Data in the social sciences often do not conform to what a researcher would like in order to have an unproblematic analysis using some form of linear regression. Especially, when data reflect counts, a number of potential problems arise. This case study discusses challenges that occurred when applying the models and examines the results. Finally, it looks at using the parameters for 'what-if' modeling by changing the input data, a step that is often implied but not performed in analysis. The approach is not that of a statistician but a statistically informed social scientist using SAS to analyze the data. [This content is provided in the format of an e-book.]