Raincloud Plots: Innovative Data Visualizations in JASP

The latest release of JASP offers a new feature for data visualization: Raincloud plots. Invented and promoted by Micah Allen and colleagues (Allen et al., 2021), the raincloud plot combines a cloud of points with a box plot and a one-sided violin plot. This combination allows the viewer to obtain a wealth of information about a variable from a single plot. Moreover, multiple “rainclouds” can be displayed next to each other to visualize the difference between levels of a grouping variable. For dependent samples, lines that connect repeated measurements of each case can be added as well.

How to Create Raincloud Plots in JASP

In JASP 0.15, raincloud plots can be created within two modules: The T-Tests and ANOVA module. The ANOVA module can be used when a grouping variable has more than two levels. Here, we focus on the T-Tests module. We first load the “Eye Movements” data set (Matzke et al., 2015) from the JASP Data Library. The data set contains the number of recalled words after subjects either fixated on a dot in the center of a screen (“Fixation”) or performed horizontal saccades (“Horizontal”). We open the T-Tests module and select the Independent Samples T-Test. Then, we drag the variable “CriticalRecall” into the “Dependent Variable” window and specify “Condition” as the “Grouping Variable”.

Now, we can tick the box “Raincloud plots” in the menu underneath to create the raincloud plot. By default, JASP displays the rainclouds with vertical orientation (horizontally next to each other).

For those who prefer the original, horizontally oriented rainclouds, we can tick the box “Horizontal display” underneath:

To illustrate the advantages of raincloud plots, we can compare them to the ordinary descriptives plots in JASP. We tick “Descriptives plots” and get the following output:

The descriptives plot suggests that the subjects in the “Horizontal” group remembered fewer words than the “Fixation” group (the confidence intervals are barely overlapping). The significant p-value of the t-test supports this suggestion. However, the raincloud plot reveals that the distribution of the data differs between the groups, such that a nonparametric analysis may be more appropriate (or at least is wise to report).

Other Raincloud Plot Options

Beyond the “Independent Samples T-Test”, JASP offers raincloud plot options for several other analyses. In the T-Tests module:

  • Select the “Paired Samples T-Test” and tick the “Raincloud plots” box to produce raincloud plots where the points of each pair are connected with lines
  • For “Paired Samples T-Tests”, the option “Raincloud difference plots” creates rainclouds of the difference scores of each pair
  • The Bayesian t-tests offer the same raincloud plot options as the classical t-tests

In the ANOVA module:

  • Select the “ANOVA” and insert fixed factor variables into the “Horizontal Axis” and “Separate Plots” windows in the section “Raincloud Plots” (under “Descriptive Plots”)
  • The “ANCOVA” analysis offers the same options as the “ANOVA”
  • For the “Repeated Measures ANOVA”, select repeated measures factors instead of the fixed factors for “Horizontal Axis” and “Separate Plots”
  • Again, the Bayesian ANOVA analyses offer the same same raincloud plot options as the classical ones

References

Allen, M., Poggiali, D., Whitaker, K., Marshall, T. R., van Langen, J., & Kievit, R. A. (2021). Raincloud plots: A multi-platform tool for robust data visualization. Wellcome Open Research, 4:63. https://doi.org/10.12688/wellcomeopenres.15191.2

Matzke, D., Nieuwenhuis, S., van Rijn, H., Slagter, H. A., van der Molen, M. W., and Wagenmakers, E.J. (2015). The effect of horizontal eye movements on free recall: A preregistered adversarial collaboration. Journal of Experimental Psychology: General, 144, e1-e15. https://doi-org.eur.idm.oclc.org/10.1037/xge0000038

About The Authors

Malte Lüken

Malte is a PhD candidate at the Centre for Contextual Psychiatry and Mind Body Research Group at KU Leuven.