Examining the role of different relational reasoning skills in mathematical and scientific reasoning

Abstract

The rising importance of STEM education motivated educators to investigate the cognitive correlates of Mathematics and Science achievement for better pedagogy and curriculum development. Relational reasoning, defined as the ability to extract and reason about meaningful patterns from streams of information, is one of the candidate correlates (Alexander et al., 2012). Four major types of relational reasoning, namely analogy, anomaly, antinomies, and antitheses, were argued to be related to scientific reasoning and learning (Dumas, 2017) and mathematical reasoning (Zhao, Alexander, & Sun, 2021). Relational reasoning has been shown to be related to mathematical reasoning. It was correlated with fractional knowledge among primary school students after controlling for IQ and math ability (Kalra, Hubbard, & Matthews, 2020) and associated with performance in math achievement test among Chinese adolescents (Zhao, Alexander, & Sun, 2021). Relational reasoning was also argued to have a role in Scientific reasoning. It was found to be related to the number of new ideas generated and their originality among undergraduate students studying Science (Dumas et al., 2018). Younger children may acquire scientific concepts more quickly after being taught to engage in relational reasoning (Gentner et al., 2016). However, there has been scarce research on the contribution of each type of relational reasoning skills. To delineate their roles, we examined the extent to which different types of relational reasoning predict Mathematical and Scientific reasoning among Junior Secondary school students, where they first learn about Science in a formal curriculum in Hong Kong. We recruited 197 Hong Kong Secondary school students (97 female; mean age = 15.37 years, SD = .65 years) from local schools. They were either at the end of Form 3 or the beginning of Form 4 at the time of testing. Due to the COVID-19 situation, participants completed the measures online with individual supervision to ensure adequate attention and motivation. Participants completed the Lawson Test of Scientific Reasoning as a measure of scientific reasoning, the Mathematical Problem-solving subtest adapted from the WIAT-III, and the test of relational reasoning (Alexander et al., 2016) for assessing their relational reasoning skills, and reasoning in Mathematics and Science respectively. They also completed measures on verbal working memory, non-verbal intelligence, visual-spatial skills, and provided demographic information. Results from multiple regression analyses showed that relational reasoning as a construct significantly predicted mathematical problem solving and scientific reasoning after controlling for age, non-verbal intelligence, verbal working memory, and visual-spatial skills. Among the four types of relational reasoning skills, only analogy and antitheses were found to be significantly predicting scientific and mathematical reasoning. The present results suggested that analogy, the ability to extract similarities in information, and antithesis, the ability to extract the opposite relation of a given process, could be the most associated to math and science reasoning among the four types of skills. Our findings shed light on the cognitive skills related to math and science, paving the way for educational intervention and curriculum improvement to help equipping our students to be better learners and thinkers in the STEM field.