Assessing complex belief structures with the triads task: Reliability and validity

Research on attitudes and beliefs often relies on assessing people's agreement with several explicit statements. We promote a different approach, focusing on the interconnectedness of beliefs and their cognitive and affective components. Here, we present the triads task as a procedure to measure complex belief networks. Leaning on similarity and difference judgments for three concepts at a time, the triads task allows us to derive similarity scores for each pair of concepts in a topic domain, which provide insights into belief networks of individuals and groups. Furthermore, we propose a procedure to translate similarity scores into visualizations of belief networks, inspired by cognitive-affective maps. For a sample of 56 participants, we show that similarity scores are test-retest reliable in a 11-to-17-day interval (averaged individual retest correlation: *r* = .78) and that participants communicatively validate our visualization procedure. We highlight the benefits of the triads task in measuring belief networks and further discuss this procedure with regards to its potential as a research and intervention tool.