Determining Cutoff Points of Factors Affecting Scores of Private Skill Tests: ROC Curve Approach

Gürol ZIRHLIOĞLU
Dr., Yuzuncu Yil University, Faculty of Education

Abstract

Problem Statement: Evaluation is carried out by a variety of factors when choosing the student to physical education and sport science departments. These factors are skill tests, University Student Selection Exam (OSS) and students’ academic success in their secondary school (AOBP). Various research have indicated that these tests scores are effective in determining academic success. Students who get a high score may be also successful in skill tests, and the grade of OSS may be effective in order to select accurate student. Problem statement of this study can be expressed as “What are cutoff points and distinctiveness level of exams those are applied for entrance to Department of Physical Education and Sports Teacher of Yuzuncu Yil University?”

Purpose of study: The aim of the study is to use the University Student Selection Exam, students’ academic success in their secondary school, shutle running and coordination test scores instead of the average academic achievement in order to distinguish successful and unsuccessful students through ROC (Receiver Operating Characteritic) analysis method.

Methods: The data used in this study were taken from the sudents’ office and the record of the physical education and sport science department. These data were obtained from 25 male and 15 female students in the first semester of the 2007-2008 academic year. Data included student selection exam, students’ academic success in their high school, and special skill test scores (i.e., including shuttle running tests and coordination test). ROC curves were used to evaluate the effect of OSS, AOBP and special skill test scores on the academic achievement score and to determine to the cut-off points of these variables. In this study, academic achievement scores of the students were used as gold standard score. This score is 60 points. Students who have 60 or more are viewed successful.

Findings and Results: ROC analysis revealed that Maximum sensitivity+specificity value were found 234.601 for OSS score. If this value is selected as the cut-off point, 91.7% of students who are successful according to academic success average estimated true. This variable is significant (p<0.05). Maximum sensitivity+specificity value was found as 78.346 for AOBP score. If this value is selected as the cut-off point, 75% of students who successful according to academic success average estimated true. This variable is significant (p<0.05). Maximum sensitivity+specificity value was found 139 for shutle runs score. If this value is selected as the cut-off point, 41.7% of students who are successful according to academic success average estimated true. This variable is not significant (p>0.05). Maximum sensitivity+specificity value was as found 29.45 for coordination score. If this value is selected as the cut-off point, 25% of students who are successful according to academic success average estimated true. This variable is not significant (p>0.05).

Keywords: Cutoff points, ROC curve, private skill tests, gold standart point.