How to promote more nuanced construct theories: An interdisciplinary perspective with pragmatic recommendations

Psychology's reliance on statistical model fit often misinterprets statistical coherence asconstruct validity. Constructs may exhibit strong model fit but lack theoretical clarity. Thispaper argues that model fit should be one of several criteria for construct validity. We presentan interdisciplinary framework that combines psychometric and philosophical insights andencourages approaches beyond structural analyses. By emphasizing predictive validity,testing (in)variance across constructs, and examining distributional properties, we proposemethods that yield theoretically robust constructs. Our aim is to move from statistical fit to anuanced understanding of construct validity, enabling models that better represent realpsychological phenomena.