The paper introduces poLCA, an R package for estimating latent class and latent class regression models for polytomous outcome variables. The basic latent class model is a finite mixture model where the component distributions are multi-way cross-classification tables with all variables mutually independent. The latent class regression model extends this by allowing covariates to predict latent class membership. poLCA uses the expectation-maximization (EM) and Newton-Raphson algorithms to estimate parameters and provides tools for model selection and goodness-of-fit assessment. The package includes functions for simulating data and visualizing model results. Two examples are provided to illustrate the use of poLCA: a basic latent class model on carcinoma data and a latent class regression model. The paper also discusses the importance of reordering latent classes and avoiding local maxima in the EM algorithm.The paper introduces poLCA, an R package for estimating latent class and latent class regression models for polytomous outcome variables. The basic latent class model is a finite mixture model where the component distributions are multi-way cross-classification tables with all variables mutually independent. The latent class regression model extends this by allowing covariates to predict latent class membership. poLCA uses the expectation-maximization (EM) and Newton-Raphson algorithms to estimate parameters and provides tools for model selection and goodness-of-fit assessment. The package includes functions for simulating data and visualizing model results. Two examples are provided to illustrate the use of poLCA: a basic latent class model on carcinoma data and a latent class regression model. The paper also discusses the importance of reordering latent classes and avoiding local maxima in the EM algorithm.