Learning statistics with interactive pictures using R Shiny: Generally preferred, but not generally advantageous

Abstract

Constructing interactive web apps has become more accessible for instructors, for example, by using the R package Shiny. Here we explored learners' preferences and the efficiency of interactive simulations versus static pictures in acquiring statistics knowledge of Cohen's d and standard normal distribution.Results revealed that students' spontaneous interaction with pictures was infrequent (pilot study, N = 26). While prompts (Exp. 1, N = 152) effectively ensured the manipulation of simulations, student exposure to interactive simulations led to longer learning times though similar test performance compared with student exposure to static pictures. Multiple interactive representations led to lower test performance than single interactive and static representations (Exp. 2, N = 117). Though no advantage was gained regarding learning outcomes, participants preferred the interactive variant (Exp. 3, N = 119). Taken together, this study demonstrates that the superiority of interactive pictures cannot be assumed to hold in general. Further work should evaluate how mental model construction can be effectively scaffolded by interactive simulations.

Choosing and using data and contexts for learning

“…The final paper in this issue [7] reports on research comparing interactive web apps and static pictures with respect to student learning and perception. No matter what description of the full statistical investigation process is preferred, the final component (before starting the cycle again) is discussion and the bringing together of investigation results or analysis within context .…”

“…No matter what description of the full statistical investigation process is preferred, the final component (before starting the cycle again) is discussion and the bringing together of investigation results or analysis within context . The discussion in context in [7] is as important as any formal analysis, setting out the nuances of most interest to all wishing to learn from the research study. These include discussion of learning outcomes, length of time students spend and student perceptions.…”