An investigation of social capital in Britain using small area estimation analysis

Social Capital is considered an important asset for development, both at local and higher levels, and has been explored across the social sciences for decades. Attempts to define and measure it in increasingly precise ways continue in order to place it at the centre of policymaking. Indeed, it is considered a precious capital in times where resources are becoming scarce. This thesis investigates the dimensions of social capital and estimates them at small area level for England and Wales in 2011. The first step is the identification of three factors measuring membership, citizenship and politics and neighbourliness dimensions using survey data and a Confirmatory Factor Analysis. The second step is to test the hypothesis that other individual characteristics and geographical characteristics may influence levels of these factors. Complex Multilevel models with individual covariates and area-level covariates from the Census and administrative sources confirm the hypothesis: the factors depend on age, gender, ethnicity, religion, marital status, socio-economic class, employment, state of health and education at individual level and ethnic diversity and economic profile at area level. Lastly, Multilevel Model results have been used as a starting point for the final synthetic estimates at small area for all the Middle Super Output Areas of the average levels of the three factors. While Membership and Citizenship and Politics social capital show higher differentials, Neighbourliness seems to be more spread and, on average, higher than the other two factors across the two countries.