Are surveys on trust trustworthy?

In this paper we explore the validity of a survey question commonly used for measuring generalized trust. Trust has become a key variable in a variety of fields, including sociology, social psychology, political science, and economics; therefore the accuracy with which it is measured has profound implications for many studies. We suggest that ambiguous wording on this survey item has led to misinterpretations concerning actual trust levels, especially in a cross-cultural context. To test this claim, we conduct an extensive survey of students at UCLA and at Hokkaido University, Japan. Results strongly suggest that the survey question measures differences in caution levels rather than in trust. Implications of this research are discussed.