NotesFAQContact Us
Search Tips
ERIC Number: ED512353
Record Type: Non-Journal
Publication Date: 2004
Pages: 14
Abstractor: As Provided
Reference Count: 25
Cross-Validation of Persistence Models for Incoming Freshmen. Professional File. Number 93, Fall 2004
Harmston, Matt T.
Association for Institutional Research (NJ1)
This study was intended to serve as an example of cross-validating results from student persistence prediction models that employed commonly available pre-college student characteristics. The study investigated whether the accuracy of predicting student persistence would vary because of the use of present-year vs. previous-year parameters on present-year data, and whether the same set of predictors would change in terms of predictive efficiency between years. Institutions with selective, liberal, and open admissions policies had consistent persistence prediction odds ratios across time regardless of the data on which model parameters were generated. As expected, all institutions demonstrated lower accuracy rates at higher persistence criteria, and accuracy rates differed between different sets of model parameters. The ACT Composite scores emerged as consistently stable, significant predictors of persistence. The paper concludes with a discussion of procedural issues to consider when using logistic regression to predict persistence. (Contains 6 tables.)
Association for Institutional Research. 1435 East Piedmont Drive Suite 211, Tallahassee, FL 32308. Tel: 850-385-4155; Fax: 850-383-5180; e-mail:; Web site:
Publication Type: Reports - Research
Education Level: Higher Education; Postsecondary Education
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: Association for Institutional Research
Identifiers - Assessments and Surveys: ACT Assessment