tseLCA
Bias-adjusted three-step estimation of latent class models with covariates and distal outcomes. The latent class measurement model is estimated first, with 'multilevLCA' (Lyrvall et al., 2025) <doi:10.1080/00273171.2025.2473935>, and held fixed; observations are then classified; and the classes are related to covariates and distal outcomes with the maximum likelihood correction of Vermunt (2010) <doi:10.1093/pan/mpq025> and Bakk, Tekle and Vermunt (2013) <doi:10.1177/0081175012470644>, or the correction of Bolck, Croon and Hagenaars (2004) <doi:10.1093/pan/mph001>. Standard errors account for the uncertainty of the measurement model (Bakk, Oberski and Vermunt, 2014) <doi:10.1093/pan/mpu003>. Includes class enumeration, modal and proportional class assignment, covariate formulas, Gaussian, Poisson, binomial, and multinomial distal outcomes, the two-step estimator of Bakk and Kuha (2018) <doi:10.1007/s11336-017-9592-7>, measurement models applied to new samples, and full-information maximum likelihood for missing indicators, standard methods for fitted models, and a data-generating process replicating the simulation design of Bakk and Kuha (2018).
- Estimating Latent Structure Models with Categorical Variables: One-Step Versus Three-Step Estimators
- Estimating the Association between Latent Class Membership and External Variables Using Bias-adjusted Three-step Approaches
- Latent Class Modeling with Covariates: Two Improved Three-Step Approaches
- Multilevel Latent Class Analysis: State-of-the-Art Methodologies and Their Implementation in the R Package multilevLCA
- Relating Latent Class Assignments to External Variables: Standard Errors for Correct Inference
- Two-step estimation of models between latent classes and external variables
This page was built for software: tseLCA
