I am trying to fit a CFA model using residual bootstrap with a weigth variable. The results seems kind of odd. The normal theory Chi-Square is highly significant (p < 0.0001) and the Bollen-Stine p-value = 1. CFI is high (0.98), RMSEA is also quite high (about 0.09) and SRMR = *****. The residual matrix is -1* the estimated matrix.
I have tried doing this combination of residual bootstrap with a weight variable with a few models using the same data and I always get similar results. If I drop the weight variable, the results are more coherent, i.e. both Chi-Squares are quite high, fit indices are mediocre and residual matrix does not have a strange pattern.
Is the residual bootstrap unusable with a weight variable?