The social capital literature attributes association members a key role in propagating inter-personal trust in the society. It has been posited that participating in associations instils pro-social attitudes in their members, and that the decline in associational membership is the main cause of the faltering rates of inter-personal trust in the US. However, the extent to which association members are indeed inclined to extend their presumed higher pro-social attitudes from within associations to the society at large is still an open question. The survey evidence on the issue is scant, and recent laboratory evidence with minimal groups assigns no value to groups as such. We investigate these issues in the first field experiment measuring trust and trustworthiness of members of real-life associations and of a demographically comparable sample of non-members. The sample was stratified with respect to age, gender, educational level. Members have played an anonymous Trust Game either with fellow members (ingroup treatment), or with people from the general population (outgroup treatment). Our main findings are: (a) Association members are significantly more trusting and trustworthy than non-members when interacting with people from the general population. (b) Association members generally trust and reward fellow members in the same way as they behave with people from the general population. Ingroup favouritism is limited to two associations out of ten with respect to trust, and to one case with respect to trustworthiness. (c) However, people who are members of only “Olsonian” associations do not trust others significantly more than people from the general population, while members of “Putnamesque” and social welfare associations trust significantly more. This confirms the conjecture that associations may have different macroeconomic impacts depending on the nature of their goals. (d) We find no evidence that growing individual involvement with associational life, measured by length of membership, hours spent volunteering, and number of associations joined, has any effect in increasing pro-social attitudes.
