ERIC Number: ED471226
Record Type: Non-Journal
Publication Date: 2002-Apr
Reference Count: N/A
A Comparison of the Relative Practical Value of a Predictive Discriminant Function Analysis and a Binary Logistic Regression Analysis of Student Success in an Innovative Alternative High School Program in South Texas: Data Set #1. Working Paper.
Kester, Donald L.; Linton, Thomas H.; Sullivan, Lynn R.
This study compared two statistical approaches, predictive discriminant function analysis and binary logistic regression analysis, to determine the characteristics of at-risk students who succeeded in comparison with those who did not succeed in a project-based alternative high school program. The study also investigated true positive hit rates for predicting correctly the students that would be successful using each statistical approach to provide information for decision making in admissions decisions for this program. Data came from the Wisconsin Youth Survey from measures of bonding with teachers, school, and peers and from a measure of educational engagement completed by 70 students. Other variables measured were family composition, student employment, grade, gender, and age. The predictive discriminant analysis approach produced a true positive hit rate of 79% compared to the remarkable 94% of the binary logistic approach. Advantages of the binary logistic approach are discussed. (SLD)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers - Location: Wisconsin