Does anyone know how to check whether heteroskedasticity is present in an ordinal logistic regression analysis? And, if present, how can it be corrected/adjusted for?

Kjell Weyde