Authorized Institutional Representative
BibTeX
@MISC{Cerrito_authorizedinstitutional,
author = {Patricia B. Cerrito and Professor Of Mathematics and Patricia B. Cerrito and Judy L. Bristow},
title = {Authorized Institutional Representative},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
The purpose of this project is to examine the role that Mathematics plays as “gatekeeper ” to the successful (or unsuccessful) completion of a college degree and to determine whether patterns of course selection impact retention and graduation. The variables to be collected in the study include demographics (race, gender, age), admission values (high school gpa, ACT scores, SAT scores), transcript information (courses, grades, and declared major) over a 6-year period. The final variable to be collected is whether the student graduates during the period under study. The database contains over 50,000 student records, and over 300,000 enrollment records. Traditional statistical methods are inadequate to handle such a large database. These techniques must be supplemented by data mining tools. The database must be investigated using the data mining process. Whereas statisticians were content to use logistic regression to identify those at higher risk of dropping out, data miners want to be able to predict accurately the sub-group of students who will not continue, and to prioritize the sub-group for interventions to increase the retention of this sub-group. The project will adapt methods used in the retail industry to examine the problem of “churn”, or customers switching to another company for goods and services. The customers at the greatest risk of leaving are clearly identified so that an intervention can be used to
Keyphrases
authorized institutional representative data mining tool student record enrollment record admission value high school gpa traditional statistical method logistic regression college degree large database final variable data mining process course selection impact retention 6-year period data miner retail industry transcript information
