Hello all

I ran an ordinal logistic regression, my DV has 3 categories, and I ran it vs. a categorical IV with 5 categories (so the computer calculated 4 dummies)

The model was NOT significant (P=0.11 - chi square test), however, one of the dummies was significant (P=0.036, chi square test).

When I did a cross-tab on this, it was NOT significant both with chi square and Fisher exact test.

My instinct say it's not significant, but how can it be that it comes out like this ?