Repository logo
Log In(current)
  1. Home
  2. Colleges & Schools
  3. Graduate School
  4. Doctoral Dissertations
  5. Prediction of persistence for college students receiving federal financial aid : a multiple discriminant function analysis
Details

Prediction of persistence for college students receiving federal financial aid : a multiple discriminant function analysis

Date Issued
December 1976
Author(s)
Sheddan, Mayme Kerley
Advisor(s)
Eugene W. Schoch
Additional Advisor(s)
Robert K. R.
Permanent URI
https://trace.tennessee.edu/handle/20.500.14382/61410
Abstract

Five multiple discriminant function analyses identified variables that predict whether college freshmen receiving federal financial aid will be persisters, transfers, or dropouts. Subjects were 340 freshman students for an eight-year period, 1966-1973, at a small, liberal arts college, located in Tennessee. The total group was divided into four subgroups by sex and by median high school grade level (below B and B or above). Cross-validations were made on data from 70 college freshmen from the class of 1974.

Fifty variables taken from records available by the end of the first week of each fall term consisted of 7 academic, 16 personality, 14 financial aid, and 13 primarily biographical variables. Variables were as follows: high school grade point average, ACT Composite score, DAT Abstract Reasoning score, Otis Quick-Scoring Mental Ability score, SRA Reading Record total score, number of semester hours taken during the first term, and persistence rate in major field; 16 personality factors taken from the Sixteen Personality Factors Questionnaire; financial aid variables of having or not having scholarships and grants, federal grants, other grants, guaranteed student loans, National Direct Student Loans, institutional loans, and work; other financial aid variables of parental contribution, percent of financial aid in grants, percent of financial aid in loans, percent of financial aid in work, percent of request met by financial aid, percent of budget met by financial aid, and percent of need met by financial aid; and other variables of age, high school class size, distance from home, declared major, declared vocation, both parents working outside the home, father's occupation, mother's occupation, marital status of parents, religious denomination, department head faculty adviser, tenured faculty adviser, and alumni faculty adviser.

The discriminating power of the predictor variables was tested by means of the Wilks' lambda statistic, with significance tested by chi-square approximations. The Discriminant program and the Oneway program of Statistical Package for Social Sciences (Nie, 1975) were utilized. Classification was based on discriminant function coefficients and the pooled within-groups covariance matrices, with size used for the priors.

Percents of correct classification for the development groups ranged from 86.67 to 100, with percent of correct classification for cross-validation groups ranging from 50 to 75. In the analysis of males with high school averages B or above 100 percent of the persisters, transfers, and dropouts were classified correctly. All types of variables made dominant contributions to the discriminant functions.

Descriptions of persisters, transfers, and dropouts in each of the five analyses were provided. Implications for counseling and policymaking were also discussed.

Degree
Doctor of Education
Major
Educational Psychology and Guidance
File(s)
Thumbnail Image
Name

Thesis76b.S3423.pdf

Size

4.39 MB

Format

Adobe PDF

Checksum (MD5)

74a535c62d740b567a2db7489bddf8b1


University Libraries

1015 Volunteer Boulevard
Knoxville, TN 37996
865-974-4351

Map & Directions
Donate to the Libraries
  • About
  • John C. Hodges Society
  • Speaking Volumes magazine
  • Outreach
  • Directory
  • Employment
  • Policies
  • Library Intranet
University of Tennessee power T logo

The University of Tennessee, Knoxville
Knoxville, Tennessee 37996
865-974-1000

Events
A-Z
Apply
Privacy
Map
Directory
Give to UT
Accessibility

Built with DSpace-CRIS software - Extension maintained and optimized by 4Science