Individual Submission Summary
Share...

Direct link:

Children’s Cognitive Reflection Predicts Successful Interpretations of Covariation Data

Wed, April 7, 11:35am to 1:05pm EDT (11:35am to 1:05pm EDT), Virtual

Abstract

Cognitive reflection is the tendency to reflect on one's own thinking, allowing a person to identify and correct judgments that are grounded in intuition rather than logic. Individual differences in cognitive reflection predict a diverse range of adult psychological and behavioral outcomes, including conceptual understanding of various scientific domains (e.g., astronomy and mechanics), endorsement for empirically-justified beliefs (e.g., evolution and vaccination), and normatively-accurate evidential reasoning. Recent research using a developmental test of cognitive reflection (the CRT-D; Young & Shtulman, 2020a; 2020b) has found elementary-school-aged children’s cognitive reflection is a strong predictor of the expression and construction of counter-intuitive scientific ideas in biology (e.g., trees are living things) and physics (e.g., air has weight). The present study extends the study of children’s cognitive reflection to another critical aspect of scientific thinking: data-interpretation skills. We asked whether children’s cognitive reflection predicted interpretation of covariation data after adjusting for their age and executive functioning.

Five- to 12-year-olds (N = 79) completed the CRT-D along with measures of inhibitory control (NIH Toolbox Flanker), set shifting (verbal fluency), and working memory (backward digit span). Children additionally judged covariation data presented in 2 x 2 contingency tables (see Figure 1). Prior research with this covariation task suggests elementary-school-aged children widely employ heuristic interpretation strategies that neglect parts of the data (e.g., by comparing just 2 of the 4 cells; Saffran et al., 2016; Saffran et al., 2019). Bayesian regression that adjusted for children’s age and measures of executive function found a positive effect of CRT-D: A 1 SD increase in CRT–D predicted a .37 SD increase in covariation judgement accuracy, 95% HDI [0.13, 0.61]. Further, Bayesian model selection suggested CRT-D performance was the most important variable for out-of-sample predictive performance. If we wanted to predict a new child’s covariation reasoning in the present task, the CRT-D would be more informative than their age or any measure of executive function included in this study.

These findings suggest cognitive reflection is broadly involved in children’s scientific thinking -- cognitive reflection supports data-interpretation skills in addition to domain-specific conceptual knowledge (e.g., Young & Shtulman, 2020a; 2020b). These findings also suggest the relationship between cognitive reflection and the use of normatively-accurate evidential reasoning strategies emerges early in development. We further discuss implications of children’s cognitive reflection for other domain-general scientific procedures (e.g., experimentation skills) and instruction.

Authors