Using Ordinal Logistic Regression Analysis in Evaluating Teachers’ Performance Level of High Schools (12th grades) in Kurdistan Regional Government

Abstract

The Ministry of Education – Kurdistan Regional Government (MOE-KRG) currently depends only on the students’ grades in the evaluation of the teachers’ performance level in all subjects as low, medium, or high performance. Relying just on one variable to determine teacher’s performance is not fair and this problem must be resolved statistically through finding new proposed statistical model. Therefore, this study tries to find some variables that are available in MOE-KRG for all high schools in order to use them in the proposed model to predict teachers’ performance level instead of the old one. This study aims to predict teacher’s performance level of high schools (12th grades) in KRG who teach Kurdish subject, and also analyze the effects of variables that have impact on the evaluation of teachers’ performance level depending on the data that are available in the MOE-KRG. In this paper Ordinal Logistic Regression (OLR) method is used to find a proposed model for evaluating teachers’ performance in Kurdish subject using the data of all scientific high schools in KRG (646 high schools). The teachers' status was analyzed by selecting nine variables related to the high schools: sector (governmental or non-governmental), geographic location, type of education, status of school (exemplary, non-exemplary), gender of student, year of school establishment, number of classes, number of teachers and student’s average marks in Kurdish subject. It is concluded that four of these variables (student’s average in Kurdish subject, number of classes, geographical location and status of school) have significant effects on teachers' evaluation in 12th class of scientific high schools, and the overall percentage of correct classification is about 87%, it that means the Ordinal logistic regression model has an ability to predict teacher’s performance level very well.