Revealing undergraduate biology students’ conception of variability and error bars within graphing

Creating and interpreting graphs are quantitative reasoning skills all science, technology, engineering, and math students need to develop. In biology, data are lacking in our understanding of how undergraduate students interpret variability across different graph types or what knowledge students have about the data error bars represent. This study analyzed 3506 student responses within a graphing assessment. We developed a code book for students' open-ended responses on the data error bars represent and used chi-square tests to look for significant differences in how students described error bars based on a graph they created and self-reported demographic data. We observed four major categories in how students described error bars: broad terms, error terms, purpose terms, and trend & analysis terms. When responses were linked with a graph the student made, those who created a bar graph with a bar for each data point showed lower understanding of error bars compared to students who themselves created a bar graph with aggregated means and error bars. We also observed differences in how students described error bars based on their major but did not observe differences based on whether students were in introductory or upper-level courses. Our results suggest that students are not receiving enough instruction on variability within graphing even though bar graphs with error bars are a common graph that students are asked to construct. More scaffolded repeated instruction is needed across biology curricula.