The two-point mixture index of fit enjoys some desirable features in model fit assessment and model selection, however, a need exists for efficient computational strategies. Applying an NLP algorithm, a program using the SAS matrix language is presented to estimate the two-point index of fit for two-class LCA models with dichotomous response variables. The program offers a tool to compute π ∗ for twoclass models and it also provides an alternative program for conducting latent class analysis with SAS. This study builds a foundation for further research on computational approaches for M-class models.
Zhang, Dongquan and Dayton, C. Mitchell
"JMASM30 PI-LCA: A SAS Program Computing the Two-point Mixture Index of Fit for Two-class LCA Models with Dichotomous Variables (SAS),"
Journal of Modern Applied Statistical Methods: Vol. 9
, Article 32.
Available at: http://digitalcommons.wayne.edu/jmasm/vol9/iss1/32