The use of imagery in statistical reasoning by university undergraduate students: a preliminary study

Many students have difficulty in grasping several concepts that are related to the solution of statistical problems. The bibliography reports how the ability of students to solve problems can be affected by the mode of the statistical problem presentation: verbal–numerical and pictorial–graphical. The dual-coding theory predicts that the graphical representation mode should enhance students’ statistical reasoning. Solving these problems requires the building, by the subjects, of a mental model, which in turn relies on visuo-spatial processing. To test this hypothesis we analysed how the ability to solve problems of 473 undergraduate students is affected by the mode of the statistical problem presentation. The study used a quasi-experimental mixed design to explore how the student’s performance is related to visuo-spatial and numerical abilities, statistical expertise, time pressure and problem representation mode (verbal/pictorial). Data analysis, based on the Hierarchical Loglinear Model and then the Logit Model, highlighted that the effect of facilitation, induced by the graphical presentation mode, would seem more likely to occur in inexperienced subjects with high visuo-spatial competence.