This study explores the use of the GORICA, a model selection method based on inequality constraints, in cross-lagged panel models (CLPM and RI-CLPM). The findings show that the GORICA can accurately identify the true model, and its performance improves as the sample size, number of measurement occasions, and differences between parameters increase. Compared to traditional methods, including the chi-square difference test, chi-bar-square difference test, and AIC, the GORICA generally performs better. It is especially effective in determining the number of random intercepts in RI-CLPM, which is often challenging in practice. The method is also applied to real data on maternal postpartum depression in Thailand, demonstrating its practical value for evaluating informative hypotheses. To support researchers, this study introduces the powerGORICA R package. This tool allows users to evaluate statistical power under different study conditions and to determine suitable sample sizes for their research.