A Simulation-Based Power Analysis for the Discordant-Sibling Random Intercept Cross-Lagged Panel Model in Family-Based Longitudinal Research

Family-based longitudinal studies are increasingly used to investigate psychological and behavioral processes, and it is imperative to understand the power to detect effects in increasingly complex statistical models. The Discordant-Sibling Random Intercept Cross-Lagged Panel Model (DS RI-CLPM) extends the Random Intercept Cross-Lagged Panel Model to estimate longitudinal bidirectional associations by separating within- and between family-level components to control stable traits and sibling-shared family-level confounding. To inform study design in family-based research, we conducted a simulation-based power analysis varying the number of families and waves, sibling discordance, random intercept variance proportion, effect sizes, and variable’s non-normality. Power increased with both increase in families and waves, although effect size was the strongest determinant. Within-family cross-lagged effects were easier to detect than between-family effects, and joint detection required most conservative designs. Practical recommendations for numbers of families and waves are provided for longitudinal research on psychological and behavioral phenomena using the DS RI-CLPM.