The relationship between error beliefs in chemistry and chemistry learning outcomes: a chain mediation model investigation

Abstract

Errors are natural elements of the learning process and provide a high potential to promote students’ learning outcomes. In recent years, there has been much research about learning from errors. However, we know little about the relationship between students’ error beliefs in chemistry and chemistry learning outcomes at present. Thus, the aim of this study was to explore the mechanisms of chemistry behavioral and cognitive engagements, adaptive reactions towards errors in chemistry and error beliefs in chemistry, and offer suggestions to the improvement of students’ chemistry learning outcomes. We assessed all variables in eight different schools in China (N = 1352 students, Grade 10) and used structural equation modelling (SEM) to check the direct and indirect relationships between four variables. Our findings revealed that (1) chemistry behavioral and cognitive engagements, adaptive reactions towards errors in chemistry and error beliefs in chemistry significantly positively predicted students’ chemistry learning outcomes; (2) both (a) adaptive reactions towards errors in chemistry and (b) chemistry behavioral and cognitive engagements acted as significant mediators between error beliefs in chemistry and chemistry learning outcomes; (3) the chain mediating effect of error beliefs in chemistry → adaptive reactions towards errors in chemistry → chemistry behavioral and cognitive engagements → chemistry learning outcomes was significant. Finally, we discussed the important findings, pointed out the educational implications, acknowledged our study's limitations and suggested directions for future study.

The relationship between self-handicapping in chemistry and chemistry academic engagement: a moderated mediation model investigation

“…The explanatory variance of the first factor in this study was 30.59%, which met the criteria of no common method bias. We then did not report how well the full model fit the data, but only considered the path coefficient (Guo et al, 2022) because our model was a saturation model (Nurkhaidarov and Shochat, 2011;Zhang et al, 2019;Huangfu et al, 2023). Subsequently, we used the Mplus 8.3 Structural Equation Model (SEM) to test this hypothetical model.…”