A Comparison of Accuracy of Selected Models for Predicting Academic Performance of Junior College Transfer Students

The possibility of increasing prediction accuracy of academic performance of junior college transfer students was explored through use of curvilinear models and models that provided for interaction effects among predictor variables. Also, categorical prediction accuracies of selected regression mode...

Full description

Saved in:
Bibliographic Details
Published in:The Journal of educational research (Washington, D.C.) D.C.), 1972-11, Vol.66 (3), p.111-114
Main Author: Marcus Nickens, John
Format: Article
Language:English
Subjects:
A priori knowledge
Academic achievement
Coefficients
Discriminants
Grade point average
Junior colleges
Modeling
Proportions
Regression analysis
Statistical variance
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The possibility of increasing prediction accuracy of academic performance of junior college transfer students was explored through use of curvilinear models and models that provided for interaction effects among predictor variables. Also, categorical prediction accuracies of selected regression models and discriminant models were compared. Curvilinear models and models that provided for possible interaction effects did not increase prediction accuracy in practical magnitudes beyond that of linear regression models. A multiple regression model, which predicted averages that correlated .60 with actual averages and had a standard error of .60, produced categorical prediction accuracies similar to discriminant models. In the case of success-failure predictions, approximately 75 percent of the students predicted to succeed were observed to succeed, and 51 percent of the students predicted to fail were observed to fail for both regression and discriminant models.
ISSN:0022-0671
1940-0675
DOI:10.1080/00220671.1972.10884421