Open AccessThe impact of visualization on flexible Bayesian reasoning
Author(s) -
Katharina Böcherer-Linder,
A. Eichler,
Markus Vogel
Publication year - 2017
Publication title -
avances de investigación en educación matemática
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.299
H-Index - 2
ISSN - 2254-4313
DOI - 10.35763/aiem.v1i11.169
Subject(s) - computer science , bayesian probability , visualization , bayesian statistics , graph , bayesian network , base (topology) , unit square , artificial intelligence , machine learning , square (algebra) , data science , bayesian inference , data mining , theoretical computer science , mathematics , mathematical analysis , geometry , combinatorics
There is wide consensus that visualizations of statistical information can support Bayesian reasoning. This article focusses on the conceptual understanding of Bayesian reasoning situations and investigates whether the tree diagram or the unit square is more appropriate to support the understanding of the influence of the base rate, which is introduced as being a part of flexible Bayesianreasoning. As a statistical graph, the unit square reflects the influence of the base rate not only in a numerical but also in a geometrical way. Accordingly, in two experiments with undergraduate students (N = 148 and N = 143) the unit square outperformed the tree diagram referring to the understanding of the influence of the base rate. Our results could inform the discussion about how to visualize Bayesian situations and has practical consequences for the teaching and learning of statistics