Predicting Suggestive Questioning from Cognitions and Emotions about Child Sexual Abuse across Three Study Paradigms

Objective: Although interviews and conversations are highly individual in nature, and suggestiveness is a major pitfall when questioning children, differential aspects of interviewer bias and suggestiveness remain understudied. We assessed the influence of Cognitions and Emotions about Child Sexual Abuse (CECSA) on suggestive questioning and a biased mindset toward the abuse hypothesis across a series of three studies with varying mock paradigms and a meta-analytical integration. Hypotheses: For all studies, we expect the scores on the three CECSA scales (Naive Confidence, Emotional Reactivity, and Justice System Distrust) to be associated with the number of suggestive questions and indicators of a biased mindset. Method: In all studies, participants read mock cases about children displaying mild behavioral symptoms that were unspecific but gave rise to suspecting sexual abuse. In Study 1, 285 human sciences students further read interview transcripts and selected questions suitable to pose to the child. In Study 2, 241 police students read interview transcripts and freely formulated questions to pose to the children. In Study 3, 148 teaching students interviewed virtual children using natural language in a Virtual Reality setting. Results: Across three studies and their meta-analytical integration, we found robust evidence that Naive Confidence and Emotional Reactivity, but not Justice System Distrust, significantly predict bias and suggestive questioning, with effect sizes of b = .14–.37. The newly developed measures to assess suggestive questioning validly captured a unidimensional trait of suggestive questioning but showed unsatisfying reliability. Conclusions: The findings enhance our understanding of individual differences in suggestive questioning and bias and can inform the development, customization, and evaluation of interviewer training programs, as well as the selection of interview personnel. We also provide recommendations to increase the reliability of the suggestiveness measures in future research.