This simulation study examined the performance of ridge and lasso regularization within a Bifactor-(S-1)-Model by varying sample size, the amount of non-informative covariates, the overall R² of the regularized regression as well as the amount of multicollinearity among predictors. For replicating the results obtained with the lslx-package (Huang et al. 2020), use the supplied materials for data generation, model estimation and model analysis. For replicating the result obtained by the StructuralEquationsModel.jl-package (Ernst & Peikert, 2022) please refer to the following repository:
https://github.com/aaronpeikert/ridge-sem-convergence