Sarstedt et al. 2022 and prior works on FIMIX-PLS suggest partitioning is ill-advised when the entropy (EN) statistic is below 0.5. However, this criterion is not foremost among others available, and only for certain conditions.

Have other researchers dealt with results where the EN is below 0.5 while other values suggest possible segmentation, and moreover, a priori segments are proposed.

What is the best practice to deal with less than clear-cut results from FIMIX-PLS?

Article Latent class analysis in PLS-SEM: A review and recommendatio...