Eric Teman posted on Wednesday, June 08, 2011 - 7:48 pm
I am working on a simulating where I generate various levels of missing data. It appears Mplus can generate missing data based on probabilities (both for MCAR and MAR) but not based on rate. For example, I can't tell Mplus to generate a dataset with 20% missing data, but I can specify the probability for missingness to be .20. Do you forsee any issues when using levels of missingness in my simulation based on probabilities versus rates?