Using Simulations to Ground Statistical Inference in Perception and Action (Poster 8): Division C - Section 2a: Cognitive and Motivational Processes, 2:35 PM

• In Event: e-Lightening Ed-Talk Session 4

Abstract

Computer-based simulations are becoming increasingly prevalent in statistics education. It is important to test whether specific design elements of these simulations have an influence on students’ learning. In the current work, we randomly assign undergraduate students (n = 120) to either a grounded or a standard simulation condition. After the inquiry activities with the simulations, students’ learning of statistical sampling concepts is tested. The results suggest that grounded simulation led to more accurate predictions during the inquiry, however, not to significantly better posttest scores. Moreover, qualitative properties of the students’ drawings (iconic vs non-iconic graphs) during the inquiry were significantly related to the accuracy of their predictions. The students’ verbal explanations, however, were a better predictor of their posttest outcomes.

Authors

• Sebahat Gok, Northwestern University

• Robert Goldstone, Indiana University