Effect of single-value response styles on latent factor model convergence and measures of fit

This research examines the effects of single-value response style contamination on measures of model fit and model convergence issues. A simulation study examines the effects resulting from percentage of contamination, number of manifest, number of reverse coded items, magnitude of standardized factor loadings, response scale granularity, and sample size. Initial results indicate that sample size, scale granularity, factor loadings and number of manifest items had little to no effect on measures of fit. Both percent contamination and number of reverse coded items had a large effect on measures of fit. Measures of fit were more readily effected by percent contamination in models with higher standardized factor loadings. Model convergence issues were most strongly related to percent contamination and factor loadings