Student 09 posted on Wednesday, February 22, 2012 - 2:28 am
Hi, when conducting a twolevel CFA with ordinal indicators a warning appears and it is suggested to reduce the number of integration points. In fact, when constraining the number of integration points to 5, the model converges fast and without problems. However, I would like to learn about the potential problems occuring when constraining the number of integration points for categorical data analysis. Are there e.g. results from simulation studies available, or is there other information specifically for Mplus?
Five integration points is most likely not enough. You should increase to at least 10 to see if your results change. See pages 418-421 of the user's guide for a brief description of numerical integration.