Title page for ETD etd-06072004-142209

Type of Document Dissertation
Author Boudreaux, Mary Allyn
Author's Email Address clboud@lsu.edu
URN etd-06072004-142209
Title Development of a Predictive Model for Student-Athlete Retention and Graduation at Louisiana State University
Degree Doctor of Philosophy (Ph.D.)
Department Human Resource Education Workforce Development
Advisory Committee
Advisor Name Title
Elwood "Ed" Holton Committee Chair
Krissana Machtmes Committee Member
Michael F. Burnett Committee Member
Reid Bates Committee Member
Amelia Lee Dean's Representative
  • college success
  • sport
  • heirarchical logistic regression
  • high school performance
  • college performance
Date of Defense 2004-05-19
Availability unrestricted
The National Collegiate Athletic Association through its member institutions has implemented academic standards governing initial athletic eligibility and has led reform initiatives tying the ability to compete athletically to student-athlete retention and graduation. Louisiana State University (LSU), like many Division I institutions, admitted its scholarship athletes using these initial eligibility standards as a minimum qualification for admission. However, as NCAA requirements have become less stringent, the admissions requirements at LSU have increased. Concerns about the retention and graduation of student-athletes and an increasing gap between the academic credentials of the student body and student-athletes led administrators to question the wisdom of this practice. There was a need to determine which variables can best predict the retention and graduation of student-athletes at LSU and whether or not these variables differed from results found in national literature. It was hoped that the predictive models could also be used to bridge the gap between NCAA and university admission standards.

This study uses hierarchical logistic regression to predict student-athlete retention and graduation using six sets of pre-college and post-enrollment variables for each dependent variable. High school performance variables, characteristics of the high school attended, achievement test scores, demographic and sport variables were used to develop a pre-college model for both retention and graduation. College performance variables that measured the student-athletes' grade point average (GPA) at three academic milestones were added to these models.

Results indicated that two different sets of variables predict retention and graduation of LSU student-athletes. The significant predictors in the pre-college retention model included: High School and English GPA, number of natural science and social science courses taken, total number of academic courses taken, math test score and sport and redshirt variables. The significant predictors in the pre-college graduation model included: High School and English GPA and total number of academic courses taken. In the development of the college performance GPA models, the researcher found that as the student-athlete progressed further in his/her academic career, the less important the pre-college variables became. However, most of the predictive power was attributed to the pre-college variables.

  Filename       Size       Approximate Download Time (Hours:Minutes:Seconds) 
 28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access 
  Boudreaux_dis.pdf 950.21 Kb 00:04:23 00:02:15 00:01:58 00:00:59 00:00:05

Browse All Available ETDs by ( Author | Department )

If you have questions or technical problems, please Contact LSU-ETD Support.