Statistical Retention Analysis of Enrollment Data at Carroll College
Loading...
Authors
Dickerson, Dustin
Date of Issue
2010-04-01
Type
thesis
Language
Subject Keywords
Other Titles
Abstract
Organizations around the world hire operations scientists and data analysts to ensure their resources are being used to their maximum potential. Colleges and universities are no exceptions. Admissions offices, in particular, store incredible amounts of data about students who attend their institution. This data includes, but is not limited to, high school GPA, SAT and ACT scores, sex, ethnicity, age, household income, and intended major. What schools hope to learn from this data is two-fold: 1. What kind of student is the school most likely/unlikely to retain until graduation? 2. How effective are new retention programs? Knowing what student is likely to complete a degree from the institution helps ensure smart recruiting spending. Similarly, being able to measure the performance of retention programs such as a mentoring program (Miller & Herreid, 2009) allows for smarter spending. The goal of this study is to analyze the enrollment data from Carroll College from fall of 2000 to spring of 2009 in order to determine what factors - major, high school GPA, gender, standardized test scores, or distance from home - contribute most to a student's retention or transfer. To accomplish this task the data must be organized and sorted and the results must be displayed in a simple, professional form. In each chapter, we analyze the effect of a different factor on a student’s retention. In chapter 3, we study retention by major. By looking at retention graphs and conducting statistics tests, we see if a student’s major offers any information on the retention of that student. Specifically, we compute the likelihood of a student spending two or fewer semesters at Carroll, and the likelihood of a student graduating from Carroll. Similar analyses are done in chapters 4, 5, 6, and 7 for high school GPA, gender, home state, and standardized test scores, respectively.