Presenting your data
Presenting your data
Data presentation is the foundation of our collective scientific knowledge, as readers’ understanding of a dataset is generally limited to what the authors present in their publications. Figures are critically important because they often show the data that support key findings (Weissgerber 2015).
Unfortunately, authors generally use simple graphs to present summary statistics, instead of providing detailed information about the distribution of the data or showing the full data. In addition, digital images are not to be considered as just nice illustrations, but are underlying data and should be treated as such (Cromey 2013). As figures are the main method of giving insight into the results of your work, researchers should strive to present their data faithfully and transparently.
ALLEA Code:
- Researchers share their results in an open, honest, transparent, and accurate manner, and respect confidentiality of data or findings when legitimately required to do so.
- Researchers report their results and methods, including the use of external services or AI and automated tools, in a way that is compatible with the accepted norms of the discipline and facilitates verification or replication, where applicable.
Reporting bias
Researchers not able to reproduce or replicate previous results are often not inclined to further pursue their investigation nor to publish the findings. Nevertheless, the resulting reporting bias might influence the correctness of the scientific literature (publication bias), with potentially canonization of false facts (Nissen et al 2016). In addition, not having the full picture might distort the conclusions that can be drawn from meta-analyses and systematic reviews and as such may lead to a biased view which in turn might impact policy decisions.
Who is involved?
Junior Researcher - Phd student
In most cases analysis and presentation of the data will be in the hands of the researchers that have collected the data.
Senior Researcher
Supervisors and promotors are responsible for checking the integrity of the collected data and making sure that the analysis, presentation and conclusions of the research are faithful to the obtained data.
Publishers should have clear journal policies in place with regards to the presentation of data into figures. In addition, these guidelines should not only be available but also be enforced.
Peer Reviewer - Editor
Peer reviewers should be sufficiently critical regarding the figures, look out for potential pitfalls and propose alternative presentation methods.
When to think about this?
Thinking about data presentation is most relevant during the research phase with regards to correct data management and data analysis, during the publication process when preparing the figures and reviewing the data, and during post-publication, for example in relation to questions related to the validity of the presented data.
Illustrating research with graphs
Although it has become standard practice to illustrate continuous data using bar graphs, thereby presenting the data using summary statistics as averages and deviations, researchers are advised not to use this approach and instead look for other ways to present the data, including its distribution, especially when the results are based on a low number of observations.
When generating graphs, you should ask yourself whether the presentation accurately presents the research findings and does not in any way mask particularities that could affect the way the data is perceived. The data should be convincing by itself, not because of the presentation!
Some good practices when presenting (continuous) data:
- Try to provide as much data as possible when presenting data allowing others to interpret the distribution of the data.
- Avoid bar graphs for continuous data, especially when working with low n-values. Examples of alternative graph presentation are dot and/or box plots (examples can be found in the Weissgerber 2019 paper).
- Show outliers, indicate whether or not these have been included in the statistical analysis and explain why.
- In addition to listing the different statistical tests used in the methods section, please also indicate the test used for each particular dataset.
The figure below illustrates that different data distributions can lead to the same bar graph. Having access to the full data may, however, suggest different conclusions from the summary statistics (Weissgerber 2015).
Weissgerber TL, Milic NM, Garovic VD (2015) Beyond Bar and Line Graphs: Time for a New Data Presentation.
This illustrates the importance of having access to the full data set, as this will show potential outliers or unequal distribution of the data, which might in turn affect the statistical analysis and/or the perception of the data.