This is what drives measurement invariance tests for ordinal items, and partial invariance involves allowing certain thresholds to differ by group while holding the rest equal.

Ordinal items do not have intercepts in the usual sense, so constraining them often has no effect on degrees of freedom or fit. Instead, look at the item thresholds in your output and constrain or free those thresholds across groups.

A preprint of our chapter, Stalking the Wild Quincunx, is here. To be published by Cambridge University Press. drive.google.com/file/d/1bfSW...

what should i do if a reviewer demands an a priori power analysis for an experiment i already did … ;)