ERIC Number: ED371014
Record Type: RIE
Publication Date: 1994-Apr
Reference Count: N/A
A Comparison of Decision Theoretic Approaches to Adjusting Cut Scores.
Druva-Roush, Cynthia Ann; And Others
Methods of adjusting cut scores used in placement decisions are examined empirically. Admission and performance variables are used to study alternate methods of adjusting cut scores for placement in standard and accelerated rhetoric courses in a large university setting, with the predicted variable being success or failure as measured by end-of-semester course grades. Data from 4,045 freshmen entering a university in the fall semesters of 1991 and 1992 were collected (1,956 for setting cut scores, and 2,089 for validation). A comparison of predicted placement errors using cross validation groups is made between three decision theoretic approaches to setting cut scores: (1) contingency table analysis; (2) discriminant analysis; and (3) logistic regression. The effectiveness of each method is evaluated psychometrically, as is the ease of using each method in a practical setting. Two-equation discriminant analysis appears to be superior to logistic regression with more than one predictor variable, and it outperforms contingency table analysis. When a single index method is used, logistic regression and contingency table analysis are quite similar, and the contingency table analysis seems to be much more interpretable by nonstatisticians. (Contains 16 references, 6 tables, and 5 figures.) (SLD)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers: Contingency Tables; Decision Theory
Note: Paper presented at the Annual Meeting of the American Educational Research Association (New Orleans, LA, April 4-8, 1994). Figures 1-5 are faint and will not reproduce well.