I am doing a six-group measurement invariance study. I compared the configural model with the fully constrained model, finding out that there is no full measurement invariance. Then I am detecting partial invariance and try to find out which item(s) shows measurement noninvariance by comparing the configural model to a scalar model by freeing an item at a time (altogether I have 21 items), but I keep getting warning messages as follows:
CORRELATION MATRIX (WITH VARIANCES ON THE DIAGONAL) ITEM21 ________
WARNING: THE BIVARIATE TABLE OF ITEM17 AND ITEM8 HAS AN EMPTY CELL.
WARNING: THE BIVARIATE TABLE OF ITEM17 AND ITEM14 HAS AN EMPTY CELL.
WARNING: THE BIVARIATE TABLE OF ITEM17 AND ITEM12 HAS AN EMPTY CELL.
THE MODEL ESTIMATION TERMINATED NORMALLY THE CHI-SQUARE COMPUTATION COULD NOT BE COMPLETED BECAUSE OF A SINGULAR MATRIX.
I do not know how I should handle this. Any suggestions would help.