Measures of interpersonal trust

Interpersonal trust (i.e. trust in other people) is an issue of high interest to both policy-makers and researchers seeking to understand what drives social and economic outcomes. However, for trust to usefully inform policy and analysis it is necessary to have valid and reliable measures of it. Despite a large body of evidence on the relationship between trust and other social and economic outcomes, evidence on the validity of trust from experimental data is conflicting. In particular, while many studies find no correlation between survey measures of trust and experimental measures at an individual level, other studies suggest a significant, if modest, correlation at the country level. This article examines the relationship between survey and experimental measures of trust in others using a large dataset containing aggregate experimental and survey measures of trust from 167 studies conducted in 36 countries. Importantly, the dataset also includes individual measures of both survey and behavioural trust in seven countries, and data from two panel studies with repeated survey measures of trust. Using these multiple data sources, the paper investigates the degree to which survey measures of interpersonal trust are valid at both an individual and cross-country level. The paper shows the existence of a significant correlation between survey and experimental measures of interpersonal trust at the country-level. Evidence on measurement errors in existing small-scale studies underscores the importance of developing better quality data from both surveys and experiments.