I am attempting to run two bifactor CFAs, one for the 12-item and the other for the 21-item version of the Depression, Anxiety, and Stress scales. (DASS). Both CFAs on the same sample (N = 292).
For the DASS-12 bifactor CFA, I get the following error message:
NO CONVERGENCE. NUMBER OF ITERATIONS EXCEEDED.
Upon checking the residual variance, it turns out that indicator/item #12 has a negative residual variance of -3.180 (i.e., Heywood case?).
For the DASS-21 bifactor CFA, I get the following error message:
THE MODEL ESTIMATION TERMINATED NORMALLY. THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES COULD NOT BE COMPUTED. THE MODEL MAY NOT BE IDENTIFIED. CHECK YOUR MODEL. PROBLEM INVOLVING THE FOLLOWING PARAMETER: Parameter 85, ANX21
In contrast with the 12-item DASS bifactor model, however, I do not see any negative residual variance on any of the items/indicators.
Iíve also done the following procedures to try to rule out possible reasons for the lack of convergence.
(1) Checked syntax for errors (2) Made sure both models are not underidentified (3) Increased number of iterations
Are there any other things I rule out before I have to decide that the model is simply a bad fit for the data?
I would greatly appreciate any help/feedback with this issue