2. Rerun the regression; save standardized and studentized residuals, leverage, Cook’s Distance, and standardized DF Betas. Check any outliers and unusually influential cases.

Do these cases look okay on these and other variables? What do you propose to do?

Discuss your options and decisions in class. (To do this analysis, you may want to create a new variable equal to the case number [e.g., COMPUTE CASENUM=$CASENUM in SPSS]. You can then sort the cases based on each regression diagnostic to find high values, but still return the data to their original order.)