Overview
The bar graph has become a mainstay for reporting comparative data. Whether it’s a research report, college paper, or presentation, bar graphs have helped others understand data and analyses for centuries. Even here at Lyngo Lab, the vast majority of our graphs are bar graphs.
Although bar graphs have traditionally been formatted using vertical bars, more recent formats have included bars that run horizontal. Although both display the same information, it remains an open question whether using a vertical vs. horizontal bar graph makes the data easier to understand or affects readers’ perceptions of rigor.
Given our own frequent use of bar graphs, we designed a simple experiment to test whether these two formats affect how understandable, rigorous, or interesting a graph is.
The Experiment
We recruited 1,200 people from the research platform Prolific to take part in a research study. Participants viewed one of two graphs, either a vertical bar graph or a horizontal bar graph, along with accompanying instructions. Unbeknownst to participants, we randomized which bar graph they saw. After viewing the graph, each participant answered a series of survey questions to measure how understandable, rigorous, and interesting the bar graph seemed to them.
Participants were provided a description of the bar graph and the study it illustrated, followed by the bar graph itself (randomly assigned as either vertical or horizontal). This description and the bar graphs are depicted below.
Below is a graph of some survey results about how much men and women enjoy different types of games. 400 people from Amazon MTurk were surveyed. Each was asked "To what extent do you enjoy playing the following types of games? (1 = Not at all, 7 = Very much)." Answer options were on a 1-7 survey scale.
Please look at this graph and answer the questions that follow.
Please look at this graph and answer the questions that follow.
Below is a graph of some survey results about how much men and women enjoy different types of games. 400 people from Amazon MTurk were surveyed. Each was asked "To what extent do you enjoy playing the following types of games? (1 = Not at all, 7 = Very much)." Answer options were on a 1-7 survey scale.
Please look at this graph and answer the questions that follow.
Please look at this graph and answer the questions that follow.
Given that many bar graphs also include lines for margin of error (i.e., error bars), we also randomized whether the bar graph included error bars, along with the following sentence at the end of the description “Error bars represent the margin of error with a 95% confidence interval.” More details on the error bar component of this study, along with the findings, can be found in our Error Bar study report.
After viewing the bar graph, participants were asked three survey questions to measure our outcomes. Answer options were on a 1-7 scale (1 = Not at all, 7 = Very much). The three survey questions were as follows:
After viewing the bar graph, participants were asked three survey questions to measure our outcomes. Answer options were on a 1-7 scale (1 = Not at all, 7 = Very much). The three survey questions were as follows:
- “To what extent is this graph easy to understand?”
- “How scientifically rigorous is this graph?”
- “How interesting is this graph?”
Results
The results revealed no significant differences between the vertical and horizontal bar graph formats, at least on average, for perceived understandability (p = 0.077), rigor (p = 0.604), or interestingness (p = 0.759). Comparisons using our own bar graphs are illustrated below.
Interestingly, adding error bars to visualize the margin of error on each column of the bar graph significantly increased the perceived rigor of the horizontal bar graph, by about 16%, or 0.58 points on a 1-7 scale (p < 0.001). Error bars had no such effect on the vertical bar graph (p = 0.627). An OLS regression analysis showed a significant interaction effect, supporting the findings (p = 0.003). No such effect was found for understandability (p = 0.548) nor interestingness (p = 0.102). For more about the effects of adding error bars to your bar graph, check out our Error Bar study.
Conclusion
It doesn’t seem to matter too much whether you format your bar graphs using a vertical or horizontal orientation. However, if you do use a horizontal bar graph, including margin of error bars should increase the perceived rigor (and actual rigor) of your reporting.
Methods Note
For our analyses, two-sample t-tests were used to test for significant differences in how understandable, rigorous, and interesting each bar graph seemed to be. Ordinary Least Squares (OLS) regression analyses with interaction terms were used to test whether the results differed based on other factors, in this case error bars. Our statistical significance threshold was a p-value below 0.05. The data and survey materials used for this study are available upon request.