ruthjlee posted on Wednesday, July 30, 2014 - 10:41 pm
We are planning an LCA using a single dataset comprising parent report data. Many of the questions are on different scales, and we will probably end up using a mix of ordered-categorical and continuous data (and possibly also dichotomous data). It's clear from other discussions on this forum that we need not standardize the scores. However, using standardized scores would help us to present the data clearly for publication. While it is unnecessary to standardize scores in a mixture model, is it acceptable to do so? Or does standardizing scores before running the analysis cause specific problems?