Existing empirical research into the causes of interpersonal trust identifies various forms of what can be loosely termed ‘social connectedness’ as important causes of generalized trust. However, the majority of existing investigations are limited by their reliance on cross-sectional data. In this paper, high-quality repeated measures data from the British Household Panel Survey are used to examine the effect of informal social networks and formal associational membership on trust over a ten year period (1998-2008). A fixed effects specification is used to control for the time-invariant characteristics of individuals that might spuriously link the explanatory variables to subsequent changes in trust. The results of the fixed effects model are contrasted with cross-sectional and random effects specifications, which do not control for unobserved individual heterogeneity. While the cross-sectional and random effects models show substantial effects of associational membership and a range of life events on subsequent levels of trust, the fixed effects specification show weak but nonetheless significant effects.
