Comparing Psychology Undergraduates Performance in Probabilistic Reasoning Under Verbal-Numerical and Graphical-Pictorial Problem Presentation Format: What is the Role of Individual and Contextual Dimensions?

Abstract

This study aims to investigate about the existence of a graphical facilitation effect on probabilistic reasoning. Measures of undergraduates' performances on problems presented in both verbal-numerical and graphical-pictorial formats have been related to visuo-spatial and numerical prerequisites, to statistical anxiety, to attitudes towards statistics and to the confidence in response correctness. Psychology undergraduates in Italy and Spain with no statistical expertise (N= 676) completed a protocol under conditions of presence versus absence of time pressure. Hierarchical linear regressions and ANCOVAs with mixed design have been carried out separately for each sample. The best predictor of performance in both formats has been the confidence in solution correctness under the condition of time pressure administration, which seemed to promote the commitment to the task. The findings suggest that the eventual occurrence of a graphical facilitation could be the result of a multifactorial interaction among contextual and individual dimensions, rather than being strictly related to the problem presentation format.

What’s about the Calibration between Confidence and Accuracy? Findings in Probabilistic Problems from Italy and Spain

“…Nevertheless, concerning the assessment of the Bias Index, we note several specificities. In particular, consistent with a previous work (Agus et al, 2015a(Agus et al, , 2015b in which it was highlighted that the performance in probabilistic reasoning was improved in Italian undergraduates working in time pressure, we speculate that the Bias of calibration improves in the presence of time pressure, specifically in the problems in the Graphical format. This effect is highlighted only for Italian undergraduates and not for Spanish undergraduates.…”

“…The confidence in the accuracy of a response has been recognized as a relevant and strong predictor of both achievement (Stankov, 2013;Stankov, Lee, Luo, & Hogan, 2012;Tempelaar, 2009) and probabilistic reasoning (Agus et al, 2015b).…”

“…To investigate these aspects, following Schraw's advice 2009, two measures have been computed to assess the calibration between confidence and accuracy: one index of the Absolute Accuracy (the Bias index) (Boekaerts & Rozendaal, 2010; and one index of the Relative Accuracy (Confidence-Judgment Accuracy Quotient; CAQ) (Jackson & Kleitman, 2013;Shaughnessy, 1979). All indices were computed on the set of items on probabilistic reasoning in the Numerical format and, separately, on the set of items in the Graphical format (Agus et al, 2015b).…”

“…All these dimensions were examined separately for Italian and Spanish undergraduates in psychology; indeed, we assume that Italian undergraduates' calibration indices might differ from the Spanish ones in conditions of time pressure. In fact, in the literature it has been suggested that Italian students show better performances in probabilistic reasoning in presence of time pressure, when compared to Spanish students (Agus et al, 2015b).…”

“…We independently analysed the samples of Italian and Spanish undergraduates, extracted from different populations. The protocols were assessed in relation to the potential effect of order and sequences, but no significant effect was identified in the Italian and Spanish universities (Agus et al, 2015a(Agus et al, , 2015b.…”