ERIC Number: ED402843
Record Type: Non-Journal
Publication Date: 1996-Oct-31
Reference Count: N/A
Predicting Student Loan Defaults. ASHE Annual Meeting Paper.
Flint, Thomas A.
The failure of students to repay federally insured loans has led to an increased emphasis on default prevention and threatens institutions with high default rates with exclusion from federal student aid programs. Prior studies of default prevention using theoretical constructs based on economics, sociology, and psychology have yielded mixed results. This large-scale multi-college study based on the Student Loan Recipient Survey (SLRS) of the 1987 National Postsecondary Student Aid Study sought to analyze the many variables that might influence loan repayment behavior. Data on 1,117 borrowers from 510 institutions were analyzed. The model used for the study grouped variables into three blocks: precollege, college-related, and postcollege, and included student background, school choice, student academic achievement, loan counseling, exit counseling, and point-of-survey (postcollege) variables. The model correctly predicted repayment status of about 87 percent of all cases, with the strongest correlation being with student background. Overall, the study found that while economic factors played a modest role in repayment behavior, the most significant influences were psychological. Three tables summarize data, and include: record loss during sample selection; description of the variables included in the model; and a regression analysis of defaulted loans for the six groups of variables. (Contains approximately 135 references.) (CH)
Descriptors: Demography, Educational Finance, Federal Programs, Financial Aid Applicants, Higher Education, Legal Responsibility, Loan Default, Loan Repayment, Models, National Surveys, Paying for College, Predictive Measurement, Predictive Validity, Predictor Variables, Student Characteristics, Student Financial Aid, Student Loan Programs, Student Responsibility
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A