The main topic of this independent study is to find ways to predict NCAA Division 1 FBS football All-Americans and predict players' draft values. Quarterbacks, wide receivers, and running backs were the position groups we performed this study on. For All-Americans we looked at college statistics to determine which were most significant. We looked at players' draft value because we wanted to see if the All-Americans would have a better draft value than players that were not All-Americans. We also wanted to see which stats would be important in deciding draft value. In order to predict correctly, we had to used several mathematical models with an end goal of finding the most important factors for both deciding All-Americans and players' NFL draft value. The techniques we used were logistic regression in order to predict the All-Americans and multiple linear regression in order to predict NFL draft values. These models led to the the assumption that being in a power-five conference was most important for All-Americans, and being an All-American was most important for a player's draft value.


Pasteur, R. Drew




Applied Statistics | Categorical Data Analysis | Statistical Models


Football, Statistical Models, NFL Draft

Publication Date


Degree Granted

Bachelor of Arts

Document Type

Senior Independent Study Thesis



© Copyright 2020 D'Andre E. Brown