^{1}

^{1}

^{‡}

^{2}

^{3}

^{1}

^{2}

^{‡}

The authors have declared that no competing interests exist.

‡ These authors are joint senior authors on this work.

Data presentation for scientific publications in small sample size studies has not changed substantially in decades. It relies on static figures and tables that may not provide sufficient information for critical evaluation, particularly of the results from small sample size studies. Interactive graphics have the potential to transform scientific publications from static reports of experiments into interactive datasets. We designed an interactive line graph that demonstrates how dynamic alternatives to static graphics for small sample size studies allow for additional exploration of empirical datasets. This simple, free, web-based tool (

This article examines the potential for interactive graphics to transform scientific papers from static publications into interactive datasets and provides a web-based tool for creating interactive line graphs.

Scientific and technological advances have enhanced our ability to study the biology of health and disease. They have also changed the way that we access and share scientific information. Study preregistration websites, data repositories, reporting guidelines and recommendations, and checklists for statistical analysis are all designed to promote transparency and enhance the reproducibility of scientific results. Data presentation for scientific publications has not changed substantially, however, despite this growing emphasis on transparency and reproducibility. Scientists rely on static figures and tables that may not provide sufficient information for critical evaluation, particularly of the results from small sample size studies.

This paper aims to explore the potential of interactive graphics to transform scientific publications from static reports of an experiment into interactive datasets narrated by the authors. Small sample size studies offer excellent opportunities to explore interactive visualizations, as small datasets generally rely on a few key types of figures. These studies commonly use bar and line graphs that show summary statistics for continuous data and scatterplots that examine the relationship between two variables. Offering interactive alternatives to these static graphs may be a simple and effective strategy for promoting widespread use of interactive graphics. We have designed and present an interactive line graph as an alternative to the static graph for small sample size studies that allows for additional exploration of empirical datasets. In addition to demonstrating the overall concept, this simple, web-based tool may encourage utilization of interactive graphics and address growing demands to show individual-level data [

A recent systematic review of original research articles published in top physiology journals demonstrated that 61% of papers contain at least one line graph, making this the second most common type of figure used to present continuous data [

Panels A–C use traditional line graphs to present a simulated dataset as mean and standard error (Panel B) or mean and standard deviation (Panels A and C). While Panels A and C clearly indicate that there is overlap between groups, it difficult to assess the magnitude of the overlap. The error bars for Groups 2 and 3 overlap, while those for Group 1 go in the opposite direction. Panels D–F show selected figures that were created using our web-based tool for making interactive line graphs. Readers can view the interactive versions by uploading

The common practice of displaying summary statistics can be misleading, as many different data distributions can lead to the same graph (

The line graph (mean ± standard error) provides no information about whether changes are consistent across individuals (Panel A). The scatterplots shown in the Panels B–D reveal very different patterns of change, even though the means and standard errors differ by less than 0.3 units. The lower scatterplots showing the differences between measurements allow readers to quickly assess the direction, magnitude, and distribution of the changes. The solid lines show the median difference. In Panel B, values for every subject are higher in the second condition. In Panel C, there are no consistent differences between the two conditions. Panel D suggests that there may be distinct subgroups of “responders” and “nonresponders.” Adapted from Weissgerber et al. [

While several alternatives to the line graph have been proposed [

Interactive line graphs may provide additional information needed to interpret longitudinal data in small studies. We developed a simple, free, web-based tool (

View different summary statistics: the base graph shows the central tendency and variation in each group for each condition or time point. The user can adjust the graph to view the mean, mean and standard deviation, mean and standard error, mean and 95% confidence interval, median, median and interquartile range, or median and range. Measures of variation for each group are shown as a semitransparent shaded region, allowing one to assess the magnitude of the overlap among observations from different groups.

Display lines for some or all individuals in each group: the line for each participant or sample in the dataset can be turned on or off individually, allowing one to view any subset of individuals in the dataset.

View a subset of groups, conditions, or time points: these options allow the viewer to focus on a subset of groups, conditions, or time points.

View change scores for any two conditions or time points: the “Difference Plot” tab displays a univariate scatterplot that shows change scores for each individual in the dataset. This allows for comparisons of the magnitude, direction, and consistency of changes across groups.

Interactive line graphs can be quickly created using a web-based application that does not require any programming expertise or specialized skills—users simply enter or upload data and customize the graph axes and labels. The insight gained from an interactive line graph will depend on the empirical dataset. In addition to enhancing readers' understanding of the data, the interactive line graph may help authors to select static graphs that most effectively illustrate key findings for print publication. A simulated dataset is provided to illustrate these points (

The tool allows for both (1) the integration of static graphics into a publication as a .tiff file and (2) downloading of a data file for a customized interactive graphic, which can be presented in the paper supplement. As color coding is used to present different groups, the tool includes a color blind working mode. All interactive line graph features can be viewed in a color blind–safe color scheme. A black-and-white mode is also included for less complex graphs.

A recent editorial highlighted the static nature of data presentation as a major limitation of scientific publications [

This paper presents a "proof of concept" example that demonstrates how interactive alternatives to static graphics for small sample size studies allow for additional exploration of empirical datasets and illustrates the types of tools that are needed to promote widespread use of interactive graphics. The principles described above can be applied to other types of figures and tables, including those applicable to big datasets. Most scientists use electronic devices to access scientific publications, yet the interactive potential of these technologies remains untapped. Exploring more dynamic alternatives is crucial as we enter an era of transparent and open science.

This example can be viewed by uploading S1 Data into the web-based tool (

(XML)

Panel A: Data for each individual in Group 2 from the dataset presented in

(TIF)

The response for each individual is presented as the change from the baseline value. Horizontal small multiples are used to highlight differences in the magnitude of the response among individuals.

(TIF)

Panel A: Individual responses for the dataset shown in

(TIF)

Changes in placental growth factor were examined longitudinally in women who had normotensive pregnancies (

(TIF)

(DOCX)