A critical assumption of LCA is conditional independence and violations can result in extraction of more classes than are required. To address this concern, we examined a factor mixture model where a single factor was nested within each latent class, relaxing the conditional independence assumption. The fit of this model was compared to the LCA. As the objective of this study was to classify individuals (i.e. person-centered) rather than symptoms (i.e. item-centered), we did not conduct factor analyses in the absence of LCA.
