More than ability: A multilevel analysis on how non-cognitive factors predict mathematics achievement in the PISA 2022 for Greece

Non-cognitive factors explain a substantial part of students' behavior at school and shape an important part of their school performance. The present study examines the extent to which a number of these factors (in particular, anxiety, self-efficacy and attitudes) predict students' achievement in mathematics, using research data from the recent 2022 cycle of the Programme for International Student Assessment for Greece. Capitalizing on the strengths and advantages provided by multilevel (hierarchical linear) modelling for assessments where the data is structured hierarchically, indicators relating to non-cognitive factors were isolated from the assessment, and their relationship with academic performance in mathematics was examined. The analysis conducted shows that most of the non-cognitive factors examined significantly predict students' achievement in mathematics, with self-efficacy in formal and applied mathematics and anxiety playing the largest role in predicting mathematics achievement, even after accounting for sociodemographic variables. Results further indicate that while self-efficacy is the strongest predictor of success, high anxiety acts as a barrier that dampens the positive effects of other motivational factors. Finally, the study highlights the importance of strengthening the integration of non-cognitive factors in the educational process and systems by adopting targeted interventions, prioritizing mastery experiences and training teachers through programs, to enhance student performance.