What are you assessing when you measure "trust" in scientists with a direct measure?

This article analyzes three publicly available datasets focused on trust in science and scientists. It specifically seeks to understand what direct measures of trust (i.e. questions that directly ask respondents how much they trust scientists) assess in terms of discrete measures of trustworthiness (i.e. perceptions of scientists' ability, integrity, and benevolence). Underlying the analyses is a concern that direct measures of trust are a poor substitute for differentiating between discrete trustworthiness perceptions and behavioral trust in the form of a specific willingness to make oneself vulnerable. The research concludes that it is unclear what direct trust measures are capturing in any given context and suggests that researchers should better use trust-related theory when designing surveys and trust-focused campaigns. The secondary data used come from the General Social Survey, Gallup, and the Pew Research Center.