Single or dual processing in reasoning development? An application of state-trace analysis to a systematic database of studies

Abstract

Influential dual-process theories of higher cognition posit that two qualitatively different processes underlie human reasoning. In contrast, single-process theories postulate that reasoners draw upon common cognitive mechanisms when making inferences. To test the competing theories, a rigorous method – state-trace analysis – has been proposed and proven to be a useful tool for beginning to diagnose the number of underlying psychological processes. This approach has been previously applied in exclusively adult based populations, suggesting that single-process theories of reasoning cannot be ruled out. However, to date it remains unclear whether such results hold across the period of child development. Therefore, the current study aimed to build a database of published developmental reasoning studies and to re-evaluate the data using state-trace analysis, to determine whether they best support the single-process or dual-process accounts. An electronic search of the PsycINFO and Scopus databases was undertaken to obtain empirical studies that have applied dual-process theories to examine reasoning in children or young adolescents (6-15 years). Two screening processes identified 10 papers that provided suitable summary data, forming a database of 78 datasets. State-trace analysis was applied to each dataset. Much of the developmental reasoning data were found to be consistent with a single-process account with one underlying latent variable, thus providing limited evidence for dual-process accounts of reasoning. More targeted experimental design and more stringent statistical tools are recommended for future research, to better understand the cognitive mechanisms underlying reasoning and its development.