랜덤 포레스트 기법을 적용한 초등교원의 교사효능감 영향요인 탐색

Abstract

The purpose of this study is to identify the influencing factors on teacher efficacy using「Korean Teacher Longitudinal Study-Elementary School Teacher」data from a more comprehensive perspective by analyzing all available items, which is the strength of the random forest technique. The data from 2,794 elementary school teachers were used, and teacher efficacy, which is the response variable of this study, was divided into three sub-domains (instruction/assessment efficacy, student guidance/class management efficacy, and future education efficacy). A random forest regression was conducted for each sub-domain. The main results of this study are as follows. First, the influencing factor with the highest importance index in all teacher efficacy was job satisfaction. Second, among the top 25 factors (approximately 10% of the total explanatory variables), 15 items, which account for 60%, were commonly found to be major influencing factors in all teacher efficacy. Third, among the top 25 factors, there were 4 to 7 major influencing factors unique to each teacher efficacy, and these factors appeared to represent the unique characteristics of each teacher efficacy. Fourth, the variables that appeared to be mainly related to teacher efficacy were all individual teacher-level factors, and school-level factors did not appear. In this study, it was able to examine the relationship between individual teacher efficacy and influencing factors by dividing teacher efficacy into three sub-domains (instruction/assessment efficacy, student guidance/class management efficacy, and future education efficacy) due to the strength of data. Also, this study examined the relationship between teachers’ practical teaching/learning activities and teacher efficacy, which was difficult to examine in previous studies.