This paper tests the hypothesis that generalized trust helps solve large-n collective action problems in a rationalistic framework. A rigors test is employed in that the hypothesis is tested in two different choice situations; a pure public good situation (whether to recycle or not) and a joint-product situation (whether to buy organic food or not). This design provides a possibility to detect whether any effect of generalized trust is due to strategic behavior or due to altruistic motives among trusters. The results show that generalized trust enhances recycling significantly whereas it only has a very week positive effect on organic food consumption. Hence, generalized trust only affects public good provision when a collective action problem is present. The analyses thus sustain the hypothesis. It is furthermore concluded that the effect of generalized trust is due to strategic behavior and not altruistic motives.