Comparing two approaches to examine the temporal relationship between psychological need satisfaction and conspiracy beliefs: cross-lagged panel models versus random intercept cross-lagged panel models

Cross-lagged panel models (CLPM) are often used to examine longitudinal data, yet they may produce biased estimates when studying within-person processes because they confound stable between-person differences and within-person change. Random intercept cross-lagged panel models (RI-CLPMs) address this limitation by separating within- and between-person variance. Using data from over 55,000 adults in New Zealand, we demonstrate that CLPMs and RI-CLPMs yield different conclusions when examining longitudinal associations between psychological need satisfaction and conspiracy beliefs. Although CLPMs suggest strong cross-lagged effects, RI-CLPMs reveal substantially weaker associations, indicating that much of the observed effect is driven by stable between-person differences rather than within-person change. These results highlight the need to align model choice with one’s research question when drawing conclusions about psychological processes over time. Failing to account for between-person stability may overstate temporal associations in psychological research and potentially misinform policy interventions.