Experimental evaluation of students' performance in judging statistical visualizations

Arendt, A. (2016) Experimental evaluation of students' performance in judging statistical visualizations.

Abstract:Constant misinterpretation of the p-value and the absence of an assessment of the underlying assumptions of statistical models have led to alarmingly low replicability in, among others, the fields of psychology and ecology. A suggestion by Zuur, Ieno and Elphick (2010) was that any student who had a standard statistical education could make inferences about the underlying assumptions with the help of good visualizations. In this research we have focused on the assumptions of normality of residuals and homogeneity of variance. An experiment was conducted where participants were asked to assess normality and homogeneity of variance with the help of 100 histograms or 100 conditional boxplots, respectively. The participants received feedback after each trial. The sample consisted of 33 Dutch and German students between the age of 19 and 32, who had their statistical education at the University of Twente. The results did not meet the expectations. The objective measures which the participants should have extracted from the visualizations did not influence the participants’ response and there was great variation in how the individual stimuli influenced the response. Also, the feedback did not elicit a learning effect. These findings are discussed with respect to the design of the stimuli, the experiment and the education of the participants. It is concluded that it is unlikely that both plots did not convey any meaningful information and an advice for fine-tuning the experiment is given. The possibility that the current practice in education caused the low performance is proposed together with possible alternatives.
Item Type:Essay (Bachelor)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:77 psychology
Programme:Psychology BSc (56604)
Link to this item:http://purl.utwente.nl/essays/70064
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page