ERIC Number: ED445116
Record Type: Non-Journal
Publication Date: 2000-Jul
Reference Count: N/A
Effects of Differential Prediction in College Admissions for Traditional- and Nontraditional-Aged Students. ACT Research Report Series.
This study determined whether using ACT Assessment Composite score, high school grade average, or both predictors jointly for college admissions would result in differential prediction or differential outcomes for traditional-aged and nontraditional-aged students. The analyses were based on data from 112,341 students at 79 postsecondary institutions. Both total-group and within-age group regression models were estimated by institution. Linear regression models were used to predict students' first-year college grade average. Corresponding logistic regression models predicted students' chances of a 2.5 or higher or 3.0 or higher grade point average (GPA). Validity statistics were calculated for each model by institution, and then summarized across institutions. The results for the total-group and within-group models and optimal cutoffs were then compared. The total-group ACT Composite and high school grade average models slightly overpredicted the first-year GPAs of nontraditional-aged students, and they underpredicted the GPAs of traditional-aged students. When used jointly as predictors, virtually no differential prediction or differential effects were found for the two age groups. (Contains 5 tables and 16 references.) (SLD)
Descriptors: Age Differences, College Admission, College Entrance Examinations, Higher Education, Nontraditional Students, Prediction, Regression (Statistics), Scores, Test Use
ACT Research Report Series, P.O. Box 168, Iowa City, IA 52243-0168.
Publication Type: Reports - Research
Education Level: N/A
Authoring Institution: American Coll. Testing Program, Iowa City, IA.
Identifiers - Assessments and Surveys: ACT Assessment