Cluster Analysis of Social Capital in Low Wealth Neighborhood

Background: In 2014, 46.7 million Americans lived in poverty. The zip code where Americans live impacts lifelong health and is considered a stronger indicator of life expectancy than genetic codes. Social capital is an important strategy to addressing neighborhood-level poverty. Though considerable research on neighborhood-level social capital exists, little is known about social capital variations among individuals within low-wealth neighborhoods. Research Question: What are the various typologies of social capital within a low-wealth community?

Methods: This study was part of a HUD Choice Neighborhood initiative. A community-engaged approach involved collaboration with residents in the design of the project, including instrument development. Measures included adapted versions of the Harvard Social Capital Community Benchmark Survey, and the National Survey of Black Americans.

Data were collected through face-to-face household-level surveys. Residents could also participate via telephone or mail. The subset of respondents completing the social capital module were included in the analysis (n=55; 76% female, 94% Black, 60% single, separated or divorced, and 76% high-school graduates).

To identify a typology of individual social capital, Ward’s method of hierarchical cluster analysis was employed using squared Euclidian distances to identify initial clusters.  The four summary social capital variables included measures of trust, civic engagement, social ties, and reciprocal relationships. The dendogram and means of each summary variable for each cluster were examined. ANOVAs determined if the variable means differed across clusters. This and the number of residents in each cluster were used to identify the optimal number of clusters. 

Results: A four cluster solution was determined to be optimal. It demonstrated significant differences in the means of three of the social capital variables. Reciprocal relationships were not significantly different in any of the cluster solutions. Four patterns of individual social capital were evident: 1) High Social Capital group had high levels of civic engagement, social ties and trust; 2) Civic-Social Dominant group had very high civic engagement and social ties, and low levels of trust; 3) Social Tie Dominant group had low levels of civic engagement and very low levels of trust; and 4) Low Social Capital group had low levels of all remaining dimensions.

Implications: Understanding patterns of social capital can help social workers plan strategies to enhance engagement in low-wealth communities. The High Social Capital group would be likely to engage first in collaborative revitalization efforts, as the only group with high levels of trust. The Civic-Social Dominant group would also likely engage but need to be approached carefully by outsiders due to low levels of trust. The Social Tie Dominant group could be reached through personal relationships but low trust levels will require targeted communication. Given their social tie resources, they would be useful in increasing numbers of participants. The Low Social Capital group would likely be most difficult to engage because of exceptionally low levels of trust and apparent isolation. Future research should identify factors that predict which type of social capital patterns individuals exhibit which could help further develop strategies to enhance individual social capital for members of all groups.