Which tests are suitable for OLM?

[Whitefeather]

Hello Econometricians!

We are trying to check how much of the credit ratings (divided into four groups) that can be explained by 11 financial ratios. We have run a Ordered Logit Regression and now have two questions.

1. Now we want to do statistical tests and have tested for multicollniearity and are now unsure regarding if we should/need to control for hetreoscedacity and to check the normal distribution of the error terms (Jarque-Bera). Would really appreciate your guidance since we can't find this in the literature.

2. We have received a goodness of fit measure of "57% correctly predicted" (prediction evaluation) from our regression in Eviews. We interpret this as in 57 cases out of 100 we will be able to estimate the correct credit rating based on the given financial key ratios (independent variables). Is this the same thing as to say that 57%of the variance in the credit ratings is a result of the variance in the financial key ratios (independent variables)?

Many thanks from SWEDEN! :D