(Converted as ePoster, See Poster Gallery) Assessing the Modified Neighborhood Cohesion Index Among African American Adolescents Living in Public Housing: An Exploratory and Confirmatory Factor Analysis

Background and Purpose: African American youth are disproportionately overrepresented in segregated, low-resourced neighborhoods. Consequently, they experience greater adverse community exposures (e.g., violence, deviant peer groups, and urban hassles). Neighborhood cohesion - the trusting social relationships, shared values, and norms among neighborhood residents - is an important protective factor for youth in such environments. However, the concept remains underdeveloped among African American youth. Extensive research indicates that Black youth may interpret scale items differently compared to youth of other racial and ethnic groups. This study addressed this gap in knowledge by assessing the performance of the modified Buckner (1988) Neighborhood Cohesion Index (NCI) in a sample of African American youth living in urban public housing. The study advances two questions: 1) What are the psychometric properties of the Buckner (1988) Neighborhood Cohesion Scale among urban African American youth; 2) What is the model fit using the NCI with this understudied population?

Methods: The sample for this study consists of 235 African American youth aged 11-20 residing in public housing communities in two large Northeastern cities. Youth participants were recruited using flyers, recruitment cards, and announcements at local community centers located near their public housing development. All potential participants met in a community center and were screened for their capacity to give informed consent. Youth who demonstrated the capacity to give informed consent, and who provided parental consent and youth assent, where needed, were invited to complete the study survey. A two-step approach was used to address our research question. First, an exploratory factor analysis (EFA) was conducted to explore whether the observed variables were adequate indicators of the latent variable as the scale was modified from its original version. Second, a confirmatory factor analysis (CFA) was conducted to test how well the observed variables represent the latent construct. We obtain the best model fit using the WLSMV estimator. Analyses were performed using Mplus v 8.3.

Results: Results of the EFA demonstrate that a one factor structure was appropriate for the NCI, with two items removed from the modified NCI due to loading below the .3 threshold. The 13-item scale had an internal consistency of .821. Results of the CFA recommended three error covariances. Model fit statistics found that even with the error covariances fit indices for chi-square and REMSEA were not acceptable ([X2[62, N = 235] = 188.632, P<.0; RMSEA = 0.093 [90% CI: 0.078, 0.109]), while model fit indices were acceptable for CFI (.90) and SRMR (.06).

Conclusion and Implications: The modified NCI is not the best measure for assessing neighborhood cohesion in this understudied population. In working toward racial and social justice in the development of psychometric measures for social science, future research should prioritize psychometric testing coupled with cognitive interviewing to provide a contextualized measure of neighborhood cohesion among African American youth living in public housing.