Do Classroom and School Characteristics Matter in Teachers' Feedback Behavior? An Application of Random Forest

• In Event: Methodological Innovations and Advanced Applications Using National and International Databases

Abstract

This study aims to investigate which teacher-level and school-level factors explain teachers’ feedback behavior frequency using a random forest, one of popular machine learning algorithms. We analyzed the data from Teaching and Learning International Survey (TALIS) 2018 which Korean sample included a total of 1,571 teachers. Findings from our preliminary analyses showed that not only individual teacher’s instructional strategies but also preparedness for teaching different groups of students clearly predicted the frequency of teacher feedback behavior. Interestingly, the classroom characteristics/compositions and collaborative climates of school currently working were strong predictors of the frequency of feedback behavior. These results suggest that teacher feedback behavior needs to be understood along with environmental factors or contexts beyond individual professionalism.

Authors

• Meereem Kim, Korea Institute for Curriculum and Evaluation
  Presenting Author

• Eun Hye Ham, Kongju National University
  Presenting Author