LAUSR

The current study explored the associations between non–cognitive science-related variables, i.e., science interest, utility, self-efficacy, science identity, and science achievement in a serial mediation model. The study also further explored the potential heterogeneity in the model parameters using one of the data-mining techniques, which is the structural equation model (SEM) Tree. Data on 14,815 high school students were obtained from a large-scale database High School Longitudinal Study of 2009 (HSLS:09). The results highlighted science interest and science utility positively influencing science achievement through a sequential pathway of mediators, including science self-efficacy and science identity. The strength of direct effects considerably varied across students, resulting in classifying them into four subgroups. For instance, among females with a low SES subgroup, developing substantial science interest would result in better science self-efficacy and science identity that flourish science achievement. These valuable findings provide fruitful tailored recommendations, elevating the science achievement in the subgroups (146 words).