This article explores how generalized trust increases or decreases over time as a result of students’ individual experiences of informal school institutions. However, since experiences in local settings are likely to have varying effects on trust in different groups of strangers, we disentangle the concept of generalized trust and separate between different group-trust items. In the article, we show that three of these measure in-group generalized trust while one taps out-group generalized trust. Drawing on original data, collected in seven public schools in Palermo, Italy, between 2002 and 2005, we conclude that experiences of the school climate are transferred to both in-group and out-group generalized trust. However, our second main conclusion is that involvement in associations outside of school tends to generate distrust in people who resemble the respondents’ in-group. By distinguishing between in-group and out-group generalized trust, we thus show that school experiences generate trust in unknown people in general, whereas other localized experiences are exclusively transferred to trust, or distrust, in familiar groups of people. We conclude by arguing these findings are of potentially great importance to trust research since they suggest that generalized trust is less stable and more multi-faceted than previously thought. Moreover, since Palermo is a critical case, we argue that our results are encouraging. If schools can generate trust here, then schools are likely to have a similar effect elsewhere, in more favourable settings.