Wai Chow posted on Monday, June 26, 2006 - 4:49 pm
I was wondering, if you had suggestions on how to deal with this problem cited in the error message ("Group does not contain all values of categorical variable.") I am trying to use multiple group analysis to test model invariance of bifactor model across independent groups and separately across time. However, frequencies reveal that no one in the sample at T3 endorsed one of Likert-points on at least one item and as a result Mplus outputted this message. How do I move forward with these analyses under these circumstances? Is there a way to estimate threshholds when not all the points were endorsed?
Use maximum likelihood with KNOWNCLASS for the multiple gorups and the CATEGORICAL option with the * specification. See the user's guide. For further help on this, send your input, data, output, and license number to firstname.lastname@example.org.