I am a new Mplus user. I am trying to complete EFA for dichotomous data and receive the following messages.
1) WARNING: THE BIVARIATE TABLE OF A5 AND A3 HAS AN EMPTY CELL. I used 999 to denote missing data, and do not think I have any missing data if this is the problem here?
2) WARNING: THE SAMPLE CORRELATION OF A29 AND A1 IS 0.999. DUE TO ONE OR MORE ZERO CELLS IN THEIR BIVARIATE TABLE. INFORMATION FROM THESE VARIABLES CAN BE USED TO CREATE ONE NEW VARIABLE. I am just unclear of how I should act on this message.
When a bivariate table between two binary variables has an empty cell, this implies a correlation of one. Both variables should not be used in the analysis because they are not statistically distinguishable.
Aaron Batty posted on Thursday, November 22, 2012 - 8:13 pm
I have the same issue, but it's reporting that 33 of the 48 items on this test (dichotomous) are involved in "empty cells." As a result, I don't get model fit indices, and can't compare a unidimensional to a multidimensional model (I'm getting similar warnings on both models).
I've never seen this before with other datasets, and I don't understand how I'm getting a correlation of 1 anyway.