By Samin Hemani (University of Illinois)
Information
Statistical thinking rests on “Accepting Uncertainty” (Wasserstein et al., 2019). However, some individuals may fundamentally struggle to accept uncertainty. Tolerance of Uncertainty (TU) is a domain-stable personality trait that is related to but distinct from Openness, one of the “Big 5” personality characteristics. As statistics requires grappling with uncertainty, is TU related to students’ statistics learning? Additionally, does the relationship between students’ TU and their learning change depending on whether the curriculum is more probability-based, or more focused on making sense of numbers (NASEM, 1975)? We recruited 744 students from three sections of an introductory-level general education statistics course; 145 students were taught a traditional curriculum in an in-person section, while the other students were taught a curriculum focused on the concepts of accepting uncertainty, 257 in an online section and 342 in an in-person section. All sections had the same course number, and students were not told of the difference when registering. We compared the relationship between students’ TU as measured by the MSTAT2 assessment, and their performance on exams (α > .70 in both sections). Only in the probability-based curriculum, students with very low TU and very high TU outperform students with moderate TU. These results highlight potential psychosocial determinants related to students’ learning, and raise the question whether high-performing students are able to ‘accept uncertainty’ when thinking statistically.