Teacher Response to Student Performance Predicting Future Performance and Student Enjoyment: Findings and Methodological Guidelines for Longitudinal Analyses

• In Event: Exploring Links Between Instructional Feedback and Emotions

Abstract

At least since the genesis of behaviourism, psychologists have known that rewarding (or punishing) behaviour increases (or decreases) the probability of showing it. Moreover, whereas the former is more prone to eliciting a positive emotional reaction, the latter is more prone to eliciting a negative one. In this study we transfer these fundamental psychological concepts to an educational setting by investigating the relationship between student-reported teacher praise, student-reported performance (i.e., marks), as well as student-reported enjoyment. Although the core concepts of reward learning are well established little research has investigated the relationship between teacher praise or punishment and emotional responses of students. Moreover, most studies linking teacher feedback to student performance have solely relied on cross-sectional data. As effects of teacher feedback could both be detrimental or facilitating to student performance or emotional well-being it is important to relate teacher behaviour directly to these student outcomes (i.e., within-subject effects) rather than comparing average effects of teacher behaviour across students (i.e., between-subject effects). Thus, going beyond previous research, here we investigated effects of teacher praise, student performance, and student enjoyment at a within student level with a longitudinal design and controlling for previous performance and enjoyment. To do this we analysed the rich PALMA dataset tracking 3530 students in 48 German secondary schools across 6 measurement points from 5th to 10th grade. We conducted two Bayesian cross-classified (i.e., with random effects at both a student and school class level) multilevel models with student-centred marks (or student-centred enjoyment for model 2), student-centred teacher praise, and measurement point predicting marks (or enjoyment for model 2) at the next respective time point. We hypothesised teacher praise would predict a gain in student performance as well as student enjoyment. Nonetheless, whereas we hypothesised a main effect for student enjoyment, student performance should increase when praise followed “better” marks (i.e., 1 SD above the student’s average) but not “worse” marks (i.e., 1 SD below the student’s average). Preliminary analyses confirmed these hypotheses showing a main effect of teacher praise on student performance, ß = 0.02, [95% CI: 0.00;0.05], and enjoyment, ß = 0.02, [95% CI: 0.00;0.04]. Moreover, in line with our prediction the effect of student performance increased for “better” marks, ß = 0.04, but disappeared for “worse” marks, ß = 0.00. In sum, this research provides novel findings relating teacher praise to student performance as well as student enjoyment at a within-subject level. Effects of teacher punishment as well as other achievement-relevant emotions such as anxiety or boredom will be discussed. Finally, as we believe within-subject level analyses of longitudinal data are an important step forward for educational research we provide interested researchers with guidelines on how to handle this research’s many challenges such as non-convergence of multilevel models, changing group membership of students (e.g., students changing classes), distinguishing within- from between-subject effects, missing data, or lagged dependent variables (e.g., controlling for previous performance).

Authors

• Felix Suessenbach, University of Munich

• Reinhard Pekrun, University of Essex