Thesis title: Students’ Abstract Reasoning Ability: The Role of Prior Knowledge and Experience and Its Relationship with Academic and Task Performance
The present study explores a range of conceptual and empirical issues concerning abstract reasoning (AR) among students, particularly in its nonverbal/figural forms, a construct widely employed in educational and professional assessment yet still characterised by theoretical ambiguity and practical inconsistency. Although AR tests are often presented as universal and content-free indicators of cognitive potential, and frequently serve as selection or screening tools in academic and occupational contexts, accumulated evidence suggests that reasoning performance is shaped by prior knowledge, experience, and task familiarity, although such effects may be domain-specific and not readily transferable across contexts. Furthermore, test-specific characteristics, training effects, and the subjective experiences of test-takers call into question the assumption of AR tests as “pure” measures of fluid cognitive ability and highlight the need for a more integrated understanding of what AR tasks actually measure and under which conditions they validly reflect cognitive ability.
Building on these considerations, the present study examined the multifaceted nature of AR and its variability across contexts. Specifically, it investigated whether students perform differently on two AR measures, Raven’s Standard Progressive Matrices (RSPM) and Hudson’s Abstract Reasoning Ability Test (HA-RAT), and what these differences may reveal about the cognitive processes underlying performance. The research further explored differences in students’ subjective perceptions of these two tests, alongside variations in performance profiles across AR, formal reasoning, and creative task-solving among academic programs, taking into account prior experience, academic specialization, and background knowledge. The study also assessed the predictive role of AR measures for academic performance, academic achievement, and novel problem-solving tasks, alongside the contribution of individual differences and prior knowledge.
The empirical investigation involved 267 undergraduate students from HSE University (Moscow, Russia), representing three academic programs: Psychology, Media, and Applied Mathematics and Information Science (AMIS). Participants completed a battery of cognitive, self-report, and custom-designed study instruments, including the RSPM, the HA-RAT, a formal (propositional) reasoning test, and a modified version of De Bono’s creative thinking task, as well as scales assessing personality traits and a set of behavioral-psychophysiological characteristics.
Results revealed significant differences between the two AR measures and clear variation in performance across academic programs. Students scored higher on the RSPM than on the HA-RAT, with AMIS students consistently outperforming peers from other programs. Subjective perceptions of the tests also differed, with the HA-RAT generally associated with a more demanding and less favorable experience, indicating distinct affective and experiential profiles across measures. Multivariate analyses showed that reasoning and task-solving patterns varied significantly across academic programs, with the strongest effects observed for formal reasoning, followed by abstract reasoning and creative thinking.
The predictive validity analyses yielded mixed results: while the HA-RAT demonstrated stronger associations with academic performance in some contexts, neither measure consistently predicted academic performance across groups. Furthermore, AR performance was positively associated with prior experience and domain-relevant skills, suggesting that familiarity and background knowledge play a meaningful role, although transfer effects across domains were limited. Personality traits and behavioral-psychophysiological characteristics showed generally weak and inconsistent relations to cognitive outcomes, with the exception of Conscientiousness, which emerged as a reliable predictor of academic performance.
Overall, the findings indicate that abstract reasoning, rather than representing a context-free indicator of general cognitive ability, particularly fluid intelligence, reflects a context-dependent interplay of cognitive, experiential, and task-specific factors. By integrating psychometric, empirical, and experiential perspectives, the present study contributes to a more nuanced understanding of the nature and role of abstract reasoning in academic and applied settings. The results also highlight the importance of considering domain relevance, prior experience, and test characteristics when interpreting AR performance, and point to directions for future longitudinal and cross-institutional research examining its cognitive and contextual dimensions.