By Nicolas Robinson and David Trumpower, University of Ottawa
This study explored the effect of a scaffolding exercise designed to make salient the relevance of within-group variation on students’ reasoning during a subsequent informal analysis of variance task. Graduate students (n=20) in an introductory-level statistics course were presented with several datasets that varied with respect to within-group variation and were asked to consider factors that could account for the variance in a scaffolding exercise. Afterwards, students were shown additional datasets that differed with respect to both within and/or between-group variability, and were asked to rate the strength of evidence provided by each dataset in support of a hypothetical theory. Consistent with prior research (e.g. Masnick and Morris, 2008; Obrecth et. al, 2007; Trumpower and Fellus, 2008), many students placed an overemphasis on between-group variability while minimizing the importance of within-group variation. However, the scaffolding exercise was effective in highlighting the relevance of within-group variation for a subset of students.