Points of contention in measure evaluation can arise from the use of divergent validity frameworks: A reply to Conway et al. (2025)

This reply to Conway et al. (2025) illustrates how points of contention in the evaluation of mindreading (or Theory of Mind) measures can arise from the use of divergent validity concepts. The construct validity model used in Wendt and colleagues’ (2024) empirical study contrasts with the perspective implicit in Conway and colleagues' commentary, which is more consistent with Lennon's content validity model. This is reflected in the authors' conception of the nature of what is to be measured (i.e., the measurand), the criterion for what makes a measure superior (i.e., validity), and the proposed methods for judging this (i.e., validation). The mismatch between the validity concepts adopted by the respective authors has three major implications: First, Conway and colleagues' critique does not fully address the specific goals, assumptions, and intricacies of construct validation methodology. Second, their approach to measuring mindreading should not be confused with, or considered as an alternative to, construct validation, but is valuable in its own right. Third, the two validity frameworks mentioned offer unique opportunities for different phases of the research process. While a content validity approach can be valuable for describing an empirical phenomenon that seems worthy of explanation (e.g., real-world mindreading), a construct validity approach helps identify the theoretical constructs that might help explain it (e.g., mindreading ability).