Extracting Conditional Proportions from Bar Graphs

Abstract

jats:pBar graphs are often touted as a presentation format that improves reasoning when faced with conditional probability scenarios. However, research suggests that extracting information from bar graphs in order to make conditional proportion comparisons may be challenging. Furthermore, the configuration of bar graphs can influence their underlying message. The purpose of this paper is to explore two undergraduate students’ reasoning when they are asked to select appropriate bar graph representations to answer comparison questions that they posed, involving two categorical variables. Results from this small pilot study suggest that there was some confusion in deciphering the comparison of interest, and bar graph configuration contributed to conflict in conclusions drawn.</jats:p>