How mathematics homework types relate to mathematics academic performance via interest? A chain mediation analysis in the Chinese context.

The aim of this study is to investigate the relationship between eighth-grade students' mathematics homework types and their combinations and their Mathematics Academic Performance (MAP) in the context of China's "Double Reduction" policy, emphasizing the chain mediating role of mathematics homework interest and learning interest. A total of 18,202 eighth-grade students completed a cross-sectional survey administered in a mixed online-offline mode. We adopted a mediation analysis method with multicategorical independent variables, using Practice Homework (PH) as the reference category. The results showed that homework combinations containing Test-Preparation Homework (TPH) tend to improve MAP through direct effects, while homework combinations including Applied/Project-Based Homework (APH)-in particular PH + APH-are more likely to enhance students' interest and improve MAP indirectly. In addition, homework interest and learning interest mediated the link between specific homework types (and their combinations) and MAP, primarily through a sequential chain mediation pathway in which homework interest predicted learning interest, which in turn predicted MAP. These findings clarify the mechanisms through which homework design-considering both types and combinations-relates to students' MAP in a "Double Reduction" policy context, offering practical guidance for educators on assigning homework that enhances learning motivation and academic performance.