Spurious effects in random-intercept cross-lagged panel models: Results from simulations and reanalyses of data on self-esteem and problematic eating behaviors used by Beckers et al. (2023).

Abstract

The random-intercept cross-lagged panel model (RI-CLPM) is an extension of the traditional cross-lagged panel model. The RI-CPLM specifically addresses prospective effects within individuals. In the present simulations, we found that the RI-CLPM is susceptible to spurious findings when observed scores on the two variables are affected by common auto-correlated state factors. In reanalyses of empirical data, we found contradictory decreasing, increasing, and null prospective effects between within-individual levels of self-esteem and problematic eating behaviors among Dutch teenagers (N = 1856). These contradictory findings indicated that previously reported prospective effects may have been statistical artifacts. Caution is advised when using the RI-CLPM, as it may produce misleading results. Researchers are recommended to validate findings using alternative analyses, e.g., by examining person-mean centered scores.