gloria posted on Thursday, July 02, 2015 - 6:29 am
Good morning! I have imputed 10 data sets using Mplus (mix of categorical and continuous) and I want to report a correlation matrix that takes into account the different types of data that I have.
Using the 10 imputed data sets, I'm able to read my data and produce a correlation matrix. However, when I define the categorical variables using the categorical command, I do not get the correlation matrix - only covariance. Is there an easy way to get the correlation matrix from the covariance? Will that work if I somehow provide mplus with the SD of the variables?
p.s. The reason I'm not using the original correlation matrix output during the actual imputation is because I'm including extra variables in the auxiliary statement and i need to those correlations too.