Thought Paper Or Reflection Paper "OPINION"

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Ethnic Diversity and Its Effects on Social Cohesion Tom van der Meer1 and Jochem Tolsma2 1 Department of Political Science, University of Amsterdam, 1012 DL Amsterdam, The Netherlands; email: [email protected] 2 Department of Sociology, Radboud University, 6500 HE Nijmegen, The Netherlands; email: [email protected]

Annu. Rev. Sociol. 2014. 40:459–78

First published online as a Review in Advance on June 16, 2014

The Annual Review of Sociology is online at soc.annualreviews.org

This article’s doi: 10.1146/annurev-soc-071913-043309

Copyright c© 2014 by Annual Reviews. All rights reserved

Keywords

ethnic heterogeneity, social capital, neighborhood effects, constrict theory, ethnic threat, anomie

Abstract

Recent years have seen a sharp increase in empirical studies on the constrict claim: the hypothesized detrimental effect of ethnic diversity on most if not all aspects of social cohesion. Studies have scrutinized effects of different measures of ethnic heterogeneity in different ge- ographical areas on different forms of social cohesion. The result has been a cacophony of empirical findings. We explicate mechanisms likely to underlie the negative relationship between ethnic heterogeneity and social cohesion: the homophily principle, feelings of anomie, group threat, and social disorganization. Guided by a clear conceptual frame- work, we structure the empirical results of 90 recent studies and observe three patterns. We find that (a) there is consistent support for the con- strict claim for aspects of social cohesion that are spatially bounded to neighborhoods, (b) support for the constrict claim is more common in the United States than in other countries, and (c) ethnic diversity is not related to less interethnic social cohesion. We discuss the implications of these patterns.

459

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

INTRODUCTION

As cross-national immigration increased sharply after the 1960s, advanced Western societies became ever more diverse. When aim- ing to deal with migration, policy makers fre- quently refer to the potential consequences of ethnic diversity for the cohesiveness of society. There is, however, little scholarly agreement on the relationship between the ethnic compo- sition of communities and the social cohesion within them. The research that investigates the diversity-cohesion relationship was initially dominated by economists (e.g., Knack & Keefer 1997, Alesina & La Ferrara 2000, Costa & Kahn 2003). The scholarly interest of soci- ologists and political scientists boomed more recently after a presentation of findings on the relationship between ethnic diversity and social cohesion by Robert Putnam at Uppsala Uni- versity in Sweden in 2006 (published in 2007).

Putnam’s (2007) findings were widely inter- preted as cause for concern. He argued that liv- ing in an ethnically heterogeneous environment was harmful to interpersonal trust and under- mined social connections between and within ethnic groups. Faced with ethnic diversity, peo- ple would tend “to hunker down—that is, to pull in like a turtle” (Putnam 2007, p. 149), or in common language, to retreat from social life. This claim came to be known as the con- strict proposition or constrict claim. If it were true, ongoing immigration would erode social cohesion. Putnam’s conclusion received wide attention in the media and among policy mak- ers, serving as input to public policy debates in various countries (cf. Hallberg & Lund 2005, Cheong et al. 2007).

Scholars rushed forward to test whether the constrict proposition held in a broad range of countries using many different indicators of social cohesion. In the six years after the publication of Putnam’s (2007) “E Pluribus Unum” article, more than 65 new empirical studies have been published, and many more are under way. Yet empirical findings on the effects of ethnic diversity are themselves highly diverse. Some studies confirm that ethnic

diversity undermines social cohesion; others reject the claim altogether or find a positive relationship between the two phenomena (cf. Portes & Vickstrom 2011). Although the cacophony of empirical findings has led to heated scholarly debates, one fundamental question has remained unanswered: Is ethnic diversity harmful to social cohesion?

The lack of consensus in the scientific com- munity is rooted in the fundamental problem that the constrict proposition lacks theoretical substantiation. The mechanisms through which the ethnic composition of one’s surroundings affects social connections have not been fleshed out. There are also serious discrepancies as to how the two core concepts (ethnic heterogene- ity and social cohesion) are understood and op- erationalized using available data. There are similar discrepancies in what constitutes a rele- vant social environment within which to expect diversity effects. To steer away from the cur- rent theoretical dead end, we do not perform yet another empirical analysis in yet another ge- ographical setting. Rather, after deriving theo- retical ideas about the mechanisms that may un- derlie negative diversity effects, we undertake a review of relevant empirical studies guided by three basic questions,

1. To what extent does ethnic heterogeneity affect different indicators of social cohe- sion differently?

2. To what extent do different indicators of ethnic heterogeneity affect aspects of so- cial cohesion differently?

3. To what extent does ethnic heterogene- ity in different geographical areas affect aspects of social cohesion differently?

THE PHENOMENA UNDER CONSIDERATION

Social Cohesion

Social cohesion may be regarded as the degree of interconnectedness between individuals that is both a result and cause of public and civic life. It encompasses feelings of commitment,

460 Van der Meer · Tolsma

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

trust, and norms of reciprocity and is demon- strated by participation in networks and civic organizations (cf. Chan et al. 2006). We use the label social cohesion and not social capi- tal because it is a more neutral and less con- voluted concept. Social capital has been fun- damentally redefined by consecutive authors as the resources that individuals extract from their networks (Bourdieu 1987), the number and density of social ties within a community (Coleman 1990), and as a public good (Putnam et al. 1994), leading to radically different def- initions and operationalizations (Fischer 2005, Portes & Vickstrom 2011). Social cohesion de- fined as interconnectedness remains an overly broad and obscure concept, just like the concept of social capital. For a better empirical under- pinning, we break down social cohesion into four dimensions: formality, mode, target, and geographical scope.1

First, a common distinction in the field is the one between formal and informal social cohesion, based on the level of institutionaliza- tion of social relationships. Pichler & Wallace (2007, p. 424), for instance, distinguish be- tween participation in “formally constituted organizations or activities” and informal bonds “[that are] particularistic, tied to particular people and social groups.” Unlike Pichler & Wallace (2007), we do not see generalized trust as an indicator of formal social cohesion, as it is not institutionalized.

Second, the interconnectedness or ties be- tween individuals may be formed by at least two modes: attitudinal (e.g., dislike, trust, and fear) and behavioral (e.g., contact, association).2

Hooghe (2007, p. 712), recognizing this dis- tinction, argues that “increasing dissimilarities within society will render it more difficult to de-

1 These four dimensions are not the only possible relevant dimensions. However, we are unable to classify papers based on other dimensions, such as bridging or bonding social co- hesion (Putnam 2001) or strong and weak ties (Granovetter 1983). 2 Our attitudinal indicators encompass cognitional aspects (e.g., stereotypes) and meta attitudes (e.g., perceived trust- worthiness).

velop trusting relations [e.g., attitude] (Macedo, 1999), but the same cannot be said about other, more structural components of social capital [e.g., behavior].”

Third, social cohesion can be distinguished by its target—the alters or the (group of ) person(s) with whom the respondent (ego) is connected. We distinguish three targets: the in-group (people with the same ethnic background as ego), the out-group (people with a different ethnic background from ego), and the general population. In the constrict proposition of Putnam (2007, p. 148), diversity erodes positive bonds not only between ethnic groups but also within one’s own group.

Lastly, geographical scope refers to the ge- ographical boundaries within which social ties are bound. Many forms of social cohesion may cross neighborhood, municipality, or even country boundaries, whereas others are explic- itly confined to specific geographical areas (e.g., contact with neighbors, perceptions of shared norms in one’s municipality, or trust in fellow countrymen). Wallman Lundåsen & Wollebæk (2013) stress that a spatial dimension should complement the commonly used generalized- particularized continuum of trust.

In short, the question of social cohesion is not the traditional “Who is connected?” or even “Who is connected to whom?” Rather, social cohesion fundamentally concerns the question, “Who is connected to whom, where, and how?” Lacking clear, theoretical, underlying mechanisms that link ethnic heterogeneity to social cohesion, scholars generally have not at- tempted to derive specific hypotheses about the relationships between ethnic heterogeneity and various indicators of social cohesion. Most have simply considered one specific aspect of social cohesion, commonly generalized social trust, which they regard as the key indicator of social cohesion (e.g., Alesina & La Ferrara 2002, Hooghe et al. 2009, Dinesen & Sønderskov 2012). Others have combined very different in- dicators of social cohesion to form one common scale with the aim of increasing measurement reliability and boosting statistical power to find corroborative evidence (e.g., Hero 2003,

www.annualreviews.org • Ethnic Diversity 461

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Lancee & Dronkers 2011), but this is prob- lematic because indicators of social cohesion generally correlate rather weakly with each other and because it clouds theoretical and conceptual differences between the indicators. Finally, scholars have tested the constrict hypothesis using a range of indicators of social cohesion, but without making convincing arguments about why relationships are to be expected (e.g., Costa & Kahn 2003, Gesthuizen et al. 2009, Gijsberts et al. 2012). Here we investigate the extent to which inconsistent results are due to the different indicators of social cohesion investigated.

Given that we define social cohesion as ties between individuals, we do not incorporate studies that (solely) relate ethnic diversity to relationships with organizations or policies— such as political participation, public goods pro- vision, institutional trust, and cross-national trust. We also do not discuss studies that exclu- sively focus on interethnic relations or neigh- borhood disorder. Both lines of research are so extensive that they would dominate our analyses.

Ethnic Heterogeneity

The ethnic composition of a geographical area can be characterized in many different ways. In the constrict literature, ethnic diversity is most commonly assessed by measures such as the fractionalization index (which is equivalent to the complement of the Herfindahl-Hirschman index). This measure can be interpreted as the chance that two randomly picked individuals living in the same geographical area have a dif- ferent ethnic background. The main alternative approach is the use of relative ethnic group size, such as the percentage of migrants or minority group members. Research on the link between ethnic heterogeneity and interethnic relations has favored such out-group measures. Finally, the ethnic composition of a locality can be de- scribed by its level of segregation. The index of dissimilarity has been the conventional mea- sure of segregation (Massey & Denton 1988). Henceforth, we use ethnic heterogeneity as the

overarching term to refer to diversity, relative group size, and segregation.

To date, theory has not specified whether the ethnic composition of the locality af- fects social cohesion through ethnic diversity, ethnic group size, segregation, or various com- binations thereof. But the impact of ethnic di- versity and relative group size differs theoreti- cally. A community with 80% whites and 20% blacks is as ethnically diverse as a community with 80% blacks and 20% whites. However, for individual community members, the out-group size differs strongly in these two communities. It is exactly for this color blindness of ethnic di- versity that some scholars favor relative group size as an indicator for the ethnic composition (cf. Laurence & Heath 2008). It may also mat- ter whether within the same geographical area groups intermingle or live in segregated spatial units (Rothwell 2012).

The conception of ethnic heterogeneity de- pends on the level of detail at which ethnic- ity is categorized. Ethnic categories may form a dichotomy (e.g., black versus white, native versus foreign) or they may be distinguished at a more detailed level (e.g., specific ethnic groups by generational status). Although we routinely speak of ethnic heterogeneity, over the years scholars have defined ethnic-racial groups based on combinations of very different criteria, including race, ethnicity, ancestry, cit- izenship, migration status, and distinctiveness (e.g., color, religion, or language). Which cri- teria are chosen seems to depend more on data availability and country idiosyncrasies than the- ory. This review exclusively covers studies that have used objective measures rather than indi- vidual perceptions of heterogeneity. Subjective evaluations of heterogeneity are likely to be in- fluenced by out-group attitudes and hence in- troduce endogeneity problems.

Geographical Area

Authors scrutinizing Putnam’s constrict propo- sition have used countries, regions, munic- ipalities, and neighborhoods to define the geographical areas within which cohesion is

462 Van der Meer · Tolsma

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

hypothesized to be affected. However, there are few convincing arguments on what con- stitutes a relevant geographical context. On the one hand, most daily activity takes place within relatively small geographical areas and residents are more likely to be aware of the level of ethnic heterogeneity in their immedi- ate social context. This argument drove many scholars to focus on relatively small geograph- ical contexts such as neighborhoods and mu- nicipalities. On the other hand, selective res- idential mobility—frequently mentioned as a possible reason why no detrimental effects of ethnic diversity are observed—is less likely to occur between larger geographical units. The latter argument has driven some authors to opt for regions or countries as their unit of analysis. However, whereas some localities (most notably countries) have clear physical, so- cial, and administrative boundaries, others do not (such as neighborhoods defined by blocks, tracts, or zip codes). This boundary problem and more generally the modifiable areal unit problem means that analytical results are sen- sitive to the definition of the geographical unit (Fotheringham & Wong 1991), leading some scholars to use “egohoods”—overlapping con- centric circles with the (residence of each) sur- vey respondent in the center of his or her own geographical space—as an alternative defini- tion of individuals’ social environment (e.g., Hipp & Boessen 2013, Dinesen & Sønderskov 2013).

We categorize study results based on the geographical scale of the higher-level unit(s). We distinguish country-level studies from stud- ies with localities larger than municipalities, municipality-level studies, and studies with lo- calities smaller than the municipality, although we recognize that the equivalence of these lev- els across countries, regions, and even stud- ies within the same country and/or region is questionable. We limit ourselves to ethnic heterogeneity effects of geographical localities and do not cover functional environments, as they invoke different constraints and different self-selection mechanisms. Studies on possible detrimental effects of ethnic diversity within

classrooms, the workplace, or associations re- main beyond the scope of this review.

THE ONE-SIZE-FITS-ALL MODEL

With these building blocks in place, we aim to assess why ethnic heterogeneity might under- mine social cohesion between and within eth- nic groups. The constrict claim was introduced as a logical, empirical possibility but lacked ex- tensive theoretical motivation (cf. Lee 2008). We construct a one-size-fits-all model in which different pathways explain how diversity, rela- tive group size, and segregation at different ge- ographical levels may cause different indicators of social cohesion to erode.

The first pathway relates conflict the- ory (e.g., Blumer 1958, Blalock 1967, Bobo & Hutchings 1996) to social disorganization theory (Shaw & McKay 1942, Sampson & Groves 1989). We use the label “conflict the- ory” freely to encompass group threat theory (Quillian 1995, 1996), ethnic competition the- ory (Scheepers et al. 2002), and integrated threat theory (Stephan & Stephan 2000). In this approach, trust in ethnically dissimilar others (or its opposite, ethnic hostility) plays a pivotal role. According to conflict theory, the ethnic out-group size within a geographical locality af- fects actual or perceived competition between ethnic groups over scarce material and imma- terial resources such as jobs, housing, power, safety, morality, and identity. This (perceived) ethnic group competition increases feelings of threat that undermine interethnic relations. In- terethnic distrust may subsequently lead to gen- eral distrust, by mechanisms specified in social disorganization theory. Because people distrust ethnic others, they avoid interethnic contact, and as a consequence of this retreat from social life, contact among coethnic residents decreases as well, with less social control, more general distrust, and fear of crime as a result.

The second pathway relates feelings of anomie—individual anxiety about the existence of shared societal norms and moral values—to social disorganization of the environment. According to this line of reasoning, ethnic

www.annualreviews.org • Ethnic Diversity 463

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

diversity within one’s social environment and concurrent linguistic diversity and diversity in social norms induce feelings of anomie. Blocked communication and a lack of reliable knowledge about shared social norms stimulate feelings of exclusion and aimlessness (Seeman 1959, Smith & Bohm 2008). As Lewicki & Bunker (1996, p. 123) put it, “Identification- based trust develops as one both knows and predicts the other’s needs, choices, and pref- erences and also shares some of those same needs, choices, and preferences as one’s own.” Once people experience anomie and no longer know how to behave in public, they are hesitant to meet and mingle with others, regardless of the ethnicity of their coresidents. Once again, as a consequence of less contact, social control will decrease, feeding general distrust, anxiety, and fear of crime.

Both mechanisms implicitly assume the homophily principle that people prefer to interact with others similar to themselves, even without a dislike for ethnic others (Lazarsfeld & Merton 1954, McPherson et al. 2001). “There is some consensus on the hypothesis that trust is developed more easily between actors resembling one another, who are familiar with one another and who have abundant access to information about the other’s previous track record or about the other’s trustworthiness” (Hooghe 2007, p. 717). However, the two mechanisms follow different pathways. In the threat mechanism, rising relative out-group and decreasing in-group sizes primarily stim- ulate feelings of ethnic group threat. Feelings of anomie find their origin in diverse environ- ments; the anomie mechanism stresses the lack of common language, identities, and values regardless of the size of the in-group itself.

Despite repeated references to the group threat mechanism in the constrict literature, there are reasons to be skeptical. Although sev- eral scholars have shown that interethnic rela- tions deteriorate with out-group size (especially within relatively large geographical areas), per- ceptions of out-group size, and feelings of ethnic competitive threat (e.g., Quillian 1995, 1996; Scheepers et al. 2002; Semyonov et al.

2004), studies have failed to demonstrate con- vincingly the crucial link between out-group size and ethnic threat (e.g., Wagner et al. 2006, Schlueter & Scheepers 2010, Savelkoul et al. 2011; but see Schlueter & Wagner 2008).3

Concurrently, interethnic contact increases with larger out-group sizes for both minor- ity and majority populations (Martinović 2013). Given that interethnic contact is negatively related to perceptions of ethnic threat (Schlueter & Wagner 2008, Pettigrew et al. 2010, Savelkoul et al. 2011) and positively related to interethnic trust (Allport 1954 [1979], Pettigrew & Tropp 2006), the threat and the anomie mechanisms are more likely to operate when there are fewer interethnic con- tact opportunities, and it is here that residential segregation becomes relevant (Stolle et al. 2013; see also Laurence 2011). It is precisely the level of segregation that affects contact opportunities; in segregated communities, ties between ethnic groups are eroded (Uslaner 2011a, Rothwell 2012). Segregation might thus moderate the relationship between hetero- geneity, anomie, and threat. Segregation may also directly impact feelings of threat as “seg- regation markedly enhances the visibility of a group; it makes it seem larger and more men- acing than it is” (Allport 1954 [1979], p. 269). Finally, according to Massey & Denton (1993, p. 138), segregation induces behaviors “that violate norms that are widely shared—by both blacks and whites—about what constitutes a good and desirable neighborhood,” so that “residents modify their routines and increas- ingly stay indoors.” Figure 1 graphically depicts this theoretical model.

The level of segregation is not the only likely moderator of the heterogeneity effect. The

3 The second crucial link in the competitive threat model be- tween (worsening) economic macro conditions and percep- tions of ethnic threat has met more consistent support. Also note that ethnic competition theory, which integrates conflict theory with social identity theory (Tajfel & Turner 1979), links ethnic heterogeneity not only to negative out-group bias (e.g., interethnic distrust) but also to positive in-group bias (e.g., in-group solidarity). This already goes against the constrict proposition.

464 Van der Meer · Tolsma

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Ethnic heterogeneity

Social cohesion

(all aspects)

Relative out-group size

Ethnic (group) threat

Interethnic social cohesion

Social disorganization

Ethnic diversity

Ethnic segregation

Anomie Intralocal

social cohesion Social

disorganization

+

+

+

– – –

+ – – –

Figure 1 A graphical summary of the one-size-fits-all model.

more members of ethnic out-groups are con- ceived to be different, the more they will be re- garded as a threat to the dominant culture, and the more uncertain residents will be in how to interact with them. Hence, ethnic heterogene- ity may be especially harmful if ethnic cleavages are deep, such as when they are cemented by an overlap with other socioeconomic cleavages (cf. Finseraas & Jakobsson 2012). Concurrently, in- terethnic contact between groups of differing status will reduce prejudice and increase in- terethnic trust less than contact between equal status groups (Allport 1954 [1979]). Admit- tedly, ethnic inequality may not only catalyze heterogeneity effects but also dampen them. According to the conflict literature, economic threat will be most severe between groups com- peting for the same resources.

SETTING UP THE REVIEW

In testing the validity of the constrict propo- sition, we restrict ourselves to articles written in English. We also exclude articles that we considered to be earlier versions of a later, more finalized article. To deal with possi- ble publication bias, we included (publicly available) unpublished conference papers (6), other working papers (3), reports (3), and book chapters (1), along with journal articles. Most of the conference papers over the past few years that we were aware of got published or are in press. The nine working and conference papers support and reject the constrict claim at equal rates. Hence, there is no indication for a file drawer problem. In all, we collected 90 unique studies (see the complete list

in the Supplemental Material; follow the Supplemental Material link from the An- nual Reviews home page at http://www. annualreviews.org).

Drawing on these studies, we assessed which indicators of social cohesion and heterogene- ity were studied and at which geographical level the analysis was specified. Results fall into three categories: support, reject, and mixed. In our scheme, studies provide support for the claim that heterogeneity hinders social cohe- sion when all indicators of ethnic heterogeneity in all investigated localities are significantly and negatively related to all scrutinized indicators of social cohesion. Conversely, when ethnic het- erogeneity is consistently not significantly and negatively related to any of the investigated in- dicators of social cohesion, the study uniformly falsifies the central claim and is labeled as re- ject. We label studies as mixed when findings are inconclusive, that is, when the overall con- clusions vary with different indicators of social cohesion and/or ethnic heterogeneity and/or at different localities. Studies are also labeled as mixed when conclusions vary for different sub- groups within the population.4 An overview of

4 We base our classification on the most rigorous models presented. We follow the significance criteria of the origi- nal publications. We do not categorize tests that were not presented in tables. When interaction effects with hetero- geneity are not significant, or when the significance of the diversity effect for each specific subgroup cannot be deter- mined, the classification of results is based on models with- out these cross-level interactions. When multiple hetero- geneity measures are included in the same analysis, we base our conclusions on models in which they are included sepa- rately (if available) to avoid collinearity issues. We disregard

www.annualreviews.org • Ethnic Diversity 465

Supplemental Material

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

0

5

Year of publication

N u

m b

er o

f p u

b lic

at io

n s

10

15

20

20 13

W or

ki ng

p ap

er s

20 12

20 11

20 10

20 09

20 08

20 07

20 06

20 05

20 04

20 03

20 02

20 01

20 00

19 99

19 98

19 97

1 1 1 1 1 1 1

1 1

1

1

1

3

3

1

3

3

2 2 2 2

3

4 3

4

4

2

3

4

7

7

5

5

6

Support (26)

Mixed (39)

Reject (25) Cross-

national: Europe

Cross- national:

Africa

Cross- national:

world

United States

United Kingdom

The Netherlands

Canada

Germany

Australia

Sweden

D enm

ark

Be BH

Es Fr

N O

N Z

R F

S L

U S

+ C

A U

S +

U K

6 E

U c

itie s

Figure 2 Site of study and degree of support for the heterogeneity claim, by year of publication (see the Supplemental Material for the full list of studies). Three studies used two different data sources. For the construction of Figure 2, we used the overall conclusion. In the remainder, we disaggregate tests of the constrict claim by data source. (Abbreviations: Be, Belgium; BH, Bosnia-Herzegovina; CA, Canada; Es, Spain; Fr, France; NO, Norway; NZ, New Zealand; RF, Russian Federation; SL, Sierra Leone.)

the information of all studies is available as Sup- plemental Tables (as are all tables belonging to the analyses we report below).

FIRST RESULTS: DIVERSE STUDIES, DIVERSE OUTCOMES

Cacophony of Findings

The left panel of Figure 2 considers literature that encompasses cross-national as well as within-country studies. Although research on the constrict proposition began mainly in the United States, by now at least one within-country study has been performed in 16 different countries. Research has been most prolific on the United States (18 studies), the United Kingdom (11), the Netherlands

interaction effects between macro-level variables (e.g., the diversity effect under different multicultural policies). All ar- ticles have been coded by the authors, and codings have been checked by three research assistants.

(10), and Canada (7), and on comparisons of countries (21). The right panel of Figure 2 shows how the literature boomed following the seminal works of Alesina & La Ferrara (2000, 2002) and Putnam (2007). More importantly, Figure 2 demonstrates the cacophony of empirical findings. For almost every study that finds uniform support for the constrict claim (26 in total), there is one that consistently rejects it (25 studies in total). Moreover, a plurality of the studies (39) finds mixed effects. The lack of consensus on the heterogeneity effects is thus staggering.

The Relevance of Methodological Rigor

The studies that offer the most rigorous tests of the constrict claim are those that control for ethnicity at the individual level, that control for alternative economic explanations at the macro level, and that take the nested structure of their data into account. It is important to control

466 Van der Meer · Tolsma

Supplemental Material

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

for respondents’ ethnicity, of course, as ethnic minority groups and immigrants are gener- ally less likely to trust others and less likely to participate in formal organizations than are ethnic majority groups and native populations (Putnam 2007). Communities with a larger share of ethnic minorities will on average show lower levels of trust and participation and higher levels of informal contact, even if the ethnic composition of the community itself has no effect on the prosocial attitudes and behav- iors of its residents.

In addition, high levels of ethnic hetero- geneity often go hand in hand with economic inequality and low levels of economic pros- perity (Letki 2008, Phan 2008). As economic deprivation and inequality may mediate the observed relationships between ethnic hetero- geneity and social cohesion, studies in which economic characteristics are included in the explanatory model can make a firmer claim that they are not reporting spurious effects of heterogeneity.5 Finally, individual-level studies that take the nested data structure (individu- als living in geographical areas) into account can deal, at least to some extent, with spatial autocorrelation. Two common alternatives fail to estimate the heterogeneity effect correctly: Aggregate-level studies cannot pull the indi- vidual effect of ethnicity and the contextual effect of ethnic heterogeneity apart, whereas individual-level studies without cluster correc- tions underestimate the standard error of the heterogeneity effect.

In light of these methodological considera- tions, we compare the most rigorous studies to those that failed to meet at least one of our crite- ria. The percentage of studies that find support- ive evidence is indeed substantially and signif- icantly lower among those studies with a more rigorous methodological design (44% versus 18%). Nevertheless, even among the method-

5 Admittedly, if ethnic heterogeneity is causally prior to eco- nomic poverty (e.g., if entrepreneurs leave heterogeneous lo- calities) or income inequality (e.g., if immigrants with low incomes try to settle in affluent localities), studies underesti- mate the heterogeneity effect.

ologically most rigorous studies, there is little consistency in the evidence (18% support, 24% reject, 58% mixed). Hence, it remains relevant to continue decomposing the evidence from a more theoretical perspective.

RESULTS

Most studies, especially more recent ones, do not rely on one single indicator of social co- hesion, one single measure of heterogeneity, one single contextual level of analysis, or even one data source. Instead, many studies contain multiple tests of the constrict proposition. In this section, we decompose study results based on the three crucial concepts we introduced above: cohesion, heterogeneity, and geograph- ical area.6

Different Indicators of Social Cohesion

We start by disaggregating studies based on the formality of the indicators used to measure so- cial cohesion (Table 1). Alesina & La Ferrara (2000) were the first to investigate the relation- ship between diversity and formal aspects of so- cial cohesion in their analysis of involvement in voluntary associations. The only subsequent study unequivocally supporting these findings is the one by Healy (2007), who assessed the link between diversity and volunteering in Melbourne, Australia. All other studies on the relationship between heterogeneity and formal social cohesion were either labeled as mixed (N = 8) or as reject (N = 12). No fewer than 82 studies analyzed informal indicators of so- cial cohesion. Corroborative evidence is found in half of these studies, although this figure drops to just over 20% if we exclude studies that did not meet all our methodological criteria. The overall conclusion seems to be that formal

6 Some studies cannot be categorized along all dimensions, for instance when they use an alternative diversity measure (such as the Theil entropy index), a different locality (such as chiefdoms), or a social capital index (such as those that combine behavioral and attitudinal aspects).

www.annualreviews.org • Ethnic Diversity 467

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Table 1 Support for the constrict claim, by dimension of social cohesion

Formality Mode Target Scope

Conclusion Formal Informal Behavior Attitude In-group Out-group

Intra- neighbor-

hood Other Support 9.1% 34.1% 14.3% 36.0% 20.0% 9.1% 50.0% 18.9% Mixed 36.4% 35.4% 42.9% 28.0% 60.0% 36.4% 41.2% 37.8% Reject 54.5% 30.5% 42.9% 36.0% 20.0% 54.5% 8.8% 43.2% N 22 82 35 75 5 11 34 37 chi-square 6.602∗ 5.771 4.613 12.958∗∗

Somer’s D −0.33∗∗ −0.18 NA 0.46∗∗∗

∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. NA, not applicable.

aspects of social cohesion are less likely to be undermined by the level of heterogeneity of one’s residential environment than are informal aspects.

Next, we turn to the mode dimension. In line with the expectation of Hooghe (2007), we find that heterogeneity negatively affects behavioral aspects of cohesion (14%) significantly less of- ten than it negatively affects attitudinal aspects (36%). Note, however, that the behavioral indi- cators overlap considerably with the formal in- dicators: All indicators of formal social cohesion are behavioral. Formal and behavioral aspects of cohesion may be less sensitive to changes in the ethnic composition of one’s living environ- ment or simply less malleable in general. How- ever, even for attitudinal aspects of cohesion, the evidence is far from convincing; just as many studies support the claim as reject the claim.

The third dimension refers to the target of social cohesion. Although an important innovation of the constrict claim lies in the suggestion that heterogeneity erodes the bonds between and within ethnic groups, only five studies included indicators of intraethnic social cohesion. These studies provide insufficient in- formation to draw firm conclusions: Evidence both in favor and against the constrict claim is weak. On the one hand, the scarce supportive evidence is based on one working paper using bivariate statistics. Although Putnam (2007) also presented a negative bivariate relationship between ethnic homogeneity of US communi-

ties and intraracial trust, and claimed that this effect “passes this same stringent multivariate, multilevel test” (p. 153, italics in original), this supportive evidence of the constrict claim is not included in Table 1 because our coding scheme focuses on the most rigorous tables presented in each manuscript. Even if we include Putnam’s study, however, the amount and frequency of corroborative evidence is very weak.

On the other hand, three of the four stud- ies that did not find consistent support focused on intraethnic contact. Huijts et al. (2014a,b) and Vervoort et al. (2011) demonstrate that higher levels of heterogeneity are related to less intraethnic contact among majority groups and to more intraethnic contact among mi- nority groups. For minority (majority) mem- bers, higher levels of heterogeneity generally imply more (less) opportunities to have con- tact with in-group members. Apparently, these contact opportunities are more important ex- planations of intraethnic contact than the threat and anomie mechanisms, although we cannot rule out the simultaneous existence of the lat- ter. Finally, in the fourth study that did not find consistent support [the German lost letter field experiment of Koopmans & Veit (2013)], the ethnicity of possible egos was unknown. They tested whether lost letters were more likely to be returned in homogeneous or heterogeneous neighborhoods and whether this depended on the ethnicity of the target (the addressee). How- ever, the nature of the experiment precluded

468 Van der Meer · Tolsma

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

assessing the ethnicity of the retriever or of the ones who decided to let it lie on the streets.

Eleven studies investigated heterogeneity effects on interethnic social cohesion. In gen- eral, the level of ethnic heterogeneity is not re- lated to less interethnic social cohesion. Only one (aggregate and bivariate) study finds sup- port for this part of the constrict claim. Six studies uniformly reject the constrict claim, four of which even demonstrate an inverse effect: that heterogeneity stimulates intereth- nic cohesion. These findings are very much in line with the wider literature on intereth- nic relations that commonly established posi- tive relationships between out-group size and interethnic attitudes and interethnic contact (Pettigrew & Tropp 2006, Wagner et al. 2006), and they demonstrate that for intereth- nic social cohesion—just as for intraethnic so- cial cohesion—contact opportunities trump the threat and anomie mechanisms.

The final dimension of social cohesion summarized in Table 1 is geographical scope. Effectively, along this dimension we can classify only those studies that investigated indicators of social cohesion that are inherently bound to neighborhoods—that is, contact with neighbors, trust in neighbors, and a wider feeling of shared norms and identity within the neighborhood—and compare them to studies of indicators of social cohesion that are not restricted to specific geographical areas. We find that intraneighborhood social cohesion is most consistently negatively affected by ethnic heterogeneity. Studies on intraneighborhood indicators of cohesion find considerable sup- port (50%) and very few rejects (9%) for the constrict claim. The support rates are hardly affected by controls for methodological rigor. Moreover, they are significantly different from studies on indicators of social cohesion that are not bound to the neighborhood, on which we find merely 19% support and 43% reject.

Our findings on intraneighborhood social cohesion tend to explain some earlier puzzles. First, the relatively strong support with respect to informal indicators of social cohesion can be traced back to their overlap with intraneigh-

borhood social cohesion: The category of informal social cohesion encompasses all intraneighborhood measures in our study. Sec- ond, the scholarly disagreement (differential findings) within the United Kingdom and the Netherlands is to a large extent explained by the coverage of intraneighborhood social cohesion in the analyses (cf. Gijsberts et al. 2012). Stud- ies in the United Kingdom, especially, now paint a very consistent picture. Letki (2008) concludes that in the United Kingdom ethnic heterogeneity is harmful only to neighbor- hood attitudes and not to non-neighborhood indicators of social cohesion (e.g., sociability, organizational involvement, and informal help relations). Nearly all other British studies that focused exclusively on intraneighborhood indicators of social cohesion found supportive evidence (Andrews 2009, Twigg et al. 2010, Bailey et al. 2012, Finney & Jivraj 2013; but see Pennant 2005), at least among one ethnic group (Fieldhouse & Cutts 2010), on one indicator of ethnic heterogeneity (Laurence & Heath 2008, Bécares et al. 2011), or conditional on interethnic contacts (Laurence 2011).

To what extent does ethnic heterogeneity affect different indicators of social cohesion differently? Ethnic heterogeneity most consis- tently negatively affects intraneighborhood co- hesion. Heterogeneity is not negatively related to interethnic cohesion.

Different Indicators of Ethnic Heterogeneity

Although ethnic diversity, group sizes, and seg- regation are theoretically very different con- cepts and may trigger distinct theoretical path- ways, relative group sizes often cannot be pulled apart empirically from diversity (Hooghe et al. 2009, Uslaner 2011b, Gijsberts et al. 2012). This strong empirical relationship between di- versity and group size may explain why we do not observe substantial differences in support for the constrict claim (see Table 2). Surpris- ingly, however, scholars who performed tests with different operationalizations of the ethnic composition on the same data, same dependent

www.annualreviews.org • Ethnic Diversity 469

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Table 2 Support for the constrict claim, by measure of heterogeneity

Heterogeneity

Conclusion HI % Support 29.8% 31.9% Mixed 35.1% 36.2% Reject 35.1% 31.9% N 57 47 chi-square 0.122 Somer’s D −0.04

Tests were not significant at p < 0.05.

variable(s), and same units of analysis also re- veal results that are not very robust, despite the (presumably) high correlations between them. Gustavsson & Jordahl (2008), Hooghe et al. (2009), and Hou & Wu (2009) find no negative effects of ethnic diversity but do find such ef- fects when relative group sizes are used instead.

We cover three studies that included an effect of segregation, two of which find corrob- orative evidence that segregation directly harms social cohesion. Unfortunately, Koopmans & Veit (2013; labeled as reject) and Rothwell (2012; labeled as support) do not report the impact of segregation without simultaneously controlling for ethnic diversity. Uslaner (2011a; labeled as support) shows only bivariate rela- tionships between segregation and cohesion indicators. All in all, studies that include main effects of segregation in their explanatory mod- els are rare (we did not include the studies of Uslaner 2011b or 2012 because we were not able to determine the main effect of segregation).

To what extent do different indicators of ethnic heterogeneity affect aspects of social co- hesion differently? Effects of ethnic diversity and the size of specific ethnic groups are em- pirically hard to pull apart. An answer requires data in which these measures are less entangled. Even so, results are not very robust as to how heterogeneity is operationalized.

Different Geographical Areas

The geographical level at which heterogene- ity is measured is neither significantly nor con-

sistently linked to support for the constrict claim (Table 3). Support is not more com- mon the more closely we zoom in on the lo- cality. In general, the greatest support is found at the regional level (38%) and the neighbor- hood level (32%), whereas most rejections are at the country level (55%) and the municipality level (43%), although these differences are not significant. Nonetheless, it is interesting that all of the cross-national studies that do find consis- tent support for the constrict claim fail to meet our criteria for methodological rigor. If selec- tive residential mobility hinders the discovery of heterogeneity effects within relatively small geographical units, the presumed absence of se- lective residential mobility at the country level does not help in discovering such effects. Sup- port in these cross-national studies does not depend on the set of countries or the survey included.

Finally, we grouped the within-country studies and find tentative evidence for Amer- ican exceptionalism (Table 3), although the number of studies gets rather small in many countries. Support for the constrict proposition is more common in the United States (50%) than in other Anglo-Saxon countries (17%) or in Europe (26%). This pattern remains largely unaffected once we focus on the methodologically robust studies (respectively 36%, 0%, and 16% support). As discussed above, most evidence for the constrict claim is found using measures of intraneighborhood social cohesion. However, in the United States, consistent support for the constrict claim is also found in two of the five studies (40%) that focused on elements of cohesion not bound to the neighborhood. By contrast, there is less consistent support for the constrict claim on indicators of social cohesion that are not bound to the neighborhood in other old immigration countries such as Canada, Australia, and New Zealand (20%) and in Europe (14%).

To what extent does ethnic heterogeneity in different localities affect aspects of social cohesion differently? Even though neighbor- hood heterogeneity erodes intraneighborhood cohesion, we observe no significant differences

470 Van der Meer · Tolsma

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Table 3 Support for the constrict claim, by locality and site

Site

Locality Within country Cross-national

Conclusion Country Region Munici-

pality Neighbor-

hood United States

Canada, Australia,

NZ Europe World Europe Support 27.3% 37.5% 21.4% 31.9% 50.0% 16.7% 25.6% 36.4% 11.1% Mixed 18.2% 43.8% 35.7% 46.8% 30.0% 50.0% 59.0% 18.2% 22.2% Reject 54.5% 18.8% 42.9% 21.3% 20.0% 33.3% 15.4% 45.5% 66.7% N 22 16 14 47 20 12 39 11 9 chi-square 10.805 7.240 Somer’s D −0.13 NA

Tests were not significant at p < 0.05. NA, not applicable.

between various levels of analysis. Ethnic het- erogeneity erodes intraneighborhood social co- hesion across the globe. There is some indica- tion that heterogeneity affects other indicators of social cohesion—most notably generalized trust—in the United States but not elsewhere.

SUMMING UP

In 2011, Alejandro Portes and Erik Vickstrom deliberately refrained from a detailed analysis of contradictory findings on the social effects of ethnic diversity and instead asked the more fundamental question: What is the fuss re- ally about? They concluded, “Although cries of alarm about declining social capital in the face of diversity have undoubtedly struck a chord, it is doubtful that the vast research program spawned by such fears has made American so- ciety any better or its public policies any more effective” (Portes & Vickstrom 2011, p. 477). Since their review, dozens of new studies on the relationship between ethnic diversity and social cohesion have been published. Yet a lack of theoretical substantiation on the mechanisms behind that supposed relationship has only in- creased the cacophony of seemingly contrary empirical findings. In this review, we set out to identify the trees within the forest of empiri- cal results using a theoretical model and clear definitions of the core concepts as our guides.

Three patterns structure the seemingly inconsistent findings observed across 90 empir- ical studies. First, ethnic heterogeneity is not consistently negatively related to interethnic cohesion. This finding goes against the first theoretical mechanism proposed under the threat hypothesis, but it is in line with empirical findings derived from both contact theory (cf. Pettigrew & Tropp 2006) and conflict theory. Increased interethnic contact opportuni- ties stimulate interethnic contact (Martinović 2013), which stimulates out-group trust directly and in-group trust and trust in neighbors indi- rectly via perceived threat (Schmid et al. 2013). Although ethnic threat itself leads to intereth- nic distrust, feelings of ethnic threat certainly do not always originate from living in hetero- geneous environments (Stolle et al. 2013).

The second empirical regularity is that in- traneighborhood cohesion is quite consistently eroded by the level of ethnic heterogeneity in neighborhoods. However, these negative het- erogeneity effects on trust in and contact with neighbors do not consistently spill over to other forms of social cohesion not bound to neigh- borhoods. There is “considerable lack of clar- ity on the causal mechanism regarding the way in which experiences with people one knows or with whom one is familiar can be transferred to people one does not know or with whom one is unfamiliar” (Marschall & Stolle 2004, p. 128).

www.annualreviews.org • Ethnic Diversity 471

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

The third regularity is the exceptional case of the United States. Across all countries we found strong support for the constrict claim on intraneighborhood social cohesion, but in the United States most consistently so. Moreover, there are hardly any negative effects of ethnic heterogeneity in Australia, Canada, New Zealand, or Europe on indicators of social cohesion other than intraneighborhood social cohesion. In the United States, however, we found some evidence that heterogeneity also erodes other aspects of social cohesion. This seems to be a case of American exceptionalism that is commonly suggested in the social capital literature.

BACK TO THE THEORETICAL DRAWING BOARD

Threat and Anomie

We began this review by raising the question of whether ethnic diversity is harmful to social cohesion. The answer appears to be yes, but only under very specific conditions. Constrict effects occur on intraneighborhood social cohesion, but otherwise they are highly context dependent. There is little evidence for them outside the United States. The group threat mechanism appears to be an unlikely explanation for the constrict claim, as ties between ethnic groups are generally stronger, not weaker, in ethnically heterogeneous envi- ronments. The main alternative mechanism we proposed—that heterogeneity harms social co- hesion by creating feelings of anomie—seems to support the constrict claim with respect to intraneighborhood social cohesion and may account for the lack of spillover effects outside the neighborhood. Uncertainty about how to behave and whom to trust in the neighborhood need not be extrapolated to general behavior or attitudes: “Most people are able to distinguish between how they conceive their immediate surroundings and how they regard the world as such” (Wallman Lundåsen & Wollebæk 2013, p. 303). Any retreat from (social) life in the neighborhood may even be compensated

by affirming ties outside the neighborhood (Aizlewood & Pendakur 2005).

Nevertheless, there are at least two rivaling explanations besides the anomie explanation as to why more support is found with respect to intraneighborhood social cohesion. Recall that social cohesion is a relational concept, fundamentally about ego (the respondent in a survey) and alters (his or her neighbors). Consequently, there are no less than four possible effects that have to do with ethnicity: (a) the ethnicity of ego, (b) the ethnicity of alter, (c) the ethnic combination of the dyad constituted by ego-alter (which is mixed or homogeneous), and (d ) the ethnic composition of the locality in which ego and alter are situated. Only the fourth is theoretically linked to the constrict claim; the other three basically indicate composition effects. The studies in our review cover only the first (ethnicity of ego) and the fourth (ethnic composition of the locality) effect. Yet in ethnically diverse localities, one’s neighbors are more likely to have a different ethnic background. Because to date studies have ignored the ethnicity of neighbors and the degree of overlap in neighbors’ ethnic back- grounds, and thereby the homophily principle, the studies we considered were biased toward finding negative effects of ethnic diversity on intraneighborhood social cohesion. An important venue for future research is the inclusion of these two alternative mechanisms in our analyses (such network approaches have been previously applied in school contexts; cf. Quillian & Campbell 2003, Tolsma et al. 2013). More fundamentally, the literature calls for a focus on mechanisms in our empirical models.

Changing Geographical Areas: White Flight and Nonlinear Effects

Advocates of the constrict claim argue that most research is biased against finding corroborative evidence for the threat mechanism owing to white flight (Putnam 2007): Residents who are dissatisfied with the ethnic composition of their local environment are hypothesized to move to more homogeneous settings. Consequently,

472 Van der Meer · Tolsma

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

even if ethnic heterogeneity were to have a harmful effect on interethnic social cohesion, this relationship may not be observed because people have left the environments in which they had lost connections to ethnic others. White flight is at best an incomplete counterargu- ment. Although it assumes that increasing het- erogeneity in one’s neighborhood erodes social cohesion and prompts those with the most ad- verse attitudes to move, decreasing heterogene- ity (either because of a changing composition of one’s neighborhood or a change of neighbor- hood) would presumably not stimulate social cohesion.

Several authors have tried to assess the va- lidity of the selective residential mobility argu- ment indirectly, by controlling for residential mobility at the geographical locality under in- vestigation and by assuming that selective resi- dential mobility is more likely among its afflu- ent residents (and hence that negative effects of heterogeneity are more likely among the poor) (e.g., Tolsma et al. 2009). Results have not been very convincing, however. Ultimately, we need panel data on residents who change neighbor- hoods and on residents within changing neigh- borhoods to assess how much positive and neg- ative changes in heterogeneity affect levels of cohesion.

In essence, the constrict claim is dynamic by nature: Increasing heterogeneity would lead to a deterioration of cohesion. Yet few authors deal with this dynamic nature by investigating the effects of dynamic measures of heterogene- ity (migration rates) or of longitudinal changes in heterogeneity (for exceptions see, among others, Gesthuizen et al. 2009, Hooghe et al. 2009, Kesler & Bloemraad 2010, Dinesen & Sønderskov 2012). Threat and anomie are likely to be triggered by recent substantial increases rather than by stable levels of hetero- geneity. The impact of stagnant heterogeneity is likely to flatten out owing to familiarization. The effect of (changing) ethnic heterogeneity is unlikely to be linear anyway, and thus threshold effects are likely (cf. Wagner et al. 2006; some authors have investigated nonlinear effects by including a quadratic term, e.g., Dincer 2011,

Reeskens & Hooghe 2009). A small increase in the size of the out-group would not necessarily constitute a threat or lead to feelings of anomie. The resulting hypothesized s-shaped heterogeneity effect has implications for both the operationalization of heterogeneity and for the sampling of the higher-level units. Repre- sentative samples would be unlikely to detect heterogeneity effects in most countries, as many localities will be rather homogeneous, domi- nated by natives or the largest ethnic group.

Changing Ethnic Heterogeneity: Ethnic Distinctions and Cross-Cutting Cleavages

Not only do social environments change over time, but even the nature of ethnic heterogene- ity itself changes. In contemporary studies on the United States, it hardly makes sense to dis- tinguish between, among others, Italian Amer- icans, Irish Americans, and Polish Americans. Rather, US-based studies tend to emphasize the racial boundaries (increasingly including the Hispanic category) that form a relevant societal cleavage nowadays. Similarly, studies in the United Kingdom generally distinguish Pak- istani and Indians as ethnic groups but not the Welsh or the Scots. Canadian studies often fo- cus on visible minorities. Dutch statistics offers an ethnic classification that separates first- and second-generation immigrants from Morocco, Turkey, Suriname, and the Antilles, but not the equally large groups from Western countries (Belgium, Germany, the United Kingdom, the United States), let alone native minority groups such as the Frysians. More generally, most studies tend to rely on classifications of ethnic groups along criteria that are socially relevant, i.e., from which the strongest, neg- ative heterogeneity effects are expected. That makes the lack of consistent support for the constrict claim even more staggering. Recent studies picked up this problem (e.g., Finseraas & Jakobsson 2012, Morales & Echazarra 2013). The cross-national study of Finseraas & Jakobsson (2012) is especially intriguing. They find that ethnic heterogeneity is harmful

www.annualreviews.org • Ethnic Diversity 473

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

only to generalized trust if ethno-linguistic and religious distinctions overlap, that is, if there are cross-cutting cleavages.

American Exceptionalism? Multiculturalism and Segregation

What can account for the remarkable dis- tinction between the United States on the one hand and other (Western) countries— including neighboring Canada—on the other? In the United States, the public reacts most negatively to diversity. Although the United States is a traditional immigration country, no similar consistent link is observed in other traditional countries of immigration such as Australia, Canada, and New Zealand (cf. Hagendoorn 2009). Thus, we have little reason to expect that recent migration countries, such as those in Europe, will necessarily experience a similar setback in social cohesion at some point in the future.

American exceptionalism may be due to unique historical forces. The racial divide and racial mixing have a different connotation in the United States than ethnicity and ethnic diver- sity have in many other countries. The salience and nature of race in the United States are contingent on its legacy of slavery until the late nineteenth century, and the subsequent Jim Crow laws that effectively barred some racial groups from exercising their social and politi- cal rights until the 1960s. Such tentative histor- ical explanations should, however, be framed in a wider theoretical framework to be tested in other environments.

Various scholars have explored whether cross-national differences in heterogeneity effects may be linked to multicultural policies (e.g., Lupo 2010, Reeskens 2010, Zimdars & Tampubolon 2012). Kesler & Bloemraad (2010, p. 336) show that the constrict effect is moderated by “institutional structures and state policies”: Heterogeneity is more likely to have a negative effect on associational membership in countries without multicultural policies. These moderating effects of multicultural policies may be due to the interethnic contact opportunities

they induce, which stimulate broader intereth- nic trust under such a favorable canopy (Allport 1954 [1979]). This brings us back to the moder- ating effect of segregation and ethnic inequality. Given relative group sizes, segregation and ethnic inequality undermine favorable contact opportunities. American exceptionalism may be linked to relatively high levels of heterogeneity combined with the pronounced segregation of cities in the United States in comparison with other Western countries ( Johnston et al. 2007) and the persistence of ethnic inequalities. Heterogeneity alone is unlikely to erode (all elements of ) cohesion, either in the short or the long run. Rather, policy makers should aim to prevent the potentially explosive mix of heterogeneity, segregation, and inequality.

Implications

This review has implications beyond the con- strict literature. Conceptually, it is striking that scholars have been unaware that there is far less disagreement than is apparent at first glance. The major obstacle has been the use of broad concepts, such as social capital, social cohesion, and social trust, to denote widely different em- pirical phenomena. Because the research object was presented as a monolithic block, differences between indicators were obscured. Evidently, this problem was exacerbated by the lack of the- oretical refinement: All aspects of social cohe- sion were supposed to be negatively affected by ethnic diversity.

Especially after the publication of Robert Putnam’s “E Pluribus Unum,” there was widespread attention in national media for the supposed negative social consequences of eth- nic heterogeneity and immigration. We found evidence to support this claim, but only to a very limited extent. Heterogeneity merely under- mines intraneighborhood social cohesion: Peo- ple in ethnically heterogeneous environments are less likely to trust their neighbors or to have contact with them. However, this does not spill over to generalized trust, to informal help and voluntary work, or to other forms of prosocial behavior and attitudes, at least not in Europe.

474 Van der Meer · Tolsma

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Rather, heterogeneity is positively related to interethnic contact and (consequently) to in- terethnic trust. Given the speed with which

apocalyptic claims were embraced by journal- ists and policy makers, these conclusions are by themselves rather sobering.

DISCLOSURE STATEMENT

The authors are not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS

This work was supported by the Netherlands Organisation for Scientific Research (NWO; grant numbers 016.115.264 to T. van der Meer and 016.135.161 to J. Tolsma). The authors thank Violet Benneker, Mirjam Fischer, and Maria Kranendonk for providing checks on the coding. The authors thank the participants of the conference “The Political Incorporation of Immi- grants: Progress, Prospects, and Pitfalls in Europe and North America” (University of California, Berkeley, March 4–5, 2011) and of the session “Immigration-Related Diversity, Social Cohesion, and the Civic Culture” at the 2011 European Consortium for Political Research (ECPR) Gen- eral Conference (Reykjavik, Iceland, August 24–27, 2011) for their valuable comments on earlier drafts of this article. The authors are also grateful to the reviewers and editors for their valuable comments.

LITERATURE CITED

Aizlewood A, Pendakur R. 2005. Ethnicity and social capital in Canada. Can. Ethn. Stud. 37:77–102 Alesina A, La Ferrara E. 2000. Participation in heterogeneous communities. Q. J. Econ. 115:847–904 Alesina A, La Ferrara E. 2002. Who trusts others? J. Public Econ. 85:207–34 Allport GW. 1954 (1979). The Nature of Prejudice. New York: Basic Books Andrews R. 2009. Civic engagement, ethnic heterogeneity, and social capital in urban areas: evidence from

England. Urban Aff. Rev. 44:428–40 Bailey N, Kearns A, Livingston M. 2012. Place attachment in deprived neighbourhoods: the impacts of

population turnover and social mix. Hous. Stud. 27:208–31 Bécares L, Stafford M, Laurence J, Nazroo J. 2011. Composition, concentration and deprivation: exploring

their association with social cohesion among different ethnic groups in the UK. Urban Stud. 48:2771–87 Blalock HM. 1967. Toward a Theory of Minority-Group Relations. New York: Wiley Blumer H. 1958. Race prejudice as a sense of group position. Pac. Sociol. Rev. 1:3–7 Bobo L, Hutchings VL. 1996. Perceptions of racial group competition: extending Blumer’s theory of group

position to a multiracial social context. Am. Sociol. Rev. 61:951–72 Bourdieu P. 1987. Distinction. A Social Critique of the Judgement of Taste. London: Routledge Chan J, To H-P, Chan E. 2006. Reconsidering social cohesion: developing a definition and analytical frame-

work for empirical research. Soc. Indic. Res. 75:273–302 Cheong PH, Edwards R, Goulbourne H, Solomos J. 2007. Immigration, social cohesion and social capital: a

critical review. Crit. Soc. Policy 27:24–49 Coleman J. 1990. Foundations of Social Theory. Cambridge, MA: Belknap Costa DL, Kahn ME. 2003. Civic engagement and community heterogeneity: an economist’s perspective.

Perspect. Polit. 1:103–11 Dincer OC. 2011. Ethnic diversity and trust. Contemp. Econ. Policy 29:284–93 Dinesen PT, Sønderskov KM. 2012. Trust in a time of increasing diversity: on the relationship between ethnic

heterogeneity and social trust in Denmark from 1979 until today. Scand. Polit. Stud. 35:273–94 Dinesen PT, Sønderskov KM. 2013. Ethnic diversity of the micro-context and generalized trust: evidence from

Denmark. Work. Pap., first presented at 6th Eur. Consort. Polit. Res. (ECPR) Gen. Conf., Reykjavik, Iceland, Aug. 25–27, 2011

www.annualreviews.org • Ethnic Diversity 475

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Fieldhouse E, Cutts D. 2010. Does diversity damage social capital? A comparative study of neighbourhood diversity and social capital in the US and Britain. Can. J. Polit. Sci. 43:289–318

Finney N, Jivraj S. 2013. Ethnic group population change and neighbourhood belonging. Urban Stud. 50:3323–41

Finseraas H, Jakobsson N. 2012. Trust and ethnic fractionalization: the importance of religion as a cross- cutting dimension. Kyklos 65:327–39

Fischer CS. 2005. Bowling alone: What’s the score? Soc. Netw. 27:155–67 Fotheringham AS, Wong DW. 1991. The modifiable areal unit problem in multivariate statistical analysis.

Environ. Plan. A 23:1025–44 Gesthuizen M, Van der Meer T, Scheepers P. 2009. Ethnic diversity and social capital in Europe: tests of

Putnam’s thesis in European countries. Scand. Polit. Stud. 32:121–42 Gijsberts M, Van der Meer T, Dagevos J. 2012. ‘Hunkering down’ in multi-ethnic neighbourhoods? The

effects of ethnic diversity on dimensions of social cohesion. Eur. Sociol. Rev. 28:527–37 Granovetter M. 1983. The strength of weak ties: a network theory revisited. Sociol. Theory 1:201–33 Gustavsson M, Jordahl H. 2008. Inequality and trust in Sweden: some inequalities are more harmful than

others. J. Public Econ. 92:348–65 Hagendoorn L. 2009. Ethnic diversity and the erosion of social capital? APSA Newsl. 20:12–14 Hallberg P, Lund J. 2005. The business of apocalypse: Robert Putnam and diversity. Race Class 46:53–67 Healy E. 2007. Ethnic diversity and social cohesion in Melbourne. People Place 15:49–64 Hero RE. 2003. Social capital and racial inequality in America. Perspect. Polit. 1:113–22 Hipp JR, Boessen A. 2013. Egohoods as waves washing across the city: a new measure of “neighborhoods.”

Criminology 51:287–327 Hooghe M. 2007. Social capital and diversity: generalized trust, social cohesion and regimes of diversity. Can.

J. Polit. Sci. 40:709–32 Hooghe M, Reeskens T, Stolle D, Trappers A. 2009. Ethnic diversity and generalized trust in Europe: a

cross-national multilevel study. Comp. Polit. Stud. 42:198–223 Hou F, Wu Z. 2009. Racial diversity, minority concentration, and trust in Canadian urban neighborhoods.

Soc. Sci. Res. 38:693–716 Huijts T, Kraaykamp G, Scheepers P. 2014a. Ethnic diversity and informal intra- and inter-ethnic contacts

with neighbors in the Netherlands: a comparison of natives and ethnic minorities. Acta Sociol. 57:41–57 Huijts T, Sluiter R, Scheepers P, Kraaykamp G. 2014b. Ethnic diversity and personal contacts at work and at

school in the Netherlands: a comparison of natives and ethnic minorities. J. Int. Migr. Integr. 15:277–98 Johnston R, Poulsen M, Forrest J. 2007. The geography of ethnic residential segregation: a comparative study

of five countries. Ann. Assoc. Am. Geogr. 97:713–38 A dynamic study showing that cross-national changes in diversity are not (everywhere) accompanied by reductions in collective-mindedness.

Kesler C, Bloemraad I. 2010. Does immigration erode social capital? The conditional effects of immigration-generated diversity on trust, membership, and participation across 19 countries, 1981–2000. Can. J. Polit. Sci. 43:319–47

Knack S, Keefer P. 1997. Does social capital have an economic payoff ? A cross-country investigation. Q. J. Econ. 112:1251–88

Koopmans R, Veit S. 2013. Cooperation in ethnically diverse neighborhoods: a lost-letter experiment. Polit. Psychol. In press

Lancee B, Dronkers J. 2011. Ethnic, religious and economic diversity in Dutch neighbourhoods: Explaining quality of contact with neighbours, trust in the neighbourhood and inter-ethnic trust. J. Ethn. Migr. Stud. 37:597–618

Laurence J. 2011. The effect of ethnic diversity and community disadvantage on social cohesion: a multi-level analysis of social capital and interethnic relations in UK communities. Eur. Sociol. Rev. 27:70–89

Laurence J, Heath A. 2008. Predictors of Community Cohesion: Multi-Level Modelling of the 2005 Citizenship Survey. London: Dep. Commun. Local Gov.

Lazarsfeld PF, Merton RK. 1954. Friendship as a social process: a substantive and methodological analysis. Freedom Control Mod. Soc. 18:18–66

Lee T. 2008. Race, immigration, and the identity-to-politics link. Annu. Rev. Polit. Sci. 11:457–78

Demonstrates via structural-equation models that geographical scope is a distinct cohesion dimension. Letki N. 2008. Does diversity erode social cohesion? Social capital and race in British neighbourhoods.

Polit. Stud. 56:99–126

476 Van der Meer · Tolsma

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Lewicki RJ, Bunker BB. 1996. Developing and maintaining trust in work relationships. In Trust in Organi- zations: Frontiers of Theory and Research, ed. RM Kramer, TR Tyler, pp. 114–39. Thousand Oaks, CA: Sage

Lupo G. 2010. Is immigration detrimental for social trust in the European Union? A three-level model of cultural heterogeneity and citizenship regime as social capital predictors. Int. J. Soc. Inq. 3:67–96

Macedo S. 1999. Diversity and Distrust. Cambridge, MA: Harvard Univ. Press Marschall MJ, Stolle D. 2004. Race and the city: neighborhood context and the development of generalized

trust. Polit. Behav. 26:125–53 Martinović B. 2013. The inter-ethnic contacts of immigrants and natives in the Netherlands: a two-sided

perspective. J. Ethn. Migr. Stud. 39:69–85 Massey DS, Denton NA. 1988. The dimensions of residential segregation. Soc. Forces 67:281–315 Massey DS, Denton NA. 1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge,

MA: Harvard Univ. Press McPherson M, Smith-Lovin L, Cook JM. 2001. Birds of a feather: homophily in social networks. Annu. Rev.

Sociol. 27:415–44 Morales L, Echazarra A. 2013. Will we all hunker down? The impact of immigration and diversity on local

communities in Spain. J. Elect. Public Opin. Parties 23:343–66 Pennant R. 2005. Diversity, trust and community participation in England. Findings 253, Home Off., Res. Dev.

Stat. Dir., London Pettigrew TF, Tropp LR. 2006. A meta-analytic test of intergroup contact theory. J. Personal. Soc. Psychol.

90:751–83 Pettigrew TF, Wagner U, Christ O. 2010. Population ratios and prejudice: modelling both contact and threat

effects. J. Ethn. Migr. Stud. 36:635–50 Phan MB. 2008. We’re all in this together: context, contacts, and social trust in Canada. Anal. Soc. Issues Public

Policy 8:23–51 Pichler F, Wallace C. 2007. Patterns of formal and informal social capital in Europe. Eur. Sociol. Rev. 23:423–35 Portes A, Vickstrom E. 2011. Diversity, social capital, and cohesion. Annu. Rev. Sociol. 37:461–79 Putnam RD. 2001. Bowling Alone: The Collapse and Revival of American Community. New York: Simon &

Schuster Putnam’s constrict claim—that heterogeneity erodes in-group and out-group solidarity— reinvigorated the debate.

Putnam RD. 2007. E pluribus unum: diversity and community in the twenty-first century. The 2006 Johan Skytte Prize Lecture. Scand. Polit. Stud. 30:137–74

Putnam RD, Leonardi R, Nanetti RY. 1994. Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton Univ. Press

Quillian L. 1995. Prejudice as a response to perceived group threat: population composition and anti- immigrant and racial prejudice in Europe. Am. Sociol. Rev. 60:586–611

Quillian L. 1996. Group threat and regional change in attitudes toward African-Americans. Am. J. Sociol. 102:816–60

Quillian L, Campbell ME. 2003. Beyond black and white: the present and future of multiracial friendship segregation. Am. Sociol. Rev. 68:540–66

Reeskens T. 2010. Ethnic-cultural diversity, migrant integration policies and social cohesion in Europe: investigating the conditional effect of ethnic-cultural diversity on generalized trust. Presented at Migration: A World in Motion, Feb. 18–20, Maastricht, Neth.

Reeskens T, Hooghe M. 2009. Is local ethnic diversity harmful for social capital? A multilevel analysis of social capital in Belgian communities. Presented at Eur. Consort. Polit. Res. (ECPR) edition:5, Sept. 10–14, Potsdam, Ger.

Rothwell JT. 2012. The effects of racial segregation on trust and volunteering in US cities. Urban Stud. 49:2109–36

Sampson RJ, Groves WB. 1989. Community structure and crime: testing social-disorganization theory. Am. J. Sociol. 94:774–802

Savelkoul M, Gesthuizen M, Scheepers P. 2011. Explaining relationships between ethnic diversity and informal social capital across European countries and regions: tests of constrict, conflict and contact theory. Soc. Sci. Res. 40:1091–107

www.annualreviews.org • Ethnic Diversity 477

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40CH21-VanDerMeer ARI 10 July 2014 6:40

Scheepers P, Gijsberts M, Coenders M. 2002. Ethnic exclusionism in European countries: public opposition to civil rights for legal migrants as a response to perceived ethnic threat. Eur. Sociol. Rev. 18:17–34

Schlueter E, Scheepers P. 2010. The relationship between outgroup size and anti-outgroup attitudes: a theo- retical synthesis and empirical test of group threat and intergroup contact theory. Soc. Sci. Res. 39:285–95

Schlueter E, Wagner U. 2008. Regional differences matter examining the dual influence of the regional size of the immigrant population on derogation of immigrants in Europe. Int. J. Comp. Sociol. 49:153–73

Schmid K, Al Ramiah A, Hewstone M. 2013. Diversity and its consequences for outgroup, ingroup, and neighborhood trust: indirect effects via intergroup contact and threat. Presented at Ethnic Diversity and Social Capital: Mechanism, Conditions and Causality, Soc. Sci. Res. Cent., May 24–25, Berlin

Seeman M. 1959. On the meaning of alienation. Am. Sociol. Rev. 24:783–91 Semyonov M, Raijman R, Tov AY, Schmidt P. 2004. Population size, perceived threat, and exclusion: a

multiple-indicators analysis of attitudes toward foreigners in Germany. Soc. Sci. Res. 33:681–701 Shaw CR, McKay HD. 1942. Juvenile Delinquency and Urban Areas. Chicago: Univ. Chicago Press Smith HP, Bohm RM. 2008. Beyond anomie: alienation and crime. Crit. Criminol. 16:1–15 Stephan WG, Stephan CW. 2000. An integrated threat theory of prejudice. In Reducing Prejudice and Discrim-

ination, ed. S Oskamp, pp. 23–45. Mahwah, NJ: Erlbaum Interethnic neighborhood contact cushions negative effects of neighborhood heterogeneity on generalized and out-group trust.

Stolle D, Petermann S, Schmid K, Schönwälder K, Hewstone M, et al. 2013. Immigration-related diversity and trust in German cities: the role of intergroup contact. J. Elect. Public Opin. Parties 23:279–98

Tajfel H, Turner JC. 1979. An integrative theory of intergroup conflict. In The Social Psychology of Intergroup Relations, ed. WC Austin, S Worchel, pp. 33–47. Monterey, CA: Brooks/Cole

Tolsma J, Van der Meer T, Gesthuizen M. 2009. The impact of neighbourhood and municipality character- istics on social cohesion in the Netherlands. Acta Polit. 44:286–313

Tolsma J, van Deurzen I, Stark TH, Veenstra R. 2013. Who is bullying whom in ethnically diverse primary schools? Exploring links between bullying, ethnicity, and ethnic diversity in Dutch primary schools. Soc. Netw. 35:51–61

Twigg L, Taylor J, Mohan J. 2010. Diversity or disadvantage? Putnam, Goodhart, ethnic heterogeneity, and collective efficacy. Environ. Plan. A 42:1421–38

Uslaner EM. 2011a. Contact, diversity, and segregation. SULCIS Work. Pap. 2011:5, Stockholm Univ. Linnaeus Cent. Integr. Stud., Stockholm, Swed.

Uslaner EM. 2011b. Trust, diversity, and segregation in the United States and the United Kingdom. Comp. Sociol. 10:221–47

Uslaner EM. 2012. Segregation and Mistrust: Diversity, Isolation, and Social Cohesion. New York: Cambridge Univ. Press

Vervoort M, Flap H, Dagevos J. 2011. The ethnic composition of the neighbourhood and ethnic minorities’ social contacts: three unresolved issues. Eur. Sociol. Rev. 27:586–605

Wagner U, Christ O, Pettigrew TF, Stellmacher J, Wolf C. 2006. Prejudice and minority proportion: contact instead of threat effects. Soc. Psychol. Q. 69:380–90

Wallman Lundåsen S, Wollebæk D. 2013. Diversity and community trust in Swedish local communities. J. Elect. Public Opin. Parties 23:299–321

Zimdars A, Tampubolon G. 2012. Ethnic diversity and European’s generalised trust: how inclusive immigra- tion policy can aid a positive association. Sociol. Res. Online 17:15

478 Van der Meer · Tolsma

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40-FrontMatter ARI 8 July 2014 6:42

Annual Review of Sociology

Volume 40, 2014Contents

Prefatory Chapter

Making Sense of Culture Orlando Patterson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Theory and Methods

Endogenous Selection Bias: The Problem of Conditioning on a Collider Variable Felix Elwert and Christopher Winship � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �31

Measurement Equivalence in Cross-National Research Eldad Davidov, Bart Meuleman, Jan Cieciuch, Peter Schmidt, and Jaak Billiet � � � � � � � � �55

The Sociology of Empires, Colonies, and Postcolonialism George Steinmetz � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �77

Data Visualization in Sociology Kieran Healy and James Moody � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 105

Digital Footprints: Opportunities and Challenges for Online Social Research Scott A. Golder and Michael W. Macy � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 129

Social Processes

Social Isolation in America Paolo Parigi and Warner Henson II � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 153

War Andreas Wimmer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 173

60 Years After Brown: Trends and Consequences of School Segregation Sean F. Reardon and Ann Owens � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 199

Panethnicity Dina Okamoto and G. Cristina Mora � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 219

Institutions and Culture

A Comparative View of Ethnicity and Political Engagement Riva Kastoryano and Miriam Schader � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 241

v

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40-FrontMatter ARI 8 July 2014 6:42

Formal Organizations

(When) Do Organizations Have Social Capital? Olav Sorenson and Michelle Rogan � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 261

The Political Mobilization of Firms and Industries Edward T. Walker and Christopher M. Rea � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 281

Political and Economic Sociology

Political Parties and the Sociological Imagination: Past, Present, and Future Directions Stephanie L. Mudge and Anthony S. Chen � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 305

Taxes and Fiscal Sociology Isaac William Martin and Monica Prasad � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 331

Differentiation and Stratification

The One Percent Lisa A. Keister � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 347

Immigrants and African Americans Mary C. Waters, Philip Kasinitz, and Asad L. Asad � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 369

Caste in Contemporary India: Flexibility and Persistence Divya Vaid � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 391

Incarceration, Prisoner Reentry, and Communities Jeffrey D. Morenoff and David J. Harding � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 411

Intersectionality and the Sociology of HIV/AIDS: Past, Present, and Future Research Directions Celeste Watkins-Hayes � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 431

Individual and Society

Ethnic Diversity and Its Effects on Social Cohesion Tom van der Meer and Jochem Tolsma � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 459

Demography

Warmth of the Welcome: Attitudes Toward Immigrants and Immigration Policy in the United States Elizabeth Fussell � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 479

Hispanics in Metropolitan America: New Realities and Old Debates Marta Tienda and Norma Fuentes � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 499

Transitions to Adulthood in Developing Countries Fatima Juárez and Cecilia Gayet � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 521

vi Contents

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

SO40-FrontMatter ARI 8 July 2014 6:42

Race, Ethnicity, and the Changing Context of Childbearing in the United States Megan M. Sweeney and R. Kelly Raley � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 539

Urban and Rural Community Sociology

Where, When, Why, and For Whom Do Residential Contexts Matter? Moving Away from the Dichotomous Understanding of Neighborhood Effects Patrick Sharkey and Jacob W. Faber � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 559

Gender and Urban Space Daphne Spain � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 581

Policy

Somebody’s Children or Nobody’s Children? How the Sociological Perspective Could Enliven Research on Foster Care Christopher Wildeman and Jane Waldfogel � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 599

Sociology and World Regions

Intergenerational Mobility and Inequality: The Latin American Case Florencia Torche � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 619

A Critical Overview of Migration and Development: The Latin American Challenge Raúl Delgado-Wise � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 643

Indexes

Cumulative Index of Contributing Authors, Volumes 31–40 � � � � � � � � � � � � � � � � � � � � � � � � � � � 665

Cumulative Index of Article Titles, Volumes 31–40 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 669

Errata

An online log of corrections to Annual Review of Sociology articles may be found at http://www.annualreviews.org/errata/soc

Contents vii

A nn

u. R

ev . S

oc io

l. 2

01 4.

40 :4

59 -4

78 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

A ri

zo na

S ta

te U

ni ve

rs it

y on

0 1/

02 /1

8. F

or p

er so

na l

us e

on ly

.

1. ar:
1. logo: