Exploring the Garden of Forking Paths in Empirical Software Engineering Research: A Multiverse Analysis

TL;DR

This paper demonstrates how multiple analytical choices in empirical software engineering can lead to vastly different results, emphasizing the need for robustness checks and better justification of methodological decisions.

Contribution

It applies multiverse analysis to an SE study, revealing the impact of analytical decisions on outcomes and proposing a structured classification model for methodological transparency.

Findings

Only 0.2% of analysis pipelines reproduced original results

Most analysis variants produced different or opposite conclusions

Multiverse analysis can improve robustness and reproducibility in SE research

Abstract

In empirical software engineering (SE) research, researchers have considerable freedom to decide how to process data, what operationalizations to use, and which statistical model to fit. Gelman and Loken refer to this freedom as leading to a "garden of forking paths". Although this freedom is often seen as an advantage, it also poses a threat to robustness and replicability: variations in analytical decisions, even when justifiable, can lead to divergent conclusions. To better understand this risk, we conducted a so-called multiverse analysis on a published empirical SE paper. The paper we picked is a Mining Software Repositories study, as MSR studies commonly use non-trivial statistical models to analyze post-hoc, observational data. In the study, we identified nine pivotal analytical decisions-each with at least one equally defensible alternative and systematically reran all the 3,072 resulting analysis pipelines on the original dataset. Interestingly, only 6 of these universes (<0.2%) reproduced the published results; the overwhelming majority produced qualitatively different, and sometimes even opposite, findings. This case study of a data analytical method commonly applied to empirical software engineering data reveals how methodological choices can exert a more profound influence on outcomes than is often acknowledged. We therefore advocate that SE researchers complement standard reporting with robustness checks across plausible analysis variants or, at least, explicitly justify each analytical decision. We propose a structured classification model to help classify and improve justification for methodological choices. Secondly, we show how the multiverse analysis is a practical tool in the methodological arsenal of SE researchers, one that can help produce more reliable, reproducible science.