Module 1: Discussion

Identifying the Urban: Resident Perceptions of Community Character and Local Institutions in Eight Metropolitan Areas Chase M. Billingham* Wichita State University

Shelley McDonough Kimelberg University at Buffalo

What does the term “urban” signify as a descriptor of contemporary communities in the United States? We investigate this question using data from the Soul of the Community survey, examining how people within eight metropolitan areas char- acterize their communities. A substantial disjunction exists between where within their regions respondents live and how they describe those areas. Many central-city residents label their communities “suburban” or “rural,” while many outlying resi- dents label their communities “urban.” We contend that people’s experiences with important local institutions—specifically, local schools and the local public safety apparatus—shape their understanding of their communities. Logistic regression models support this contention. Controlling for where within their regions respon- dents live, they are more likely to label their communities “urban” if they perceive local schools to be low in quality and their neighborhoods to be unsafe. Notably, these effects are not consistent across racial and ethnic groups.

Attempts to define, describe, and distinguish “the urban” have animated the study of ur- ban sociology for generations. From Simmel’s ([1903] 1995) discussion of the intense city-specific stimuli that generate a “blasé outlook” (p. 35) and shape social relationships among urban individuals through the Chicago School’s call for an empirical approach to the observation and explanation of spatial patterns and differentiation (Park and Burgess [1925] 1967), the issue of what constitutes the city was a central theme of the subfield’s origin. Evaluative labels such as “urban,” “suburban,” and “rural,” however, are applied to places by scholars and government agencies in ways that may not match the concep- tions of place held by those communities’ own residents. Given the significance that place labels can have for the development of geographically tailored policies and resource allo- cation, as well as for economic development strategies, local governance structures, and the “sentiment and symbolism” (Firey 1945) that contribute to community formation, it is imperative to understand how residents conceive of “urbanism” and “suburbanism” and how they define where they live. In this article, we examine the extent to which

∗Correspondence should be addressed to Chase Billingham, Department of Sociology, Wichita State University, 1845 Fairmount St., Wichita, KS 67260-0025; [email protected].

City & Community 17:3 September 2018 doi: 10.1111/cico.12319 C© 2018 American Sociological Association, 1430 K Street NW, Washington, DC 20005

858

IDENTIFYING THE URBAN

individuals’ characterizations of geographic spaces and boundaries may be understood as a function of their experiences within and interpretations of specific places.

To do so, we utilize data from the 2010 Soul of the Community (SOTC) survey, a joint effort of the Knight Foundation and Gallup “focused on the emotional side of the connection between residents and their communities” (Knight Foundation 2017) in 26 metropolitan regions of the United States. While specifically designed to explore the factors associated with residents’ loyalty to and satisfaction with their communities, the SOTC project also yielded data that allow for an analysis of how people describe the communities they inhabit. We first compare the labels that individuals attach to their residential communities (“urban,” “suburban,” “rural,” etc.) to a categorization of those communities based solely on ZIP code designation, exploring the extent to which people whose ZIP codes reflect a central city, suburban, or rural residence actually char- acterize their communities as urban, suburban, or rural. As we demonstrate, the data indicate a fair amount of disjunction, with approximately one-third of respondents em- bracing a residential identity different from that suggested by their ZIP code.

Next, we use logistic regression models to predict the odds that, controlling for whether respondents’ homes are located within or outside of the central city of their region, they describe the place they inhabit as “urban” or use an alternative label like “suburban” or “rural.” We find that residents’ perceptions of the quality of local institutions — partic- ularly, local schools and the local public safety apparatus — significantly influence their characterizations of their communities.

Finally, to investigate how prevailing views of what constitutes the “urban” vary across different social groups, we break down these analyses by race and ethnicity to compare the relative sensitivity of black, Hispanic, and white respondents’ characterizations of their communities to different attitudes toward public safety and local school quality. We conclude with a discussion of the contingent and evolving meaning of the term “urban.”

PERCEPTIONS OF THE URBAN

Motivating the concern of early scholars who focused on the consequences of urban liv- ing, or urbanism as “a way of life,” was the need to make sense of the rapidly urbanizing and modernizing society of the early to mid-20th century (Wirth 1938). Even as later gen- erations of urban sociologists have expanded their focus to include a wide range of issues, the field has continued to wrestle with the very core questions from which it emerged. The “urban question” posed by Castells (1977) and the “community question” framed by Wellman (1979), for example, both force a reconsideration of the appropriate object of study for urban sociologists. As Gans (2015) proposed in his essay about America’s “two urban sociologies” (p. 239), a central interest of what he terms “object-centered researchers” is the basic question of “what makes communities urban or suburban” (p. 239).

Indeed, one of the primary objectives of urban sociology has been the identification and explanation of observed spatial, economic, political, social, and racial differences between the central city, the suburbs, and the outlying hinterland. Over time, these per- ceived distinctions have become reified in our conceptions of urban, suburban, and rural communities, at least in part due to the field’s historically narrow focus on a few cities in the Northeast and Midwest that displayed fairly predictable, significant patterns of

859

CITY & COMMUNITY

difference (e.g., the “Chocolate City, Vanilla Suburbs” pattern identified in the Detroit metropolitan region; see Farley et al. 1978). Even as some scholars have issued pleas for research on “diverse destinations” and “regional cities” in an effort to attend to the vari- ety and complexity of urban spaces and the relationships within them (Billingham 2015a; Norman 2013; Oakley 2015), the term “urban” continues to conjure very powerful — though sometimes ambiguous or contradictory — images in the minds of individuals.

On the one hand, a vast literature associates “the urban” with perceived negative at- tributes, including violence, decay, poverty, and disorder (Auletta 1982; Banfield 1974; Sampson and Groves 1989). Given the nation’s intense and persistent patterns of res- idential segregation and concentrated poverty (Sharkey 2013), urban spaces are often highly racialized; consequently, they are imbued with the stigmas associated with the racial or ethnic groups that inhabit them while also serving to reinforce those very stig- mas (Gotham 2002; Loury 2002; Sampson and Raudenbush 2004). In this way, we see the malleability of the term “urban”: It is used to refer not only to places, but also to people (namely, those of certain racial or socioeconomic backgrounds), to problems believed to be associated with those people (e.g., urban crime, urban poverty), and to the insti- tutions that serve those people (e.g., urban schools). In all cases, the term “urban” acts not simply as a modifier denoting a specific geographic space, but also as a proxy for a perceived set of negative characteristics. As Rusk (1999) notes, “by far the strongest ob- jection that suburbanites raise [to the proposal for regional metropolitan development policies] concerns what they call ‘city problems.’ Just what are city problems? Most often they refer to high crime, and more tellingly, poor schools. . . . What most suburban critics implicitly mean by the city’s ‘poor schools,’ I find, is that the city system has many schools with a lot of poor children in them” (p. 131; emphasis in original).

Meanwhile, a very different discourse utilizes “the urban” to establish cachet and to communicate an air of desirability, trendiness, or sophistication. Particularly salient in the literature on gentrification, this use of “urban” is less a description of people and institutions than a connotation of a particular lifestyle boasting a range of amenities, au- thentic experiences, and entertainment (Brown-Saracino 2009, 2010; Clark et al. 2002; Lloyd 2002; see also Zukin 2010 on urban “terroirs”). The inherent ambiguity and ten- sion bound up in the use of the word “urban” can be seen in the challenges faced by gentrifiers, who often find that their actions upon taking up an urban lifestyle simulta- neously undermine the very features (e.g., authenticity, edginess) that drew them to the city in the first place (Brown-Saracino 2009; Schlichtman and Patch 2014; Tissot 2015; Zukin 2010), or that their pursuit of an urban lifestyle forces confrontation with urban institutions (e.g., schools) that they find less desirable (Billingham and Kimelberg 2013; Cucchiara 2013; Hankins 2007; Kimelberg 2014; Posey-Maddox 2014).

Urban schools and education provide a useful window into the flexible, dynamic, and broad uses of the term “urban.” As Milner (2012) observes, “[p]eople across the U.S. classify schools in different parts of the country as urban because of characteristics as- sociated with the school and the people in them, not only based on the larger social context where the schools and districts are located” (p. 557). Similarly, Gadsden and Dixon-Román (2017) note, references to urban schools are “as likely to evoke images of low-income students and families of color (images that disproportionately point to deficits or highlight their cultural and social capital) as to highlight the restrictions to opportunity and access that institutional structures and social hierarchies impose” (pp. 431–432). As such, urban schools and the families they serve are alternately seen as

860

IDENTIFYING THE URBAN

problems in need of fixing, or as institutions and groups to be avoided (Hirschman 1970). In either case, Gadsden and Dixon-Román (2017) argue, references to urban schools are “rarely neutral,” but instead “have special meanings that may increase the potential for marginalization . . . or, in more problematic terms, add to real or perceived risk for chil- dren and families” (p. 451).

The flip side of the application of the “urban” label to connote dysfunction or failure in certain schools or among certain populations is the underidentification of problems in schools or among populations that are presumed to be successful by virtue of their sub- urban location. As Posey-Maddox (2016) explains, “‘[s]uburban’ is commonly conflated with middle-class Whites and treated as the normative reference group against which ‘urban’ schools and students are evaluated” (p. 225). Indeed, “much of the education discourse and policy continues to reflect monolithic framings of the ‘urban’ and ‘subur- ban,’ often treating these settings — and the populations they serve — as quite distinct and relatively undifferentiated” (Posey-Maddox 2016:227), thus potentially masking pat- terns of inequality, as well as politicizing and complicating efforts at school reform. This tendency is particularly concerning given the vast demographic changes taking place in suburban schools in recent decades (Frankenberg and Orfield 2012; Siegel-Hawley 2016).

The need to rethink the definitions of, and the relationships between, urban and sub- urban communities extends beyond the issue of education. Social scientists have high- lighted the limitations of the familiar narrative of the struggling (mostly nonwhite) city and the prosperous (mostly white) suburb (Hanlon et al. 2009). For example, the rapid gentrification of many U.S. neighborhoods has remade the economic and social land- scape of numerous cities (Brown-Saracino 2010; Lees et al. 2008). At the same time, the increasing suburbanization of poverty (Holliday and Dwyer 2009; Kneebone and Berube 2013; Raphael and Stoll 2010) and the growth of immigration outside of urban areas (Wilson and Svajlenka 2014) have underscored the economic and racial diversity in sub- urbs around the country. Hanlon (2009) identifies five distinct types of inner-ring sub- urbs, “largely differentiated by issues of class, race, and ethnicity” (p. 221). Many of those inner-ring suburbs have experienced significant declines over the past few decades (Short et al. 2007). These shifts have been so pronounced, according to Ehrenhalt (2012), that the typical relationship between city and suburb in the U.S. is becoming “inverted.” The current article contributes to this discourse by examining the extent to which individuals’ descriptions of their residential communities comport with the rigid geographic divisions delineated by municipal boundaries, or reflect a more nuanced and variable understand- ing of what constitutes an urban or suburban lifestyle.

CONCEPTUALIZING THE NEIGHBORHOOD AND THE URBAN

Wirth’s (1938) identification of urbanism as a “way of life” prompted decades of debate over how individuals experience life differently as a product of their geographic loca- tion. Gans ([1962] 1995) questioned the contention that distinct modes of being were evident among urban and suburban communities, arguing instead that social class and life cycle stage were more relevant than spatial relation to the central city. He called for refocusing attention to a more meaningful unit of analysis, the neighborhood, observing

861

CITY & COMMUNITY

that “ . . . people do not live in cities or suburbs as a whole but in specific neighborhoods” (p. 181). Subsequent studies took up the issue of “neighborhood saliency” by exploring whether neighborhoods are only readily identifiable when they have some clear historic significance, or whether neighborhoods are more relevant for urban residents than sub- urban residents (Haney and Knowles 1978; Keller 1968).

As the focus on “neighborhood effects” has assumed a prominent place in 21st-century urban sociology (Sampson 2012; Sampson et al. 2002; Sharkey and Faber 2014), research into how individuals perceive their neighborhoods has proliferated. Scholars have ex- amined the influence of neighborhood perceptions on numerous behaviors and condi- tions, including safety and crime (Sampson and Raudenbush 2004), residential decision- making (Krysan et al. 2009), health outcomes (Wen et al. 2006), parenting practices (Dahl et al. 2010), and school choice (Billingham 2015b; Posey-Maddox 2014). Efforts to quantify the impact of neighborhoods, however, are complicated by the challenge of operationalizing the unit of analysis (Sampson et al. 2002). Two issues in particular are common: (a) inconsistencies between researcher-identified and resident-identified neighborhood boundaries; and (b) variation across similarly located individuals in terms of how neighborhoods are constructed, bounded, and labeled.

Evidence points to a disjuncture between the administrative boundaries typically em- ployed by researchers as neighborhood proxies (e.g., census tracts, ZIP codes) and those delineated by residents (Kolko 2015; Spilsbury et al. 2009). While convenient due to data availability and consistency, the use of official boundaries often yields neighborhoods that differ in size, constituent elements, and precise location from those identified by in- dividuals (Coulton et al. 2001; Coulton et al. 2013; Hart and Waller 2013). Consequently, reliance on administratively defined boundaries increases the risk of misspecification, potentially distorting the effects of the neighborhood context on the variable of interest (Foster and Hipp 2011; Sampson et al. 2002). In an effort to more accurately represent those ecological contexts that are most salient and meaningful for residents (Chaskin 1997), researchers have proposed a variety of alternative conceptual approaches and mea- surement techniques (Coulton 2012; Coulton et al. 2001; Foster and Hipp 2011; Hart and Waller 2013; Hipp et al. 2012). These attempts share a desire to move away from the “sim- plifying assumptions about boundaries . . . [inherent in the use of] census geography or political jurisdictions to operationalize neighborhood units” and to capture the nuance and fluidity that characterize residents’ perceptions of the places they inhabit (Coulton 2012:231).

A second issue concerns the reliability of neighborhood perceptions — in other words, the degree to which individuals residing in close proximity define their neighborhood in similar terms. A number of studies point to variation in perceived neighborhood bound- aries, size, and composition as articulated by residents of the same spatial area (Lee and Campbell 1997; Logan and Collver 1983; Guest and Lee 1984; Sastry et al. 2002). Findings suggest that some of these differences can be explained, in part, by individual characteris- tics or neighborhood characteristics; that is, there appears to be some systematic variation in the social construction of neighborhoods. Sastry et al.’s (2002) study of Los Angeles, for example, revealed that residents’ educational attainment, income, immigrant status, and social ties within the neighborhood influenced their perceptions of neighborhood size. Likewise, neighborhood-level characteristics, such as population density and socioe- conomic status, and the number of vacant buildings in the area, were also associated

862

IDENTIFYING THE URBAN

with how large or small residents believed their neighborhood to be (Sastry et al. 2002). Studies of other cities have also found individual- and neighborhood-level effects on residents’ perceptions of neighborhood scale (Campbell et al. 2009; Coulton et al. 2013).

Disparities in neighborhood identification on the basis of age and race are especially worthy of note. Parents and children, for example, often define their neighborhoods differently (Campbell et al. 2009; Spilsbury et al. 2009), suggesting that policy or social service interventions targeted at families must be sensitive to the discrete ways that adults and children may interpret and interact with their residential environment. Similarly, Hwang’s (2016) study of a gentrifying neighborhood highlights racial differences in the construction of neighborhoods. Through the use of cognitive maps, Hwang (2016) re- veals that black residents define their neighborhoods in broader, more inclusive, and more conventionally recognized terms than do white residents, whose perceptions of their neighborhoods are more selective and idiosyncratic, omitting areas that they be- lieve are less desirable due to crime and social class characteristics (see also Campbell et al. 2009; Rich 2009).

Consequently, researchers must be aware of the reasons why disparate notions of neigh- borhoods may exist, as well as the potential implications of those distinctions. As Chaskin (1997) argues, “the delineation of boundaries is a negotiated process; it is a product of in- dividual cognition, collective perceptions, and organized attempts to codify boundaries to serve political or instrumental aims” (p. 539). In other words, neighborhood perceptions matter not only because they influence those who live within them, but also because, as Florida (2015) explains, they “have real world consequences. Reinvestment and redevel- opment funds, for example, are sometimes distributed based on informal neighborhood boundaries.”

Our study is motivated by similar concerns, albeit on a larger scale. Individuals’ willing- ness to support a given community-based policy or program, for example, is likely influ- enced, in part, by what they consider that community to be, and whether they consider themselves to be members of it. Thus, a clearer understanding of how individuals iden- tify the community they inhabit — and the factors associated with that identification — is necessary.

In this article, we address three sets of questions. First, to what degree do people’s per- ceptions of the urbanicity of their own communities reflect city/suburb divides created by municipal boundaries (see Kolko 2015)? Do inner-city dwellers consistently charac- terize their communities as “urban,” while residents of surrounding municipalities de- scribe their areas as “suburban” or “rural,” or is there ambiguity in the degree to which urbanicity corresponds to municipal divisions? Second, if municipal location does not fully predict the likelihood that residents identify their communities as “urban,” “subur- ban,” or “rural,” what other factors matter in making that determination? To what ex- tent are institutional factors — in particular, residents’ perceptions of the quality of local institutions — salient? Third, do different racial and ethnic groups characterize their communities differently, and if so, do they vary in the degree to which specific character- istics of those communities influence their descriptions? More broadly, do members of different racial and ethnic groups vary in their conceptions of “urban” life?

863

CITY & COMMUNITY

DATA AND METHOD

To address these questions, we used data from the 2010 SOTC survey, sponsored by the Knight Foundation and conducted by Gallup.1 SOTC surveyed residents of 26 U.S. regions, which were selected as “Knight Foundation communities” because they once served as the homes of newspapers owned by the Knight-Ridder Corporation. The orig- inal purpose of the survey was to examine the nature of community attachment, with questions designed to assess respondents’ satisfaction with various civic amenities and other social characteristics of their communities in order to provide information to com- munity leaders on how to “attract and keep talented workers” (Knight Foundation 2017). Respondents were randomly contacted by phone to generate samples representative of each of the 26 regions. In 2010, roughly 400 respondents were surveyed in 18 of the 26 regions, while larger samples ranging from 1,300 to 1,800 individuals were surveyed in the remaining eight regions. Interviews were conducted by telephone (both landlines and cell phones) in English and Spanish; they consisted of 73 questions and lasted, on average, 16 minutes.

Because our main focus was upon salient distinctions between life in the central cities and in the outlying communities within large American metropolitan areas, we limited the number of communities that we included in our analyses. Specifically, we consid- ered only respondents living within one of the 100 most populous metropolitan areas of the United States; for that reason, we eliminated all respondents from 13 less populous regions. Furthermore, we included only those living in regions in which there existed one and only one clearly dominant central city; for that reason, we excluded respon- dents from four regions in which multiple cities arguably compete for the designation of metropolitan urban hub. Finally, we included only regions in which both central-city and outlying residents were surveyed; for that reason, we excluded respondents from one re- gion where only central-city residents were surveyed. The final analyses included respon- dents from eight large metropolitan regions with clear city-suburb distinctions centered around one and only one primary city: Akron, OH, Charlotte, NC, Columbia, SC, Detroit, MI, Miami, FL, Philadelphia, PA, San Jose, CA, and Wichita, KS.2

Although the SOTC project was originally designed to gauge residents’ attitudes re- garding their communities and the amenities that they provide (see Fitz et al. 2016; Neal and Neal 2012), a unique characteristic of the questionnaire makes this data set especially useful for examining residents’ characterizations of places. The SOTC data set contains data on where within their regions respondents actually live, as well as data regarding how — in terms of degree of urbanicity — they perceive the communities they inhabit. Two variables in the data set are of key importance. First, respondents were asked to pro- vide their ZIP code. While the raw ZIP code data were not made available to researchers in the data set, Gallup staff used the ZIP code data to create a dichotomous variable that identifies whether each respondent lived in the central city of his or her region or in another community (i.e., in a nearby suburban or rural area). Second, respondents were asked, in the following question, to offer their own view of their communities: “How would you describe the area where you live? Would you say you live in . . . ?” Response options included “a city or urban area,” “a suburb,” “a rural area,” or “something else.” We drew on responses to the latter question to construct a dichotomous variable mea- suring urban identification, with respondents who described their community as “a city

864

IDENTIFYING THE URBAN

or urban area” coded 1 and all other respondents coded 0. Although the vast majority of respondents coded 0 self-identified as living in a “suburb,” we use more general terms like “nonurban” to refer to this group in the remainder of this article to avoid mischar- acterizing those who self-identified as “rural” or “something else.” This urban/nonurban identification variable serves as the key dependent variable in the analyses that follow.

As one of our primary aims was to determine the extent to which respondents’ percep- tions of their communities are moderated by their beliefs about key institutions within them, we include as our main independent variables respondents’ evaluations of two elements of daily life that are often perceived to vary substantially between urban and nonurban communities: the quality of the schools and the level of personal safety in the community. Respondents to the SOTC survey were asked the following question: “On a five-point rating scale, where 5 means very good and 1 means very bad, how would you rate the overall quality of public schools in your community?” Later in the survey, re- spondents were also asked to rate “how safe you feel walking at night within a mile of your home” on a similar scale. We focus here on the effects that respondents’ evaluations of schools and neighborhood safety have on their perceptions of the urbanicity of their communities because concerns over education and safety have frequently been cited as key drivers of urban flight in the United States. In addition, connotations of the “urban” in general, and of urban social problems specifically, are often bound up with images of stigmatized inner-city public school systems and unsafe inner-city neighborhoods.

Along with the ZIP code-identified location variable and the institutional evaluation variables, we included in our models a range of individual characteristics that are likely to affect the geographic location of respondents, as well as their perceptions of their communities. These characteristics include income, employment status, family structure, homeownership, age, educational attainment, and racial/ethnic identity. In addition, we incorporated city- and county-level demographic and economic characteristics of the places that respondents inhabit (such as population, population density, racial compo- sition, homeownership rate, poverty rate, and the percentage of the population living in “urban” areas, as determined by the U.S. Census Bureau) in order to account for some of the social and visual cues that, net of respondents’ attitudes toward local institutions, are likely to influence their perceptions of place (Sampson and Raudenbush 2004). Cases with missing data were excluded by listwise deletion.

RESULTS

Table 1 presents descriptive statistics for the variables incorporated into later analyses. On both their perceptions of local school quality and their feelings of safety, respondents were generally optimistic. On a scale ranging from 1 to 5, the mean on both variables exceeded 3; just under half of the respondents gave an answer of 4 or 5 when asked to rate the quality of local schools, and a solid majority gave an answer of 4 or 5 when asked to evaluate local safety. The majority of the sample was married, but most had no chil- dren living at home. About half of the respondents had at least a college degree, and most were homeowners. On average, respondents had lived in their current communi- ties for over 30 years. About three-quarters of the respondents identified as non-Hispanic white, with 12 percent identifying as non-Hispanic black, 10 percent as Hispanic, and 3 percent as Asian. This sample is somewhat older, more highly educated, and more

865

CITY & COMMUNITY

T A

B L

E 1.

D es

cr ip

ti ve

St at

is ti

cs

M ea

n o

r P

ro p

o rt

io n

SD M

in M

ax

R es

p o

n d

en t

R ep

o rt

s L

iv in

g in

a C

it y

o r

U rb

an A

re a

0. 36

0 1

R es

p o

n d

en t

L iv

es in

C en

tr al

C it

y G

eo gr

ap h

ic al

ly 0.

38 0

1 E

va lu

at io

n o

f L

o ca

l Sc

h o

o l

Q u

al it

y (1

= ve

ry b

ad ;5

= ve

ry go

o d

) 3.

37 1.

22 1

5 E

va lu

at io

n o

f L

o ca

l Sa

fe ty

(1 =

ve ry

b ad

;5 =

ve ry

go o

d )

3. 56

1. 32

1 5

In co

m e

(i n

T h

o u

sa n

d s

o f

D o

ll ar

s) 66

.9 0

37 .2

6 15

12 0

E m

p lo

ye d

0. 50

0 1

M ar

ri ed

0. 55

0 1

N u

m b

er o

f C

h il

d re

n U

n d

er 6

L iv

in g

at H

o m

e 0.

18 0.

54 0

7 N

u m

b er

o f

C h

il d

re n

6– 12

L iv

in g

at H

o m

e 0.

20 0.

56 0

7 N

u m

b er

o f

C h

il d

re n

13 –1

7 L

iv in

g at

H o

m e

0. 17

0. 47

0 3

M al

e 0.

46 0

1 H

o m

eo w

n er

0. 79

0 1

L ev

el o

f Sa

ti sf

ac ti

o n

w it

h C

o m

m u

n it

y (1

= n

o t

at al

l sa

ti sf

ie d

;5 =

ve ry

sa ti

sf ie

d )

3. 64

1. 06

1 5

A ge

52 .8

6 17

.9 5

18 85

Ye ar

s L

iv in

g in

C o

m m

u n

it y

33 .3

6 20

.4 4

1 85

E d

u ca

ti o

n al

A tt

ai n

m en

t G

ra d

e Sc

h o

o l

o r

L es

s 0.

01 0

1 So

m e

H ig

h Sc

h o

o l

0. 04

0 1

H ig

h Sc

h o

o l

G ra

d u

at e

0. 17

0 1

So m

e C

o ll

eg e

o r

T ec

h n

ic al

Sc h

o o

l 0.

30 0

1 C

o ll

eg e

G ra

d u

at e

0. 25

0 1

P o

st gr

ad u

at e

0. 23

0 1

R ac

e/ E

th n

ic it

y W

h it

e 0.

73 0

1 B

la ck

0. 12

0 1

H is

p an

ic 0.

10 0

1 A

si an

/ P

ac if

ic Is

la n

d er

0. 03

0 1

N at

iv e

A m

er ic

an 0.

01 0

1 O

th er

0. 02

0 1

N 5,

65 3

866

IDENTIFYING THE URBAN

heavily white than the general population of the metropolitan areas from which it is drawn, and it contains a higher percentage of homeowners. To the extent that these fac- tors are related to residents’ perceptions of their communities, the sample characteristics have the potential to affect the generalizability of the findings. To account for some of these discrepancies, the data were weighted according to race, gender, and age distribu- tions within each region (for more information on the survey, see Knight Foundation 2017).

There were some notable differences in sample characteristics between those who in- habited central-city ZIP codes and those who lived in ZIP codes outside of central cities. City residents tended to earn slightly lower incomes, but were more likely to have an advanced degree. The proportion of white respondents was lower in central cities, but the samples were remarkably similar on other characteristics, including age, length of residence, homeownership, and employment status. They differed substantially, though, in their evaluations of their local communities. Among those inhabiting central-city ZIP codes, the mean school quality rating was 3.10; among respondents residing outside of the central city, the mean school quality rating was 3.54 (t = –13.29, p < 0.001). A similar pattern emerged in the assessment of local safety: The mean safety rating for central-city respondents was 3.25; for respondents residing outside the central city, the mean safety rating was 3.75 (t = –14.25, p < 0.001).3

About 36 percent of the 5,653 respondents reported that they lived in the central city of their metropolitan area; the remaining 64 percent stated that they lived in a suburb, a rural area, or “something else,” or that they did not know. Similarly, about 38 percent of respondents actually did inhabit central-city ZIP codes, with the remainder living in ZIP codes not contained within central city municipal boundaries. Though the proportion of respondents living in central cities was similar to the proportion reporting that they lived in urban areas, these groups were not identical. Table 2 presents, for each of the metropolitan areas, the distribution of respondents’ geographic location, as well as their reports about the types of communities they inhabited.

As Table 2 illustrates, the degree to which respondents’ perceptions of the type of com- munities they inhabited matched the labels suggested by their ZIP codes varied widely across the eight regions. The top two rows portray the rate of coordination between re- spondents’ geographic address and their perceptions of community, while the bottom two rows portray the degree to which there was ambiguity in respondents’ ideas about where they lived. Across all regions, there was ambiguity in approximately one-third of cases, with 15 percent of respondents reporting that they lived in a city or urban area despite a suburban or rural ZIP code, and 17 percent of respondents reporting that they lived in a suburban or rural area despite a central-city ZIP code.

That level of ambiguity varied across the eight metropolitan areas. It was lowest in the older industrial cities of Philadelphia and Detroit. These are among the nation’s largest metropolitan areas, and they feature stark city-suburb divides (see, e.g., Ehrenhalt 2012; Sugrue 1996, chapter 6). Among respondents in the Detroit region, only 20 percent characterized their communities in ways that contradicted what their ZIP codes would suggest, and in the Philadelphia region, the rate of disjuncture was below 10 percent. By comparison, ambiguity was higher in the other regions, and was especially high in the Miami region.

As Table 2 indicates, then, whether people live within the municipal boundaries of a region’s central city or in its suburbs or rural periphery does not, by itself, determine

867

CITY & COMMUNITY

T A

B L

E 2.

G eo

gr ap

h ic

L o

ca ti

o n

s an

d D

es cr

ip ti

o n

s o

f R

es id

en ti

al C

o m

m u

n it

y am

o n

g R

es p

o n

d en

ts in

E ig

h t

M et

ro p

o li

ta n

A re

as

A kr

o n

C h

ar lo

tt e

C o

lu m

b ia

D et

ro it

M ia

m i

P h

il ad

el p

h ia

Sa n

Jo se

W ic

h it

a T

o ta

l

R ep

o rt

s L

iv in

g in

an U

rb an

C o

m m

u n

it y,

L iv

es in

C en

tr al

C it

y 15

.9 %

17 .4

% 16

.0 %

5. 0%

34 .8

% 25

.7 %

18 .3

% 42

.8 %

20 .6

% R

ep o

rt s

L iv

in g

in a

N o

n u

rb an

C o

m m

u n

it y,

L iv

es O

u ts

id e

C en

tr al

C it

y 51

.0 %

48 .7

% 51

.2 %

74 .7

% 10

.4 %

64 .6

% 39

.4 %

24 .0

% 47

.0 %

R ep

o rt

s L

iv in

g in

an U

rb an

C o

m m

u n

it y,

L iv

es O

u ts

id e

C en

tr al

C it

y 20

.4 %

16 .4

% 8.

9% 18

.2 %

11 .7

% 5.

4% 25

.1 %

9. 2%

15 .4

% R

ep o

rt s

L iv

in g

in a

N o

n u

rb an

C o

m m

u n

it y,

L iv

es in

C en

tr al

C it

y 12

.7 %

17 .6

% 24

.0 %

2. 0%

43 .1

% 4.

3% 17

.2 %

24 .0

% 17

.0 %

N 84

9 83

2 33

8 83

4 82

7 83

0 81

8 32

5 5,

65 3

868

IDENTIFYING THE URBAN

how their community appears or how it is experienced. This is not surprising; indeed, communities’ physical, social, and economic characteristics are arguably stronger indi- cators of “urbanism” than the often arbitrary divides of municipal boundaries (Gieryn 2000). If, as Table 2 shows, individuals’ geographic location cannot entirely explain their characterization of their communities, then what other factors matter in shaping that understanding?

The results in Table 3 begin to address that question. Table 3 presents logistic regres- sion models predicting the odds of a respondent reporting that she or he resided in “a city or urban area” within her or his metropolitan area.4 In Model 1, we examine, by itself, the effect of respondents’ geographic location on the likelihood of reporting an urban residence. As expected, the effect was strong and highly statistically significant. Compared to those living outside the boundaries of their regions’ central cities, the odds that those with central-city addresses would report that they lived in an urban area were nearly four times higher.

However, while still substantial and highly significant, the effect of geographic location was somewhat attenuated when the safety and school quality variables were taken into consideration. As Model 2 indicates, with the safety and school quality variables in the equation, the odds of central-city residents reporting an urban home remained nearly four times higher than those of noncity residents. Their evaluations of local safety mat- tered, too, though in Model 2 the impact of evaluations of local schools was not statisti- cally significant. Controlling for whether they actually resided in the central city, those who perceived their local community to be safe were significantly less likely to report that they lived in an urban environment. For every one-point increase on the five-point safety scale, the odds of reporting to live in an urban area declined by more than 20 percent.

In Model 3, we introduce a range of control variables regarding respondents’ charac- teristics, while retaining the key variables of interest, respondents’ geographic location and their evaluation of local schools and safety. In Model 3, residents’ geographic loca- tion remained a strong predictor of their perceptions of their communities. The school and safety evaluation variables also maintained their direction, but interestingly, the in- troduction of the control variables rendered the impact of school evaluations statistically significant (p < 0.01). The results in Model 3 indicate that, controlling for their geo- graphic location and other individual-level factors, every one-point increase in respon- dents’ evaluation of school quality on the five-point scale was associated with a 10 percent drop in the odds of identifying their location as urban.

The control variables themselves offer important information about the factors con- tributing to urban identity, as well. All else equal, homeowners and those with higher incomes were less likely to report that they resided in an urban area. Those with very young children (under six years old) were significantly more likely than their peers with no children or with older children to identify as urban residents. Interestingly, control- ling for other factors, the odds of identifying as urban were about 25 percent higher for men, compared to women. Respondents’ levels of education had no significant effect upon their characterizations of their communities, but their racial and ethnic identities did. White respondents were the least likely to identify as urban residents. Compared to whites, the odds of describing their communities as urban were 58 percent higher for black respondents and 54 percent higher for Hispanic respondents.

These findings make intuitive sense. Racial and ethnic minorities, renters, and lower- income people are more likely than whites, homeowners, and higher-income people in

869

CITY & COMMUNITY

TABLE 3. Odds Ratios from Logistic Regression Predicting the Odds of Reporting to Live in a City or Urban Community (N = 5,653)

Model 1 Model 2 Model 3

Respondent Lives in Central City Geographically

3.92*** 3.75*** 3.83** (3.00, 5.11) (2.87, 4.89) (2.93, 5.00)

Evaluation of Local School Quality (1 = very bad; 5 = very good)

0.97 0.90** (0.90, 1.05) (0.83, 0.97)

Evaluation of Local Safety (1 = very bad; 5 = very good)

0.79*** 0.85*** (0.74, 0.85) (0.79, 0.91)

Income (in Thousands of Dollars) 0.99*** (0.99, 1.00)

Employed 1.08 (0.89, 1.31)

Married 0.85 (0.69, 1.05)

Number of Children Under 6 Living at Home

1.29** (1.11, 1.51)

Number of Children 6–12 Living at Home

0.99 (0.84, 1.16)

Number of Children 13–17 Living at Home

0.89 (0.73, 1.07)

Male 1.25** (1.05, 1.50)

Homeowner 0.74* (0.58, 0.93)

Level of Satisfaction with Community

1.07 (0.97, 1.17)

Age in Years 1.01 (1.00, 1.01)

Years Living in Community 1.00 (0.99, 1.00)

Educational Attainment (Reference Group = Graduate or Professional Degree) Grade School or Less 1.45

(0.77, 2.72) Some High School 1.02

(0.64, 1.60) High School Graduate 1.15

(0.87, 1.53) Some College or Technical

School 1.08

(0.86, 1.36) College Graduate 0.90

(0.72, 1.12) Race/Ethnicity (Reference Group = Non-Hispanic White)

Black 1.58** (1.19, 2.10)

Hispanic 1.54** (1.16, 2.05)

Asian/Pacific Islander 1.33 (0.86, 2.08)

Native American 2.60** (1.37, 4.93)

Other 0.85 (0.44, 1.62)

Note: 95% confidence intervals in parentheses. Models also contain city- and/or county-level controls for population, population density, racial composition, homeownership rate, poverty rate, and proportion of the population living in census-designated “urban” areas. Complete results are available from the authors upon request. *** p < 0.001, **p < 0.01, *p < 0.05 (two-tailed).

870

IDENTIFYING THE URBAN

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

City Dweller, Low Safety and School Rating

Non-City Dweller, Low Safety and School Rating

City Dweller, High Safety and School Rating

Non-City Dweller, High Safety and School Rating

FIG. 1. Predicted probability of respondent characterizing community as “urban.”

the same regions to live in dense communities near the urban core (i.e., in places that feel more “urban”), regardless of whether their addresses actually place them physically within the boundaries of the city. Elements of the built environment — including the physical condition of structures, the presence of security apparatus like bars on windows or video surveillance, and visible “signs of disorder” — may all serve as symbolic indi- cators of a place’s “urban” character (Sampson and Raudenbush 2004). These charac- teristics, not captured in the SOTC data set, therefore serve as potential unmeasured confounders.5

Figure 1 portrays predicted probabilities that illustrate this point more clearly. This figure, drawing upon the results from Model 3 of Table 3, holds values of all variables constant,6 with the exception of ZIP code-determined geographic location and evalua- tion of local schools and safety. These factors are varied to produce four unique hypo- thetical scenarios; for each scenario, we predict the probability that a respondent with these characteristics would identify the community that she inhabits as “urban.” In the four scenarios, we compare the probability of a person whose ZIP code places her in the city against one whose ZIP code places her outside the city boundaries, and a person who gives a rating of five out of five to local schools and local safety against a person who gives those community characteristics a rating of one out of five.

As Table 1 indicated, about 36 percent of the total SOTC sample reported that they lived in an urban area. The scenarios presented in Figure 1 suggest that having an address within the municipal boundaries of a central city was an important, but often

871

CITY & COMMUNITY

TABLE 4. Odds Ratios from Logistic Regression Predicting the Odds of Reporting to Live in a City or Urban Area, Disaggregated by Respondents’ Race/Ethnicity

Black Hispanic White

Respondent Lives in Central City Geographically 1.02 10.25*** 4.78*** (0.45, 2.28) (3.63, 28.97) (3.48, 6.57)

Evaluation of Local School Quality (1 = very bad; 5 = very good) 0.91 1.08 0.85** (0.73, 1.14) (0.87, 1.35) (0.78, 0.94)

Evaluation of Local Safety (1 = very bad; 5 = very good) 0.81* 0.77** 0.87** (0.67, 0.98) (0.64, 0.93) (0.80, 0.96)

N 657 540 4,127

Note: 95% confidence intervals in parentheses. Models also include all other variables shown in Table 3. Complete results are available from the authors upon request. ***p < 0.001, **p < 0.01, *p < 0.05 (two-tailed).

insufficient, criterion for reporting an urban identity. The probability that our hypotheti- cal respondent would report that she inhabited an urban area was about 59 percent when her address was located in the central city and when she gave low ratings to both local schools and safety. The probability that that same respondent would report that she in- habited an urban area, however, dropped to just 33 percent when she gave high ratings to local schools and safety, as portrayed in the third bar of Figure 1. In fact, the probability that she would report that she inhabited an urban area under those circumstances was only slightly higher than the probability would be (28 percent) if she lived outside the cen- tral city but gave low ratings to local schools and safety, as portrayed in the second bar of Figure 1. As shown in the fourth bar, the probability of identifying as an urban resident was lowest (11 percent) when our hypothetical respondent was given a noncentral-city address and high ratings of local schools and safety.

The results in Table 3 and Figure 1 reflect data from all of the respondents taken to- gether. But they do not address the important possibility that conceptions of the urban may vary systematically across important social divisions. Given the persistence of racial segregation in American metropolitan areas (Sharkey 2013), we turn our attention in the following set of analyses to the varying ways in which black, Hispanic, and white respon- dents’ characterizations of their communities are sensitive to their geographic location and their perceptions of local safety and school quality.7 Table 4 reproduces the results of Model 3 in Table 3, broken down by racial and ethnic group. This table allows us to examine the degree to which black, Hispanic, and white respondents differ in the factors that they take into account when forming a perception of their local communities. In to- tal, 56 percent of black respondents, 52 percent of Hispanic respondents, and 31 percent of white respondents self-identified as residents of “urban” communities, yet as Table 4 indicates, there was a substantial divergence among blacks, Hispanics, and whites in the influence of geographic location on the labels that they applied to their communities. For Hispanic respondents who resided within the central city of their metro areas, the odds that they would describe their communities as “urban” were more than 10 times higher compared to Hispanics living outside the city limits. The odds of characterizing their communities as “urban” were nearly five times higher for white central city residents compared to their peers outside the central city. For black respondents, however, there was essentially no effect of geography on the likelihood of describing their communities as “urban.”

872

IDENTIFYING THE URBAN

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

City Dweller, Low Safety and School Rating

Non-City Dweller, Low Safety and School Rating

City Dweller, High Safety and School Rating

Non-City Dweller, High Safety and School Rating

African American Respondents Hispanic Respondents Non-Hispanic White Respondents

FIG. 2. Predicted probability of respondent characterizing community as “urban,” by racial/ethnic identification.

Another important racial/ethnic difference is evident in the effect of school evalua- tions. Among black and Hispanic respondents, perceptions of local school quality had no significant effect on the odds of describing their communities as “urban.” For white respondents, though, school evaluations were a significant factor in shaping their charac- terizations. Controlling for other factors, every one-point increase in whites’ perceptions of school quality was associated with a 15 percent decrease in the odds that they would describe their areas as “urban”. Evaluations of neighborhood safety were consistent across all racial groups, with increasing perceived safety leading to significantly lower odds of an “urban” classification.

The varying effects of geographic location, school evaluations, and feelings of safety thus complicate the picture of what it means to be “urban.” In Figure 2, we reproduce the four hypothetical scenarios that we used in Figure 1 to generate predicted probabili- ties of a given respondent (see note 6) characterizing her local community as “urban.” In Figure 2, though, we use the results from Table 4 to disaggregate these predicted proba- bilities by race and ethnicity.

The evidence in Figure 2 points to different experiences of urban and nonurban life for black, Hispanic, and white Americans. For both Hispanic and white respondents, there are strong indications that locational differences go a long way toward explaining their understanding of the character of their communities. For these two groups, the likelihood that respondents would describe their areas as “urban” was higher for city

873

CITY & COMMUNITY

dwellers than for residents living beyond the city limits, regardless of whether they gave high or low ratings to local schools and safety. Within those groups, though, schools and safety did matter. A Hispanic city dweller who gave low ratings to local schools and safety had the highest predicted probability in the sample of characterizing her community as “urban” (77 percent). Meanwhile, a white respondent living outside the central city who gave high ratings to local schools and safety had the lowest probability of thinking of her area as “urban” (10 percent). For both whites and Hispanics, though, city residents with positive opinions of local schools and safety had a higher probability of reporting an “urban” identity than noncity residents with negative opinions of these institutions.

That pattern did not hold for black respondents. As Figure 2 indicates, black respon- dents’ geographic location played no significant role in predicting whether they would describe their local communities as “urban”; instead, these respondents’ evaluations of local safety (and, to a lesser extent, school quality) were a far more decisive factor. The probability that a hypothetical black respondent who gave local schools and safety low ratings would identify as an “urban” resident was nearly identical if her ZIP code placed her within or outside the city limits (44 percent in both cases); regardless of her address, the probability was also identical if she gave schools and safety a high rating (19 percent in both cases). The evidence of Table 4 and Figure 2 thus suggests that, for Hispanic and non-Hispanic white Americans, there is a stark difference in the experience of urban and nonurban life. For blacks, though, the lived experience of the city and its outlying communities is comparatively more similar.

These findings are in line with contemporary research on the characteristics of black suburban communities. In recent years, and particularly in the wake of the shooting of Michael Brown in suburban Ferguson, Missouri, scholars, journalists, and commentators have devoted increased attention to segregation, inequality, and poverty within inner-ring suburbs (Howell and Timberlake 2014; Jargowsky et al. 2014; Kneebone 2014; Lichter et al. 2015). While suburban communities blossomed in the 20th century as destinations for upwardly mobile whites seeking to escape the problems of the inner city (Jackson 1985), they always have been, and continue to be, diverse and segregated places. Be- cause suburbs are not monolithic, “the economic profile of individual suburbs is vitally important to understanding the quality of life of its citizens” (Howell and Timberlake 2014:94). While many black suburbanites enjoy affluence and comfort that match their white counterparts (Lacy 2007), on average black suburbs are characterized by higher lev- els of crime, violence, and poverty, resembling many predominantly black communities within central cities (Semuels 2015).

DISCUSSION

“Urban” is an imprecise term, open to multiple interpretations and contingent upon a variety of physical, demographic, and social factors. The label that a government bu- reaucrat or social scientist attaches to a given community does not necessarily reflect what those who inhabit that community believe about their geographic identity. Similarly, next-door neighbors might disagree about whether they live in an urban, suburban, or exurban area. Municipal boundaries matter, of course. Overall, our findings indicate that a postal address that places an individual within the official city limits is the best predic- tor of whether that individual identifies his or her community as “urban.” Yet municipal

874

IDENTIFYING THE URBAN

boundaries alone cannot account for the wide variation in individuals’ perceptions of their communities. When most people characterize their communities as “urban,” “sub- urban,” or “rural,” they do so not by pulling out a map, but by reflecting on how they experience daily life in that community.

As the analyses presented here indicate, two factors in particular — individuals’ as- sessments of the local schools and how safe they feel in their neighborhood — play a significant role in the identity ascribed to place. A person residing outside the borders of a region’s central city, but in a community where she felt unsafe and had little faith in the local schools, was about equally likely to say that she lived in an urban area as someone with the same characteristics who lived within the city borders, but who felt safe in her neighborhood and had high confidence in the local schools.

Importantly, however, the understanding of place also varies by race. Even when they inhabit similar parts of their respective metropolitan regions, black, Hispanic, and white Americans have different experiences and report different community identities. Most U.S. metropolitan areas no longer resemble the stark “Chocolate City, Vanilla Suburbs” pattern (Farley et al. 1978) that prevailed in the late 20th century. The lived experience of community is still racialized, however, even as racial and ethnic minorities increas- ingly settle in suburban communities and gentrification brings new cohorts of whites into central-city neighborhoods that their peers avoided in previous generations. For blacks, the geographical divide, at least as operationalized by ZIP code designation, is far less salient than it is for Hispanics and non-Hispanic whites. Rather, our analyses suggest that blacks see the distinction between urban and nonurban living more as a function of com- munity characteristics, especially personal safety. These social factors influence the per- ceptions of place for all respondents, but they are particularly meaningful for blacks.

Several limitations are worthy of mention. First, only eight U.S. metropolitan areas are represented in these data. While they are diverse in terms of size, character, and geogra- phy, they cannot reflect the vast array of urban, suburban, and exurban experiences in the United States, and therefore the generalizability of these findings is limited. Second, the SOTC data include only respondents from the state in which the principal city is located, potentially limiting our understanding of community perceptions in metropolitan areas (e.g., Philadelphia) that cross state lines. Our results for the determinants of urban and nonurban identity should therefore be interpreted with caution.8 Third, the sample con- tained disproportionate shares of homeowners, college graduates, and white residents, and these sample characteristics may pose an additional challenge to the generalizability of the findings.

Perhaps most importantly, there are many aspects of urban, suburban, and rural life that are not captured in the SOTC questionnaire. Because the data set does not pro- vide respondents’ raw ZIP code information, we do not know which central-city neigh- borhoods or outlying communities respondents inhabit within each metropolitan area. Moreover, there are no indicators of the characteristics of the housing, infrastructure, or built environment in the neighborhoods. We have tried to account for these place- level characteristics by incorporating county-level covariates, but it is likely that signif- icant levels of place-level variation remain unmeasured here. As previous research has demonstrated (Krysan et al. 2009; Sampson and Raudenbush 2004), visual cues re- lated to an area’s physical characteristics have a strong influence on the mental impres- sions that people attach to that area. So, too, do place-level density and the socioeco- nomic and demographic characteristics of the other people that one encounters. It is

875

CITY & COMMUNITY

reasonable to expect that these visual cues would also influence residents’ beliefs about local school quality and safety, the main factors we used to predict respondents’ labels for their communities. If so, the analyses presented here may be affected by omitted variable bias.

Finally, while the SOTC survey was originally designed to answer questions related to the community characteristics that influence resident satisfaction and community attach- ment, we have used the data to address a separate set of questions in this article. Thus, it is important to consider the potential for bias in our findings. The biggest concern in this regard is whether the original objective of the survey affected who chose to par- ticipate. If urban and suburban residents were unequally motivated to answer questions about community attachment, or if residents who were especially satisfied or dissatisfied with their communities were more motivated to participate than residents with a neutral view of their communities, then it is possible that the results reported here may have been skewed. However, there is no indication in the data to suggest a self-selection mech- anism of this sort. If the survey questions themselves are valid, if the survey was conducted properly, and if the data were coded accurately (all of which we believe to be true), then we believe it reasonable to use these data to investigate relevant questions that are not precisely within the scope of interest of the survey’s creators.

Having noted these concerns about the generalizability of the findings, we maintain that our research offers several insights for sociologists and policymakers alike. First, this study highlights the enduring relevance of the questions posed by early community the- orists concerning the importance of spatial conditions to the shaping of human experi- ences. While Simmel ([1903] 1995), and later Park and Burgess ([1925] 1967) and Wirth (1938), were primarily concerned with explaining the consequences for individuals and communities associated with the large-scale migration from low-density rural areas to high-density urban areas, their efforts to identify the distinctive characteristics associated with cities are echoed in contemporary studies of the obverse trend and its meaning: the population shift to suburban, exurban, or rural life (Jackson 1985; Salamon 2002; Walks 2013). By demonstrating the degree to which residents’ conception of what constitutes an urban or nonurban place is informed by their observations of and experiences in their own community, our findings reaffirm the need for scholars to attend to the psychologi- cal and relational components of physical spaces in their research.

More to the point, by examining the labels that individuals attach to the communities they inhabit, this study underscores the need to better understand how individuals per- ceive their residential environments. The disjuncture between where some individuals say they live and what a simple ZIP code analysis would indicate is particularly noteworthy for methodological reasons. Parallel to the idea of construct validity within the realm of sur- vey design and experimental research, investigators who explore the spatial patterning of behaviors or outcomes should give greater attention to the disparity between perceived (i.e., respondent-identified) location and assigned (i.e., researcher-identified) location when considering how geographic variables are measured.

Furthermore, to the extent that variation in respondents’ understanding of what con- stitutes an “urban,” “suburban,” or “rural” environment is patterned by race or ethnicity (or, for that matter, socioeconomic status or similar individual-level factors), researchers who are not sensitive to these differences run the risk of introducing bias into their anal- yses, or drawing unsubstantiated inferences from their findings. Such concerns under- lie the use of community-based participatory research and similar approaches designed

876

IDENTIFYING THE URBAN

to ensure that the language choices and perspectives of a given community’s mem- bers are reflected in a study’s research design and methods (Cheatham-Rojas and Shen 2003).

Our findings likewise demonstrate the influence that local organizations and institu- tions have on individuals’ perceptions of place. As Allard and Small (2013) argued in their call for an organizational perspective in urban sociology, “ . . . understanding the conditions of . . . highly disadvantaged populations requires a focus not only on individ- uals and their neighborhoods but also, and perhaps more importantly, on the organiza- tions that structure their lives, the systems in which those organizations are embedded, and institutions that regulate the operation of both” (p. 8). According to McQuarrie and Marwell (2009), organizations are a “missing dimension” in the study of urban sociol- ogy, critical to understanding “the link between institutional transformations and urban neighborhoods” (p. 247).

Indeed, from a policy perspective, the lack of support for metropolitan-level policies in housing, education, transportation, and the like is often attributed to a desire to avoid not simply “the urban” in general but, more specifically, urban institutions such as schools (Katz and Liu 2000; Rusk 1999). Individuals’ perceptions of the quality of urban schools and the safety of urban neighborhoods are highly racialized. This is due, in large part, to school and housing policies that maintain segregated spaces and institutions, and thus effectively reify and perpetuate stereotypes of the urban and the nonurban. Advocates for regional approaches to school assignment and housing vouchers, for example, note that strategies that bridge city-suburb divides can produce far more equitable outcomes for individuals (see, e.g., Berube and Holmes 2015; Siegel-Hawley 2016); our findings suggest that they may also serve to weaken the mental associations that individuals draw between “the urban” and lower-quality institutions by addressing the unequal distribution of resources that create and sustain those associations in the first place.

This research affirms what is already widely suspected: The lived experience of our cities and regions varies across different groups, and broad labels like “urban” and “sub- urban” obscure the nuanced variations in place character in our metro areas. We con- tend, however, that it remains critically important to document and explain how people think about their communities. Particularly as traditional notions of the “urban,” the “suburban,” and the “exurban” continue to shift, it will become increasingly necessary to pay attention to how people interpret where they live — and, in turn, how researchers operationalize and study these places. Our findings lend support to the idea that the notion of the “urban” is as much a social and cultural label — and often a stigmatized one — as it is a physical or geographic label. How the boundary work engaged in by resi- dents or visitors within a community maps onto (or, perhaps, does not clearly map onto) the actual physical boundaries of that place may have significant implications for the future of inequality, migration, segregation, and gentrification within U.S. metropolitan regions.

Acknowledgments

Earlier versions of this article were presented at the annual meeting of the Midwest Socio- logical Society (Chicago, IL, March 23, 2016) and at the annual meeting of the American Sociological Association (Seattle, WA, August 20, 2016). The authors thank Zachary Neal, Amelia McNamara, Adam Thal, and Dawn Royal for guidance on questions related to the

877

CITY & COMMUNITY

data, as well as Russell Arben Fox for constructive feedback regarding the nature of com- munity identity.

Notes

1SOTC data were obtained via the ICPSR and are available for download for researchers wishing to repli-

cate or expand upon our analyses (https://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/35532?q=soul+ of+the+community).

2Detailed information about the regions included in these analyses, as well as the excluded regions, is avail-

able from the authors upon request. 3Detailed tables breaking down the descriptive statistics by city/noncity residence, by urban/nonurban iden-

tification, and by metropolitan region are available upon request, as are analyses that decompose the “nonur-

ban” respondents into “suburban,” “rural,” and “other” categories. 4In these models, the data are weighted using sampling weights assigned by Gallup. 5To address this possibility of confounding, all models in Tables 3 and 4 include region-level control vari-

ables measuring various demographic and economic characteristics of the places that respondents inhabited.

These include population, population density, the black proportion of the population, the proportion of owner-

occupied homes, the poverty rate, and the proportion of the population living in census-designated “urban”

areas. These data, drawn from the 2010 census, were applied to each case based on the smallest geographic

level of analysis that we have available for each respondent: their county (and, for central-city residents, their

city). Respondents whose ZIP code placed them within the central city of their region were assigned place char-

acteristics associated with that city. All other respondents were assigned place characteristics associated with

their county. This is an imperfect means of capturing local dynamics that contribute to people’s perceptions

of place, as city- and county-level data fail to measure important variations that occur within individual cities

and counties. Still, especially for people in places near the extremes of the urban-rural continuum, these place-

level differences may influence respondents’ perceptions of their communities. While not adequate measures

of local dynamics, in the absence of more specific geographic identifiers for the respondents to the SOTC sur-

vey, these regional variables are the best tools at our disposal for measuring potential place-level confounders

in respondents’ views of their communities. In the analyses presented in Table 3, these variables are almost

all nonsignificant. Though they are omitted here, complete results, including these region-level variables, are

available from the authors upon request. 6The hypothetical respondent portrayed in Figure 1 is an employed married white female homeowner with

no children who has attended some college but does not have a college degree. Age, years living in the com-

munity, and income are set at their mean values (see Table 1), while level of community satisfaction is set at its

modal value (4 out of 5). 7Similar analyses disaggregating the sample into income categories are available upon request. 8To examine the possibility that the selection of these eight regions substantially affected the results, we

reran all analyses included in this article using a sample that included respondents from all 26 SOTC regions

(N = 12,736). The results did not differ substantially from the findings reported here, but these supplementary analyses are available upon request.

REFERENCES

Allard, Scott W., and Mario L. Small. 2013. “Reconsidering the Urban Disadvantaged: The Role of Systems, Institutions, and Organizations.” Annals of the American Academy of Political and Social Science 647(1):6–20.

Auletta, Ken. 1982. The Underclass. New York, NY: Random House. Banfield, Edward C. 1974. The Unheavenly City Revisited. Boston, MA: Little, Brown & Co.

878

IDENTIFYING THE URBAN

Berube, Alan, and Natalie Holmes. 2015. “Affirmatively Furthering Fair Housing: Considerations for the New Geography of Poverty.” Brookings. Available at: https://www.brookings.edu/opinions/affirmatively- furthering-fair-housing-considerations-for-the-new-geography-of-poverty/ [Accessed 31 July 2017].

Billingham, Chase M. 2015a. “The Broadening Conception of Gentrification: Recent Developments and Av- enues for Future Inquiry in the Sociological Study of Urban Change.” Michigan Sociological Review 29:75–102.

———. 2015b. “Parental Choice, Neighbourhood Schools, and the Market Metaphor in Urban Education Re- form.” Urban Studies 52:685–701.

——— and Shelley M. Kimelberg. 2013. “Middle-class Parents, Urban Schooling, and the Shift From Consump- tion to Production of Urban Space.” Sociological Forum 28:85–108.

Brown-Saracino, Japonica. 2009. A Neighborhood That Never Changes: Gentrification, Social Preservation, and the Search for Authenticity. Chicago, IL: University of Chicago Press.

———. 2010. The Gentrification Debates. New York, NY: Routledge. Campbell, Elizabeth, Julia R. Henly, Delbert S. Elliott, and Katherine Irwin. 2009. “Subjective Constructions

of Neighborhood Boundaries: Lessons from a Qualitative Study of Four Neighborhoods.” Journal of Urban Affairs 31:461–90.

Castells, Manuel. 1977. The Urban Question: A Marxist Approach. Cambridge, MA: MIT Press. Chaskin, Robert J. 1997. “Perspectives on Neighborhood and Community: A Review of the Literature.” Social

Service Review 71:521–47. Cheatham-Rojas, Ann, and Eveline Shen. 2003. “Community-based Participatory Research with Cambodian

Girls in Long Beach, California: A Case Study.” In Meredith Minkler and Nina Wallerstein (eds.), Community Based Participatory Research for Health: From Process to Outcomes, pp. 121–35. San Francisco, CA: John Wiley & Sons.

Clark, Terry Nichols, Richard Lloyd, Kenneth K. Wong, and Pushpam Jain. 2002. “Amenities Drive Urban Growth.” Urban Affairs Review 24:493–515.

Coulton, Claudia. 2012. “Defining Neighborhoods for Research and Policy.” Cityscape: A Journal of Policy Develop- ment and Research 14:231–36.

Coulton, Claudia J., Jill Korbin, Tsui Chan, and Marilyn Su. 2001. “Mapping Residents’ Perceptions of Neigh- borhood Boundaries: A Methodological Note.” American Journal of Community Psychology 29:371–83.

———, M. Zane Jennings, and Tsui Chan. 2013. “How Big is My Neighborhood? Individual and Contextual Effects on Perceptions of Neighborhood Scale.” American Journal of Community Psychology 51:140–50.

Cucchiara, Maia Bloomfield. 2013. Marketing Schools, Marketing Cities: Who Wins and Who Loses When Schools Become Urban Amenities. Chicago, IL: University of Chicago Press.

Dahl, Trayci, Rosario Ceballo, and Marisela Huerta. 2010. “In the Eye of the Beholder: Mothers’ Perceptions of Poor Neighborhoods as Places to Raise Children.” Journal of Community Psychology 38:419–34.

Ehrenhalt, Alan. 2012. The Great Inversion and the Future of the American City. New York, NY: Vintage Books. Farley, Reynolds, Howard Schuman, Suzanne Bianchi, Diane Colasanto, and Shirley Hatchett. 1978. “Chocolate

City, Vanilla Suburbs: Will the Trend Toward Racially Separate Communities Continue?” Social Science Research 7:319–44.

Firey, Walter. 1945. “Sentiment and Symbolism as Ecological Variables.” American Sociological Review 10:140–48. Fitz, Brittany M., Larry Lyon, and Robyn Driskell. 2016. “Why People Like Where They Live: Individual- and

Community-level Contributors to Community Satisfaction.” Social Indicators Research 126:1209–24. Florida, Richard. 2015. “How Gentrifiers Change the Definition of a Neighborhood.” CityLab. Available at:

http://www.citylab.com/housing/2015/03/how-gentrifiers-change-the-definition-of-a-neighborhood/ 386986/ [Accessed 3 May 2017].

Foster, Kirk A., and J. Aaron Hipp. 2011. “Defining Neighborhood Boundaries for Social Measurement: Ad- vancing Social Work Research.” Social Work Research 35:25–35.

Frankenberg, Erica, and Gary Orfield. 2012. The Resegregation of Suburban Schools: A Hidden Crisis in American Education. Cambridge, MA: Harvard Education Press.

Gadsden, Vivian L., and Ezekiel J. Dixon-Román. 2017. “‘Urban’ Schooling and ‘Urban’ Families: The Role of Context and Place.” Urban Education 52:431–59.

Gans, Herbert J. [1962] 1995. “Urbanism and Suburbanism as Ways of Life: A Re-evaluation of Definitions.” In Philip Kasinitz (ed.), Metropolis: Center and Symbol of Our Times, Chapter 11, pp. 170–95. New York, NY: NYU Press.

———. 2015. “America’s Two Urban Sociologies.” City & Community 14:239–41. Gieryn, Thomas. 2000. “A Space for Place in Sociology.” Annual Review of Sociology 26:463–96.

879

CITY & COMMUNITY

Gotham, Kevin Fox. 2002. Race, Real Estate, and Uneven Development: The Kansas City Experience, 1900–2000. Albany, NY: SUNY Press.

Guest, Avery M., and Barrett A. Lee. 1984. “How Urbanites Define Their Neighborhoods.” Population and Envi- ronment 71:32–56.

Haney, Wava G., and Eric S. Knowles. 1978. “Perception of Neighborhoods by City and Suburban Residents.” Human Ecology 6:201–14.

Hankins, Katherine. 2007. “The Final Frontier: Charter Schools as New Community Institutions of Gentrifica- tion.” Urban Geography 28:113–28.

Hanlon, Bernadette. 2009. “A Typology of Inner-ring Suburbs: Class, Race, and Ethnicity in U.S. Suburbia.” City & Community 8:221–46.

———, John Rennie Short, and Thomas Vicino. 2009. Cities and Suburbs: New Metropolitan Realities in the U.S. New York, NY: Routledge.

Hart, Timothy C., and Jeremy Waller. 2013. “Neighborhood Boundaries and Structural Determinants of Social Disorganization: Examining the Validity of Commonly Used Measures.” Western Criminology Review 14(3):16– 33.

Hipp, John R., Robert W. Faris, and Adam Boessen. 2012. “Measuring ‘Neighborhood’: Constructing Network Neighborhoods.” Social Networks 34:128–40.

Hirschman, Albert O. 1970. Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cam- bridge, MA: Harvard University Press.

Holliday, Amy L., and Rachel E. Dwyer. 2009. “Suburban Neighborhood Poverty in U.S. Metropolitan Areas in 2000.” City & Community 8:155–76.

Howell, Aaron J., and Jeffrey M. Timberlake. 2014. “Racial and Ethnic Trends in the Suburbanization of Poverty in U.S. Metropolitan Areas, 1980–2010.” Journal of Urban Affairs 3679–98.

Hwang, Jackelyn. 2016. “The Social Construction of a Gentrifying Neighborhood: Reifying and Redefining Identity and Boundaries in Inequality.” Urban Affairs Review 52:98–128.

Jackson, Kenneth T. 1985. Crabgrass Frontier: The Suburbanization of the United States. New York, NY: Oxford Uni- versity Press.

Jargowsky, Paul, Debra J. Rog, and Kathryn A. Henderson. 2014. “Suburban Poverty and Racial Segregation.” Urban Publications. Paper 1307.

Katz, Bruce, and Amy Liu. 2000. “Moving Beyond Sprawl: Toward a Broader Metropolitan Agenda.” Brookings Review 18(2):31–34.

Keller, Suzanne. 1968. The Urban Neighborhood: A Sociological Perspective. New York, NY: Random House. Kimelberg, Shelley McDonough. 2014. “Middle-class Parents, Risk, and Urban Public Schools.” In Annette,

Lareau and Kimberly Goyette (eds.), Choosing Homes, Choosing Schools, Chapter 7, pp. 207–36. New York, NY: Russell Sage Foundation.

Kneebone, Elizabeth. 2014. Ferguson, Mo. Emblematic of Growing Suburban Poverty. The Avenue Blog, The Brookings Institution. Available at: https://www.brookings.edu/blog/the-avenue/2014/08/15/ferguson- mo-emblematic-of-growing-suburban-poverty/[Accessed 30 November 2016].

Kneebone, Elizabeth, and Alan Berube. 2013. Confronting Suburban Poverty in America. Washington, DC: Brook- ings Institution Press.

Knight Foundation. 2017. About Knight Soul of the Community. Available at: https:// knightfoundation.org/sotc/about-knight-soul-community/ [Accessed 29 August 2017].

Kolko, Jed. 2015. “How Suburban are Big American Cities?” FiveThirtyEight May 21. Available at: https://fivethirtyeight.com/features/how-suburban-are-big-american-cities/ [Accessed 3 May 2017].

Krysan, Maria, Mick P. Couper, Reynolds Farley, and Tyrone Forman. 2009. “Does Race Matter in Neighborhood Preferences? Evidence from a Video Experiment.” American Journal of Sociology 115:527–59.

Lacy, Karyn. 2007. Blue-Chip Black: Race, Class, and Status in the New Black Middle Class. Berkeley, CA: University of California Press.

Lee, Barrett A., and Karen E. Campbell. 1997. “Common Ground? Urban Neighborhoods as Survey Respon- dents see them.” Social Science Quarterly 78:922–36.

Lees, Loretta, Tom Slater, and Elvin Wyly. 2008. Gentrification. New York, NY: Routledge. Lichter, Daniel T., Domenico Parisi, and Michael C. Taquino. 2015. “Toward a New Macro-Segregation? De-

composing Segregation Within and between Metropolitan Cities and Suburbs.” American Sociological Review 80:843–73.

Lloyd, Richard. 2002. “Neo-Bohemia: Art and Neighborhood Redevelopment in Chicago.” Journal of Urban Affairs 249:517–32.

880

IDENTIFYING THE URBAN

Logan, John R., and O. Andrew Collver. 1983. “Residents’ Perceptions of Suburban Community Differences.” American Sociological Review 48:428–33.

Loury, Glenn C. 2002. The Anatomy of Racial Inequality. Cambridge, MA: Harvard University Press. McQuarrie, Michael, and Nicole P. Marwell. 2009. “The Missing Organizational Dimension in Urban Sociology.”

City & Community 8(3):247–68. Milner, H. Richard. 2012. “But what is Urban Education? Urban Education 47:556–61. Neal, Zachary P., and Jennifer Watling Neal. 2012. “The Public School as a Public Good: Direct and Indirect

Pathways to Community Satisfaction.” Journal of Urban Affairs 34:469–86. Norman, Jon R. 2013. Small Cities USA: Growth, Diversity, and Inequality. New Brunswick, NJ: Rutgers University

Press. Oakley, Deirde. 2015. “Broadening our Focus: Regional Cities as a New Frontier of Urban Sociology.” City &

Community 14:249–53. Park, Robert E., and Ernest W. Burgess. 1925 [1967]. The City: Suggestions for Investigations of Human Behavior in

the Urban Environment. Chicago, IL: The University of Chicago Press. Posey-Maddox, Linn. 2014. When Middle-class Parents Choose Urban Schools: Class, Race, and the Challenge of Equity

in Public Education. Chicago, IL: University of Chicago Press. ———. 2016. “Challenging the Dichotomy between ‘Urban’ and ‘Suburban’ in Education Discourse and Pol-

icy.” The Educational Forum 80:225–41. Raphael, Stephen, and Michael A. Stoll. 2010. Job Sprawl and the Suburbanization of Poverty. Washington, DC: The

Brookings Institution. Rich, Meghan Ashlin. 2009. “‘It Depends On How You Define Integrated’: Neighborhood Boundaries and

Racial Integration in a Baltimore Neighborhood.” Sociological Forum 24:828–53. Rusk, David. 1999. Inside Game/Outside Game: Winning Strategies for Saving Urban America. Washington, DC: Brook-

ings Institution Press. Salamon, Sonya. 2002. Newcomers to Old Towns: Suburbanization of the Heartland. Chicago, IL: University of Chicago

Press. Sampson, Robert J. 2012. Great American City: Chicago and the Enduring Neighborhood Effect. Chicago, IL: University

of Chicago Press. ———, and W. Byron Groves. 1989. “Community Structure and Crime: Testing Social-Disorganization Theory.”

American Journal of Sociology 94:774–802. ———, Jeffrey D. Morenoff, and Thomas Gannon-Rowley. 2002. “Assessing ‘Neighborhood Effects’: Social Pro-

cesses and New Directions in Research.” Annual Review of Sociology 28:443–78. ———, and Stephen W. Raudenbush. 2004. “Seeing Disorder: Neighborhood Stigma and the Social Construc-

tion of ‘Broken Windows.” Social Psychology Quarterly 67:319–42. Sastry, Narayan, Anne R. Pebley, and Michela Zonta. 2002. Neighborhood Definitions and the Spatial Dimension of

Daily Life in Los Angeles. Los Angeles, CA: California Center for Population Research. Schlichtman, John Joe, and Jason Patch. 2014. “Gentrifier? Who, me? Interrogating the Gentrifier in the Mir-

ror.” International Journal of Urban and Regional Research 38:1491–508. Semuels, Alana. 2015. “The Destruction of a Black Suburb.” The Atlantic. Available at: http://www.theatlantic.

com/business/archive/2015/07/lincoln-heights-black-suburb/398303/ [Accessed 3 May 2017]. Sharkey, Patrick. 2013. Stuck in Place: Urban Neighborhoods and the End of Progress toward Racial Equality. Chicago,

IL: University of Chicago Press. ———, and Jacob W. Faber. 2014. “Where, When, Why, and for Whom do Residential Contexts Matter? Moving

Away from the Dichotomous Understanding of Neighborhood Effects.” Annual Review of Sociology 40:559–79. Short, John Rennie, Bernadette Hanlon, and Thomas Vicino. 2007. “The Decline of Inner Suburbs: The New

Suburban Gothic in the United States.” Geography Compass 1:641–56. Siegel-Hawley, Genevieve. 2016. When the Fences Come Down: Twenty-First Century Lessons from Metropolitan School

Desegregation. Chapel Hill, NC: University of North Carolina Press. Simmel, Georg [1903] 1995. ”The Metropolis and Mental Life.” In Philip Kasinitz (ed.), Metropolis: Center and

Symbol of Our Times, pp. 30–45. New York, NY: NYU Press, Chapter 3. Spilsbury, James C., Jill E. Korbin, and Claudia J. Coulton. 2009. “Mapping Children’s Neighborhood Percep-

tions: Implications for Child Indicators.” Child Indicators Research 2(2):111–31. Sugrue, Thomas J. 1996. The Origins of the Urban Crisis: Race and Inequality in Postwar Detroit. Princeton, NJ:

Princeton University Press. Tissot, Sylvie. 2015. Good Neighborhoods: Gentrifying Diversity in Boston’s South End. New York, NY: Verso. Walks, Alan. 2013. “Suburbanism as a Way of Life, Slight Return.” Urban Studies 50(8):1471–88.

881

CITY & COMMUNITY

Wellman, Barry. 1979. “The Community Question: The Intimate Networks of East Yorkers.” American Journal of Sociology 84:1201–31.

Wen, Ming, Louise Hawkley, and John Cacioppo. 2006. “Objective and Perceived Neighborhood Environment, Individual SES and Psychosocial Factors, and Self-Rated Health: An Analysis of Older Adults in Cook County, Illinois.” Social Science and Medicine 63:2575–90.

Wilson, Jill H., and Nicole Prchal Svajlenka. 2014. Immigrants Continue to Disperse, with Fastest Growth in the Suburbs. Washington, DC: The Brookings Institution.

Wirth, Louis. 1938. “Urbanism as a Way of Life.” American Journal of Sociology 44:1–24. Zukin, Sharon. 2010. Naked City: The Death and Life of Authentic Urban Places. New York, NY: Oxford University

Press.

882