The trustworthiness of trust measurement. Predicting trusting behaviour in trust games by survey-measured trust

Abstract

Trust has found to be of great influence on social and economic transactions of everyday life (Fehr, 2009). Although the topic has caught the attention of scholars within many fields, the question how trust can be measured has not been answered properly (Nannestad, 2008). Previous studies have shown that the main measurement methods of trust – the survey based generalized trust question and the experimental trust game – do not correlate (Ermisch et al., 2009; Fehr et al., 2003; Glaeser et al., 2000). Data and methods on which these results are based can be fairly criticized. I hypothesise that survey-based trust can predict trusting behaviour in trust games. In this study, generalized trust is perceived by the behaviouristic definition. I argue that measurement on generalized trust can be improved by using a homogeneous sample, by using a random and anonymous matching procedure in the experiment and by creating a survey-measure, formed by a multiple-item trust scale. The data consists of experiments from Switzerland (2010 and 2014), the United Kingdom (2011) and Spain (2012), including both trust game and survey. I conclude that survey-measured trust, when used in a multiple-item scale, successfully predicts trusting behaviour in trust games.