The document presents the Stata command pcm designed for Partial Credit Models (PCM) within Rasch Measurement Theory (RMT), aimed at analyzing questionnaire data with polytomous items. The paper details the pcm command's use in assessing health literacy determinants among homeless shelter residents, particularly examining the "Social Support for Health" dimension from the Health Literacy Questionnaire (HLQ).

The pcm command addresses RMT limitations in existing Stata tools, offering quick parameter estimation, automatic generation of fit indices, and graphical outputs to simplify analysis. Key features include the estimation of latent trait levels, fit assessment using INFIT and OUTFIT indices, and visualization through Item and Test Characteristic Curves (ICCs and TCCs). The command also supports latent regression, enabling analysis of external predictors' impact on the latent trait.

An empirical illustration using data from the 2021 ECHO study demonstrates the pcm command's capabilities. It highlights the role of health perception, language proficiency, and social factors in health literacy among the homeless, with results indicating good model fit and reliability (PSI 0.78). Variables such as French language skills, sex, and perceived health status were significant predictors of health literacy levels.