Question

While sociologists have been concernedwith urban residents and their neighbor- hoods since the birth of the discipline (e.g., DuBois [1899] 1996; Park and Burgess 1925; Shaw 1929), the publication of Wilson’s The Truly Disadvantaged (1987) and Massey and Denton’s American Apartheid (1993) sparked a renewed interest among scholars of urban inequality in the role of neighborhood context in the intergenerational transmission of pover- ty. These works directly linked urban sociolo- gy to stratif ication, mobility, and race. In theoretical terms, neighborhoods became an

important context for the social processes driv- ing stratif ication and racial inequality, and neighborhood context was viewed as a causal force in the lives of youth and adults. These works set off a sustained effort to understand the effects of neighborhood context on individual outcomes, particularly for youth. Although there is mounting evidence that neighborhood effects are real causal effects (see Harding [2003] for a review and evidence on this issue), social sci- entists have only begun to uncover the mecha- nisms by which such effects operate.

Cultural Context, Sexual Behavior, and Romantic Relationships in Disadvantaged Neighborhoods

David J. Harding University of Michigan

When culture is invoked to understand the consequences of growing up in disadvantaged

neighborhoods, the isolation of ghetto residents from mainstream institutions and

mainstream culture is often emphasized. This article attempts to reorient current

theorizing about the cultural context of disadvantaged neighborhoods, particularly when

it comes to adolescent decision making and behavior. I argue that rather than being

characterized by the dominance of “oppositional” or “ghetto-specific” cultures,

disadvantaged neighborhoods are characterized by cultural heterogeneity: a wide array

of competing and conflicting cultural models. I apply this conception to sexual behavior

and romantic relationships among adolescents using survey data from Addhealth.

Analyses show that disadvantaged neighborhoods exhibit greater heterogeneity in

cultural frames and scripts and that, in more heterogeneous neighborhoods, adolescents’

frames and scripts are poorly predictive of their actual behavior.

AMERICAN SOCIOLOGICAL REVIEW, 2007, VOL. 72 (June:341–364)

Direct correspondence to David J. Harding, Population Studies Center, University of Michigan, 426 Thompson St., Ann Arbor, MI 48106-1248 ([email protected]). Funding for this research was provided by the National Science Foundation (SES-0326727), The William T. Grant Foundation, the American Educational Research Association, the MacArthur Foundation Network on Inequality and Economic Perfor mance, and by the Har vard Multidisciplinary Program on Inequality and Social Policy. An NICHD Post-Doctoral Fellowship at the University of Michigan Population Studies Center provided additional support. Katherine Newman,

Christopher Winship, Michèle Lamont, Robert Sampson, Christopher Jencks, Jal Mehta, Arland Thornton, and four ASR reviewers provided helpful comments on previous versions of this article. This research uses data from Add Health, a program proj- ect designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris, and funded by the National Institute of Child Health and Human Development, with cooperative funding from 17 other agencies. Persons interested in obtaining data files from Add Health should contact Add Health, Carolina Population Center, 123 W. Franklin Street, Chapel Hill, NC 27516-2524 ([email protected]).

Culture has been largely ignored in the empir- ical effort to identify and understand the social processes underlying neighborhood effects. Indeed, scholars studying race and poverty have shied away from discussing culture since the early 1970s, after critics of scholars such as Frazier (1966), Moynihan (1965), and Lewis (1969) argued that cultural explanations of poverty ignore structural barriers and blame the victim (e.g., Valentine 1968; see Wilson 1987 for a review). In the interim, however, cultural sociology has moved away from con- ceptions of culture as an internally coherent set of values and toward a view of culture as frag- mented and composed of “disparate bits of information and .|.|. schematic structures that organize that infor mation” (DiMaggio 1997:263). These ideas, however, have only slowly found their way into the sociology of dis- advantaged neighborhoods. As I discuss below, new concepts in cultural sociology such as frames, scripts, and repertoires have the poten- tial to illuminate the social processes at work in neighborhood effects.

When culture is invoked to help us to under- stand the consequences of growing up in dis- advantaged communities, emphasis is often placed on the isolation of ghetto residents from mainstream social networks and mainstream culture. Wilson (1987, 1996) argues that the out-migration of the black middle class and the decline of manufacturing lead to neighborhoods in which life is no longer organized around work. Social interaction in isolated neighbor- hoods leads to the development of cultural reper- toires that are “oppositional” or “ghetto specific,” adaptations to blocked opportunities in the labor market and society generally. Anderson (1999) invokes an alternative status system among adolescents from underclass neighborhoods to understand high rates of teenage pregnancy and single parenthood. Massey and Denton (1993) argue that racial segregation and the concentration of poverty lead to an “oppositional culture” in inner cities—an oppositional culture that upends con- ventional norms and values in response to blocked opportunities. “As intense racial isola- tion and acutely concentrated poverty have con- tinued, ghetto values, attitudes, and ideals have become progressively less connected to those prevailing elsewhere in the United States. More

and more, the culture of the ghetto has become an entity unto itself, remote from the rest of American society and its institutions, and drift- ing ever further afield” (Massey and Denton 1993:172).

This article aims to reorient current thinking about culture in disadvantaged neighborhoods and how it relates to adolescent decision mak- ing and behavior. I reintroduce an idea that was once a staple of theorizing about urban neigh- borhoods—disadvantaged neighborhoods are characterized by cultural heterogeneity (Shaw and McKay 1969).1 I apply these ideas, coupled with recent theoretical advances in cultural soci- ology, to the analysis of adolescent sexual behavior and romantic relationships. Using sur- vey data from the National Longitudinal Study of Adolescent Health (Addhealth) on frames regarding teenage pregnancy and scripts for romantic relationships, I show that disadvan- taged neighborhoods exhibit greater hetero- geneity in scripts and frames. I also investigate the consequences of cultural heterogeneity for two adolescent behaviors, premarital sexual activity and the sequencing of events in roman- tic relationships.

CULTURE, INEQUALITY, AND NEIGHBORHOOD EFFECTS

Most recent work at the nexus of inequality and cultural sociology has focused on the analysis of class, broadly speaking (e.g., Bourdieu and Passeron 1977; Bryson 1996; DiMaggio 1982; Erickson 1996; Lamont 1992, 2000). Increasingly, however, poverty scholars have also employed cultural analysis, particularly in documenting the cultural world of the urban poor, describing how they understand their options and make decisions with regard to work, welfare, schooling, parenthood, and marriage (Anderson 1999; Carter 2005; Edin and Kefalas 2005; Newman 1999; Young 2004). Recent lit- erature on the role of neighborhood context in the intergenerational transmission of poverty, however, has less explicitly incorporated culture.

342—–AMERICAN SOCIOLOGICAL REVIEW

1 Shaw and McKay did not use the term cultural heterogeneity but instead referred to “different sys- tems of values” and “different forms of organization.” As the term “different systems of values” suggests, they also employed a values conception of culture.

Two mutually compatible theories provide explanations of neighborhood effects on indi- viduals, and cultural concepts have been incor- porated into them to different degrees. Social organization theory argues that neighborhood disadvantage leads to difficulties establishing and maintaining order. Lack of resources, racial and ethnic heterogeneity, and population turnover result in fewer social ties and therefore diminished social control—communities lose the ability to regulate the behavior of their mem- bers (Park and Burgess 1925; Shaw 1929). Communities with denser social networks are better able to articulate and enforce common norms and values. In addition, local formal and informal institutions affect the ability of neigh- bors to maintain social control by influencing norms and expectations and by providing con- texts within which social ties are created and strengthened. External institutions, such as police, city government, and markets affect the resources that are available for social control (Bursik and Grasmick 1993). Collective effi- cacy, defined as the “social cohesion among neighbors combined with their willingness to intervene on behalf of the common good,” (p. 918) mediates the relationship between con- centrated structural disadvantages (residential instability, ethnic or racial heterogeneity, and pover ty) and crime rates (Sampson, Raudenbush, and Earls 1997). Collective effi- cacy has been used almost exclusively to explain neighborhood differences in crime, violence, and delinquency outcomes (exceptions include Browning, Leventhal, and Brooks-Gunn [2005] on neighborhood collective efficacy and sexu- al initiation).

Though social organization models are not usually thought of as cultural models, they do incorporate cultural elements. In the classical formulation, social organization matters because socially organized neighborhoods are better able to enforce common values. In addition, collective efficacy can be thought of as a cul- tural concept, insofar as it is measured as resi- dents’ commonly held expectations or beliefs about how others around them will behave when faced with nonnormative behavior. Finally, Small (2004) shows how cultural frames regard- ing a neighborhood and its origins can impact residents’ willingness to engage in the collec-

tive activity and institution building that lead to collective efficacy and social organization.

While social organization theory focuses pri- marily on processes internal to the neighbor- hood, social isolation theory emphasizes social and cultural disconnections between neighbor- hood residents and the outside world. Social isolation theory suggests that residents of con- centrated poverty neighborhoods are more like- ly to be isolated from middle-class or mainstream social groups, organizations, and institutions (Wilson 1987). The joblessness endemic to high poverty areas means many res- idents are not connected to the mainstream labor market, an important tie to the culture of mid- dle-class life (Wilson 1996). The lack of resources in high poverty neighborhoods makes sustaining neighborhood institutions more dif- ficult, further isolating neighborhood residents from associated mainstream institutions. The result is that social interaction in isolated neigh- borhoods leads to the development of cultural repertoires that differ from the mainstream. Youth are socialized into a cultural environ- ment that promotes behaviors, such as early sexual behavior and dropping out of high school, that are viewed as detrimental in the outside world.

Most research grounded in social isolation theory investigates the social connections of neighborhood residents, finding that neighbor- hood poverty predicts organizational participa- tion and network ties to employed or college-educated individuals, net of individual characteristics (Fernandez and Harris 1992; Rankin and Quane 2000; Tigges, Browne, and Green 1998). Meanwhile, the cultural predic- tions of social isolation theory have been left largely uninvestigated. Social isolation theo- rists have relied heavily on the notion of “oppo- sitional culture” from ethnographic research on racial differences in educational performance (Fordham and Ogbu 1986), extending the con- cept to domains other than education (e.g., Massey and Denton 1993). Fordham and Ogbu (see also Ogbu 2004), for instance, argue that poor black students develop an oppositional culture in which behaviors that promote aca- demic achievement, such as speaking standard English, doing homework, and engaging in class discussion, become defined as “acting white,” as a response to inferior schools, discrimination,

NEIGHBORHOOD CULTURAL CONTEXT—–343

and blocked opportunities.2 However, survey research has rejected the claim that black stu- dents are disproportionately sanctioned by their peers for academic effort (Ainsworth-Darnell and Downey 1998; Cook and Ludwig 1998). Carter (2005) shows that behaviors unconnect- ed to school achievement are at the heart of notions of “acting white” among poor black and Latino youth. Moreover, cultural isolation and the development of a “ghetto-specific” or “oppositional culture” in poor neighborhoods is further challenged by both survey-based and ethnographic research on attitudes among the poor that finds very strong support for conven- tional or traditional views about education, work, welfare, and marriage (Carter 2005; Dohan 2003; Duneier 1992; Edin and Kefalas 2005; Goldenberg et al. 2001; Newman 1999; Solorzano 1992; Young 2004).

In short, there is little evidence that cultural isolation is an accurate description of the cul- tural context of poor neighborhoods. This arti- cle proposes a new conception of the cultural context of disadvantaged neighborhoods—a conception that emphasizes the cultural het- erogeneity of such neighborhoods and the con- sequences of that heterogeneity for adolescent decision making and outcomes.3

CULTURAL HOMOGENEITY/ HETEROGENEITY IN URBAN NEIGHBORHOODS

The concept of neighborhood cultural hetero- geneity has roots in previous work in urban sociology and is consistent with much ethno- graphic research on urban poverty. However,

cultural heterogeneity has been the subject of only limited explicit theorizing, perhaps because much foundational work on urban neighbor- hoods rests largely on a view of neighborhoods as culturally homogeneous. For example, Park and Burgess (1925) and Gans (1962) viewed urban neighborhoods as immigrant receiving areas. Differences between neighborhoods were viewed largely as the consequence of cultural differences between immigrants’ home coun- tries. The culture that immigrant groups brought with them was the basis of local neighborhood cultures. These analyses, like those of many of their contemporaries, viewed culture as rela- tively homogenous within local contexts.

Not all urban sociologists, however, viewed neighborhoods as composed of homogenous subcultures. Shaw and McKay (1969) argued that socially disorganized slum neighborhoods present youth with a wide array of “competing and conflicting moral values,” both conven- tional and unconventional, creating a break- down of social control that leads to higher rates of delinquency in such neighborhoods. Urban ethnographers have also complicated the stark divisions of “ghetto culture” and “mainstream culture,” tending to see culture in disadvan- taged neighborhoods as derived from main- stream culture but modified or reinterpreted to serve local needs and in response to blocked opportunities (e.g., Anderson 1978; Bourgois 1995; Duneier 1992; Liebow 1967). Nevertheless, because these works focused on particular groups within urban neighborhoods, there is little emphasis on cultural heterogene- ity beyond subjects’ attempts to distinguish themselves from culturally defined others lower in local status hierarchies. Suttles (1968), for instance, highlighted cultural differences with- in disadvantaged neighborhoods, but he focused on those between ethnic groups that use differ- ent communication devices and have different cultural practices.

Hannerz (1969) was one of the first to rec- ognize cultural heterogeneity not just across groups residing together in a single neighbor- hood but also in the actual use of culture by indi- viduals. Though he draws on classic cultural concepts such as norms and symbolic meanings, he also introduces the concept of “cultural reper- toire.” For Hannerz, there are multiple forms of culture: norms and values, meanings, and modes of action, and each individual has a repertoire

344—–AMERICAN SOCIOLOGICAL REVIEW

2 Fordham and Ogbu (1986) are often interpreted as arguing that academic success itself comes to be defined as acting white within an oppositional cul- ture, but in a recent article, Ogbu (2004) clarifies that their argument was that behaviors that lead to aca- demic achievement are what are defined as acting white by poor black adolescents.

3 There are undoubtedly cultural dimensions on which middle-class neighborhoods have comparable or greater cultural heterogeneity than do poor neigh- borhoods, such as political views, fashion prefer- ences, or religious beliefs. The focus of this article is on cultural dimensions related to individual out- comes typically studied by poverty and inequality researchers, such as sexual behavior.

of these. Local cultures can add to or substitute for items in the mainstream cultural repertoire and, thus, provide adaptations and reactions to a given structural situation. Like Liebow (1967) and Anderson (1978), Hannerz sees local cul- ture as helping individuals come to terms with contradictions between the wider society’s cul- ture and the individual’s position in the social structure. Yet, Hannerz makes it clear that “ghet- to culture” is not a monolithic entity but rather a heterogeneous mix of fluid ideal-type lifestyle groups (“mainstreamers,” “swingers,” “street families,” and “street corner men”). In ghetto neighborhoods, members of these groups live in close physical proximity, which often leads them to construct exaggerated social hierar- chies and distinctions (see also Newman 1992). Countering the divisive moral judgments between lifestyle groups, however, are the fam- ily ties and spatial proximity that pull individ- uals with divergent lifestyles into regular contact and confrontation.

Consistent with the points above, some recent research has demonstrated that within disad- vantaged neighborhoods, there are multiple cul- tural models available.4 For example, Newman (1999) shows that even in neighborhoods with high levels of joblessness, the majority of peo- ple pursue activities consistent with mainstream ideologies, such as working or going to school. Anderson (1999) documents the presence of both “street” and “decent” orientations among those living in disadvantaged neighborhoods. Though they are in the numerical minority, those with a “street” orientation dominate pub- lic space and public life in inner-city neighbor- hoods.

The concept of cultural heterogeneity is also broadly consistent with classic research tradi- tions within the study of social stratification focusing on class inequality. Parkin (1971), in his analysis of the normative order of the work-

ing class, describes this order as composed of a number of competing meaning systems—the dominant, subordinate, and radical—each pro- viding a different model of class stratification. The result is a state of “normative ambivalence,” in which the working class draws from a “reser- voir of meaning” that is fed by these three dif- ferent “streams.” Rodman (1963) develops the concept of the “lower-class value stretch” to explain how the lower classes can have a some- what distinctive value system that is still derived from the dominant value system.

In sum, prior research suggests that the cul- tural context of disadvantaged communities can be thought of as derived from mainstream cul- ture but modified or reinterpreted to serve local needs and in response to blocked opportuni- ties. Most analysts recognize the presence of multiple competing lifestyle groups (to use Hannerz’s terminology) or orientations (to use Anderson’s) within urban neighborhoods. Culture within poor neighborhoods is not a sin- gle entity but rather a heterogeneous mix of lifestyles or orientations that individuals move between or draw upon as necessary. Such a con- ception is far from consistent with the view of poor inner-city neighborhoods as places of cul- tural isolation from middle-class culture and institutions.

However, analyses of the role of culture in explaining behavior in poor neighborhoods still largely rest on social isolation models. These analyses tend to identify subcultures that pro- mote or justify particular behaviors as expla- nations for those behaviors. For example, Anderson (1999) explains adolescent sexuality, gender relations, and teenage pregnancy in dis- advantaged neighborhoods with the develop- ment of a subculture in which early sexual activity and early parenthood are normative. The coexistence of these two incompatible mod- els (one that describes poor neighborhoods as containing a mix of nondiscrete cultural groups and another that relies on cultural subgroup explanations of behavior) creates an analytical conundrum: If individuals draw from multiple cultural lifestyle models, how can subcultures hold such sway over behavior, action, or deci- sion making? I propose that recent efforts to incorporate new cultural concepts such as frames, scripts, or repertoires into theorizing about culture in urban neighborhoods provide an important theoretical bridge toward better

NEIGHBORHOOD CULTURAL CONTEXT—–345

4 I follow Quinn and Holland (1987) in my use of the term “cultural models,” which they define as, “presupposed, taken-for-granted models of the world that are widely shared (although not necessarily to the exclusion of other, alternative models) by the mem- bers of a society and that play an enormous role in their understanding of that world and their behavior in it” (p. 4). I consider frames and scripts to be two types of cultural models.

understanding the relationship between culture and behavior among adolescents in disadvan- taged neighborhoods.

CONCEPTUALIZING CULTURAL HETEROGENEITY

In any social context, from the perspective of any individual, there are multiple cultural mod- els available from which to choose (Fuller et al. 1996; Holloway et al. 1997; Quinn and Holland 1987; Swidler 1986). These models may be overlapping or contradictory, and they reflect ideas about how the world works, what appro- priate goals are, and how to go about accom- plishing things.5 To further unpack these issues, I rely on three concepts: culture as repertoire or tool kit (Hannerz 1969; Swidler 1986, 2001), culture as frame (Benford and Snow 2000; Goffman 1974; Small 2004), and culture as script.

Swidler (1986, 2001) draws upon the concept of cultural repertoire to develop a general con- ception of culture that allows it to play a causal role in influencing action.6 She sees culture as a “tool kit” of symbols, stories, and worldviews that people use to solve different problems. Under this model, culture is not a unified sys- tem but a repertoire from which to draw. Culture provides the components used to construct “strategies of action” or “persistent ways of ordering action through time” and can thereby have a causal role. The elements that make up one’s tool kit come not just from direct experi- ence or social interaction but also from the wider culture through institutions such as the media, schooling, and religion. The ability of culture to predict behavior in the tool kit model comes from variation in repertoires across cul- tural groups or across individuals.7

In the analysis below, I measure two types of cultural objects that may be present in an indi- vidual’s or group’s cultural repertoire: frames and scripts. Frames are ways of understanding “how the world works” (Young 2004). They encode expectations about consequences of behavior and how various parts of the social world relate or do not relate to one another. A frame is a lens through which one interprets events, and it therefore impacts how one reacts. Benford and Snow (2000) emphasize that frames are collectively constructed, often unin- tentionally but sometimes intentionally. Frames identify problems and assign blame, provide solutions or strategies, and provide a rationale for engaging in action. Like repertoires, frames allow for cultural heterogeneity. Individuals can have multiple contradictory or competing frames that they deploy in different situations, and frames may have various levels of speci- ficity. Small (2004) shows that individuals in the same neighborhood can and often do employ distinct frames. He shows that frames are not fixed, as young people’s neighborhood frames change through interaction with neighborhood activists of an earlier generation.

Scripts provide cultural templates for the sequencing of behaviors or actions over time. They are akin to Swidler’s “strategies of action” in that they show how to solve problems or achieve goals. There need not be consistency across various scripts or frames, as individuals are often able to live with many contradictions and inconsistencies. Therefore, one should not think of frames and scripts as necessarily hier- archically nested. Instead, individuals or groups may possess or employ multiple contradictory frames and scripts. People of a common culture do not share a coherent, monolithic culture, but rather a set of available frames and scripts, objective structural conditions, and knowledge of what others do and think. Sewell (1999) char-

346—–AMERICAN SOCIOLOGICAL REVIEW

5 A conception of culture as heterogeneous and contradictory appears elsewhere as well. For exam- ple, Sewell notes that “social actors are capable of applying a wide range of different and even incom- patible schemes” (1992:17). He also describes cul- tures as contradictory, loosely integrated, contested, subject to constant change, and weakly bounded (Sewell 1999).

6 Tilly (1978) is also credited with the development of the concept of repertoire in his work on “repertoires of collective action.”

7 Swidler’s tool kit metaphor has been criticized for

failing to specify how individuals choose which ele- ments of their tool kits to employ in different situa- tions (Lamont 1992; Lamont and Thévenot 2000). Lamont (1992) suggests that both proximate and remote structural conditions influence such selec- tions. For an example of comparison of repertoires across cultural groups, see Lamont and Thévenot (2000).

acterizes this level of coherence as “thin coher- ence.”

Though cultural heterogeneity is viewed here through what one might describe as macro- or mesolevel cultural concepts, it is important to note that my conception is also consistent with micro or interactionist perspectives on culture. Implicit in Goffman’s (1974) analysis of frames is the notion that multiple frames can apply in any interaction, and that participants must work to maintain the dominance of a particular frame and often come into conflict over which frame should govern a particular situation (see also Goffman 1959). His discussion of “out of frame activity” acknowledges that multiple frames can be present simultaneously, even if they are not mutually compatible. The cultural hetero- geneity approach also shares with interaction- ist approaches the criticism of subcultural analyses as assuming homogeneity and stasis and defining culture in terms of values (Fine and Kleinman 1979). Furthermore, interactionist perspectives emphasize the role of local or group cultures in mediating the relationship between the wider social environment and indi- vidual action (e.g., Fine 1979). As will be dis- cussed below, the consequences of neighborhood cultural heterogeneity are theo- rized to occur in part through microlevel inter- actions with neighbors with varying cultural orientations or lifestyles and through local inter- pretations of cultural information diffused through the media.

NEIGHBORHOOD DISADVANTAGE AND CULTURAL HETEROGENEITY

Why are adolescents in disadvantaged neigh- borhoods presented with a more heterogeneous array of cultural models from which to fashion their beliefs and actions than are those in more advantaged neighborhoods? Among disadvan- taged neighborhoods, all but the most extreme- ly poor neighborhoods contain a mix of people—some are working and some are on public assistance or involved in crime; some individuals are high school dropouts and some have attended college; families with middle- class incomes live near families that are strug- gling below the poverty line. Though many middle-class blacks left inner-city neighbor- hoods in the 1970s, many also remained there (Patillo-McCoy 1999). This mix means that

adolescents will come into contact with people with a wide array of lifestyles.8

In addition, cultural models do not just come from immediate ecological factors or interper- sonal interactions. Institutions that are decou- pled from everyday interactions such as the media, religion, and politics also contribute to cultural repertoires and are another mechanism by which youth in disadvantaged neighbor- hoods are exposed to mainstream or middle- class culture. For example, youth draw role models from television and radio (Carter 2005, chap. 5), the American dream provides a cultural template for many of the poor (Hochschild 1995), feminism shapes how young mothers think about economic independence and mar- riage (Edin and Kefalas 2005), and religion provides a repertoire for constructing strategies of action in black communities (Patillo-McCoy 1998).

Residents of disadvantaged neighborhoods are also often less able to control public behav- ior or regulate those who deviate from main- stream lifestyle choices. Consensus is weaker on which behaviors are appropriate and inappro- priate. A lack of strong social ties in such neigh- borhoods means that social control is diminished—local institutions are weak and collective behavior is more difficult (Shaw and McKay 1969). Residents of poor neighborhoods also typically lack confidence that their neigh- bors will intervene to stop public disorder and so hesitate to do so themselves (Sampson et al. 1997). Such breakdowns of order can be self- perpetuating, as residents lose touch with neigh- bors as they retreat from the public spaces, leading to further weakening of neighborhood social ties (Venkatesh 2000). Similarly, Wilson (1996) argues that the lack of social organiza- tion in these neighborhoods makes “ghetto-

NEIGHBORHOOD CULTURAL CONTEXT—–347

8 There are of course many factors that determine the capacity of culture to influence behavior. Schudson (1989) outlines five “dimensions of cultural power”: retrievability, rhetorical force, resonance, institutional retention, and resolution. I emphasize retrievability, the availability or accessibility of a cultural model or cultural object. Schudson suggests that retrievability is highest when a cultural model or object is physically present, institutionalized in com- mon practice or public memory, or more salient because it is more recent or more dramatic.

related behaviors” more acceptable, while the lack of opportunities makes them necessary for survival. As the number of individuals pursuing “ghetto-related” behaviors such as early par- enthood, reliance on public assistance, or street hustling increases, the stigma attached to these behaviors declines.

Of course, the mere presence of a mix of mainstream and alternative lifestyles in disad- vantaged communities does not necessarily mean that adolescents will interact with those who engage in these various lifestyles. In dis- advantaged neighborhoods, lifestyles outside the mainstream may have more salience than sheer numbers would lead us to expect. In con- trast to those with mainstream lifestyles who leave the neighborhood for work or school, peo- ple with “street orientations” have a daily pres- ence in the neighborhood. In this vein, Anderson (1999) documents the power of drug dealers and others who use violence to control public space and flaunt material goods such as cars, jewelry, and clothing. While ethnographers such as Suttles (1968) and Hannerz (1969) have described the strong divisions along ethnic and lifestyle lines that occur when groups in phys- ical proximity define themselves as fundamen- tally different or better than one another, today’s young people may be more forgiving of racial and ethnic differences than those of the 1960s (see, for example, Carter 2005). In addition, Hannerz’s neighborhood hierarchies are creat- ed and maintained largely by adults, not ado- lescents. Indeed, he describes how “mainstream” parents go to great pains to keep their children away from “street families.”

These factors related to neighborhood dis- advantage suggest my first hypothesis:

Hypothesis 1: Adolescents in more disadvan- taged neighborhoods will exhibit greater heterogeneity of cultural frames and scripts.

THE CONSEQUENCES OF CULTURALLY HETEROGENEOUS NEIGHBORHOODS

Adolescents may be particularly influenced by cultural heterogeneity because their develop- mental life stage is associated with a strong focus on social identities and peer influences, as well as greater experimentation and risk tak- ing. What are the consequences for adolescents

who face a wider array of cultural models in their social environments? If an individual’s repertoire is constructed from what one observes among the people with whom one interacts and from the broader cultural ideas to which one is exposed through media and institutions, then the average adolescent in a disadvantaged neigh- borhood will have a wider range of scripts and frames in his repertoire. From an interactionist perspective (e.g., Fine 1979), significant cultural fluidity is created when cultural models diffuse across small groups through local (neighbor- hood) and extended interaction networks.

Consider the multiple frames around teenage pregnancy that exist in disadvantaged neigh- borhoods and inform adolescents of the conse- quences of becoming a father or mother at a young age. One frame defines early parenthood as a hindrance to schooling and future success. Another highlights the adult status that comes from parenthood and the attention and admira- tion that a new baby brings (Anderson 1999). A third implies that having a child too late in life will mean one will be too old to have the ener- gy to enjoy the child and to be a good parent. A fourth highlights the power of a child to boost a young man’s masculinity or a young woman’s femininity (Anderson 1999).

Consider also the multiple scripts for roman- tic relationships that are available to an adoles- cent in a disadvantaged neighborhood. One such script may follow a conventional path start- ing with casual dating, the exchange of gifts, exclusive dating, statements of love, sexual activity, marriage, living together, and preg- nancy. Another script may involve little dating and a fast transition to sexual activity and cohab- itation, followed by pregnancy, and far in the future, possibly marriage (Edin and Kefalas 2005). A third may involve early statements of love to multiple partners, nonexclusive dating, and sexual activity with multiple partners and the potential for unplanned—or at least unavoid- ed—pregnancy (Anderson 1999).

When it comes, then, to choosing a course of action from among available relationship scripts or to considering the various frames that define the risks and benefits of a particular behavior such as teenage sexual activity, adolescents in disadvantaged neighborhoods have much more to consider than do those in more affluent areas. Not only do disadvantaged neighborhoods pro- vide a wider array of models, but—unlike more

348—–AMERICAN SOCIOLOGICAL REVIEW

middle-class communities—there is also social support for both mainstream and alternative models. One can observe neighbors engaging in behaviors that conform to various cultural mod- els. In addition, when there are multiple mod- els, the advantages and disadvantages of various options are more poorly defined. The social environment provides a much weaker signal about what option is best because there is social support from others for many different options. We can imagine two consequences of this prob- lem.

First, because there is less consensus or agree- ment where there is greater cultural hetero- geneity, it will be harder for an adolescent to choose between competing options. The result may be weaker commitment to the chosen option and a lower likelihood of follow-through. With a weaker commitment, when a particular option does not seem to be working out, it is eas- ier to shift course because another option is available—with local approval. Obstacles or setbacks may be more likely to push the ado- lescent off track. For example, when a roman- tic partner is not serious about contraception, one can downplay the “pregnancy as roadblock” frame and adopt the “parenthood as status” frame to understand one’s options and behaviors. There are others one can look to who appear to be surviving or even thriving while engaging in behavior consistent with various cultural mod- els.

The second consequence arises at the aggre- gate level. In a social environment in which many options are present and visible, there will be fewer people who have enacted a particular option. Because fewer people have followed a particular script to successful completion, there will be fewer examples of how to do so. In other words, the details of the available scripts will be less clearly defined, which may make it more difficult to successfully enact the script. On the one hand, this can be thought of as an infor- mation problem. For example, when fewer neighbors have successful long-term romantic relationships, it will be less clear how to devel- op such a relationship. On the other hand, it can also be thought of as a social support prob- lem. In a culturally heterogeneous environment there will be considerably less than unanimous social and cultural support for particular frames or scripts. For instance, one’s frame about the advantages and disadvantages of early parent-

hood may face frequent challenge from others with different frames.

These arguments suggest that in the context of neighborhood cultural heterogeneity, ado- lescents will be less likely to act in ways con- sistent with the frames and scripts that they articulate.9 This leads to my second hypothesis:

Hypothesis 2: In culturally heterogeneous neigh- borhoods, there will be a weaker relation- ship between the scripts or frames that adolescents adopt and their corresponding future behavior.

METHODS

DATA

I use data from the National Longitudinal Survey of Adolescent Health (Addhealth; Harris et al. 2003). The Addhealth survey initially sam- pled about 150 middle schools, high schools, and junior high schools. The first wave of data was gathered in the 1994–1995 school year, the second wave in 1996, and the third wave in 2001 and 2002. Students were in grades 7 to 12 in Wave 1. The first wave of data includes a short questionnaire completed by school administra- tors about school characteristics and policies, an in-school questionnaire completed by almost every eligible student in the sample schools, and longer in-home student and parent inter- views with a subsample of about 20,000 stu- dents. Wave 2 followed the in-home students and includes another in-home interview with the student (but not the parent) and another school administrator questionnaire. Structural neigh- borhood characteristics from the 1990 census

NEIGHBORHOOD CULTURAL CONTEXT—–349

9 These arguments are also consistent with social psychological literature on the link between attitudes and behavior (though frames and scripts are differ- ent from attitudes—the degree of overall favorabil- ity toward a psychological object [Ajzen 2001]). According to the Fishbein/Ajzen model, discrepan- cies between attitudes and behavior can be produced by moderating third variables, for example, group norms (Fishbein and Ajzen 1975), or in this case, neighborhood context. Attitude-behavior discrepan- cy can also be a product of attitude specificity. For instance, concrete attitudes predict black adoles- cents’ educational behavior, whereas abstract atti- tudes do not (Mickelson 1990).

are available for in-home respondents. The strengths of the Addhealth data for this analy- sis are the nesting of respondents within neigh- borhoods, the availability of neighborhood structural characteristics, and the availability of survey measures of frames and scripts.

MODELS

To examine Hypothesis 1, that disadvantaged neighborhoods exhibit greater cultural hetero- geneity, I construct neighborhood level meas- ures of cultural heterogeneity for two cultural concepts, pregnancy frames and romantic rela- tionship scripts (described below). I then regress each measure of neighborhood cultural hetero- geneity on neighborhood disadvantage, con- trolling for other neighborhood characteristics that might also produce cultural heterogeneity, such as racial and ethnic diversity or residential instability. More disadvantaged neighborhoods are expected to have greater cultural hetero- geneity.

To examine Hypothesis 2, that adolescents in more heterogeneous neighborhoods are less likely to act in accordance with their articulat- ed frames and scripts, I estimate a series of multilevel regression models, one set for preg- nancy frames and one set for relationship scripts. The pregnancy frame model examines the rela- tionship between pregnancy frames (F

~ ) and

engaging in sexual activity (Y) using an inter- action between individual pregnancy frame and neighborhood level heterogeneity in pregnancy frame. If we index individuals with i, neigh- borhoods with j, and schools with k, we can write a three-level model (Raudenbush and Bryk 2001).10 The individual level model is:

Pr (Yijk = 1) log ( 1 – Pr (Yijk = 1)) = �0jk +

�1jkFijk + �2jkXijk + eijk

Because Y is binary, I use a logit link. F is a measure of the individual’s pregnancy frame (described below), and X is a set of control vari- ables measuring individual and family charac-

teristics (and �2 is a vector of coefficients). There are two neighborhood level equations:

�0jk = �00k + �01kF ~

jk + �02kWjk + rjk

�1jk = �10k + �11kF ~

jk + �12kVjk

The first equation models the intercept from the individual level model as a function of neigh- borhood cultural heterogeneity in the pregnan- cy frame (F

~ ) and a set of neighborhood control

variables (W), including the neighborhood mean pregnancy frame. The second equation can be thought of as adding cross-level interaction terms. It models the coefficient capturing the relationship between the individual frame (F) and sexual activity (Y) from the individual level model as a function of neighborhood cultural heterogeneity in the pregnancy frame (F

~ ) and

a set of neighborhood control variables (V), again including the neighborhood mean preg- nancy frame. The coefficient on (F

~ ), �11k, in the

second neighborhood level equation tests Hypothesis 2 in this model. Finally, there is a school-level equation that serves to control for a set of school characteristics, Z:

�00k + �000 + �001Zk + uk

The analysis proceeds by estimating a set of increasingly more complicated models that build up to the model just described. Earlier models set various control variable coefficients to zero (effectively leaving out the control variables).

The romantic relationship scripts model is somewhat simpler. In this model, I measure directly the difference between one’s ideal romantic relationship and one’s subsequent actu- al romantic relationship (described in more detail below), denoted Y. This is the dependent variable. The primary predictor of interest is heterogeneity in ideal romantic relationship scripts (also described in more detail below), denoted R

~ . Again, we can write a three-level

model. The individual-level equation is:

Yijk = �0jk + �1jkXijk + eijk

The difference between one’s ideal and actu- al romantic relationship script is modeled with an intercept and a series of control variables for individual and family characteristics (X). The neighborhood-level equation is:

�0jk = �00k + �01kR ~

jk + �02kWjk + rjk

350—–AMERICAN SOCIOLOGICAL REVIEW

10 Though schools are not of analytical interest here, school is included as a level in the model because of the structure of the data and to allow school characteristics to be used as control variables.

This equation models the individual level intercept as a function of neighborhood cultur- al heterogeneity in ideal romantic relationship scripts (R

~ ) and neighborhood control variables

(W). The coefficient on R ~ , �01k, is the test of

Hypothesis 2 for relationship scripts. It indicates the degree to which neighborhood heterogene- ity in relationship scripts predicts the differ- ence between individuals’ ideal relationship scripts and their subsequently realized rela- tionships. The school-level equation serves to control for a set of school characteristics, Z:

�00k = �000 + �001Zk + uk

Again, the analysis proceeds by building up to a model with the full set of neighborhood and school control variables.

VARIABLES

STRUCTURAL NEIGHBORHOOD DISADVANTAGE. Neighborhoods are measured as census tracts. As is common in neighborhood effects research (e.g., Sampson et al. 1997), neighborhood dis- advantage is measured by a scale constructed from a series of neighborhood structural char- acteristics. The neighborhood disadvantage scale is the mean of the following standardized items: the census tract’s family poverty rate, percent single mother households, percent youth, male unemployment rate, percent black, percent of those older than 25 who are college graduates, percent of workers in managerial or profes- sional occupations, and percent affluent fami- lies (those with incomes above $75,000 per year), with the last three reversed in polarity. These data come from the 1990 census. Individuals’ residence at the Wave 1 in-home interview, which was conducted in spring or summer of 1995, is their census tract. The aver- age inter-item correlation for this scale is .52 and Cronbach’s alpha is .90.

The Structural Neighborhood Disadvantage Scale (hereafter, Neighborhood Disadvantage) measures the economic and social characteris- tics of the families that make up the neighbor- hood.11 Five of these variables (poverty, single

mother households, percent youth, male unem- ployment, and percent black) indicate the pres- ence of disadvantaged families.12 The remaining three (college graduates, managerial and pro- fessional workers, affluent families) indicate the absence of middle-class families since their polarity is reversed. While some researchers (e.g., Brooks-Gunn et al. 1993) argue that the absence of middle-class families is more impor- tant than the presence of disadvantaged families, there are high inter-item correlations across all eight variables in these data. This suggests that these two sets of measures capture the same underlying neighborhood SES concept but sim- ply focus on the presence of families at oppo- site ends of the SES distribution as indicators of a neighborhood’s position in that distribu- tion.13

PREMARITAL SEX. Premarital sex between Wave 1 (when almost all respondents are age 13 to 18) and Wave 2 (about a year later on aver- age) measures whether the adolescents report having had sexual intercourse outside of mar- riage between the first and second waves of the survey.14

RO M A N T I C R E L AT I O N S H I P S C R I P T S . The Addhealth in-home data contain a series of questions about adolescents’ ideal and actual romantic relationships. Ideal relationships are

NEIGHBORHOOD CULTURAL CONTEXT—–351

11 Neighborhood disadvantage is not meant to directly proxy for neighborhood cultural context as is sometimes done when direct measures of cultural context are not available. One of the goals of this arti- cle is to investigate the relationship between struc-

tural and cultural characteristics, and therefore cul- tural characteristics of neighborhoods are measured directly, as described below.

12 Percent youth roughly captures the number of adults per child to supervise or monitor young peo- ple, both within families and in the neighborhood in general.

13 A correlation matrix of these variables is avail- able from the author upon request.

14 Researchers studying adolescent sexual behav- ior from a more demographic perspective often focus on the transition to first sex. In this case, because of the emphasis on the effect of neighborhood context, conditioning on prior sexual behavior would be con- ditioning on an endogenous variable, one that is also affected by neighborhood context. In addition, lim- iting the sample to those who never had sex prior to Wave 1 would result in a large reduction in statisti- cal power. Models estimated on this more limited sample show substantively similar results but with larger standard errors (models not shown).

measured at Wave 1 by asking the adolescents to choose the events that would be part of their ideal romantic relationships and then to arrange the events in their preferred chronological order. For the ideal relationship scripts, respondents choose from and then order the following list of 17 events:

• We would go out together in a group • I would meet my partner’s parents • I would tell other people that we were a couple • I would see less of my other friends so I could

spend more time with my partner • We would go out together alone • We would hold hands • I would give my partner a present • My partner would give me a present • I would tell my partner that I loved him or her • My partner would tell me that he or she loved me • We would think of ourselves as a couple • We would talk about contraception or sexually

transmitted diseases • We would kiss • We would touch each other under our clothing • We would have sex • My partner or I would get pregnant • We would get married

In Wave 2, each respondent is asked to describe up to three past romantic relationships using the same procedure. However, in Wave 2 the menu of possible events is slightly different. The two events involving telling each other that you love one another are collapsed into one event, as are the two events involving giving and receiving gifts. Getting married is not included in the menu of events in Wave 2, and an addi- tional event, “You touched each others genitals (private parts),” is added. I consider the ideal relationship from Wave 1 to be a measure of the relationship script for the respondent’s ideal romantic relationship. I then use the first rela- tionship listed at Wave 2 that started after the Wave 1 interview as the actual relationship script.

The analysis requires the construction of two measures, one that captures heterogeneity in ideal relationship scripts within a neighborhood and another that captures the degree to which individual respondents’ actual relationship scripts differ from their ideal relationship scripts. Hypothesis 2 predicts that adolescents in neigh- borhoods with more heterogeneity in ideal rela- tionship scripts will have actual relationships

that are more different from their ideal rela- tionship scripts.

The string of events in the relationship scripts can be thought of as sequence data, and sequence analysis methods can be used to meas- ure the resemblance or difference between two scripts. Optimal matching provides a method for comparing two sequences based on the inser- tions, deletions, and substitutions required to transform one sequence into the other. Each insertion, deletion, or substitution carries a cost, and the optimal matching algorithm finds the minimum total cost to transform one sequence into the other. This cost is a measure of the dif- ference between two sequences. To measure heterogeneity at the neighborhood level in ideal relationship scripts, I take the mean of the dif- ferences between ideal romantic relationship scripts for all possible pairs of individuals in each neighborhood. The reliability of this meas- ure depends heavily on the number of Addhealth respondents in the neighborhood, so all models are weighted by this reliability. Further infor- mation on the romantic relationship scripts and sequence analysis is provided in the Online Supplement on the ASR Web site: http:// www2.asanet.org/journals/asr/2007/toc057. html.

PREGNANCY FRAMES. The degree to which respondents agree or disagree with the state- ment, “It wouldn’t be all that bad if you got preg- nant [or got someone pregnant] at this time in your life,” measures their frame regarding teenage pregnancy. Above I conceptualized frames as ways of thinking about the world and about the consequences of one’s actions. While ideally one would use a measure that distin- guished between multiple frames regarding pregnancy (such as early parenthood as status, early parenthood as hindrance, or early parent- hood as leading to a closer relationship between parent and child), no such measures are avail- able in these or any other survey data. However, this survey question is a reasonable measure of the respondent’s frame regarding teenage preg- nancy in that it captures both the valence and strength of an adolescent’s thinking about the consequences of teenage pregnancy. By includ- ing the phrase, “at this time in your life,” it cues respondents to think of consequences, and in particular of the consequences for themselves

352—–AMERICAN SOCIOLOGICAL REVIEW

rather than teenagers in general.15 Because of this practical orientation, this measure can be distinguished from more general attitudinal measures that tap values regarding the moral sta- tus of teenage sexual behavior or early parent- hood.16

To measure an individual’s adherence to this “pregnancy would have negative consequences for me” frame, I assign responses the numbers one through five such that higher values indi- cate a stronger feeling that pregnancy would have negative consequences. Measurement at the neighborhood level of pregnancy frame het- erogeneity is based on a variance measure appropriate for ordinal variables and is described in the Online Supplement.

NEIGHBORHOOD CONTROLS. In models that examine the relationship between neighbor- hood disadvantage and neighborhood cultural heterogeneity, it is necessary to control for other potential sources of cultural heterogeneity. Social organization theory points to three other potential sources of cultural heterogeneity: immigration, residential instability, and ethnic and racial diversity. To control for immigra- tion, I use the percent of neighborhood resi- dents who are foreign born. To control for residential instability, I control for percent of housing units that are owner-occupied and per- cent of housing units that have housed the same

household for the last five years.17 To control for ethnic diversity, I control for percent Hispanic and percent Hispanic squared. Finally, to con- trol for racial diversity, I use Simpson’s Interaction Index, described in the Online Supplement.

INDIVIDUAL/FAMILY CONTROLS. To minimize selection bias due to differences between indi- viduals across neighborhoods, individual and family control variables are required. Measured at Wave 1, these controls include race and eth- nicity indicators, age, gender, adolescent immi- grant status, language spoken at home, log family income, single parent household, step- parent or other household, mother’s age at birth, low birth weight, and for the primary parent (mother or female caregiver if available, father or male caregiver if not) immigration status, education, professional/managerial occupation, disability, and welfare receipt. Control vari- ables are described in more detail in the Online Supplement.

SCHOOL CONTROLS. Since previous research has found that school characteristics are impor- tant predictors of sexual and romantic behavior (e.g., Teitler and Weiss 2000), I include school controls in the multilevel models. School con- trols include indicators for private school, Catholic school, and rural/suburban/urban, and measures of school size, percent of students in a college-prep program, and cumulative dropout rate. Finally, continuous variables are grand mean centered in multilevel models.

RESULTS

Table 1 shows descriptive results on the degree to which Addhealth adolescents in different types of neighborhoods agree that getting preg- nant or getting someone pregnant would “not be all that bad” at this time in their lives. As we

NEIGHBORHOOD CULTURAL CONTEXT—–353

15 Contrast this question with questions in the atti- tude measures in the General Social Survey: “There’s been a lot of discussion about the way morals and atti- tudes about sex are changing in this country. If a man and woman have sex relations before marriage, do you think it is always wrong, almost always wrong, wrong only sometimes, or not wrong at all?” and, “What if they are in their early teens, say 14 to 16 years old? In that case, do you think sex relations before marriage are always wrong, almost always wrong, wrong only sometimes, or not wrong at all?”

16 This interpretation is also based on qualitative interviews with 60 adolescent boys in three neigh- borhoods in Boston in which this survey question was administered to each boy and then both their under- standing of the question and their responses were dis- cussed (these data are discussed in more detail in Harding 2005). When discussing how they thought about answering this question, the boys focused on how their lives would be different were they to

become a parent in the near future and how that would affect their future plans and life chances.

17 These two variables are obviously somewhat collinear, but their role in the model is to control for potential confounding due to residential instability. I am not interested in the magnitude or statistical sig- nificance of their coefficients.

would expect, adolescents in more advantaged neighborhoods are more likely to strongly dis- agree that getting pregnant would not be all that bad. More surprising from the perspective of social isolation theory, though, we find that even in the most disadvantaged neighborhoods, over 70 percent of adolescents disagree or strongly disagree that a pregnancy would not be all that bad. It would be very difficult to con- clude from Table 1 that the majority of adoles- cents in disadvantaged neighborhoods express cultural models radically at odds with main- stream views on teenage pregnancy. Yet Table 1 also reveals the greater heterogeneity of respons- es in disadvantaged neighborhoods. Descriptively, disadvantaged neighborhoods are distinguished from advantaged neighborhoods by the heterogeneity of their responses. Whereas more advantaged neighborhoods have consid- erable consensus on teenage pregnancy, in more disadvantaged neighborhoods there is a sizable minority of adolescents who do not subscribe to the dominant view. Most research on neigh- borhood context measures neighborhood char- acteristics using measures of central tendency (such as the mean), but Table 1 reveals that internal variation (or what I am calling cultur- al heterogeneity) also differs across neighbor- hoods and that neighborhood mean and neighborhood cultural heterogeneity are empir- ically related.

NEIGHBORHOOD DISADVANTAGE AND CULTURAL HETEROGENEITY

Table 2 examines the relationship between neighborhood disadvantage and neighborhood cultural heterogeneity, controlling for other

neighborhood characteristics that may be sources of spuriousness. It examines Hypothesis 1, that disadvantaged neighborhoods will exhib- it greater cultural heterogeneity. Table 2 shows one OLS regression model each for hetero- geneity in pregnancy frames and relationship scripts. The unit of analysis is the neighbor- hood (census tract). The key predictor is neigh- borhood disadvantage.18 Controls are included for percent Hispanic, racial diversity, percent foreign born, and residential stability, though their coefficients are not of substantive interest for testing Hypothesis 1.19 As these control vari- ables are highly correlated with one another, their individual coefficients should not be inter- preted as indicating their ability to predict cul- tural heterogeneity.20 Each model is weighted by the reliability of the neighborhood level het- erogeneity measure. Descriptive statistics for the

354—–AMERICAN SOCIOLOGICAL REVIEW

Table 1. Pregnancy Frame by Neighborhood Disadvantage

“It wouldn’t be all that bad if you got pregnant [or got someone pregnant] at this time in your life.”

Neighborhood Neither Agree Strongly Disadvantage Scale Strongly Agree Agree Nor Disagree Disagree Disagree Quintiles (percent) (percent) (percent) (percent) (percent) Row Obs

1st 1.1 04.0 05.2 31.7 58.1 01,899 2nd 1.6 05.3 07.9 32.1 53.1 03,508 3rd 1.2 06.8 09.3 31.7 50.9 03,880 4th 2.9 09.7 10.1 32.6 44.8 02,689 5th 4.0 12.4 11.1 35.2 37.4 01,640 All 2.0 07.4 08.7 32.4 49.5 13,616

Note: Estimates account for Addhealth sample design and Addhealth in-home Wave 1 weight.

18 Entering the neighborhood structural charac- teristics that make up the Neighborhood Disadvantage Scale in the model individually produces similar results (not shown).

19 Percent Hispanic is given a quadratic function- al form in these models because one would expect the association between percent Hispanic and cultural heterogeneity to increase as the proportion of Hispanics increases from zero, but to peak and then decrease as the proportion Hispanic approaches one.

20 Multiple control variables are entered into the models individually so that they account for the max- imum potential omitted variable bias in the relation- ship between structural disadvantage and cultural heterogeneity. Drawing conclusions about racial diversity, immigrant density, and residential stabili- ty as causes of cultural heterogeneity would require a different model specification.

variables in these models are provided in Table S2 in the Online Supplement.

Both models show a strong and statistically significant association between neighborhood disadvantage and neighborhood cultural het- erogeneity, net of controls, and these associa- tions are of the same order of magnitude across domains. In the pregnancy frame model, a one standard deviation increase in the neighbor- hood disadvantage scale is also associated with an increase of about one-fifth of a standard deviation in pregnancy frame heterogeneity.21 In the relationship script model, a one standard deviation increase in the neighborhood disad- vantage scale is associated with an increase of a little less than one-sixth of a standard devia- tion in neighborhood cultural heterogeneity. These results show that disadvantaged neigh- borhoods exhibit greater cultural hetero- geneity.22

CULTURAL HETEROGENEITY AND ADOLESCENT OUTCOMES

Hypothesis 2—adolescents in more culturally heterogeneous neighborhoods will be less like- ly than those in more homogenous ones to act in accordance with their own frames and scripts—predicts that the relationships between individual frames or scripts and individual out- comes will be weaker in neighborhoods with greater cultural heterogeneity. Table 3 shows various versions of the multilevel logit models predicting premarital sexual activity described above.23 (Descriptive statistics for the variables used in this model are provided in Table S3 in

NEIGHBORHOOD CULTURAL CONTEXT—–355

Table 2. OLS Models of Neighborhood Cultural Heterogeneity

Model 1 Model 2

Independent Variables Pregnancy Frame Relationship Script

Nhood Disadvantage Scale .06786* .15083* (.00872) (.02704)

Percent Hispanic –.00014 –.00069 (.00125) (.00398)

Percent Hispanic Squared .00001 –.00002 (.00014) (.00004)

Racial Diversity Index .00102* .00102 (.00043) (.00138)

Percent Foreign Born .00074 .00429 (.00080) (.00264)

Percent Owner-Occupied .00026 –.00283* (.00041) (.00133)

Percent Units Occupied 5 Years .00031 –.00087 (.00055) (.00178)

Constant .56636* .14836 (.04751) (.15386)

N Neighborhoods 1,322 1,367 R2 .086 .057

Note: Each model weighted by reliability of neighborhood heterogeneity measure. Descriptive statistics available in Table S2 in the Online Supplement. * p < .05.

21 The neighborhood disadvantage variable is stan- dardized because it is a scale with no inherent met- ric. The standard deviation of the pregnancy frame heterogeneity variable is .3285—see Table S2 in the Online Supplement.

22 One might question the relatively low R2 values produced by the models in Table 2. These are likely

the result of measurement error in the cultural het- erogeneity measures due to the low reliability of measurement in neighborhoods with few respon- dents. In OLS, measurement error in the dependent variable does not bias coefficients but does increase error variance.

23 Cases were selected for the analysis in Table 3 if they had nonmissing data on individual pregnan- cy frame and neighborhood heterogeneity in preg- nancy frames. Respondents under the age of 15 were not asked the survey question measuring the preg- nancy frame.

the Online Supplement. Individual, family, and school characteristic control variable coeffi- cients are provided in Table S4.) In Model 1, the individual pregnancy frame is the primary pre- dictor.24 This variable is interacted with dummy

variables for quartiles of the distribution of neighborhood heterogeneity in pregnancy frames. These quartile dummies are also entered into the model directly. Converting the neigh- borhood frame heterogeneity measure to quar-

356—–AMERICAN SOCIOLOGICAL REVIEW

Table 3. Three-Level Logit Models Predicting Sex Between Waves I and II

Independent Variables Model 1 Model 2 Model 3

Individual Pregnancy Frame –.4384* –.4185* –.4975* (.0873) (.0988) (.1030)

Nhood Mean Pregnancy Frame –.1504 –.0930 (.1128) (.1123)

Individual Frame � Nhood Mean Pregnancy Frame –.0196 .0477 (.0830) (.0868)

Pregnancy Frame Nhood Heterogeneity Quartiles: —2nd Quartile Versus 1st Quartile .0595 .0200 .0351

(.1006) (.1104) (.1115) —3rd Quartile Versus 1st Quartile .0517 –.0360 –.0368

(.1019) (.1285) (.1289) —Top Quartile Versus 1st Quartile .1503 .0144 .0159

(.1648) (.1986) (.1947) Individual Frame � NHood Frame Heterogeneity Quartiles: —2nd Quartile Versus 1st Quartile .2100* .1949 .1727

(.1001) (.1055) (.1040) —3rd Quartile Versus 1st Quartile .3745* .3514* .3144*

(.0902) (.1161) (.1189) —Top Quartile Versus 1st Quartile .4335* .4163* .4059*

(.1381) (.1560) (.1654) Nhood Disadvantage Scale Quartiles: —2nd Quartile Versus 1st Quartile .0965

(.0767) —3rd Quartile Versus 1st Quartile .1408

(.0881) —Top Quartile Versus 1st Quartile .2160*

(.1057) Individual Frame � Nhood Disadvantage Scale Quartiles: —2nd Quartile Versus 1st Quartile .1361

(.0729) —3rd Quartile Versus 1st Quartile .0783

(.0765) —Top Quartile Versus 1st Quartile .2419*

(.0809)

Constant –.5372* –.4742* –.5926* (.1521) (.1714) (.1824)

N Individuals 9,281 9,281 9,281 N Neighborhoods 1,357 1,357 1,357 N Schools 0,142 0,142 0,142

Note: Models weighted by reliability of neighborhood frame means. Robust standard errors in parentheses. Descriptive statistics available in Table S3. Individual, family, and school control variable coefficients available in Table S4 in the Online Supplement. Missing values on control variables imputed using multiple imputation with 10 replications. * p < .05.

24 I experimented with entering this variable into the model as a series of dummy variables, but the rela-

tionship with the outcome was fairly linear, so to reduce model complexity, I enter it as a linear term.

tiles allows its effects to be nonlinear. The indi- vidual frame variable captures the relationship between one’s pregnancy frame and premarital sexual behavior within low heterogeneity neigh- borhoods. Not surprisingly, among adolescents in these neighborhoods, there is a strong rela- tionship between adolescents’ individual preg- nancy frames and their sexual behavior. The interaction terms capture the difference in this relationship across neighborhoods with differ- ent levels of frame heterogeneity. These coef- f icients show that as cultural heterogeneity within a neighborhood increases, the association between an adolescent’s own pregnancy frame and sexual behavior declines dramatically.

Neighborhoods with greater heterogeneity in pregnancy frames also exhibit an average frame that views teenage pregnancy less nega- tively (see Table 1). In other words, there is an association between central tendency and vari- ation at the neighborhood level for pregnancy frames. Thus, one potential source of spurious- ness in Model 1 of Table 3 is a neighborhood’s mean pregnancy frame. Model 2 adds the con- trol for the neighborhood mean pregnancy frame. Its coefficient is in the expected direc- tion, but it is not statistically significant. Its inclusion changes only slightly the primary coefficients of interest, the interaction terms. This model also adds a term for the interaction between the individual pregnancy frame and the neighborhood mean pregnancy frame. Its coefficient is small and statistically insignifi- cant. These results suggest that the smaller rela- tionship between adolescents’ own pregnancy frames and their sexual behavior in more het- erogeneous neighborhoods is not merely a prod- uct of greater average permissiveness toward teenage pregnancy in such neighborhoods.

Moreover, the comparatively small magni- tudes and lack of statistical significance of the neighborhood mean pregnancy frame coeffi- cients in predicting sexual activity suggest that once cultural heterogeneity is taken into account, a neighborhood oppositional cultural environ- ment is not an important cause of differences in sexual behavior across neighborhoods. This is further evidence against the social isolation view of the cultural context of disadvantaged neighborhoods. Were social and/or cultural iso- lation affecting individual adolescent behavior, one would expect neighborhood mean preg-

nancy frames to be strong predictors of indi- vidual sexual activity.

It is also possible that the importance of neighborhood cultural heterogeneity is a spuri- ous result of the strong association between neighborhood cultural heterogeneity and neigh- borhood disadvantage. Without controlling for neighborhood disadvantage, we cannot elimi- nate the possibility that the effects of cultural heterogeneity at the neighborhood level are sim- ply capturing the effects of neighborhood dis- advantage. Model 3 adds controls for neighborhood disadvantage that parallel the variables measuring frame heterogeneity. These include dummy variables for quartiles of the dis- tribution of neighborhood disadvantage and interactions between these quartiles and the individual pregnancy frame. As we would expect, inclusion of the neighborhood disad- vantage variables reduces the magnitudes of the frame heterogeneity interaction coefficients, but they remain large and statistically signifi- cant.

The coefficients in Model 3 show that in the lowest frame heterogeneity neighborhoods, the relationship between individual frame and behavior is strong, but that the importance of the respondent’s individual frame declines as neigh- borhood frame heterogeneity increases. In the lowest heterogeneity neighborhoods, a one cat- egory increase in the individual pregnancy frame multiples the odds of sexual behavior by about .61 (e–.4975 = .6080), while in the highest het- erogeneity neighborhoods, a similar move in the individual pregnancy frame multiples the odds of sexual behavior by a more modest .91 (e[–.4975 + .4059] = .9124). (Recall that since expo- nentiated coefficients from logit models are odds ratios, values closer to 1 represent small- er associations.) Adolescents in higher frame heterogeneity neighborhoods are much less like- ly to act in accordance with their own articulated frames regarding teenage pregnancy than are their counterparts in more culturally homoge- nous neighborhoods.

One can also illustrate the interactions in these models graphically. Figure 1a shows the relationship between predicted probabilities of sexual activity and individuals’ pregnancy frames based on estimates from Model 3 in Table 3. Separate lines are plotted by quartiles of neighborhood frame heterogeneity. Other variables are fixed at either zero (indicator vari-

NEIGHBORHOOD CULTURAL CONTEXT—–357

ables) or their means (continuous variables). Among adolescents living in neighborhoods with the least cultural heterogeneity (those in the 1st and 2nd quartiles of heterogeneity), indi- vidual pregnancy frames are strongly predictive of sexual activity. Among those in neighbor-

hoods in the 3rd quartile, there is still a nega- tive sloping line, but the slope is considerably more modest. Among those in neighborhoods with the highest pregnancy frame heterogene- ity (the 4th quartile), the relationship between one’s own pregnancy frame and one’s sexual

358—–AMERICAN SOCIOLOGICAL REVIEW

Figure 1a. Relationship Between Individual Pregnancy Frame and Sexual Activity (by Neighborhood Pregnancy Frame Heterogeneity Quartiles)

Figure 1b. Relationship Between Neighborhood Pregnancy Frame Heterogeneity and Sexual Activity (by Individual Pregnancy Frame)

1 2 3 4 Heterogeneity Quartiles

behavior is almost flat. Figure 1b rearranges the same data points to show the relationship between neighborhood pregnancy frame het- erogeneity and sexual activity, with separate lines plotted by the adolescents’ own individual pregnancy frames. The importance of neigh- borhood cultural heterogeneity is also evident here. As cultural heterogeneity at the neigh- borhood level increases, the predicted proba- bilities of sexual activity converge for all adolescents, no matter what their stated frame regarding teenage pregnancy.

I now turn to the analysis of the romantic relationship scripts. Recall that the models are slightly simpler here, as the failure to realize one’s ideal relationship script is measured direct- ly. Table 4 examines the relationship between neighborhood heterogeneity in ideal relationship scripts and the difference between the ideal and actual relationship scripts of an individual.25

(Descriptive statistics for the variables used in this model are provided in Table S5, and con- trol variable coefficients are provided in Table S6 in the Online Supplement.) Model 1 shows that, net of individual, family, and school char- acteristics, neighborhood script heterogeneity is a significant predictor of the difference between

one’s ideal and actual romantic relationships. Models 2 adds controls for neighborhood dis- advantage. This control reduces the hetero- geneity coefficient, but it remains statistically significant. These models provide further evi- dence in support of Hypothesis 2, suggesting that greater neighborhood cultural heterogene- ity leads to a lower likelihood of realizing one’s ideal relationship. To the extent that romantic relationship partners are selected from among one’s neighbors, part of this association is like- ly related to the process of partner selection. In more heterogeneous neighborhoods, it is more difficult to find a romantic partner whose ideal relationship script is similar to one’s own. When two relationship scripts are incompatible, at least one partner must fail to realize his or her ideal relationship if the relationship continues.26

CONCLUSION

This article has attempted to reorient current thinking about the cultural context of disad- vantaged neighborhoods, particularly as it relates to adolescent outcomes. The canonical account of neighborhood culture in the urban poverty literature (Wilson 1987) is one in which

NEIGHBORHOOD CULTURAL CONTEXT—–359

Table 4. Three-Level Models Predicting Difference Between Ideal Relationship and Actual Relationship Scripts

Independent Variables Model 1 Model 2

Nhood Heterogeneity Ideal Relationship Script .0539* .0473* (.0154) (.0156)

Nhood Disadvantage Scale .0237 (.0128)

Constant 2.0518* 2.0509* (.0310) (.0310)

N Individuals 5,916 5,916 N Neighborhoods 1,299 1,299 N Schools 0,143 0,143

Note: Models weighted by reliability of neighborhood script heterogeneity means. Robust standard errors in parentheses. Descriptive statistics available in Table S5 in the Online Supplement. Individual, family, and school control variable coefficients available in Table S6. Missing values on control variables imputed using multiple imputation with 10 replications. * p < .05.

25 Cases were selected for analysis in Table 4 if they had nonmissing data on the ideal relationship script and neighborhood relationship script heterogeneity variables and had at least one romantic relationship that started after the Wave 1 in-home interview (dur- ing which data on the ideal relationship script were collected).

26 It is likely that there are important gender dif- ferences in adolescents’ ideal romantic relationship scripts. Unfortunately, creating gender-specific meas- ures of relationship script heterogeneity requires double the subject density per neighborhood to main- tain measurement reliability. It is not possible to cre- ate such gender-specific measures in these data.

the departure of middle-class blacks from inner- city neighborhoods has left behind a socially iso- lated underclass with values that both depart significantly from middle-class or mainstream culture and are relatively uniform and coherent. Massey and Denton (1993) take this a step fur- ther, writing that “ghetto culture has become an entity unto itself ” (p. 172). Theoretical notions based on this premise, such as “oppositional culture,” now dominate much scholarship on inner-city, disadvantaged communities. Accordingly, in considering the impact of neigh- borhood culture on adolescents, most previous work examining adolescent outcomes has employed a subculture model, in which negative outcomes are the product of participation in a particular subculture whose values promote negative behavior. Paradoxically, this work also emphasizes the multiple and overlapping lifestyle groups or orientations in disadvan- taged neighborhoods.

I have proposed an alternate view of the cul- tural context of disadvantaged neighborhoods, a view that emphasizes their cultural diversity. The urban poverty literature has lost track of the insights developed by Shaw and McKay (1969), in which poor communities are characterized by conflicting cultural systems. Though Shaw and McKay were concerned with crime and delin- quency, these ideas have the potential to further our understanding of the cultural context of dis- advantaged neighborhoods as it relates to other domains as well. I have argued that an impor- tant way in which disadvantaged neighborhoods differ from more advantaged neighborhoods is cultural heterogeneity, the presence of a wide array of competing and contradictory cultural models.

The above empirical analyses show that cul- tural heterogeneity helps to predict adolescent outcomes regarding sexual activity and roman- tic relationships. This article has examined neighborhood cultural heterogeneity with regard to frames about teenage pregnancy and scripts for romantic relationships. Using data from Addhealth, I have shown that more disadvan- taged neighborhoods exhibit greater cultural diversity in pregnancy frames and relationship scripts. This article also shows that adolescents in more culturally heterogeneous neighborhoods are less likely to act in accordance with their own articulated scripts and frames. In neigh- borhoods with a greater diversity of frames

regarding the consequences of teenage preg- nancy, adolescents are less likely to act in accor- dance with their own frames. And where there is greater variation in scripts for romantic rela- tionships, adolescents are less likely to realize their own ideal relationships in the actual rela- tionships that they pursue.

This research has theoretical implications for both social isolation and social organization explanations of neighborhood effects. The importance of cultural heterogeneity calls into question one of the fundamental ideas of social isolation theory, that disadvantaged neighbor- hoods are isolated from mainstream or middle- class culture. Instead, the results suggest that disadvantaged neighborhoods contain a mix of cultural models and that this heterogeneity impacts neighborhood adolescents. In contrast, the findings in this study are generally sup- portive of neighborhood effects models based on social disorganization theory and move us toward understanding the mechanisms under- lying these models. Social disorganization the- ory highlights the limited capacity of residents in disadvantaged communities to regulate the behaviors of their neighbors. Heterogeneity in cultural lifestyles or orientations can be under- stood as the failure of more middle-class or mainstream residents of disadvantaged neigh- borhoods to regulate behavior in their commu- nities. The result of this failure is that neither “oppositional” nor “mainstream” behavior is dominant in disadvantaged communities.

More broadly, this article’s findings chal- lenge the underclass concept as a fruitful frame- work for the analysis of life in poor urban neighborhoods. This study suggests that rather than being disconnected from mainstream soci- ety, poor urban residents live in cultural envi- ronments that include both local elements and elements from the wider society and that these disparate elements are often contradictory and competing. This means that, as has been described in the ethnog raphic literature (Anderson 1999; Newman 1999), cultural con- flict and cultural disagreement are widespread. I have argued here that such conflict may have further consequences for well-being and deci- sion making. In addition, a theoretical frame- work that incorporates cultural heterogeneity allows both for variation in individual outcomes among those who experience the same structural conditions and for greater agency, as individu-

360—–AMERICAN SOCIOLOGICAL REVIEW

als choose between and evaluate the cultural models that exist in their cultural environments and in their repertoires.

The cultural heterogeneity view is both more optimistic and in some ways more pessimistic than the underclass view. The fact that there is considerable support for conventional norms in culturally heterogeneous environments means that the seeds of positive change are likely pres- ent in the vast majority of disadvantaged neigh- borhoods. On the other hand, neighborhood cultural heterogeneity can be created when even a relatively small minority of individuals or families display alternative or oppositional behaviors or lifestyles and thereby disrupt, con- fuse, or weaken the dominance of convention- al cultural models. Cultural heterogeneity, and its consequences, will likely persist until the most disadvantaged residents of poor neigh- borhoods are empowered with more positive ways to adapt to and rise out of poverty.

Though this study has not set out to explic- itly test competing hypotheses regarding cultural theories, it does suggest some implications for cultural sociology. First, findings regarding the consequences of neighborhood cultural hetero- geneity show how local or group cultural con- texts can matter for behavior, even when those local cultures are not starkly distinct or separate from other local cultures or the wider culture. In other words, different behavioral outcomes do not require distinct subcultures. Second, this article provides another example of the utility of cognitive cultural concepts such as frames, scripts, and repertoires. In doing so, it also sug- gests a further specification of how repertoires operate. The activation of elements from one’s repertoire depends not just on one’s structural position (Lamont 1992) but also on the rela- tionships between various elements in the reper- toire and on local social support for those elements. Third, the importance of cultural het- erogeneity suggests that not only are the mean- ings attached to cultural models important, but how those meanings are organized and related to one another also has implications for behav- ior. The degree of harmony or conflict among available cultural models is another dimension along which cultural contexts or cultural reper- toires should be evaluated.

While the results in this study are strongly suggestive of the importance of cultural het- erogeneity for understanding disadvantaged

neighborhoods, further research is required. Better measures of frames and scripts would provide stronger tests of the hypotheses exam- ined here. For example, the measure of hetero- geneity of pregnancy frames does not capture the presence of multiple frames but rather vari- ation in allegiance to a particular frame.27

Further empirical predictions can be generated to test the utility of cultural heterogeneity as a sociological concept. For example, the hypoth- esized weaker commitment to frames or scripts that comes from exposure to greater cultural variations would suggest that adolescents in more heterogeneous neighborhoods should exhibit greater temporal variation in their artic- ulated frames and scripts. Finally, further research is needed to examine the degree of cultural heterogeneity in other domains such as education or health and the ability of cultur- al heterogeneity to predict other outcomes.

David J. Harding is Assistant Professor in the Department of Sociology and Assistant Research Scientist at the Population Studies Center, University of Michigan. With Katherine Newman, Cybelle Fox, Jal Mehta, and Wendy Roth, he is coauthor of Rampage: The Social Roots of School Shootings (Basic Books, 2004). His recent work has also appeared in American Jour nal of Sociology, Sociological Methods and Research, and Sociology of Education. He is currently completing a book manuscript, titled Living the Drama, that investi- gates the mechanisms that generate neighborhood effects among adolescent boys, based on qualitative interviews in three neighborhoods in Boston.

REFERENCES

Ainsworth-Darnell, James and Douglas Downey. 1998. “Assessing the Oppositional Culture Explanation for Racial/Ethnic Differences in School Performance.” American Sociological Review 63:536–53.

Ajzen, Icek. 2001. “Nature and Operation of

NEIGHBORHOOD CULTURAL CONTEXT—–361

27 In addition, it could be argued that the survey question used here to measure pregnancy frames also taps the moral status of teenage pregnancy or that the question wording is confusing because it contains a double negative. Nevertheless, this measure is strongly predictive of behavior among adolescents in culturally homogenous neighborhoods, so measure- ment problems would have to disproportionately affect adolescents in culturally heterogeneous neigh- borhoods to call into question the larger argument.

Attitudes.” Annual Review of Psycholog y 52:27–58.

Anderson, Elijah. 1978. A Place on the Corner. Chicago, IL: University of Chicago Press.

———. 1999. Code of the Street: Decency, Violence, and the Moral Life of the Inner City. New York: W. W. Norton.

Benford, Robert D. and David A. Snow. 2000. “Framing Processes and Social Movements: An Overview and Assessment.” Annual Review of Sociology 26:611–39.

Bourdieu, Pierre and Jean-Claude Passeron. [1970] 1977. Reproduction in Education, Society and Culture. Beverly Hills, CA: Sage.

Bourgois, Phillipe. 1995. In Search of Respect: Selling Crack in El Barrio. Cambridge, England: Cambridge University Press.

Brooks-Gunn, Jeanne, Greg J. Duncan, Pamela K. Klebanov, and Naomi Sealand. 1993. “Do Neighborhoods Influence Child and Adolescent Development?” American Journal of Sociology 99:353–95.

Browning, Christopher R., Tama Leventhal, and Jeanne Brooks-Gunn. 2005. “Sexual Initiation in Early Adolescence: The Nexus of Parental and Community Control.” American Sociological Review 70:758–78.

Bryson, Bethany. 1996. “Anything But Heavy Metal: Symbolic Exclusion and Musical Dislikes.” American Sociological Review 61:884–99.

Bursik, Robert J., Jr. and Harold G. Grasmick. 1993. Neighborhoods and Crime: The Dimensions of Effective Community Control. New York: Lexington Books.

Carter, Prudence. 2005. Keepin’ It Real: School Success Beyond Black and White. New York: Oxford University Press.

Cook, Phillip and Jens Ludwig. 1998. “The Burden of ‘Acting White’: Do Black Adolescents Disparage Academic Achievement?” Pp. 375–401 in The Black White Test Score Gap, edited by C. Jencks and M. Phillips. Washington, DC: Brookings.

DiMaggio, Paul. 1982. “Cultural Capital and School Success: The Impact of Status Culture Participation on the Grades of U.S. High School Students.” American Sociological Review 47:189–201.

———. 1997. “Culture and Cognition.” Annual Review of Sociology 23:263–87.

Dohan, Daniel. 2003. The Price of Poverty: Money, Work, and Culture in the Mexican American Barrio. Berkeley, CA: University of California Press.

DuBois, W. E. B. [1899] 1996. The Philadelphia Negro: A Social Study. Philadelphia, PA: University of Pennsylvania Press.

Duneier, Mitchell. 1992. Slim’s Table: Race, Respectability, and Masculinity. Chicago, IL: University of Chicago Press.

Edin, Kathryn and Maria Kefalas. 2005. Promises I Can Keep: Why Poor Women Put Motherhood Before Marriage. Berkeley, CA: University of California Press.

Erickson, Bonnie. 1996. “Culture, Class and Connections.” American Journal of Sociology 102: 217–51.

Fernandez, Roberto M. and David Harris. 1992. “Social Isolation and the Underclass.” Pp. 257–93 in Drugs, Crime, and Social Isolation: Barriers to Urban Opportunity, edited by A. V. Harrell and G. E. Peterson. Washington, DC: Urban Institute Press.

Fine, Gary Alan. 1979. “Small Groups and Culture Creation: The Idioculture of Little League Baseball Teams.” American Sociological Review 44:733–45.

Fine, Gary Alan and Sherryl Kleinman. 1979. “Rethinking Subculture: An Interactionist Analysis.” American Journal of Sociology 85:1–20.

Fishbein, Martin and Icek Ajzen. 1975. Belief, Attitude, Intention, and Behavior: An Introduction to Theory and Research. Reading, MA: Addison- Wesley.

Fordham, Signithia and John Ogbu. 1986. “Black Students’ School Success: Coping with the Burden of ‘Acting White.’” Urban Review 18:176–206.

Frazier, E. Franklin. 1966. The Negro Family in the United States. Rev./Abridged ed. Chicago, IL: University of Chicago Press.

Fuller, Bruce, Susan D. Holloway, Marylee F. Rambaud, and Constanza Eggers-Pierola. 1996. “How Do Mothers Choose Child Care? Alternative Cultural Models in Poor Neighborhoods.” Sociology of Education 69:83–104.

Gans, Herbert. 1962. Urban Villagers. New York: Free Press.

Goffman, Erving. 1959. The Presentation of Self in Everyday Life. New York: Doubleday.

———. 1974. Frame Analysis. Cambridge, MA: Harvard University Press.

Goldenberg, Claude, Ronald Gallimore, Leslie Reese, and Helen Garnier. 2001. “Cause or Effect? A Longitudinal Study of Immigrant Latino Parents’ Aspirations and Expectations, and Their Children’s School Performance.” American Educational Research Journal 38:547–82.

Hannerz, Ulf. 1969. Soulside: Inquiries into Ghetto Culture and Community. New York: Columbia University Press.

Harding, David J. 2003. “Counterfactual Models of Neighborhood Effects: The Effect of Neighborhood Poverty on Dropping Out and Teenage Pregnancy.” American Journal of Sociology 109:676–719.

———. 2005. “Why Neighborhoods Matter: Structural and Cultural Influences on Adolescents in Poor Communities.” Ph.D. disser tation, Department of Sociology, Harvard University.

Harris, Kathleen Mullan, Francesca Florey, Joyce

362—–AMERICAN SOCIOLOGICAL REVIEW

Tabor, Peter S. Bearman, Jo Jones, and J. Richard Udry. 2003. The National Longitudinal Study of Adolescent Health: Research Design [WWW doc- ument]. Retrieved February 17, 2007 (http:// www.cpc.unc.edu/projects/addhealth/design).

Hochschild, Jennifer L. 1995. Facing Up to the American Dream: Race, Class, and the Soul of the Nation. Princeton, NJ: Princeton University Press.

Holloway, Susan D., Bruce Fuller, Marylee F. Rambaud, and Constanza Eggers-Pierola. 1997. Through My Own Eyes: Single Mothers and the Cultures of Poverty. Cambridge, MA: Harvard University Press.

Lamont, Michèle. 1992. Money, Morals and Manners: The Culture of the French and the American Upper-Middle Class. Chicago, IL: University of Chicago Press.

———. 2000. The Dignity of Working Men: Morality and the Boundaries of Race, Class, and Immigration. Cambridge, MA: Harvard University Press.

Lamont, Michèle and Laurent Thévenot. 2000. Rethinking Comparative Cultural Sociology: Repertoires of Evaluation in France and the United States. Cambridge, England: Cambridge University Press.

Lewis, Oscar. 1969. “The Culture of Poverty.” Pp. 187–200 in On Understanding Poverty: Perspectives from the Social Sciences, edited by D. P. Moynihan. New York: Basic Books.

Liebow, Elliot. 1967. Tally’s Corner: A Study of Negro Streetcorner Men. Boston, MA: Little, Brown.

Massey, Douglas and Nancy Denton. 1993. American Apartheid: Segregation and the Making of an Underclass. Cambridge, MA: Harvard University Press.

Mickelson, Roslyn A. 1990. “The Attitude- Achievement Paradox Among Black Adolescents.” Sociology of Education 63:44–61.

Moynihan, Daniel P. 1965. The Negro Family: The Case for National Action. Washington, DC: U.S. Department of Labor.

Newman, Katherine S. 1992. “Culture and Structure in The Truly Disadvantaged.” City and Society 6:3–25.

———. 1999. No Shame in My Game: The Working Poor in the Inner City. New York: Vintage and Russell Sage.

Ogbu, John U. 2004. “Collective Identity and the Burden of ‘Acting White’ in Black History, Community, and Education.” The Urban Review 36:1–35.

Park, Robert E. and Ernest W. Burgess. 1925. The City. Chicago, IL: University of Chicago Press.

Parkin, Frank. 1971. Class Inequality and Political Order: Social Stratification in Capitalist and Communist Societies. London: MacGibbon and Lee.

Patillo-McCoy, Mary. 1998. “Church Culture as a

Strategy of Action in the Black Community.” American Sociological Review 63:767–84.

———. 1999. Black Picket Fences: Privilege and Peril Among the Black Middle Class. Chicago, IL: University of Chicago Press.

Quinn, Naomi and Dorothy Holland. 1987. “Culture and Cognition.” Pp. 3–40 in Cultural Models in Language and Thought, edited by D. Holland and N. Quinn. Cambridge, England: Cambridge University Press.

Rankin, Bruce H. and James M. Quane. 2000. “Neighborhood Poverty and the Social Isolation of Inner-City African American Families.” Social Forces 79:139–64.

Raudenbush, Stephen W. and Anthony S. Bryk. 2001. Hierarchical Linear Models: Applications and Data Analysis Methods. 2d ed. New York: Sage Publications.

Rodman, Hyman. 1963. “The Lower-Class Value Stretch.” Social Forces 42:205–15.

Sampson, Robert, Stephen Raudenbush, and Felton Earls. 1997. “Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy.” Science 227:918–24.

Schudson, Michael. 1989. “How Culture Works: Perspectives from Media Studies on the Efficacy of Symbols.” Theory and Society 18:153–80.

Sewell, William. 1992. “A Theory of Structure: Duality, Agency, and Transformation.” American Journal of Sociology 98:1–29.

———. 1999. “The Concept(s) of Culture.” Pp. 35–61 in Beyond the Cultural Turn, edited by V. E. Bonnell and L. Hunt. Berkeley, CA: University of California Press.

Shaw, Clifford R. 1929. Delinquency Areas. Chicago, IL: University of Chicago Press.

Shaw, Clifford R. and Henry D. McKay. 1969. Juvenile Delinquency and Urban Areas. Rev. ed. Chicago, IL: University of Chicago Press.

Small, Mario L. 2004. Villa Victoria: The Transformation of Social Capital in a Boston Barrio. Chicago, IL: University of Chicago Press.

Solorzano, Daniel G. 1992. “An Exploratory Analysis of the Effects of Race, Class and Gender on Student and Parent Mobility Aspirations.” Journal of Negro Education 61:30–44.

Suttles, Gerald. 1968. The Social Order of the Slum: Ethnicity and Territory in the Inner City. Chicago, IL: University of Chicago Press.

Swidler, Ann. 1986. “Culture in Action: Symbols and Strategies.” American Sociological Review 51:273–86.

———. 2001. Talk of Love. Chicago, IL: University of Chicago Press.

Teitler, Julien and Christopher Weiss. 2000. “Effects of Neighborhood and School Environments on Transitions to First Sexual Intercourse.” Sociology of Education 73:112–32.

Tigges, Leann M., Irene Browne, and Gary P. Green.

NEIGHBORHOOD CULTURAL CONTEXT—–363

1998. “Social Isolation of the Urban Poor: Race, Class, and Neighborhood Effects in Social Resources.” Sociological Quarterly 39:53–77.

Tilly, Charles. 1978. From Mobilization to Revolution. Reading, MA: Addison-Wesley.

Valentine, Charles A. 1968. Culture and Poverty: Critique and Counter-Proposals. Chicago, IL: University of Chicago Press.

Venkatesh, Sudhir A. 2000. American Project. Chicago, IL: University of Chicago Press.

Wilson, William Julius. 1987. The Truly Disadvantaged: The Inner-City, the Underclass, and Public Policy. Chicago, IL: University of Chicago Press.

———. 1996. When Work Disappears: The World of the New Urban Poor. New York: Knopf.

Young, A lford A ., Jr. 2004. The Minds of Marginalized Black Men: Making Sense of Mobility, Opportunity, and Life Chances. Princeton, NJ: Princeton University Press.

364—–AMERICAN SOCIOLOGICAL REVIEW