Trust in risk information about food related-hazards may be an important determinant of public reactions to risk information. One of the central questions addressed by the risk communication literature is why some individuals and organizations are trusted as sources of risk information and others are not. Industry and government often lack public trust, whereas other sources (for example, consumer organizations, the quality media, medical doctors) are highly trusted. Problematically, previous surveys and questionnaire studies have utilized questions generated by the investigators themselves to assess public perceptions of trust in different sources. Furthermore, no account of the hazard domain was made. In the first study reported here, semistructured interviewing was used to elicit underpinning constructs determining trust and distrust in different sources providing food-related risk information (n= 35). In the second study, the repertory grid method was used to elicit the terminology that respondents use to distinguish between different potential food-related information sources (n= 35), the data being submitted to generalised Procrustes analysis. The results of the two studies were combined and validated in survey research (n= 888) where factor analysis indicated that knowledge in itself does not lead to trust, but that trusted sources are seen to be characterised by multiple positive attributes. Contrary to previous research, complete freedom does not lead to trust—rather sources which possess moderate accountability are seen to be the most trusted.