the interaction of spatial, social, and personal factors, and characterized by the adverse manifestations of stress” (Stokols 1972b, p75). He also distinguishes between non-social crowding (a person not having enough physical space for some task) and social crowding (unwanted social contact—the primary type of crowding discussed in the related literature) (Stokols 1972b). Yust (2012) defines crowding in a numerical, non-psychological way, as “the relationship between the amount of space in a housing unit to the number of individuals in the household,” and considers a dwelling unit to be “crowded” at one person per habitable room (which excludes bathrooms and storage rooms), “severely crowded” at 1.5 persons per habitable room, and overcrowded at two persons per habitable room (see also Lauster & Tester 2010). Evans (2000), on the other hand, defines crowding as “an adverse psychological response that occurs when the need for space exceeds the current supply.” Standards of crowding in one cultural context may be far different from acceptable standards in another cultural context (Lauster & Tester 2010), although Evans (2000) claims no scientific evidence exists on this point. 51 How does urban density influence sense of community? In our investigation of density’s effect upon sense of community, we should first consider the breadth of influences upon sense of community. While it is impossible to do this completely, I will show the major themes that I have found in the literature. Halamová (2016) categorizes three approaches to building sense of community, which she calls “accidental” (due to crisis (which she doesn’t recommend as a strategy)), “unintentional” (by putting people with similar interests or characteristics into close proximity), and “deliberate” (which involves purposive activities or other interventions). She argues that these approaches may be aimed at individuals, groups, or the physical environment (Halamová 2016). Jabareen & Zilberman (2017) propose a different evaluative framework of sense of community that includes three categories of factors, namely, physical typologies (objective and subjective measures of the built environment), demography and socioeconomics, and cultural perceptions (of, for example, trust). Jung et al. (2015) frame these categories as physical environment characteristics, socio-demographic characteristics, and social interaction characteristics. Kim (2007) groups sense of community influences into four domains, namely, community attachment, social interaction, community identity, and pedestrianism. A commonality among these frameworks is a consideration of how the physical and cultural environments influence how people perceive their communities. Studies have considered several influences on sense of community, including architectural design (Molana & Adams 2019), migration intentions (Wolfe et al. 2020), leisure time physical activity (Ross & Searle 2019), events (Zhao & Wise 2019), happiness (Ross et al. 2019), economic opportunity (Lardier et al. 2019), loneliness (Itzhaki & Cnaan 2019), diversity (Mannarini et al. 2017), values (Mannarini et al. 2019), walking (Wood et al. 2010), and dog-walking (Toohey et al. 2013). Researchers have found correlations between sense of community and several demographic elements, such as age, length of time in community, number of children, and education, but often separate findings contradict each other (Sense of community Partners 2004). A study by Glynn (1981, noted by McMillan & Chavis (1986)) found that length of time residents expected to live in a community, how satisfied they were with the community, and how many of their neighbors they knew by name were the strongest predictors of the residents’ sense of community (see also Cochrun 1994). In a multi-level analysis of several neighborhoods in New York City, Long & Perkins (2007) found length of residence, 52 participation, neighboring, empowerment, communitarianism, place attachment, block satisfaction, and block confidence to all predict sense of community, with place attachment being the strongest predictor. A study by Wilson & Baldassare (1996) showed positive relationships between sense of community and percentage Anglo, localism (residents’ relative interest in local issues), privacy (ability to control one’s separation from others), income, and age. A study by Kingston et al. (1999) found positive relationships between sense of community and both income and education. In contrast to Wilson & Baldassare and Kingston et al., Long & Perkins (2007) found positive relationships between sense of community and both affluence and non-white ethnicity, but no relationship with education. Sense of community is strongly related to participation in neighborhood associations, though it is difficult to know how to assign causality (Unger & Wandersman 1985). Similarly, sense of community has been linked to social control of the neighborhood and public ownership of neighborhood facilities (Talen 1999). Given the negative effect of heterogeneity, successful development of sense of community is more likely when members acknowledge and accept cultural differences rather than ignore or seek to suppress differences (Halamová 2016). While establishing common ground is an essential aspect of sense of community, such commonality must include recognition of, and respect for, differences among members in order to be genuine (Putnam 2003). Many factors may inhibit sense of community. Putman (2000) provides a detailed and compelling description of the decline of civic engagement in America over several decades and provides some speculation as to the reasons for this, including increased financial pressures, sprawl, and television watching, but fails to find any compelling evidence of correlation with any of these factors, or with any others. Wilson & Baldassare (1996) found negative relationships between sense of community and city size, city density, and home ownership. Other studies have found affluence and increased social status to be at odds with neighborhood attachment (Talen 1999). This may be due to a positive correlation between affluence and expectations of privacy. Privacy is an important complicating variable. The relationship between privacy and sense of community appears to be non-linear. Too much privacy reduces opportunities to develop one’s sense of community (which may be desirable to the individual) and too little privacy leads to withdrawal from social contact (Wilson & Baldassare 1996). However, withdrawal (or reluctance to engage) may also occur when privacy is not threatened. 53 In one neighborhood studied by Merry (1987), residents avoided interaction with neighbors, not out of hostility, but because they were “preoccupied with status, completion, individual growth and fulfillment, and constant activity.” Residents met their needs for community interaction elsewhere and considered taking time for informal neighborhood chat to be a sign of lower status and importance. In less affluent settings, safety, rather than privacy, may be a prime consideration. Lack of trust, fear of crime, and struggle for resources all make sense of community in a neighborhood difficult (Jabareen & Zilberman 2017). Demographic diversity has also been shown to hinder sense of community (Neal & Neal 2014). Cultural, ethnic, and other demographic differences can prove challenging to persons seeking to build a sense of community, leading to feelings of distress, distrust, and alienation (Halamová 2016). Ethnically homogeneous sections of a neighborhood may resist integration into the larger neighborhood as defined by spatial boundaries (Unger & Wandersman 1985). Competing communities, such as virtual environments, may reduce sense of community in other, more traditional communities, such as neighborhoods. Farahani (2016) describes a ‘virtual sense of community’ as “members’ feelings of membership, identity, belonging and attachment to a group that interacts primarily through electronic communication,” and argues that such online interaction may enhance or detract from neighborhood sense of community, but cannot exactly replace it. Researchers have claimed several associations between the built environment and sense of community. Moustafa (2009 p81-84) distinguishes between the instrumental role of the built environment in affecting sense of community (“the capacity of physical characteristics of the environment to enable or promote the occurrence of behavior”), in which the built environment operates as a tool that provides affordances for interaction, and the corresponding symbolic role (“the capacity of physical characteristics of the environment to affect perceptions about the social environment”), such as signs of neighborhood beautification or degeneration that affect residents’ pride of place or fear of lingering. Common approaches to influencing sense of community with the built environment typically involve facilitating informal social contact with the thoughtful placement and design of common public areas (Halamová 2016). How this can best be accomplished is the subject of many urban design books. Hester (2006) suggests that good public centers should concentrate multiple uses and provide opportunities for both routine activities (such as shopping) and special rituals (such as community events). Cochrun (1994), on the other hand, warns that when public institutions from several neighborhoods are concentrated 54 in one area, this may reduce opportunities for local interaction by putting the venues too far away. A study by Kingston et al. (1999) found associations between sense of community and the presence of recreational spaces, the presence of a town grocery, and the absence of auto traffic, but found no association with the presence of neighborhood-bounding arterial roads. The concept of “New Urbanism” has, for the last few decades, been central to ideas linking the built environment, and, especially the public realm, to sense of community (Hooper et al. 2020). Enhancing sense of community with the built environment is fundamental to New Urbanism. Strategies include the thoughtful integration of private and public space, clear neighborhood boundaries, pedestrian and Transit Oriented Development, and mixed land use (Talen 2000). Some communities have been built according to New Urbanist principles and researchers have evaluated some to test the claimed links with sense of community. Kim (2007) studied ten physical features of Kentlands, a New Urbanist development in Maryland, U.S.A., and found that the mixed-use nature of the development and the proximity of the local shopping center were the most significant built-environment contributors to sense of community. A study by Lund (2003) of several New Urbanist neighborhoods in Portland, Oregon found that amenities such as parks and retail shops tended to increase pedestrian travel and that people who walk in their neighborhoods were more likely to develop relationships with their neighbors. Other studies have found associations between pedestrian-friendly environments and sense of community, but the results vary depending upon whether residents are walking for leisure or for transportation (French et al. 2014). Anecdotal success stories are available, such as that of one Vancouver resident who found that by moving from a higher-level apartment to a ground-level apartment within the same building (a more New Urbanist environment), he went from having no social contact with his neighbors to knowing several of them and having an active social life (Montgomery 2013). Both within the New Urbanist movement and without, advocates of social cohesion point to the social benefits of open space. However, there are not many studies that directly relate open space to sense of community, and fewer that account for the design and quality of the open space under consideration or the frequency of its use (Francis et al. 2012). However, a few studies are instructive. A study by Kazmierczak (2013) showed that visitors to well-maintained local parks tended to have more extensive social ties within their neighborhoods. A study by Francis et al. (2012) compared six open space types (parks, plazas, sidewalks, shopping malls, 55 community centers, and schoolyards) according to ten attributes (walking paths, shade, water features, irrigated lawn, birdlife, lighting, sporting facilities, playgrounds, type of surrounding roads, and presence of nearby water) in 1,900 open space locations in Perth, Australia and found a high correlation between sense of community and what residents considered to be high-quality open space. Farahani & Lozanovska (2014) also suggests that social life in public spaces may be enhanced through improved activity-generating spaces (such as parks and plazas), planning strategies (such as incentivizing higher density and mixed land use), and design strategies (such as landscaping and outdoor seating). While high-quality open space tends to be more useful than low-quality space, a study by Cattell et al. (2008) in the ‘most ethnically diverse borough in Britain’ suggested that even mundane public spaces can act as important venues for building tolerance for neighborhood diversity. Still, we must remember the limitations of our ability to assign causality. For example, a study of two edge city communities by Schwaller (2012) found a positive relationship between the resident use of public space and resident sense of community, though there was little evidence that this sense of community is built either en route (walking) to the public space locations or by interacting at these particular locations. And, a study by Francis et al. (2012) found that, of the many possible uses people might have for open space, relaxation was the only use that corresponded to sense of community. Jacobs (2011) observed that city parks are often unsafe and unused except for crime or other unsavory endeavors. So, while public open space may correlate with sense of community in many studies, there is little basis to assume that any public space anywhere at any time will have similar effects. A significant limiting factor in the relationship between public space and sense of community is the issue of privacy. Privacy features as a mitigating factor in several studies of sense of community. A study of community-oriented housing in Finland by Helamaa (2013) showed that key features important to residents who sought out such housing included purpose-built spaces for both formal and informal encounters and the ability to control residents’ level of privacy. In seeming fulfillment of Lewis Mumford’s description of suburbia as “a collective effort to lead a private life,” typical suburban shopping malls provide an environment in which its denizens experience ‘the presence of others, but not their company’ (Putman 2000). A challenge of designing the built environment to facilitate interaction is that the built environment, by its nature, tends to be inflexible. A case study of a co-housing community in Atlanta, Georgia, in which the physical layout was designed for, and the community members were self- 56 selected for, optimal communal existence, found that the narrow sidewalks created conditions of both wanted and unwanted social contact (Brower 2011). Ideally, the built environment should provide residents with the ability to limit their contact with their neighbors without having to retreat entirely (Gehl & Birgitte 2013). A 1973 study of dormitory students by Andrew Baum showed that those who had a semi-private buffer zone between their (private) room and the (public) corridor were far less anxious and more sociable than those who had to transition directly from their rooms to the corridor space (Baum et al. in Aiello & Baum 1979). And, a study of Danish residents by Jan Gehl found that residents were most likely to chat with their neighbors when front porches were close enough to walkways to facilitate conversation but far enough away that conversation could easily be avoided (Gehl & Birgitte 2013). There may also be built environment challenges to sense of community. While some retail locations, such as pubs and cafes, may increase local social contact by providing opportunities for casual interaction among residents of a neighborhood, other retail locations, such as grocery stores or clothing stores, may decrease local social contact by filling the local sidewalks with transient, non-local shoppers (Baum et al. 1978). A study of three parallel residential streets in San Francisco, California by Donald Appleyard showed a strong negative correlation between the amount of vehicular traffic and the vitality of social life and sense of community of residents on these streets (Gehl & Birgitte 2013). Negative visual cues, such as litter, unkempt yards, and persons loitering, may lead residents to associate a neighborhood with crime and then avoid developing a (positive) sense of community in that area (Unger & Wandersman 1985). This wariness might be mitigated by physical and visual boundary markers that define outdoor private and semi-private space, thereby creating “defensible space” that can help preserve perceptions of safety, privacy, and environmental control (Unger & Wandersman 1985). Based upon their study of sense of community in Beer Sheva, Israel, Jabareen & Zilberman (2017) recommend planners seeking to improve sense of community should seek to improve neighborhood aesthetics, transportation, and accessibility, and should strive to create more compact neighborhoods. Another challenge to neighborhood sociability may simply be time. Jacobs (2011 p73) notes that “the trust of a city street is formed over time from many, many little public sidewalk contacts.” She suggests that residents in a new neighborhood may need months or even years of head nods and other small acknowledgements before they begin to commit to engaging conversations. 57 Having considered a breadth of other potential influences upon sense of community, including influences related to the built environment, let us now turn our attention to what influence density might have. Jabareen & Zilberman (2017) note that physical typologies can involve either objective elements (such as street networks, compactness, density, land-use types and mixes, transportation systems, connectivity, and aesthetic elements) or subjective elements (people’s perceptions of the objective ones). While we can benefit from studies that consider the objective measure of density, it would be even more useful to consider studies that compare residents’ perception of density (such as crowding) to their sense of community. Unfortunately, this seems to be a gap in research. Some studies do comment on some aspects of density as related to sense of community. Jung et al. (2015) compared residents’ sense of community in a pedestrian-oriented neighborhood versus an auto-oriented environment in Seoul, Korea and found negligible difference. Wilson & Baldassare (1996) found a statistically significant negative relationship between density and sense of community, but their finding was limited in that it was restricted to a low-density area and relied upon a single question to describe the dependent variable18. French et al. (2014) found a negative relationship between density and sense of community, but this was also in a low-density environment (Mean = 6.36 dwellings/acre, Standard Deviation = 3.02) and their results were statistically insignificant (p-value = 0.08). Baum et al. (in Aiello & Baum 1979) report their findings related to a high-density environment and claim that the nature of the circulation in most high-rise buildings is antithetical to meaningful neighbor contact and prevents the development of sense of community, but their study was very limited in scope and demography, making generalization tenuous. While several studies have related density to behavioral responses, and several others have related built environment factors to sense of community, I have found no studies that attempt to relate high-density residential environments, or perceptions of density, to sense of community. This leaves the relationship unresolved in the current literature. 18 “The dependent variable, which is the respondent’s overall sense of community, is derived from a question: ‘In general, would you describe your city or community as one which has a sense of community, or not?’ About 68% perceived that their city or community had a sense of community; 32% said that it did not.” (Wilson & Baldassare 1996 p34) 58 The role of public space A special focus of this study is the provision and nature of public space in high-density environments. Measurement of such spaces requires an informed understanding of their distinct nature. A premise of the study was that high-density neighborhoods and buildings would experience a lack of quality public space, which would create a lack of informal social interaction and, thereby, a reduction in residents’ sense of neighborhood community. I expected this condition to be especially notable in high-rise buildings. The literature tends to support this expectation. Several researchers have noted the lack of public space associated with high-rise developments. For example, Kim (2014) notes that in South Korea, public spaces and amenities are typically afterthoughts, shoe-horned into undeveloped sections of a lot. Shim et al. (2004), on the other hand, suggest that these spaces are moving from ground level to higher levels in mixed-use high-rises in the form of deck spaces and rooftop gardens. Still, they argue that these elevated public spaces suffer from a lack of integration with the surrounding urban fabric. Holahan (1976) found that a fundamental discontent of residents of a North American ghetto neighborhood who moved into a high-rise environment was the lack of semi-public space, and the resultant lack of opportunities for informal social exchange. Zaff & Devlin (1998), in their investigation of elderly residents in Connecticut housing developments, found that those living in garden apartments had a higher sense of community than those living in high-rise buildings, and theorize that this is due to differences in the amount of both defensible space and semi-public space for informal socializing. Other researchers have had similar findings and certainly none have associated high-rise development with excessive public space. Some municipal authorities have responded to the lack of public space in high-density development with legislation or suggested practices. In response to the poor quality of public space provided by market development in the 1960’s and 1970’s, the Singaporean government instituted a “New Town Structural Model” based on the precinct unit. Each precinct would include a center which had a playground, garden, or other amenity. The precinct public spaces were intended to promote social interaction and community awareness (Hee 2017). North American cities have also responded to the perceived deficit of public space in dense areas. For example, the City of Edmonton’s “Basic planning principles for high-rise residential infill in 59 mature neighbourhoods” (2007) include a suggestion that high-rise infill projects include both indoor amenity spaces and outdoor social and recreational spaces for residents, and the City of Ottawa’s “Urban design guidelines for high-rise buildings” (nd) suggests that public spaces associated with high-rise projects connect and integrate into existing networks of streets, parks, open space, and amenities. Such recognition is noteworthy, as cities are requesting developers to produce non-income-generating space for the public benefit. A primary challenge for high-rise public space, of course, is that the ratio of available ground-level area per resident is, by the nature of the structures, the most limited of all building typologies (March & Lehrer 2019). This has led to some developments (as noted regarding South Korea above) moving public space to higher levels. To compensate for a lack of ground-floor open space due to the intensive use of land in Singapore, developers have incorporated public spaces into high-rise buildings on raised platforms, podiums, roof deck gardens and skybridges (Menz 2014). Sky gardens in high-rise buildings may offer residents both the social benefits of informal meeting spaces and the mental and physical health benefits of green space (Chan 2005). Some high-rise projects provide internal common space in even more innovative ways. Some examples include the Mirador Building in Madrid, Reliance Tower in Mumbai, the Premier City Project in Almaty, Kazakhstan, Sapphire Residence in Istanbul (Engur 2013), Marina Bay Sands in Singapore (Safdie 2011), and the Raffles City development in Chongging, China (Wang 2017). How well the public spaces in these buildings perform in comparison to ground-based options will be a fertile subject for ongoing study. Some researchers have begun to do this. During the boom in interest in North American inner-city public housing, Holahan (1976) found that residents of a high-density, low-income neighborhood in New York City relied more on informal social spaces than on formally designed social spaces for neighborhood interaction. For example, formal (linear) seating areas and large grassy areas were far less used and offered less socializing potential than areas that offered a mix of functional, recreational, and leisure uses. (This observation comports with Zarghami & Gheydari (2015), who note that common spaces in high-rise buildings, even those designed for purely functional purposes, can afford opportunities for socialization and, thereby, increase social capital among residents.) More recently, Menz (2014), in his observation of a high-rise building in Singapore, found that the 60 most popular types of open spaces were playgrounds, open green spaces, and roof terrace gardens. The most common reasons interviewees gave for preferring a public space included • presence of facilities, • natural ventilation, • scenery, • accessibility, • density of people present, and • community networks. Based upon their observations of how residents in high-rise housing at the University of British Columbia use the public space in their buildings, Daneshpanah et al. (2015) offer the following suggestions for optimizing the quality of internal common spaces: • Clearly define whether common spaces are appropriate for loud socializing or for quiet studying, • make spaces large enough to accommodate multiple groups, • provide amenities (like a café) that facilitate interaction, • allow space users to modify their environment (for example, by moving chairs), • make the space easily accessible to all residents, • provide acoustical isolation to protect private spaces from communal space noise, and • consider the lifestyle distinctions of potential residents when designing common spaces. A keen awareness of the special needs that public space serves in high-density environments, and especially in high-rise buildings, will facilitate meaningful observation and measurement of such spaces and how well they are functioning. Measuring public space I expected that a primary moderating influence on residents’ sense of community would be the quantity and quality of public space. But, how should one quantify and describe public space? What questions should one seek to answer? Fundamental questions suggested by the literature include • How should one categorize the public spaces to be measured? • What types of evaluations are instructive? • What aspects and qualities of these spaces should one measure? 61 • What procedure should one follow to make these measurements? • What specific activities are important to record and how should they be recorded? Some authors have addressed these questions. Talen (2000) offers suggestions for measuring public open space as it relates to sense of community. She proposes that relevant factors include size, access (distance from residence to open space), “residential grain” (lot density), and “transport environment” (percentage of residential units facing arterial, collector, local, and pedestrian streets). Talen also presents a taxonomy of public spaces that includes • parks, • playgrounds, • squares/plazas, • community facilities, • commercial/retail space, • quasi-public facilities (such as religious buildings), • and streets. I believe that all of these spaces are useful to monitor in a study of public space. Mehta (2014) suggests evaluating public space according to five aspects (shown graphically as a pentagram) including inclusiveness, meaningful activity, comfort, safety, and ‘pleasurability.’ Similarly, Varna (2016) suggests evaluating the ‘publicness’ of public spaces according to five aspects (shown, though, as a star) including civility, animation, physical configuration, ownership, and control. Macdonald et al. (2017) offer a rating system that focuses on the quality of the pedestrian environment of neighborhoods. Each paper provides rationales for highlighting its aspects of choice. In their book How to Study Public Life, Gehl & Svarre (2013, p13-19) discuss strategies for evaluating public spaces. They recommend framing evaluations according to key questions such as the following: • How many? (Taking count of the number of people in a space and noting the time and circumstances of their activities.) • Who? (Demographic data, such as gender and age, can prove instructive in understanding why a space attracts some people and not others.) 62 • Where? (Every space has sub-spaces within it with their distinct characteristics.) • What? (Gehl & Svarre claim that primary public space activities include walking, standing, sitting, and playing. Activities may be categorized in many ways, but the categorization should be intentional.) • How long? (Duration of individual visits can be just as instructive as head count in an area in estimating how appealing it is.) Menz (2014), in a study of public space in dense environments in Singapore, proposes a similar list of potential observations, including • what people were doing, • where they were doing it, • how long they did it, • how they entered the public spaces in which their activity took place, • who they were (demographic data), and • what kinds of interactions they had while they were there. Further to these general questions, Jan Gehl (Gehl & Svarre 2013, p107) suggests a series of specific characteristics to note when evaluating the conditions of a site. They include • protection against traffic & accidents, • protection against crime & violence, • protection against unpleasant sense experiences, • possibilities for walking, • possibilities for standing, • possibilities for sitting, • possibilities for seeing (fenestration, views, lighting, etc.), • possibilities for hearing and talking, • possibilities for playing or unwinding, • provision of small scale services (notice boards, signs, waste bins, etc.), • provision of design for enjoying positive climate elements, and • provision of design for positive sense-experiences. Having a set list of observations for which to check will be especially important when comparing the activities of different sites according to identical metrics. 63 What procedures may be useful to make these observations? Gehl & Svarre (2013, p24-34) recommend a variety of procedures for documenting activity in public spaces, including • counting (as people at a given time, people over a set time interval, or objects), • mapping (such as locations of people in an area at various times of day), • tracing (for example, paths of pedestrian or cycling activity), • tracking (such as routes taken by specific individuals), • looking for traces (finding evidence of use after people have left an area), • photographing (to capture a depth of data that would be difficult to record otherwise), • keeping a diary (as a means of creating a focused record of conditions and events), and • test walks (to gain understanding of the actual experience of moving through a space). Many of these techniques were used by observers in Vancouver as part of the “Places for People Downtown” initiative led by Gehl and the City of Vancouver (Gehl + City of Vancouver 2018). Holahan (1976) observed people’s use of public space by recording their behavior in ten 30-second intervals as verbal interaction, non-verbal interaction, or isolated activity. He also created an activity map according to (instantaneous) observations of active recreation, leisure activity, or functional activity. It would seem that several observation procedures could provide instructive data, provided the observer is rigorous, detailed, and consistent. *** In summary, we have discussed the importance of addressing causality, as both the findings of this research and any recommendations that derive from them will imply that certain environmental conditions—and modifications to those conditions—will either effect, or make more likely, certain outcomes. This thinking falls generally into the realm of determinism. After a review of several variants of determinism, we found that no existing theoretical framework is suitable as a basis of this research, and, therefore, we saw the need for a new one. I have dubbed this new framework “Architectural Affordance” and sketched an outline of its assumptions. We have considered the nature of “sense of community” and of “density” and discussed what the current literature has to say about potential influences of density on sense of community. We have looked at the role of public space in this relationship and considered how we might evaluate it. This literature review and discussion have prepared us to move on to a consideration of the research project at hand, beginning with a discussion of the methodology. 64 Chapter 3: Methodology To address the primary research question of this study, namely, What is the relationship between urban density and sense of community?, I chose to use an explanatory sequential mixed-methods study19. I chose this approach because I wanted to inform the study with both quantitative and qualitative data. I wanted to use quantitative data because I wanted the study to be replicable and generalizable and I wanted to use qualitative data because the thing I wanted to measure and investigate is a feeling. The quantitative data came primarily from an online survey and the qualitative data came primarily from in-person interviews. I also conducted site observations. I discuss the formation, data collection and processing, and limitations of these three sources in this section. Many of the items discussed in this section are described further or have related visuals in the appendices. The purpose of this study was to investigate the relationship between urban density and residents’ sense of community (SOC). This was challenging, because it involved an independent variable, density, that has a variety of accepted metrics (though typically involving a ratio that includes a count of people in the numerator and a measure of area in the denominator), and a dependent variable, sense of community, that is a subjective concept, a feeling with no universally-accepted method of measurement. Density information for the areas included in the study was available from Canadian census data20, but linking that data to residents’ sense of community was more involved. To make this connection, I needed to measure the sense of community of a sample of residents in areas of differing densities. Since sense of community is a feeling, the only way to gain information about it is to ask people. However, simply asking people to rate their own level of sense of community (though other studies have done this) would not produce generalizable results, as people would likely have an inconsistent understanding of the question. I discuss my approach to generating sense of community scores below. 19 In the prospectus for this study, I assumed I would treat the research as a case study. I thought this would be the most appropriate methodological approach because I expected to have very few survey responses. Since I was able to reach statistical significance with virtually all of my survey questions, I dropped the idea of framing this research as a case study (but retained the use of a mixed methods approach). For a discussion of what a case study is and when its use is appropriate, see ‘Appendix L’ of this thesis. 20 This data provides residential density. A study of ‘daytime’ density, or locational densities when people are at work might also be informative, but is outside the scope of this study (and would require some methodology to gather, or estimate, the relevant temporal population data). 65 In addition to addressing this primary relationship, I wanted to control for a reasonable suite of potential confounding variables. This was necessary to validate the primary relationship and show that I wasn’t measuring the wrong thing. I also wanted to know what factors might moderate the relationship between density and sense of community. The questions I formed to investigate these issues seemed suitable to a survey format. Other questions, however, I thought would be poorly suited to a survey. For example, I wanted to know more about what people were thinking as they took the survey, such as how they understood the terms used (terms such as “neighbourhood,” “sense of community,” and “public space”). I also wanted to gather more discussion about how people used the public space in their neighborhoods than I expected I could get from a survey. I thought a few people might be willing to offer expanded answers, but many wouldn’t and would not complete the survey if I loaded it down with too many essay questions. Therefore, I chose to include personal interviews in the study, as I expected I could gain insights from this medium that I could not capture in the survey. Since interviewees would come from the pool of survey respondents, I would be able to link their survey responses to their interview responses. Finally, I conducted site observations. The chosen locations derived from the areas in which the interviewees lived. I intended to investigate the places that interviewees spoke of and conduct quantitative assessments of these sites. Really, my initial intention was for the study to be an exploratory sequential mixed methods study (qualitative, then quantitative), with the survey serving only to generate recruits for the interviews (I did not expect to generate significant levels of data with the survey), the interviews being the main source of data, then the site observations providing quantitative backing for the qualitative interview data. As it turned out, the survey produced a large quantity of significant data and the site observations did not, thus turning my exploratory sequential mixed methods study into an explanatory sequential (quantitative, then qualitative) mixed methods study. Survey The survey for this study was conducted online using Qualtrics software. I saw no other viable media for collecting survey data. 66 Creating the questions As noted above, a critical aspect of this study was the ability to link density data to sense-of-community data. I decided that the best metric to form this link was postal code data. I believed that everyone in my target area (see below) would know his or her postal code and most would be willing to share it. Although the postal code region can be small enough to limit one’s ability to maintain anonymity, especially if enough related personal data is linked to it (postal codes encompass a much smaller region than the United States equivalent, the ‘zip code’), I believed that people would be willing to share it. The finest resolution at which census density data was available was the ‘dissemination block’ level, but I had no expectation that survey respondents would know what their dissemination block label was (or what a dissemination block was). To link survey data to census data, I had to relate postal code data to dissemination block data. This proved far more challenging than I had expected. The systems are spatially related but readily linked in no other way. Fortunately, the outlines of the postal code regions generally fit within the outlines of the dissemination block outlines. By overlaying the two systems and locating the centroids of the postal code regions within the outlines of the dissemination block outlines, I was able to transfer the density data from the dissemination blocks to the postal codes (see figures 1 and 2). Note that the resolution remained at the courser dissemination block level. I don’t know why this exercise has not been performed at the national level previously and the postal code/dissemination block relationship made available for public use, but I think anyone needing this connection in the future will have to go through this process. 67 Figure 1 - Postal code outlines (blue areas) with dots showing postal code centroids (green areas show dissemination block areas with no postal code) Figure 2 - Postal code centroids shown within dissemination block outlines 68 The most important component of this study was the metric for the dependent variable, sense of community. Although much research has already been done related to measuring sense of community, including an industry-standard measure, the Sense of Community Index (SCI), I felt it was important to investigate how reliable this leading measure (and any similar measures) would be. Appendix ‘A’ discusses this investigation and details my rationale for creating and selecting the test items I used to create SOC scores for respondents. In addition to test items (questions) related to sense of community, I also wanted to test for other factors that I thought might be related and that might act as either confounding or moderating variables. I asked questions about demography, as I thought these might significantly influence respondents’ sense of community. Since public space was a focus of the study, I asked questions about respondents’ use of public space. I especially wanted to understand how use of public space related to residents’ sense of community. I also asked about what type of housing respondents live in and whether they felt crowded or unsafe. Previous research has shown a very tenuous relationship between density (an objective measure) and crowding (a negative perception of density), so I saw a need to account for both the relationship between density and sense of community and between feelings of crowding and sense of community (as well as between density and feelings of crowding). Since safety seemed to play a significant role in previous density research, I also tested for feelings of safety (I did no research at all of actual safety, ie, police records and such). Finally, I asked questions related to respondents’ previous housing experience. I felt that if respondents had had a substantially better or worse experience in their previous housing situation, this could disproportionately influence their current responses, especially related to sense of community. See Appendix D for a static (offline) representation of the questions used in the survey. Selecting the target areas Since this was a study related to high density, I wanted to target areas that included high-density sections. Obviously, this would include urban areas. I also wanted to include proximate areas of medium density for comparison. I was not particularly interested in low-density areas. I felt that the real comparison I was after was in the medium to very-high density range. I wanted to know if my data would suggest what happened to residents’ sense of community as urban areas progressed from medium to high to very high density. So, at the outset, I knew my study would focus on at least one urban area. 69 The next question to resolve was, Which urban area(s) should I try to include? Although there are some cities that naturally lend themselves to studies of density (such as Singapore, Hong Kong, Shanghai, etc.), I felt I had no way to reach potential respondents in distant areas. Besides the language barriers, I had no funds for travel (or for anything else) and no contacts in foreign cities. Although I had no support for outreach and only a poor plan for doing it myself, I decided that my best chance for success at outreach was to keep my study local. This would allow me the greatest opportunity for both initial outreach for my survey and for the subsequent interviews that I intended to conduct. So, I chose the Greater Vancouver Regional District as my general target area. I felt that I could reach any point within this area within a day’s trip. The next step was to decide on which areas within the District I would focus my outreach. I used Google Earth to locate all of the areas within the district that had high-rise buildings (see Appendix B). From this set, I looked for areas that had both very high density sections and medium and high density sections. I thought this would allow me some control over non-proximate factors when I compared the SOC scores from one density level to another. I also looked at the public space of the areas I found. My intent was to ‘rate’ the public space at the different sites and then later compare the quality of the public space to the SOC scores. I later abandoned this because I didn’t have enough qualified areas and because I could not find or create an objective rating system for the quality of the public space. As a last consideration, I filtered for areas that had a high percentage of families with small children. I was particularly interested in the life quality of this demographic living in high-density environments (where outdoor play spaces were limited). A more detailed discussion of my site selection process is provided in Appendix B. Advertising the survey I had no good venues for systematically advertising my survey to my target areas, so it was clear from the outset that I would have to rely on a sample of convenience. I created a table of the sites I wanted to target with my survey and listed any options I could find for advertising (see Appendix C). I used this as a starting point, mostly as locations to post flyers (see Appendix C) both at physical locations and online. I found very few public kiosk locations where I could post flyers. Also, I had no way to track the effectiveness of my outreach actions, so I could only speculate as to how effective any specific outreach was. 70 In addition to flyering and online posting (I found no opportunity to pass out handbills in person), I sent over one hundred emails to strata and property managers (see Appendix C). By far, the most effective approach was when one property management company (Associa) agreed to advertise my survey in their email newsletter to their residents. Until this point, I had gathered roughly 350 valid responses in over 6 months. After the assistance from Associa, I more than doubled that amount in just over one month. This was very fortunate (and due to the benevolence of one advocate out of hundreds of persons contacted). Since I had no funds for outreach by mailer (which would have cost thousands of dollars for postage alone), this was really the only way I was able to reach residents in high rise buildings. Unfortunately, they weren’t necessarily located in just my target neighborhoods, but the results were useful anyway. Of the six areas I targeted, only two (UBC and Port Moody) returned enough results to evaluate. On the other hand, due (I suppose) to paid advertising through Facebook (targeting the Vancouver area) and through help from Associa, I ended up with results all over the Greater Vancouver Regional District. Processing the data I ran the survey for nine months, from December 2018 through August 2019 (with a few early returns in November as I was proofing the survey). Qualtrics allows for download into a variety of software formats. I downloaded my data into an Excel format. Valid responses required postal code data and most SOC items completed. I calculated aggregate (average) scores for SOC, feeling of crowding, and feeling of safety. Appendix E shows the variables, response types, data types, and response options used in the survey. I used QGIS and Access to create the density scores for the postal codes and link them to the survey data. Analyzing the data After cleaning and arranging the data, I used PSPP (similar to SPSS) to calculate correlation coefficients and p-values (using the Spearman method) for the primary dependent variable (dv) and independent variable (iv) and other relationships (see Table 4 in Findings – Survey responses). I used QGIS to visually evaluate the relationship of SOC scores to density (see Figures 10 – 15 in Findings – Survey responses). I used Tableau to create scatterplots of the various relationships I evaluated (see examples in Findings – Survey responses). 71 Limitations The survey component of the study had several limitations. First, it was not possible to make this survey both randomized and statistically significant given the number of variables involved and sample size I had. Second, I had no budget to advertise it (though I did spend some money for Facebook ads). I offered four drawings for $25 prizes (all paid), but it is questionable how much of an incentive this was. Third, there was an unavoidable sample bias in the survey toward those who were willing to take it. How this willingness correlated to respondents’ SOC score (thus skewing results), I have no way of knowing, but I suspect that it would skew the results toward higher SOC scores because I imagine that people with a higher sense of community would be more likely to participate in community-related things in general and in sense-of-community-related things in particular. Since I was able to compare SOC scores over a full spectrum of densities, this potential skewing probably isn’t significant. Still, one might suggest that the sample is biased. A fourth limitation was the density resolution, which I applied to the postal code level as derived from the larger dissemination block level. I believe the sample size was large enough to overcome this error, but it still remains unaccounted for. It might be possible to produce more accurate results by manually counting densities of postal codes, but this level of effort was beyond the scope of this study. Further to this limitation was the currency of my population density data. The census data was a few years old at the time I processed my survey data. Some of my respondents lived in neighborhoods that were so new, the dissemination block data still showed a population of zero. Also, the results may be highly dependent upon the scale of areas that I used for my density data. The dissemination block was the smallest unit of density I had, but I could have chosen a larger unit, such as the dissemination area, the next unit larger. To test the potential difference in findings, I chose a subset of my results (the UBC area) to test the potential difference between dissemination block densities and the densities of the dissemination areas within which they lie. I found, as I suspected, that the densities were very dissimilar. For the area I examined, I found that the correlation coefficient between dissemination block densities and their associated dissemination area densities was only 0.26 (n = 44) (although it went up to 0.48 (n = 29) for dissemination block densities under 10,000 persons per square kilometer). Also, the correlation between sense of community scores and density for this area varied drastically, being effectively nil (0.03 (n = 35)) when using dissemination block densities 72 and substantial (0.41 (n = 35)) when using dissemination area densities. This suggests that my overall results would be very different if using densities of larger scales. Whether they would be any more accurate or meaningful are matters for theory and interpretation in future research. Finally, the survey was very long. In my interest to be comprehensive, I included many topics in the survey. Future researchers interested in conducting a similar study would do well to review the findings and consider whether some or many items could be pruned without deficit. Lessons learned Generally, the survey was successful. All of the data gathered was useful to the study and speaks to either the primary question or to issues that critics might raise against the answer to the primary question. On the other hand, after reviewing the findings, it seems that future studies of similar intent could be much shorter without losing needful context. Still, I don’t think this could have been known prior to running the survey. One clear win was the use of new test items for SOC metrics. As will be seen in the Findings – Survey results, future studies of sense of community would likely do better to use just two of the test items of this study instead of other indices used previously, based on the results of this study. Interviews The purpose of the interviews was to gain insights that would have been difficult to obtain from the survey format, such as what people were thinking as they took the survey. What did they have in mind when they answered questions about their neighborhood or the public spaces in it? What did they think sense of community was? How did their SOC score compare with their own evaluation of their sense of community? These were the kinds of things I hoped to probe in the interviews. Creating the questions Creating the survey questions was fairly straightforward once I was clear in the perspectives I wished to gather from the interviewees. These perspectives related to meanings of terms, perception of neighborhoods, perceptions of public space, opinions about crowding, density, and safety, and issues of how culture and personal connections might influence sense of community. The set of interview questions is presented in Appendix ‘F’. 73 I formed the interview questions prior to beginning the survey. Of the three common interview formats (structured, semi-structured, and open), I chose to conduct semi-structured interviews. I thought this was appropriate because the survey had already provided structured response data and open surveys would not provide sufficient control over the discussion to be able to compare different interviewee responses later. I thought that the semi-structured format would allow me to generate comparable data but also allow for discussion and divergence as I saw fit. After conducting the interviews, I believe this was the right decision. I intended to update the interview questions after reviewing the survey responses so as to better probe unresolved issues presented by the survey. However, after an initial review of the survey responses, I saw no need to amend the interview questions from what I had proposed at the outset. Selecting interviewees I recruited interviewees through the survey (which asked, at the end, if respondents would be willing to sit for a half hour to one hour interview). Initially, this was the primary purpose of the survey (as a recruitment tool). This proved effective, as 253 respondents (just over a quarter of the total) agreed to be interviewed. Of these, I selected about two dozen who were located in the sites I had selected (and a downtown area that I added because it generally met my selection criteria). Of these, I managed to successfully set up interviews with 15 people in three different sites. Since recruits entered their contact information in the survey, I could connect their survey responses with their interview information. Since I only intended to use the interview data for qualitative research (“explaining” the previously conducted quantitative data), there was no minimum sample size required. In fact, I interviewed as many people as I could get from the sites I had selected. I could have performed more interviews (of the 253 volunteers), but this would have been a burdensome addition to the study. As it was, conducting and processing the 15 interviews was a substantial amount of work and generated, I believe, adequate results. Conducting the interviews I asked the interviewees to name a public location in their neighborhoods in which to conduct the interviews. All of the interviewees knew of a local coffee shop in which to meet, so the interviews took place at coffee shops. These were pleasant venues. I offered to buy the 74 interviewees coffee and gave them the honorarium prior to the interview. I also got permission to record the interviews. Processing the data I used Rev to transcribe the interviews. While this was expensive (over $1,000), it saved me a considerable amount of time. I reviewed the transcripts for accuracy and made corrections as needed. Then I divided the transcripts according to the questions of the interviews. Using Access, I created a database that included the answers from each interviewee from both the survey and the interview. I then created profiles for each interviewee (Appendix G) and collections of responses to each interview question (Appendix H). I also created a collection of notable exchanges that I thought could be enlightening if read in their entirety (see Chapter 6). Analyzing the data Since these were semi-structured interviews, the answers were pre-sorted and did not require extensive coding, as they would have in an open survey. Also, in creating the database of interview answers, I reviewed the responses extensively and in depth and sorted them as according to content, such that further coding would have added very little organizational value. Still, I had to look for patterns and commonalities among the responses and consider how each response and set of responses spoke to the issues I was investigating. I had to evaluate the best way to show the answers and explain what I thought was relevant to the study. Limitations The primary challenge was finding potential interviewees in my target areas. I certainly had no shortage of potential interviewees generally, but the challenge of finding them in my target areas can be seen as an extension of the challenge of targeting my advertising in the sites I wanted. I was also limited by a lack of diversity in my interviewees. Of the 15 interviewees, only one was male. Also, the age of most was fairly high. Again, I did not screen by demographics--I just had a disproportionate number of middle-aged and older females. While not all were white, most were. Of course, all spoke English, but some spoke other languages as well. Also, while I maintain that the semi-structured interview format was the most appropriate, an open format, or even group interviews, might have produced broader results. 75 Lessons learned The use of a survey to recruit interviewees was very effective. It allowed me to screen interviewees beforehand. In this case, I only screened for location, but potentially, I could have screened for other factors as well. For example, if I had had more recruits than I needed, I might have screened for diversity or to include parents of small children (a demographic I was especially targeting). Also, I speculate that the survey was effective for recruitment because it allowed recruits to invest only a small amount of their time and yet become exposed to the intent of the survey. Those who found the topic engaging could then opt to make a further time (and trust) commitment of an hour for the interview. And, again, I found the format and venues to be very effective. After conducting the interviews, I think I would modify some of the questions for future interviews. Although I wanted to know how interviewees would self-rate their sense of community, I never came up with a good question for this. The question I used was “How would you describe your sense of community in your neighbourhood?” But, interviewees didn’t really know how to answer this question, even with added prompting, and it didn’t produce very useful responses. After asking “Do you speak with your neighbours in these (public) spaces?” I asked the follow-up question “What types of things do you usually discuss?” but I don’t think the latter question was useful. I asked the question “Do you wish you spent more time or less time speaking with your neighbours?” and got interesting responses, but I think it would have been better to provide context by first asking “How often do you speak with your neighbours?” Finally, after I ask “Which communities or groups do you feel connected to?” I ask the follow-up question “How would you rank the importance of your connection to these groups?” I wanted to understand whether outside connections were displacing neighborhood connections, but the question felt too invasive and produced inconsistent results. Site Observations The purpose of conducting site observations was to examine the environments described by the interviewees. Initially, I intended to conduct extensive site observations, but, in the end, I conducted only 16 site observations. While useful for many questions, I felt that the observations were not particularly instructive for this study, as they could not address the question of how people were feeling. They could address what people were doing (and where 76 and how and with whom and such) but not whether their activities were leading to an increased sense of community. Selecting sites I selected sites to observe based on the areas in which I had interviewees. In the end, this had little utility, as the sites were not so particular that other sites mightn’t have been just as instructive. But, I was following my own protocol and intention of linking sites to interviews (and, to survey data). Of the six sites I initially chose, I had interviews from only two (UBC and Klahanie (a neighborhood in Port Moody)). At UBC, interviewees came from three (or four) distinct neighborhoods, so I conducted one site observation at Klahanie and three site observations at neighborhoods in UBC. At each of these four neighborhoods, I observed activity at four different locations for an hour each. Selecting metrics For each site, I was primarily trying to answer two questions, namely How much do people interact here? and What is this place like? To address these questions, I formed a template that I could use to record both how many people were in a place and how many people were interacting in that place over five-minute intervals. For the second question, I considered work by urban researchers (such as Jan Gehl) to form a set of questions to answer about each location. Appendix I shows the template I formed for the site observations. Conducting observations It was difficult to know how much time to invest in doing the site observations. It would have been difficult to argue, no matter how much time was spent, that the observations were sufficient to describe the comings and goings of people during different times of the day, different days, and different seasons. Also, whatever time investment I arbitrarily decided was ‘sufficient’ for a site observation would have to be multiplied by 16, the number of locations I wanted to observe. I decided that one hour each would prove to be roughly as instructive as two and only take half as much time. I quickly extended this reasoning to conclude that spending more than an hour at each location would be both prohibitive and of diminished return. Less than an hour each, however, seemed useless (and lazy). I conducted the site observations over a series of Saturday mornings in August of 2019, filling out my templates as I went. 77 Processing data There was relatively little to process, other than some very basic math, after the site observations were done. I had collected a substantial amount of data about the sites, but, as I had no way to connect this data to my primary research question, I spent little time evaluating it. The results are shown in Appendix J. Limitations The two biggest limitations of this site observation process were time and applicability. The site observations are very time intensive for the amount of data they generate. Also, they are poorly suited to addressing questions about perception. Other limitations include the challenges of comparing one site to another (what makes one coffee shop ‘successful’ and another not, and how do you know?) and generalizing findings. I think with a large enough data set, one could use the templates I developed to produce generalizable arguments, but this level of involvement lay outside the scope if this inquiry. Lessons learned I think a future study of sense of community would also have limited use for site observations. Related studies, however, such as studies of who uses public space and for what could make excellent use of the templates developed for this study. I found few references that informed methods for evaluating public space and none that provided a template such as I developed. A future study that included similar site observations should provide justification for the range of times and sites that it would cover and make sure it had sufficient resources to complete such observations. 78 Chapter 4: A quantitative inquiry into the relationship between urban density and sense of community in the Greater Vancouver Regional District This section discusses the findings of the study. As this is a mixed-methods study, I will present both survey data and interview data in this section. Some findings incorporate data from both the survey and the interviews. Other findings are limited to one or the other medium. Discussions of each finding from the survey data will generally include the following topics, unless they have been addressed previously: • Reason—why I chose to produce data that would inform the finding; • Measurement—why I chose the metrics involved in the finding; • Test—why I chose the statistical test that I used to analyze the data; • Relationship—the strength and direction of the relationship between the data items; and • Relevance—the implications of the relationship I will discuss the broader implications of the findings and their relationship to previous research in the Discussion section. The relationship between urban density and sense of community As noted earlier, the primary purpose of this study is to see if a meaningful relationship exists between neighborhood sense of community and urban density. The intent of this inquiry, in particular, was to note whether residents’ sense of community (SOC) diminished at very high urban densities, and, if so, whether any factors might moderate this relationship. As noted in the Methodology section, I used a set of 26 5-point Likert-scale questions to form a composite sense of community average score for each survey participant. In order for a survey entry to be valid, a participant had to complete the sense of community questions. To measure density, I used data from the 2016 Canadian census. A particular challenge was translating density data from the census at the ‘dissemination block’ level to the Canada Post postal code level (smaller areas that generally fit within the dissemination blocks). I used Geographic Information System (GIS) software to accomplish this translation, and ended with a ‘persons per square kilometer’ value for most of the postal codes used in my study. Although the statistical test I used did not require breaking density values into categories, I thought it would be useful to set these boundaries for comparison. As noted earlier, ‘low,’ 79 ‘high,’ and other categories for density levels are subjective and relevant only to specific areas under consideration. To set these subjective levels for the area of my study, the Greater Vancouver Regional District, I used a ‘natural breaks’ function in the GIS software and rounded somewhat to achieve a range of density categories ranging from very low (less than 600 people per square kilometer) to very high (more than 6700 people per square kilometer), as noted in Table 2. The statistical test most appropriate for comparing ordinal data, such as that produced by a Likert-scale test, is the Spearman test. This is the test I used to find correlation coefficients and significance levels for my survey data. The relationship between sense of community (SOC) and urban density was very weakly negatively correlated, as shown in the scatterplot in Figure 3, by the correlation coefficient (-0.065) in Table 2, and by the maps in Figures 10 – 15. This relationship is the primary finding of the study. Additional to this general trend, I looked at the same relationship as broken into the density categories referred to previously. These relationships are shown in Figures 4 – 8. Note that while all density categories are very weakly related, it is the ‘very high’ category that is mostly responsible for the overall negative relationship found in aggregate. Also, as I was particularly interested in the possible effects of density on sense of community for families with small children, I specifically noted the density/SOC relationship for this demographic. The results (weakly positive) are shown in Figure 9. The relationship between sense of community and urban density is significant primarily for what it does not show, namely a strong correlation at any density level. While it is interesting that the relationship trends downward at the very high density category (after trending upward at the high density category), the fact that the relationship is very weak is important, and is potentially good news for advocates of increased urban densities, as it suggests that residents in very dense urban environments may experience a sense of neighborhood community that is just as high as that of residents in any other density category. 80 Figure 3 – Relationship between sense of community and urban density (SOC of 1.0 is highest) 81 General question Test Item N corre-lation p How does SOC correlate with density? among all responses with density data 634 -0.065 0.102 at the very low quintile range21 39 0.047 0.775 at the low quintile range 76 -0.048 0.677 at the medium quintile range 82 0.133 0.234 at the high quintile range 89 0.134 0.212 at the very high quintile range 353 -0.116* 0.030 for families with children aged 5-9 76 0.204 0.076 How does SOC correlate with various test items? age 886 0.278*** 0.000 (male) gender 902 0.014 0.681 number of people are in household 882 0.111** 0.001 number of children in household 896 0.136*** 0.000 home ownership 898 0.128*** 0.000 annual income 840 0.149*** 0.000 amount spent on rent or mortgage 836 -0.055 0.111 a feeling of connection to family 910 -0.148*** 0.000 a feeling of connection to co-workers or school friends 910 -0.002 0.946 a feeling of connection to a religious group 910 -0.049 0.139 a feeling of connection to a political group 910 -0.034* 0.030 a feeling of connection to a sports or hobby group 910 -0.111** 0.001 a feeling of connection to an online community 910 -0.025 0.447 length of time at address 905 0.194*** 0.000 use of a building common space 857 -0.027 0.435 use of a walkway 870 0.133*** 0.000 use of a park 874 0.218*** 0.000 use of a playground 858 0.192*** 0.000 use of a community center 868 0.215*** 0.000 use of a cafe 863 0.140*** 0.000 use of a grocery store 866 0.052 0.128 use of a non-grocery store 863 0.098** 0.004 interaction at a building common space 836 0.200*** 0.000 interaction at a walkway 857 0.494*** 0.000 interaction at a park 845 0.454*** 0.000 interaction at a playground 835 0.326*** 0.000 interaction at a community center 847 0.380*** 0.000 interaction at a cafe 842 0.392*** 0.000 interaction at a grocery store 847 0.403*** 0.000 interaction at a non-grocery store 827 0.389*** 0.000 (increasingly dense) housing type 870 -0.157*** 0.000 21 Density quintiles for this study were set as follows: very low < 600; low = 601 - 2800; medium = 2801 - 4200; high = 4201 - 6700; very high > 6700. These were calculated by rounding from the natural breaks in the data of 0, 594 , 2839, 4222, 6716, 454783. Units are people per square kilometer. Data taken from Canadian census. See Methodology section for method used to apply census data to postal code areas. See appendix for visual examples of density levels. 82 General question Test Item N corre-lation p single-family housing type 910 -0.155*** 0.000 low-rise attached housing type 910 -0.033 0.321 low-rise apartment housing type 910 0.098** 0.003 high-rise apartment housing type 910 0.074* 0.026 presence of a neighbourhood association 634 0.058 0.143 involvement in a neighbourhood association 478 0.217*** 0.000 feelings of crowding 864 -0.320*** 0.000 feelings of safety 864 0.368*** 0.000 How do various test items moderate the relationship between SOC and Density? age 624 0.038 0.348 (male) gender 637 -0.061 0.123 number of people are in household 625 -0.016 0.686 number of children in household 630 -0.007 0.866 home ownership 634 0.001 0.982 annual income 597 0.041 0.313 amount spent on rent or mortgage 593 -0.003 0.937 a feeling of connection to family 639 -0.112** 0.005 a feeling of connection to co-workers or school friends 639 -0.035 0.371 a feeling of connection to a religious group 639 0.060 0.132 a feeling of connection to a political group 639 0.021 0.601 a feeling of connection to a sports or hobby group 639 -0.030 0.443 a feeling of connection to an online community 639 0.012 0.756 length of time at address 636 -0.096* 0.015 use of a building common space 608 0.058 0.155 use of a walkway 617 0.063 0.120 use of a park 616 0.029 0.469 use of a playground 612 -0.019 0.643 use of a community center 618 0.056** 0.016 use of a cafe 613 0.083** 0.040 use of a grocery store 615 0.133** 0.001 use of a non-grocery store 611 0.140** 0.001 interaction at a building common space 597 0.068 0.096 interaction at a walkway 608 0.030 0.462 interaction at a park 601 0.085* 0.037 interaction at a playground 598 0.013 0.756 interaction at a community center 606 0.046 0.260 interaction at a cafe 601 0.070 0.088 interaction at a grocery store 605 0.084* 0.039 interaction at a non-grocery store 591 0.004 0.923 (increasingly dense) housing type 618 0.038 0.345 single-family housing type 639 0.127** 0.001 low-rise attached housing type 639 -0.056 0.157 low-rise apartment housing type 639 -0.068 0.087 high-rise apartment housing type 639 0.035 0.372 the presence of a neighbourhood association 440 0.149** 0.002 involvement in a neighbourhood association 328 -0.117* 0.034 83 General question Test Item N corre-lation p feelings of crowding 607 -0.053 0.196 feelings of safety 609 0.083* 0.040 How does SOC correlate with past experience? How does SOC correlate with previous home being a particular (increasingly dense) housing type? 847 -0.099** 0.004 How does SOC correlate with most of life lived in a particular (increasingly dense) housing type? 819 -0.077* 0.028 How does SOC correlate with use of a building common space in previous neighbourhood? 795 -0.012 0.734 How does SOC correlate with use of a walkway in previous neighbourhood? 821 0.087* 0.012 How does SOC correlate with use of a park in previous neighbourhood? 815 0.135*** 0.000 How does SOC correlate with use of a playground in previous neighbourhood? 808 0.107** 0.002 How does SOC correlate with use of a community center in previous neighbourhood? 811 0.115** 0.001 How does SOC correlate with use of a cafe in previous neighbourhood? 809 0.121** 0.001 How does SOC correlate with use of a grocery store in previous neighbourhood? 820 0.125*** 0.000 How does SOC correlate with use of a non-grocery store in previous neighbourhood? 794 0.123*** 0.000 How does SOC correlate with a feeling that current neighbourhood is safer than previous? 844 0.232*** 0.000 How does SOC correlate with a feeling that current neighbourhood is more crowded than previous? 845 -0.057 0.100 How does SOC correlate with a feeling that one has more sense of community in current neighbourhood than in previous? 843 0.538*** 0.000 How does SOC correlate with importance of having sense of community in current neighbourhood as compared to previous? 842 0.341*** 0.000 How does perception of crowding correlate with density and housing type? How does perception of crowding correlate with density? 610 0.111** 0.006 How does perception of crowding correlate with (increasingly dense) housing type? 855 0.113** 0.001 How does perception of crowding correlate with single-family housing type? 864 0.174*** 0.000 How does perception of crowding correlate with low-rise attached housing type? 864 -0.037 0.273 How does perception of crowding correlate with low-rise apartment housing type? 864 -0.059 0.085 How does perception of crowding correlate with high-rise apartment housing type? 864 -0.046 0.173 How does perception of safety correlate with density and housing type? How does perception of safety correlate with density? 610 -0.078 0.055 How does perception of safety correlate with (increasingly dense) housing type? 855 -0.126*** 0.000 How does perception of safety correlate with single-family housing type? 864 -0.156*** 0.000 How does perception of safety correlate with low-rise attached housing type? 864 -0.004 0.916 How does perception of safety correlate with low-rise apartment housing type? 864 0.077* 0.023 How does perception of safety correlate with high-rise apartment housing type? 864 0.034 0.316 Table 2 - Survey results Direction of correlations as noted. Significance indicated as * = p < 0.05, ** = p < 0.01, *** = p < 0.001. Strengths of correlation coefficients considered ‘very weak’ below 0.2, ‘weak’ between 0.2 and 0.4, ‘moderate’ between 0.4 and 0.6. ‘strong’ between 0.6 and 0.8, and ‘very strong’ over 0.8. Most correlations were very weak. 84 Figure 4 - Relationship between sense of community and urban density at very low density (SOC of 1.0 is highest) Figure 5 - Relationship between sense of community and urban density at low density (SOC of 1.0 is highest) Figure 6 - Relationship between sense of community and urban density at medium density (SOC of 1.0 is highest) Figure 7 - Relationship between sense of community and urban density at high density (SOC of 1.0 is highest) Figure 8 - Relationship between sense of community and urban density at very high density (SOC of 1.0 is highest) Figure 9 - Relationship between sense of community and urban density for parents with young children (SOC of 1.0 is highest) 85 Figure 10 - Map of density by postal codes in Kitsalano and downtown Vancouver (darker grey represents higher density) Figure 11 - Relationship between sense of community and urban density in Kitsalano and downtown Vancouver (green is highest SOC, red is lowest SOC) 86 Figure 12 - Map of density by postal codes in and near the Klahanie area (darker grey represents higher density) Figure 13 - Relationship between sense of community and urban density in and near the Klahanie area (green is highest SOC, red is lowest SOC) 87 Figure 14- Map of density by postal codes at the University of British Columbia (darker grey represents higher density) Figure 15- Relationship between sense of community and urban density at the University of British Columbia (green is highest SOC, red is lowest SOC) 88 Potential confounding variables Many factors may influence one’s sense of community. To understand the influence of urban density (the study’s primary independent variable) on residents’ sense of community (the dependent variable), it is important to consider other factors (secondary independent variables) that may also influence sense of community. While it is impossible to verify which factors may or may not influence sense of community, I sought to test factors I thought would be most likely to do so based upon my review of related literature. Figure 16 below represents these factors, along with the principal independent variable of the study, urban (residential) density. Figure 16 - Potential influences on sense of community and on the relationship between urban density and sense of community 89 To measure these secondary independent variables, I included questions related to demographics (such as age, gender, and income), use of public space, interaction in public space, housing type, presence and involvement in neighborhood association, feeling of crowding, and feeling of safety (see test items in appendix and results in Table 2, above). As noted above, I used a Spearman test to compare the independent variables to the dependent variable. This test was suitable, even though the independent variables included ratio data (such as number of people in household), ordinal data (such as household income, which was presented in brackets), and nominal data (such as gender). All of these were tested against the dependent variable, SOC, which, as noted above, was a ratio-data calculated average of several Likert-scale (ordinal data) items. The results of these tests are shown in Table 2 above and in Figures 17 through 20 below. Of the demographic variables, only age had more than a very weak relationship with SOC. As a positive relationship, this suggests that the respondents tend to have a stronger sense of community as they age. Although many factors showed only very weak relationships with sense of community, the direction of the relationships can still be instructive. For example, SOC relates positively with income, home ownership, and length of time at residence, but negatively with amount spent on housing. Of particular interest to this study was whether social contacts outside of the neighborhood would have a negative influence on neighborhood sense of community by reducing the need to satisfy one’s need for connection within the neighborhood. The negative correlations between (neighborhood) SOC and connections to groups outside of the neighborhood tend to support this supposition. The survey also investigated both use of, and reported levels of interaction in, local public space. This was a key area of investigation, as type, amount, and quality of public space are some of the very few factors that developers and city officials may be able to adjust when planning new neighborhoods. The results suggest that some types of public space are more positively related than others to residents’ sense of community. Examples of public spaces that have a greater-than-average positive relationship with SOC include parks and community centers (see Table 2 and Figure 17). 90 Even greater correlations exist for variables related to interaction in public spaces. This is reasonable, as it involves more of a behavioral component rather than simply use of public spaces (without interacting with others). Independent variables in this category involving interaction in local walkways, parks, and grocery stores scored the highest correlations in the study (see Table 2 and Figure 18). While this data provides a strong argument for the inclusion of such spaces in neighborhoods as a strategy for increasing sense of community, it is also of limited practical value, in that neighborhood planners cannot provide interactions, but only the venues for interactions. As we see in Table 2, of the three spaces noted for a ‘moderate’ correspondence in the public space interaction category, only one, parks, scores above ‘very weak’ for public space use. Figure 17 - Relationship between sense of community and use of public spaces 91 Figure 18 - Relationship between sense of community and interaction in various public spaces Housing typology is not directly related to density, but it is closely related and represents both a useful proxy and a good point of comparison. Clearly, single-family detached housing is less dense, at least in practice, than high-rise housing (the Corbusian tower-in-a-field typology is virtually unknown, likely as a result of market forces), though attached single family housing and low-rise apartment housing may have density ranges with a high percentage of overlap. This study looked at both the correlation between housing type (as a range from single-family detached to high-rise apartment) and SOC, and the individual correlations between the four housing types considered (see methodology for a discussion of how the housing types considered were selected) and SOC. At first look, there seems to be a very weak, but clear (correlation coefficient = -0.157; significance = 0.000), negative relationship between increasingly dense housing typology and sense of community (see Table 2). This would suggest that residents’ sense of community in single-family detached housing would be slightly—but definitively—higher than residents’ sense of community in high-rise apartment buildings. However, when we disaggregate the results by typology, we notice that the results for the single-family detached/SOC correlation (correlation coefficient = -0.155; significance = 0.000) are very similar to the overall results, and may, in fact, disproportionately influence the overall results. I suggest this because the two lower- 92 density types (single-family detached and low-rise attached) have a negative relationship with SOC, whereas, the two higher density types (low-rise and high-rise apartment) have a positive relationship. Thus, the individual relationships run counter to the aggregated relationship (and the primary density/SOC relationship). In any event, it seems the relationship between housing type and SOC is very weak. The study also looked at the potential influence of neighborhood associations on sense of community. A neighborhood association may take many forms. It may be effective or not, highly or poorly representative of the views of the neighbors, congenial or antagonistic, harmonious or fractured, having paid or volunteer members who are either elected or appointed and whose decisions may be either enforceable or easily ignored. It may be formed through strictly grass-roots initiative, by fiat of some higher authority, or incentivized to form and continue by some external entity. To the extent that it can be incentivized, it becomes another tool that neighborhood planners may consider if they wish to influence residents’ sense of community. I asked both whether survey respondents were aware of the presence of a neighborhood association and whether they were involved in any way with the neighborhood association (if they answered yes to the first question). Awareness of the presence of a neighborhood association had a very weak positive relationship with SOC, and involvement had a weak positive relationship. I think these results are fairly intuitive. Another special interest of the study was the relationship between sense of community and feelings of crowding. Crowding, in this sense, is a negative reaction to excessive density. I devised several Likert-scale questions (see appendix) based upon a review of relevant literature related to crowding. I averaged the responses to these questions to create a single score for respondents’ feelings of crowding. As has been shown in several studies, including this one (see below), there is a very weak correspondence between crowding and density. Still, it is fundamental to this study to understand whether the relationship between density and SOC can be framed in terms of a relationship between SOC and crowding (or whether it represents a relationship between SOC and some other aspect of density). It is particularly important to see whether the relationship between SOC and crowding is in the same direction as the relationship between SOC and density. As we can see from Figure 19, it is. As we also might expect, it is 93 stronger (weak) than the SOC/density relationship (which is very weak). (See below and Figure 21 to see the very weak but positive relationship between density and crowding.) Figure 19 - Relationship between sense of community and feeling of crowding Similar to the crowding score, I calculated a composite score for respondents’ feeling of safety in their neighborhood based upon an average of several questions that I formed after a review of relevant literature. I thought safety would be an important variable for which to control, as I thought it could have a disproportionately high influence on residents’ sense of community. Survey results suggest that while the relationship between feelings of safety and SOC is weak, it is positive (as one might expect), and it is stronger than most other factors measured. 94 Figure 20 - Relationship between sense of community and feeling of safety I believed this study would have been open to valid criticism if it had not made a reasonable effort to control for confounding variables. I made some effort to run multiple regressions, but with so many variables and such weak correlation coefficients, I decided it was more instructive to run the regressions separately to better show their particular values. This would still allow me to show whether any secondary independent variables might be overly influencing SOC in ways that might distort the apparent relationship between density and SOC. It is also worth noting these secondary relationships (secondary in importance within this study) in their own right, as someone might be interested in these other relationships more than in the primary focus of the study. Since a focus of this study was to inform practice, it is worth noting variables that positively correlate with SOC, even if they do so irrespective of density. Examples are the relatively strong relationship between SOC and interaction in some local public spaces, such as walkways, parks, and grocery stores. While one might argue the direction of the causality (perhaps people with high SOC scores just tend to talk to people more, generally), and, while neighborhood planners cannot force people to interact, still, this study suggests that providing these types of public space will have a positive influence on residents’ sense of community. 95 Potential moderating variables Additional to my interest in investigating the relationship between density and sense of community, and in considering potential confounding factors that might also influence sense of community, I wanted to test possible moderating factors that might influence the relationship between density and sense of community (see Figure 18). Why? Because, while I thought it was useful to understand the relationship between density and SOC, I thought it would be even more useful to understand which factors might influence this relationship. This is because density is not an easily manipulated variable. It would be very unlikely for anyone to adjust the density level of a project based on the results of this research. If I could, however, identify factors that reduce a negative density/SOC relationship or accentuate a positive one, that could be useful knowledge. For example, if the study were to show a strong negative relationship between very high density and SOC, except in cases in which residents had community center (or some other tested item), that might be useful to know. While a developer (who wished to produce a development in which future residents would have a high level of SOC) would be unlikely to move the project to a lower density area or reduce the number of units, she might consider adding a community center, if, indeed, the results suggested that this might be a measure that moderated a negative effect. In order to test the magnitude of the effects of the variables (Figure 18) on the relationship between density and SOC, I first had to express this relationship in terms that I could measure. To do this, I calculated Z-scores for the SOC scores of each survey respondent. I also calculated Z-scores for the density of each postal code in my study area. By multiplying these two Z-scores together, I was able to create a dependent variable that represented the magnitude of the relationship between density and SOC. Then, I used a Spearman test to evaluate various factors as independent variables to evaluate the relationships between these variables and the density/SOC relationship. As the results of these tests show (see Table 2), the correlation coefficients are very small (meaning that the influence of these items on the density/SOC relationship is very small). At first, this may seem disconcerting. After all, without some indication of what interventions may alleviate the ill effects (or accentuate the benefits) of high, low, or some other level of density, the value of the study is diminished. In any event, the lack of strength of these dependent 96 variables is explained, I think, mostly by the fact that the relationship between density and SOC itself is very weak. In other words, the effects of any given intervention on mitigating the effect of density on SOC could only be small because the effect of density on SOC is small to begin with. Thus, such small effects on a small effect leave little to discuss, even though many relationships were statistically significant. Still, a few of these mitigating relationships are of interest and subject to speculation. I was particularly interested to see whether provision of public space would positively influence the density/SOC relationship. With one exception, both use of, and interaction within, various forms of public space had a positive relationship with the density/SOC relationship, with use of cafes and local stores (grocery and non-grocery) showing a strong statistical significance. This is useful information, because, while the effect is not large, the results are significant. Also, since the dependent variables involve use of, and not just interaction within, these spaces, the results suggest that merely providing these amenities will have a beneficial influence on sense of community, irrespective of density level. Another interesting finding is that the presence of a neighborhood association seems to have a positive effect on the density/SOC relationship, while involvement with a neighborhood association seems to have a negative effect. Perhaps sometimes it’s better not to know so much about one’s neighbors. Related considerations Further to studying the direct relationship between urban density and sense of community, the relationship between potential confounding variables and sense of community, and the potential influence of moderating variables on the relationship between urban density and sense of community, I also wanted to control for other, related factors. In a sense, these could also be considered potential confounding variables, but I present them separately because I think they are further removed—tangential, but important to understand. These considerations include the potential influence of respondents’ past experience, their perception of crowding, and their perceptions of how safe their neighborhoods are. I explain the reasoning for including these factors below. Relationship between sense of community and past experience I thought it would be important to control for survey respondents’ previous experience. I imagined that if someone came from a neighborhood in which he previously had a very high or 97 very low sense of community relative to his sense of community in his current neighborhood, this could greatly skew the results. In fact, I wondered whether a test of neighborhood sense of community might really be a test of sense of community relative to one’s former neighborhood, or relative to the type of neighborhood to which one was most accustomed. Therefore, I included several questions related to respondents’ previous neighborhood experience (see Table 2). I used a Spearman test to compare the independent variables to the dependent variable, SOC score. The results suggest that there is very little correlation between past experience and current sense of neighborhood community. While most of the results in this category were statistically significant, the strength of the associations were generally very weak. The association between feeling that one’s neighborhood is safer than the previous and sense of community was notably stronger than most other associations (though still ‘weak’ at a correlation coefficient of 0.232). Also, questions directly related to feelings of sense of community had a high correlation to SOC score, but this doesn’t tell us anything particularly interesting. While the results of this section suggest that previous experience has very little effect on a person’s current neighborhood sense of community, they also serve as a useful control to show that other results were not skewed by respondents’ past experience. They also could serve as a justification for future research to leave this section out of a survey with similar objectives. Relationship between density and crowding As discussed in the literature review, research on the subject of crowding (a negative emotional response to unwanted social contact, generally associated with high population density environments) has shown a positive but very weak relationship with density. Since this study relied so heavily on understanding residents’ emotional response to urban density, I felt it was important to test this relationship, rather than rely solely on the findings of previous studies. To do this, I included several questions related to respondents’ perception of density. By comparing their responses to the density values from census data, I could compare their feelings of crowding to the level of density in their postal code. I also asked respondents to identify their housing type. This question served as a secondary test for density as a related proxy. I used a Spearman test to evaluate these relationships. 98 For both of the relationships between density and crowding and between housing type and crowding, the results showed a positive but very weak association (see Figures 21 and 22). While counter-intuitive, the results are in line with previous studies, as noted above. Of interest, though, is the fit of the line in the scatterplot in Figure 18, which shows an overall average increase in feelings of crowding at very high densities. Still, there were several respondents who lived in the highest density environments in the study and had very low levels of feelings of crowding. Of course, these are likely persons who self-selected to live in these areas and brought with them a high tolerance for close living. Alternately, it may be that these high-density environments have been purposely designed to minimize negative effects of density with strategies such as noise-resistant construction. But this is only speculation and outside the scope of inquiry for this study. The relationship between density and crowding is highly significant for this study, as I speculated that feelings of crowding would be the dominant mechanism by which density might suppress residents’ sense of community in high-density environments. By showing (as other studies have done) that feelings of crowding are largely disassociated from density levels, I was able to provide a rationalization for the lack of influence of density upon SOC. In other words, if SOC is diminished by crowding (as we’ve seen that it is, even if ‘weakly’ with a correlation coefficient of -0.320), but not so much by density (correlation coefficient of -0.065), knowing that crowding is only very weakly related to density (correlation coefficient of 0.111) helps explain why this is so. This, again, is further good news for those who advocate for higher density and wish to rebut those who suggest higher densities may be linked to a lower quality of life. 99 Figure 21 - Relationship between density and feeling of crowding Figure 22 - Relationship between feeling of crowding and type of housing 100 Relationship between density and safety I believed that feelings of safety could also be a strong confounding factor in this study. How could people feel a strong sense of community in a neighborhood in which they felt unsafe? Indeed, as shown in Table 2, the relationship between safety and SOC is statistically significant and, though weak, it is stronger (correlation coefficient = 0.368) than most variables tested. But, was safety a confounding variable? Was the test of the relationship between density and SOC really a test of safety and SOC due to a high correlation between safety and density? I used a Spearman test to compare the composite safety scores to the postal code densities. As Table 2 and Figures 23 and 24 indicate, there is an overall very weak (correlation coefficient = -0.078 and not statistically significant) relationship between density and feelings of safety. This shows that safety is not a confounding variable in this study. Also, while the relationship between density and safety, and the relationship between increasingly dense housing type and safety, are both negative, looking at the individual relationships between specific housing types and safety tells a different story, as the lower density housing types have a negative density/safety relationship, and the higher density housing types have a positive one. Again, as with the relationship between housing type and SOC, it seems that the disaggregated housing type results differ from the overall trend and, again, it seems to be the strength of the single-family house category that skews the results. In other words, respondents in the single-family house category have such a strong negative association between safety and density (correlation coefficient = -0.156) that it strongly influences the overall relationship (correlation coefficient = -0.126) more than the other categories. While this is a bit ironic, it may be that people who live in the least dense housing category are the most sensitive to perceived crime in increasingly dense environments, and they simply have nowhere less dense that they can choose to live. 101 Figure 23 - Relationship between density and feeling of safety Figure 24 - Relationship between feeling of safety and housing type 102 Effectiveness of test items As discussed in the introduction, methodologies for measuring both the primary independent variable for this study, density, and the primary dependent variable, sense of community, are poorly established. While the vagaries associated with density can generally be resolved by clearly defining the numerator and denominator used in its measurement, measuring sense of community still suffers from a lack of agreement among experts as to which test items are best suited. Therefore, I thought it appropriate to try to contribute to knowledge in the study’s methodology rather than simply accept the most popular test for sense of community (McMillan and Chavis’ Sense of Community Index (SOCI)). As discussed in depth in Appendix A, I chose a suite of test questions that included the 12-item SOCI and a set my own 12 questions based on a survey of leading sense-of-community tests published by several researchers. I also included the statement “It is important to me to feel a sense of community in my neighbourhood” at the beginning of the SOC questions and the statement “If I lost my wallet in my neighbourhood, I would probably get it back” at the end. This made a total of 26 test items. I used the average score of these 26 items to create the SOC score for each participant. I also asked five number-based (“how many...”) questions with Likert-scaled categories, but I did not include these in the SOC score. To test the effectiveness of these 31 SOC test items, I ran three types of tests. The first was a standard measure of internal consistency for a group of test items known as the Chronbach’s alpha. It is intended to show how closely related a group of test items is as a means of determining overall test validity. The Chronbach’s alpha scale ranges from 0 to 1, with scores above 0.9 considered excellent. The Chronbach’s alpha score for the 31-item SOC test I used was 0.965, which is substantially higher than similar tests by previous researchers. The second statistical test I used to evaluate the test items was the Spearman test. I simply compared each test item individually against the composite SOC scores to see how well any given item would predict the overall score. The results ranged from “very strong” to “moderate,” as shown in Table 3 below (Table 3 also shows the scores for test items related to safety and crowding as compared to their respective composite scores). Figure 25 shows scatterplots of the individual items compared to the composite score (steeper slopes show higher correlation). 103 Question number Test: SOCI N: Corre-lation Sense of Community Q3.1_17 I feel a sense of connection with many of my neighbours. 904 0.854 Q3.1_24 I feel comfortable being around my neighbours. 902 0.808 Q3.1_19 I have neighbours I can chat with when I want to. 900 0.801 Q3.1_18 I belong in my neighbourhood. 900 0.786 Q3.1_6 I feel at home in this neighbourhood. y 907 0.777 Q3.1_15 It’s easy for me to fit in with my neighbours. 905 0.777 Q3.1_7 Many of my neighbours know me. y 904 0.766 Q3.1_22 If I have an emergency, my neighbours will help me. 902 0.762 Q3.1_21 If I need to borrow something, I don’t mind asking my neighbours for it. 903 0.752 Q3.1_16 I’m glad that I live in my neighbourhood. 905 0.748 Q3.1_20 I have friends in my neighbourhood. 905 0.729 Q3.1_5 I can recognize many of the people who live in my neighbourhood. y 906 0.718 Q3.1_2 I think my neighbourhood is a good place for me to live. y 908 0.711 Q3.1_4 My neighbours and I want the same things from the neighbourhood. y 905 0.710 Q3.1_13 I would prefer to live in this neighbourhood for a long time. y 904 0.708 Q3.1_3 People in this neighbourhood share the same values. y 907 0.697 Q3.1_23 If my neighbours and I want to improve our neighbourhood, we can. 900 0.690 Q3.1_12 People in this neighbourhood generally get along with each other. y 904 0.688 Q3.1_11 It is very important to me to live in this particular neighbourhood. y 905 0.686 Q3.1_10 If there is a problem in this neighbourhood, people who live here can get it solved. y 904 0.679 Q3.1_14 My neighbours are a lot like me. 906 0.670 Q3.1_9 I can influence what this neighbourhood is like. y 904 0.664 Q3.1_26 If I lost my wallet in my neighbourhood, I would probably get it back. 900 0.616 Q3.2_1 How many of your neighbours do you know by name? 907 0.593 Q3.1_25 I feel comfortable walking around my neighbourhood. 903 0.571 Q3.2_3 If you had an emergency, to how many of your neighbours could turn for help? 907 0.561 Q3.2_4 How many of your neighbours do you consider friends? 907 0.545 Q3.2_2 From how many of your neighbours would you feel comfortable borrowing a cup of sugar? 907 0.537 Q3.1_1 It is important to me to feel a sense of community in my neighbourhood. 909 0.514 Q3.2_5 How many of your neighbours would you feel comfortable asking to care for your home while you were away on vacation? 906 0.505 Q3.1_8 I care about what my neighbours think of my actions. y 907 0.480 104 Crowding Q5.1_7 In your neighbourhood,... - how often do you wish you had a place in your neighbourhood where you could be alone? 863 0.814 Q5.1_3 In your neighbourhood,... - how often do you feel overwhelmed because you come into contact with too many people? 863 0.781 Q5.1_8 In your neighbourhood,... - how often do you feel you live in a crowded environment? 862 0.777 Q5.1_4 In your neighbourhood,... - how often do you come into contact with people you would rather avoid? 863 0.760 Q5.1_5 In your neighbourhood,... - how often do you go out of your way to avoid interacting with your neighbours? 862 0.753 Q5.1_2 In your neighbourhood,... - how often do you feel annoyed, bothered, or disturbed by the noise or activity of your neighbours? 864 0.701 Q5.1_6 In your neighbourhood,... - how often do you feel angry because people in your neighbourhood don’t leave you alone? 862 0.657 Q5.1_1 In your neighbourhood,... - how often do you feel you do not have enough privacy? 863 0.619 Safety Q6.1_7 I worry about my personal safety in this neighbourhood. 863 0.871 Q6.1_1 My neighbourhood is not safe. 864 0.820 Q6.1_8 I think I would feel safer if I moved to a different neighbourhood. 860 0.792 Q6.1_2 My building is not safe. 854 0.786 Q6.1_3 I am afraid to walk in my neighbourhood at night. 861 0.768 Q6.1_6 I worry about my personal property being damaged or stolen in this neighbourhood. 861 0.753 Q6.1_4 I am afraid that I could be attacked or harmed in my building. 856 0.750 Q6.1_5 I think parents should not feel comfortable letting their young children play in this neighbourhood with minimal supervision. 858 0.659 Table 3 - Effectiveness of various test items by Spearman test. All correlations are positive with p < 0.001. Items are listed in decreasing order of strength (within categories), with correlation coefficient strengths considered ‘very weak’ below 0.2, ‘weak’ between 0.2 and 0.4, ‘moderate’ between 0.4 and 0.6. ‘strong’ between 0.6 and 0.8, and ‘very strong’ over 0.8. Items that are part of the Sense of Community Index test (SOCI) are noted. 105 Figure 25 - Effectiveness of sense of community test items Finally, I ran a regression model to determine how much predictive power successive questions added in determining the overall score. As shown in Table 4, question 3.1_17 (“I feel a sense of connection with many of my neighbours.”) is listed first, as it is the best predictor, able to predict about 75% of the change of the overall score (as shown in the “R Square change” column). Question 3.1_6 (“I feel at home in this neighbourhood.”), although ranking fifth best predictor per the Spearman tests, was the second most predictive question after 3.1_17 in the regression model, adding another roughly 10% of predictive power. Likely, questions ranked higher than 3.1_6 by Spearman had less cumulative predictive power because they were more similar to 3.1_17 than 3.1_6 was. As Table 4 suggests, additional questions add very little predictive power beyond the 85% given by 3.1_17 and 3.1_6. So, a future survey might do well to just use those two items. As Table 3 shows, question 3.1_6 was an SOCI test item and 3.1_17 was introduced in this study. 106 Table 4 - Results of regression model of SOC test items showing predictive power of successive items I also ran the first two tests for the eight test items I used to generate the score for “crowding” and the eight items used for the score for “safety” (see Table 3 and Figures 26 and 27). The Chronbach’s alpha for the crowding items was 0.871 and the score for safety was 0.898, both “very good” scores (and close to excellent). This was heartening, as these tests were less carefully crafted that the one for SOC. In fact, Spearman testing showed that all items for these tests were either “strong” or “very strong.” 107 Figure 26 - Effectiveness of feeling of crowding test items Figure 27 - Effectiveness of feeling of safety test items The results of these methodological tests are significant. First, the Chronbach alpha scores lend a high level of credibility to the study and make a strong recommendation for these tests to be used in future studies. Second, they suggest that the leading test for SOC is not as 108 effective as the test items used here. In fact, none of the SOCI test items scored in the top four places and none scored “very strong.” On the other hand, three of the test items that I created had “very strong” scores. Finally, these results suggest that a similar future survey could be just as effective with far fewer test items. One could test for sense of community, feelings of crowding, and feelings of safety with only two or three questions each. This could greatly reduce the time needed to complete such a survey and potentially lead to a higher completion rate without substantially degrading the quality of individual results. 109 Chapter 5: Gaining a deeper understanding of residents’ sense of community through semi-structured interviews I felt that in order to adequately address the question of the relationship between density and sense of community, it would be best to use a mixed-methods approach. I thought it would be necessary to conduct an online survey to achieve an adequate breadth of information, and use a sub-set of this survey for in-person interviews to gain a deeper understanding of the issues involved. From the pool of survey respondents, I had 15 persons from my geographic areas of interest volunteer to be interviewed in person. From these semi-structured interviews, I was able to gain insights that would have been prohibitive to glean from the survey responses alone. (See appendices ‘G’ and ‘H’ for summaries of interviewees’ responses.) Perceptions of terms One of the issues I wanted to discuss in interview format was how the survey respondents/interviewees understood some of the terms I used. Examples of such terms were “neighborhood,” “sense of community,” and “public space.” I could have done extensive research to discuss definitions for all of these terms as they are understood by academic researchers, but this would not tell me what was in the minds of my survey respondents as they took the survey. Also, I thought asking survey respondents to define these terms in the survey would have made the survey prohibitively arduous, especially as it was already quite long. Thus, I saw the in-person interviews as an opportunity to gain insight into what they, and, by extension, possibly other respondents were thinking as they took the survey. Questions that addressed interviewees’ perception of terms included the following: • What do you consider to be your neighbourhood? • What do you think it means to have a sense of community? • What are the public/common spaces in your neighbourhood? I consider these next. What do you consider to be your neighbourhood? One word that I decided early on would be very difficult (and useless) to define is “neighborhood.” What is a neighborhood? What is your neighborhood? From a practical standpoint, the definition can only be subjective. Whatever your neighborhood is for you is up to 110 you to define. This is why the first question of my interview asked interviewees to define their neighborhood. Most responses to this question defined boundaries in some way, often including several city blocks. Some responses approached the definition in other ways. Nick (all names are aliases) defined his neighborhood as the people who live around him. Ineth also thinks of her neighborhood primarily as person-based, including her own building and the neighbors who live on either side of her. Dee thinks of her neighborhood as anywhere she can reach quickly by foot, bike, or bus. Similarly, Liz considers her neighborhood to be the area within walking distance of her home. Lou and Whohan both live in the Klahanie neighborhood of Port Moody, British Columbia, but Lou considers her neighborhood to include all of Central Port Moody and Whohan considers only Moody Centre (a much smaller area) to be her neighborhood. The interviewees seemed to substantiate the intuition that, while neighborhoods can have generally-accepted boundaries, there is no way to know what any individual considers her neighborhood to be without asking her. What do you think it means to have a sense of community? Another subjective phrase is “sense of community.” While it is the core concept of the study, and while I offer substantial digression on the term in the introduction, it is anybody’s guess what it means to a survey participant until one can ask him. In fact, even a simple request for a definition may evoke only a tautology rather than a meaningful working definition. I hoped the dialogic nature of a semi-structured interview would allow opportunities to draw out what interviewees were thinking of when they filled out the survey (with the hope, of course, of finding themes that could justify some extrapolation and generalization). Several words were used by multiple interviewees, such as variants of ‘belong,’ ‘connect,’ ‘safe,’ and ‘familiar.’ Hearing these descriptors helps triangulate the wording of the test items and verify that the survey questions are indeed representative of what people tend to associate with the phrase ‘sense of community.’ They also support my intuition that people would closely associate safety with sense of community. Each interviewee had his or her own take on the meaning of sense of community, but the definitions tended to form a close pattern. Nick thought of SOC as ‘a group of people residing together as a team.’ Dee talked about having a sense of place, feeling safe, and having ‘not quite a sense of ownership, but not wanting to see a place vandalized.’ Seedsaver also spoke about a 111 sense of place, safety, familiarity, and belonging. Lyla focused primarily on safety for both her and her children. Kathy talked about belonging, familiarity, being comfortable, and having things in common with neighbors. Amelia mentioned a willingness to speak up and join a community. Claudia talked of belonging and connection. Marie said, “it means to feel connected with the people...in your community, and feeling a sense of belonging and a sense of ownership.” Helen thinks it means “knowing the people that live around you and being involved.” Olivia associates SOC with being “happy going back home,” being able to greet neighbors, and feeling safe. Liz thinks of sense of community in terms of being able to stop and have conversations with people in her neighborhood. Grace discussed feelings of belonging and inclusion. Lou brought up safety and whether there are “people here that care whether you exist” and if she could “stop with a neighbor on the street and have a chat.” Ineth simply related it to ‘people who share her values.’ And, Whohan thinks SOC means “to be engaged and feel like I'm contributing to the community's spirit and growth and that the community is contributing to my growth.” What are the public/common spaces in your neighbourhood? The third term I asked interviewees to discuss was “public space.” Again, I believe this is a subjective term that people often take for granted. I wanted to know what my interviewees envisioned as they spoke about the nature and quality of the public spaces in their neighborhoods. I summarize their responses here: Nick: a small park, a reading room, a social room, and a gym Dee: two local streets that are closed to auto traffic Seedsaver: Jim Davis Square Mall, a community garden, Nelson Park, the English Bay and Cole Harbor sea walls, Stanley Park, the mini park on Butte, and the mini park on Cardero Lyla: the sidewalks, the shops, the community center, the forest, the farm, the hallways and the building lobby Kathy: Strip parks, Wesbrook Community Centre Amelia: roundabouts, sidewalks Claudia: "Everything but the houses" Marie: Old Barn Community Centre Helen: park with playground; 'Doggie Lane;' field with BBQ area and horseshoe pit area; indoor recreation area with spa, pool, lounge, library, ping-pong table, gym, and woodworking shop Olivia: community center with community room and gym, coffee shop Liz: several parks, including a children’s park with a swing and a sandbox; pool Grace: coffee shop, children's playground, condominium amenity rooms, community center, city park Lou: green space; community center with gym, movie room, dance room, lounge, pool, hot tub; city park; coffee shop Ineth: community center, building courtyard, green space (used by people with dogs), playground, creekside walkway with benches Whohan: coffee shops, street plaza, city park, 'Brewers' Row' Table 5 - Places that Interviewees considered 'public space' in their neighborhoods 112 These responses validate, for the most part, the types of spaces I chose to test in the survey, namely, • A building common space (lobby, corridor, elevator, etc.), • A walkway, • A park, • A playground, • A community center, • A café, • A grocery store, and • A store other than groceries. A few interviewees also mentioned gym space, and that would likely have been useful to include as an option in the survey. My favorite response was Claudia’s: “Everything but the houses.” The survey also offered opportunities for people to suggest public spaces other than those listed above. The survey asked both about which public spaces interviewees used and in which public spaces they interacted with others. In the “other” category for public space use, interviewees suggested several alternatives to the ones given, including the following: Places for buying goods, such as • art shop • convenience store • corner store • drug store • farmers market • gas station • mall Places for buying services, such as • banks • bar • barber • drycleaners • hair salon • health services • hospital/doctor/other medical • local breweries • pub (several respondents) • restaurant (several respondents) • walk-in clinic Places for recreation or activities, such as • beach • community garden • gardens • fitness • gym • laundry area • library (several respondents) • live theatre venue 113 • pool • recreation centre • soccer field • tennis Club • YMCA • yoga studio (several respondents) • mailboxes; property gates Places for gathering, such as • church • each other’s homes for book club meetings • local cemetery (!) • school • seniors friendship society • UBC • volunteer facility • courtyard Places for moving around, such as • bike lane • bus stops • Langley Airport • Skytrain • "The road! All our kids play on the road daily and the neighbours visit." • transit • trails As for the “other” public spaces in which people claimed to interact, the list includes the following: Places for buying goods, such as • drug store • farmers market • mall • marrijuanna store (sic) • bakery • gas station Places for buying services, such as • breweries • movies • pub (several respondents) • hair salon Places for recreation or activities, such as • community garden • gardens • gym (several respondents) • laundry room • pool (several respondents) • tennis courts • yoga studio • library (several respondents) • mail boxes; property gates 114 Places for gathering, such as • AGM • backyard • church • common outdoor space • courtyard • each other’s homes for book club meetings • front yard or street • HOA annual meeting • preschool • school • seniors friendship society • Wechat Places for moving around, such as • on the street (several respondents) • sidewalk • Skytrain • trails • transit • transit stop Based on the above two lists, additional response options that a future survey might offer include “gym,” “library,” “pool,” “pub,” “restaurant,” and “yoga studio.” A review of both the “other” categories of the survey and the responses of the interviewees offer several ideas for researchers as to what people consider to be public space and which public spaces may be more amenable to personal interaction. Perceptions of neighborhoods A central purpose of the interviews was to understand how interviewees viewed their neighborhoods and how these views connected to their sense of neighborhood community. To gain this understanding, I used the following questions: • How would you describe your sense of community in your neighbourhood? • What do you like about your neighbourhood? • If you could change anything about your neighbourhood, what would it be? • Do you wish you spent more time or less time speaking with your neighbours? • Do you consider your neighbourhood to be very dense? o Is it crowded? o Would you rather live in a less dense neighbourhood? • Tell me about how safe your neighbourhood is. o What would make it safer? 115 I will discuss the responses I received to these questions in this section. I also asked questions related to improving public space, and I will discuss the responses to those questions in a following section. How would you describe your sense of community in your neighbourhood? In addition to how interviewees described the concept of sense of community as they understood it, I asked them to describe their own sense of community in their neighborhood. My purpose in asking this question was to calibrate interviewees SOC scores from the survey to their self description of their level of SOC. Interestingly, it mostly failed in that task. In retrospect, I think I never really figured out how to ask the question properly. I still don’t know and I think there may be no way to ask it. How can someone self assess the degree to which she experiences sense of community? What could someone possibly use as a baseline? I think what I was hoping for was a description that I could use as a basis for comparing the responses to each other to see if they validated the SOC scores. In this respect, I think the answers are useful. Also, I think the responses to this question were more informative due to the interviewees queuing in to the word “describe” and providing fairly freeform answers in response. This serendipity was possible due to the semi-structured format of the interviews, as I could help guide the interviewees to make the most of the question. Still, in the spirit of my initial intent for the question, I provide the SOC scores along with the interviewees responses below. (Note that the possible SOC scores range from 1 to 5, with 1 being the highest and mean score being 2.28 for all survey respondents. The range for interviewees was 1.31 (at 1.25 standard deviations above the mean) to 3.15 (1.12 standard deviations below the mean)). Here are summaries of their responses listed in decreasing order of SOC score (lower numbers represent a higher score due to the way I coded the questions): Claudia (SOC score = 1.31) "I do feel truly connected and I do feel part of the community. I do think that we are building this community and this neighborhood with the people that are here. It's a dynamic community and sometimes I miss people that leave but then I'm always happy to connect with new neighbors and welcome them to the neighborhood." Ineth (SOC score = 1.31) "That we respect each other's privacy, that we aren't noisy." Liz (SOC score = 1.38) "One of the reasons...we chose our complex was because...there were kids playing outside, or there were obvious signs that kids were just playing outside. So, helmets and bikes all over the place....There was enough room for cars to drive by, but also 116 sort of a space in front of each of the units. So, there would be... it seemed every third or fourth house had a hockey net, and garages were open, and bikes were just thrown on the ground. It seemed like a 'lived in' place. It seemed like a place where kids could run out the door and find a bunch of friends and play in the neighborhood." Helen (SOC score = 1.42) "I can remember growing up, when my parents would have two or three tables set up in the living room and have other couples come over and play cards for an evening. That doesn't happen anymore. So I think that the sense of community is declining as people go their own way and there's so many things out there happening that everybody's got other things to do." Marie (SOC score = 1.42) "I feel very connected to my community, and I think one of the big factors in that for myself is that I actually was one of the first people to move into this neighborhood when it first started, the very first building that went in for staff and faculty. We were one of the first families to move in, so we saw the whole neighborhood grow up around us. I do feel a deep ownership to what's going on in the neighborhood, and I know a lot of the people who have lived here for a while." Lyla (SOC score = 1.46) "I have a strong sense of community. I love where I live. I love my home. I love the amenities near my home...the shops, the community center, the forest." Grace (SOC score = 1.50) "My sense of community here is that people help each other. Like on Saturday, we had a Klahanie garage sale, and although I wasn't volunteering at it, I went down there. I must have seen five, six people that I knew. You just leave feeling really good, because you had a cup of coffee with them, and you see everything from pregnant moms all the way up to elderly seniors that are there, all sort of knowing each other." Lou (SOC score = 1.54) "I feel at home here for sure and it's a comfortable environment. It's a safe environment. There's lots of people that I've met who have similar sort of outtakes on life." Whohan (SOC score = 1.77) Whohan likes her neighborhood and feels engaged in her neighborhood, but feels that her sense of community "struggles" for lack of a strong community association and lack of a good meeting venue. Dee (SOC score = 1.92) Dee's sense of community in her neighborhood is colored by her role as a property manager, which leaves her "engaged with a certain amount of reservation." Olivia (SOC score = 2.04) "Trying to interact and meeting as many people as possible in the neighborhood, that's kind of our sense of community. Relative to the earlier neighborhoods, I would say it's very strong here." Seedsaver (SOC score = 2.04) "I feel a sense of belonging--a familiarity with my neighborhood, where places are like public buildings, schools, churches, community garden, public spaces. I don't feel so much of a sense of community with the new buildings that have replaced the former old buildings that were three or four story walk up buildings. Now we have these new very large condo towers, and I feel we no longer has eyes on this street. It's more alienated because you don't know the people that live in those buildings." median of all 910 survey respondents (SOC score = 2.28) Nick (SOC score = 2.31) Nick expresses his sense of community by volunteering and being helpful and active in his neighborhood. Amelia (SOC score = 2.35) Amelia loves her neighborhood and knows many of her neighbors, but doesn't like living in a strata arrangement. Kathy (SOC score = 3.15) Kathy feels isolated in her high-rise but has found an online queer woman Facebook group in which to find company. Table 6 - How interviewees describe their sense of community 117 Not only is it challenging to try to place these responses on a spectrum (which might allow us to validate the SOC scores based on the responses), it is also difficult to parse any specific patterns that help us quantitatively differentiate the interviewees’ sense of community in their neighborhood. Clearly, Claudia is highly engaged with her neighborhood and Kathy is lonely, but how can we distinguish those who are half a standard deviation above the mean (like Dee) from those who are a full deviation above (like Marie)? So, while instructive, it is difficult to generalize the information in these responses. What do you like about your neighbourhood? I asked the interviewees what they liked about their neighborhoods. The answers tended to relate to both people and places, but usually emphasized one more than the other. People-related themes included ‘events,’ ‘kids,’ ‘diversity,’ and ‘human potential.’ Place-related themes included ‘walking,’ ‘transit,’ ‘nature,’ and ‘public spaces.’ Within the people-related themes, Nick was the only interviewee that referenced organized events and activities for this question. The interest in ‘kids’ was obviously higher among those that had them. "I like that kids can just run outside and find someone to play with--they're comfortable here,” noted Liz. “They have a bit of independence, where I don't always have to be with them. They can create their own adventures without me or my husband, which is important, I think, for them. And it's also nice for us, too, because we can just sit at home and we know they're safe. They'll be okay. We don't have to constantly be with them." Marie likes that they live on a dead-end street that children use to play hockey, and Olivia just likes hearing kids playing nearby. Several interviewees mentioned the value of cultural diversity, including Dee, Seedsaver, and Claudia. Whohan focused more on her hope for the future. "What I really, really like about Moody Center is the potential of Moody Center. I really, really like that that there is so much potential for real positive change in Moody Center through the range--social, economic, development--the full range. Like it's really just sitting the
UBC Theses and Dissertations
Dense but not crowded : maintaining a sense of neighborhood community in a world of increasing urban… Douglas, Eric 2021
Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.
- 24-ubc_2021_november_douglas_eric.pdf [ 8.25MB ]
- JSON: 24-1.0397248.json
- JSON-LD: 24-1.0397248-ld.json
- RDF/XML (Pretty): 24-1.0397248-rdf.xml
- RDF/JSON: 24-1.0397248-rdf.json
- Turtle: 24-1.0397248-turtle.txt
- N-Triples: 24-1.0397248-rdf-ntriples.txt
- Original Record: 24-1.0397248-source.json
- Full Text