Visualizing vastness: Graphical methods for multiverse analysis

Enlarge

How can research findings be visualized when they are based on hundreds, thousands, or even hundreds of thousands of models?

In their article published in PLOS One, Daniel Krähmer (LMU Munich) and Cristobal Young (Cornell University) discuss the visualization of "multiverse analyses" – analyses that consider not just a few, but all plausible model specifications. Although multiverse analyses have lately gained importance across disciplines, their results have so far been difficult to visualize adequately.

The article addresses this problem: With multiverse plots, which are introduced in the article, any number of model specifications can be displayed without losing important information.

Compared to existing methods, multiverse plots reliably illustrate which empirical conclusions a dataset can support and which researcher decisions drive variation in results. By providing software (Stata and R) for generating multiverse plots, the authors contribute to the advancement of multiverse analysis and the transparency of quantitative empirical research.

Krähmer, D., & Young, C. (2026). Visualizing vastness: Graphical methods for multiverse analysis. PLOS One, 21(2), e0339452. https://doi.org/10.1371/journal.pone.0339452