Teacher Support and Students' Everyday Academic Behavior and Performance: A Bayesian Moderated Mediation Modeling

• In Event: Examining Factors That Mediate Student Achievement

Abstract

Background: This study explored the divergent mechanisms through which teachers’ support for students affects students’ performance in low- and high-stakes assessments.
Method: Data from published studies were incorporated into primary data from Ghana (N=135), and Bayesian structural equation modeling with Markov Chain Monte Carlo algorithm were used to test mediated, unmediated, and moderated-mediated relationships.
Findings: Results show teacher support directly translates into a .75-point increase in high-stakes exam performance. A chain mediation effect also exists through student homework behavior and low-stakes everyday performance. Teacher support influenced girls’ homework behavior and performance more than boys.
Implication and conclusion: The results could aid in identifying malleable leverage points to intervene to address learning needs, gaps, and challenges in low- and high-stakes assessments.

Authors

• David Ansong, University of North Carolina - Chapel Hill
  Presenting Author

• Moses Okumu
  Non-Presenting Author

• Isaac Koomson
  Non-Presenting Author

• Jamal Appiah-Kubi
  Non-Presenting Author

• Emmanuel Owusu Amoako, University of North Carolina - Chapel Hill
  Non-Presenting Author