Identifying Latent Profiles of Neighborhood Contexts: Evidence from the National Longitudinal Study of Adolescent and Adult Health

Background and Purpose: Neighborhood contexts play an important role in the adolescent development. Prior research largely focuses on demographic and economic compositions of neighborhood and investigates them as separate indicators. Emerging studies suggest that neighborhood characteristics are interrelated. To capture the multidimensional characteristics of neighborhood contexts, this study aims to identify (1) unobserved latent classes of neighborhood contexts and (2) individual characteristics associated with the classes.

 Methods: The sample consisted of 9,191 adolescents who participated in the National Longitudinal Study of Adolescent and Adult Health. First, we conducted a Latent Profile Analysis (LPA) to identify the latent structure of neighborhood contexts. We used 13 observed variables measured during adolescence (Wave I), guided by previous research. Each variable indicated a proportion of the corresponding characteristic measured at the census-tract level: unemployment, public assistance use, no high school diploma, below the poverty line, female-headed households, foreign born, limited English proficiency, owned housing, residential mobility, housing quality, Blacks, Hispanics, and urbanity. Second, we conducted a multinomial logistic regression analysis to compare the identified classes by individual sociodemographic backgrounds.  

Results: Based on the evaluation of the fit indices and substantive criteria, the LPA identified three heterogeneous classes of neighborhood contexts. The first class was estimated to have high proportions of the unemployed, public assistance use, the income-poor, female-headed households, and Blacks. The second class demonstrated high proportions of foreign-born residents, limited English proficiency, residential mobility, Hispanics, and urbanity. The third class presented low percentages in most neighborhood disadvantages and a high percentage in owned housing. Given these features, the first class were named the low socioeconomic (SES) and segregated community (18.63%), the second class as the urban diverse community (4.7%), and the third class as the high socioeconomic (SES) community (76.6%).

Multinomial logistic analysis found significant associations between the neighborhood profiles and several characteristics of adolescents. Compared to Whites, Blacks (β=2.46 p<.001) and Native Americans (β=.95 p<.001) were more likely to reside in the low SES community (vs. high SES). Similarly, Hispanic (β=3.86 p<.001), African American (β=1.91 p<.001), and Asian (β=1.34 p<.001) were more likely to grow in the urban diverse community in adolescence (vs. high SES). Immigrants were less likely to live in the low SES community (vs. high SES), while they were more likely to live in the urban-diverse community in adolescence (vs. high SES). Moreover, adolescents grown up in the low SES community showed significantly lower levels of educational attainment and household income in young adulthood compared to those in high SES community. Interestingly, adolescents in the low SES neighborhood (β=0.32 p<.009) and the urban diverse neighborhood (β=.40 p <.009) presented higher likelihoods of experiencing community violence than those in the high SES neighborhood.

Discussion and Implications: This study contributes to better understandings of neighborhood contexts by expanding traditional approaches limited to race, economic status, and geography and their separate effects. Our findings highlight the importance of addressing multifaceted characteristics of communities. We will discuss research and practice implications for healthy development of children and youth.