Document Type

Dissertation

Abstract

This study finds predictive factors that have a significant effect on the ending Mathematics/Actuarial GPA of Actuarial majors at Bryant University. This was done to add clarity to incoming students for what it takes to do well in the actuarial major. There were 266 subjects consisting of Bryant students that graduated between the years 2009 and 2015. The data was received in the form of final transcripts upon graduation for these students. Through manipulation of these transcripts, GPAs were calculated for Mathematics/Actuarial, English/Literary Cultural Studies, History/Politics, Economics, Science, Social Sciences, Computer Information Systems, Finance, Accounting, Management, and Marketing. These GPAs were found after 2 years (4 semesters) of classes and also after 3 years (6 semesters) of classes to see if the predictive factors differ between the two years. I also ran models comparing just the Arts and Sciences as well as just the business to see how the results differ using just those subjects rather than putting them all together. I used regression and decision trees in order to generate results. In order to use regression, I needed to impute missing values with SAS Enterprise Miner 13.2 to use all of the data. I did this using the tree surrogate method to accurately impute the data. I did not include Mathematics/Actuarial in these predictive models because of the high correlation to the final Mathematics/Actuarial GPA that I am trying to predict. After comparing the different models, English, Computer Information Systems, and Economics were the highest predictors after two years and Finance English and Science were the highest predictors after three years.

Included in

Analysis Commons

Share

COinS