The article introduces PROC LCA, a new SAS procedure for latent class analysis (LCA), multiple-group LCA, and LCA with covariates. LCA is a statistical method used to identify discrete, mutually exclusive latent classes of individuals based on their responses to observed categorical variables. In multiple-group LCA, both the measurement and structural parts of the model can vary across groups, and measurement invariance can be tested. LCA with covariates extends the model to include predictors of class membership. The procedure is demonstrated using data on alcohol use behavior in a national sample of high school seniors. LCA is useful in the social and behavioral sciences for identifying latent variables, which are inferred from multiple observed items. Covariance structure analysis provides a framework for mapping items onto continuous latent variables, while LCA provides an analogous framework for categorical latent variables. In traditional LCA, two sets of parameters are estimated: class membership probabilities and item-response probabilities conditional on class membership. The latent class model estimates and removes measurement error from the vector of latent-class membership probabilities. Latent class models usually involve categorical indicators, although latent profile analysis is also used. When categorical data are used, the latent class model has the advantage of making no assumptions about the distributions of the indicators other than local independence. The latent class model has been applied in many domains, including psychology, education, and sociology. It has been used to assess temperament, depression, teaching style, poverty, and substance use behavior. Two particularly useful extensions of LCA are multiple-group LCA and LCA with covariates. In multiple-group LCA, both class membership and item-response probabilities can vary across groups, and measurement invariance can be tested. LCA with covariates extends the model to include predictors of class membership. Software for LCA includes Latent GOLD, PanMark, WinLTA, and Mplus. This article introduces PROC LCA, a new SAS procedure for latent class analysis developed for SAS Version 9.1 for Windows. The software is available for download free of charge at http://methodology.psu.edu. The article reviews features of the software and illustrates its use with a series of empirical analyses. The procedure estimates parameters such as gamma (latent class membership probabilities) and rho (item-response probabilities conditional on latent class membership). When a grouping variable is included, both sets of parameters can be conditioned on group. When covariates are included, an additional set of parameters (beta) is estimated, predicting class membership through a logistic link. The procedure uses maximum likelihood estimation with an EM algorithm to handle missing data. The article discusses model specification, including data preparation, exploratory data analysis, and basic model specification. It covers model assessment, including the use of likelihood-ratio G² statistic, AIC, and BIC. Expanded model specification includes multiple-group LCA and LCA with covariates. The article also discusses estimation options, optional output, andThe article introduces PROC LCA, a new SAS procedure for latent class analysis (LCA), multiple-group LCA, and LCA with covariates. LCA is a statistical method used to identify discrete, mutually exclusive latent classes of individuals based on their responses to observed categorical variables. In multiple-group LCA, both the measurement and structural parts of the model can vary across groups, and measurement invariance can be tested. LCA with covariates extends the model to include predictors of class membership. The procedure is demonstrated using data on alcohol use behavior in a national sample of high school seniors. LCA is useful in the social and behavioral sciences for identifying latent variables, which are inferred from multiple observed items. Covariance structure analysis provides a framework for mapping items onto continuous latent variables, while LCA provides an analogous framework for categorical latent variables. In traditional LCA, two sets of parameters are estimated: class membership probabilities and item-response probabilities conditional on class membership. The latent class model estimates and removes measurement error from the vector of latent-class membership probabilities. Latent class models usually involve categorical indicators, although latent profile analysis is also used. When categorical data are used, the latent class model has the advantage of making no assumptions about the distributions of the indicators other than local independence. The latent class model has been applied in many domains, including psychology, education, and sociology. It has been used to assess temperament, depression, teaching style, poverty, and substance use behavior. Two particularly useful extensions of LCA are multiple-group LCA and LCA with covariates. In multiple-group LCA, both class membership and item-response probabilities can vary across groups, and measurement invariance can be tested. LCA with covariates extends the model to include predictors of class membership. Software for LCA includes Latent GOLD, PanMark, WinLTA, and Mplus. This article introduces PROC LCA, a new SAS procedure for latent class analysis developed for SAS Version 9.1 for Windows. The software is available for download free of charge at http://methodology.psu.edu. The article reviews features of the software and illustrates its use with a series of empirical analyses. The procedure estimates parameters such as gamma (latent class membership probabilities) and rho (item-response probabilities conditional on latent class membership). When a grouping variable is included, both sets of parameters can be conditioned on group. When covariates are included, an additional set of parameters (beta) is estimated, predicting class membership through a logistic link. The procedure uses maximum likelihood estimation with an EM algorithm to handle missing data. The article discusses model specification, including data preparation, exploratory data analysis, and basic model specification. It covers model assessment, including the use of likelihood-ratio G² statistic, AIC, and BIC. Expanded model specification includes multiple-group LCA and LCA with covariates. The article also discusses estimation options, optional output, and