Key points are not available for this paper at this time.
Undergraduates in Tanzania and Sweden participated in one Trust game, one Dictator game, and answered a standard set of survey questions relating to trust. In both countries we detected a strong and significant relation between Dictator donations and proportions returned in Trust games, indicating that unconditional distribution preferences matter in Trust game behaviour. In Sweden incentive compatible evaluations of unrevealed Dictator donations significantly predicted the amount sent in Trust games. The predictive power of survey trust questions differed between countries: a plausible relation between survey trust and trust behaviour was found in Sweden but not in Tanzania. Copyright 2005 Royal Economic Society.
Holm et al. (Fri,) studied this question.