I used multiple imputation on SPSS to deal with missing data in my study. I then carried out multiple regression from the imputed and original data-sets, using a split-file. I now have output for each of the 5 imputed data-sets plus the original and a pooled model. I want to test the assumptions of multiple regression, but I'm not sure how.
For example, plotting studentized residuals against unstandardized predicted vales gives me 6 different scatterplots. Which one do I interpret and report? Also I have 6 Durbin-Watson values, etc.
