Abstract
Mathematical problem-solving is necessary to encounter professional, 21st-century, and everyday challenges. The relevant context of mathematical problem-solving is related to science, which is presented using natural language. Mathematical problem-solving requires both mathematical skills and nonmathematical skills, e.g., science knowledge and text comprehension skills. Thus, several internal and external factors affect success in mathematical problem-solving. In this study, we investigated the cognitive (i.e., mathematics domain-specific prior knowledge (DSPK), science background knowledge, and text comprehension skills) and socioeconomic status (SES) (i.e., parents' educational level and family income) factors that affect students' mathematical problem-solving skills. The data considered in this study included tests, documents, and a questionnaire from grade seven to nine students (n = 1067). In addition, a theoretical model was constructed using structural equation modeling. We found that this model was close to satisfying the critical values of fit indices. The model was then modified by deleting the nonsignificant paths, and the modified model exhibited a better fit. We found that most of the exploratory variables directly affected mathematical problem-solving skills, with the exception of the parents' educational levels. The strongest factor was mathematics DSPK. Both the father's and mother's educational levels indirectly influenced mathematical problem-solving skills through family income. In addition, text comprehension skills indirectly impacted mathematical problem-solving skills with science background knowledge acting as a mediator.