Abstract
This article analyses the validity of the standard question, ‘Generally speaking, would you say that most people can be trusted, or that you can’t be too careful in dealing with people?’. It builds on unique material, including survey responses (N = 356) and letters (N = 28) from students aged 18–19 years, collected by the author in seven secondary schools in Stockholm. The results suggest that the standard question is less valid as a measure on generalized trust than is commonly assumed in trust research. Rather than referring to trust in people in general, about a third of the responses expressed themes related to rational trust, that is, trust is domain-specific and person-specific. Systematic variation between different school contexts indicates that the standard question reflects social trust in a limited social sphere rather than trust in ‘most people’. The conclusion is that the standard trust question has a problem with validity among young respondents.
