poLCA
Polytomous Variable Latent Class Analysis
 
Drew A. Linzer      Jeffrey Lewis
Department of Political Science      Department of Political Science
Emory University      University of California, Los Angeles

OVERVIEW

poLCA is a software package for the estimation of latent class models and latent class regression models for polytomous outcome variables, implemented in the R statistical computing environment.

Latent class analysis (also known as latent structure analysis) can be used for identifying groups of observations in multivariate categorical data, estimating the characteristics of these groups, and predicting the probability that each observation belongs to each group. These models are also helpful in investigating sources of confounding and nonindependence among categorical variables. Typical applications include the analysis of opinion surveys; rater agreement; lifestyle and consumer choice; and other social and behavioral phenomena.

The basic latent class model is a finite mixture model in which the component distributions are assumed to be multi-way cross-classification tables with all variables mutually independent. The model stratifies the observed data by a theoretical latent categorical variable, attempting to eliminate any spurious relationships between the observed variables. The latent class regression model makes it possible for the researcher to further estimate the effects of covariates (or "concomitant" variables) on predicting latent class membership.

poLCA uses expectation-maximization and Newton-Raphson algorithms to find maximum likelihood estimates of the parameters of the latent class and latent class regression models.

DOCUMENTATION

Download user's manual (PDF). The package is also documented internally upon installation.

ACCESS

To install the package directly through R, select Packages > Install package(s)... and choose a local CRAN mirror. Then select poLCA from the list of packages and click OK.

To install poLCA from a local zip file, download poLCA version 1.1, and place the downloaded poLCA_1.1.zip file in your R library subdirectory. In R, select Packages > Install package(s) from local zip files... and navigate to the zip file. Click Open to begin the installation.

Once the installation is complete, type library(poLCA) in R to load the package into memory.

Users of poLCA are requested to cite the software package as:
Linzer, Drew A. and Jeffrey Lewis. 2007. "poLCA: Polytomous Variable Latent Class Analysis." R package version 1.1. http://userwww.service.emory.edu/~dlinzer/poLCA.
and
Linzer, Drew A. and Jeffrey Lewis. Forthcoming. "poLCA: Polytomous Variable Latent Class Analysis." Journal of Statistical Software.
poLCA is provided free of charge, subject to version 2 of the GPL or any later version.

CONTACT

Please direct all inquiries, comments, and reports of bugs to dlinzer@emory.edu.

VERSION HISTORY

1.1: Adds additional user control over ordering of latent classes and printing of model results. New functionality to automatically estimate the latent class model multiple times to locate the global maximum likelihood solution. Maintains package compatibility with R version 2.6.0 patched. (November 1, 2007)

1.0: Provides standard errors for all model parameters, and covariance matrix for regression model coefficients. Also allows users to specify the starting parameters for the estimation algorithm, to aid in convergence and increase control over model output. (April 4, 2007)

0.9: First public release. (June 1, 2006)