Automated Reproducibility Assessments in the Social and Behavioral Sciences Using Large Language Models

MCML Authors

Tobias Holtdirk

→ Group Frauke Kreuter
 Social Data Science and AI

Pietro Marcolongo

→ Group Stefan Feuerriegel
 Artificial Intelligence in Management

Sarah Ball

→ Group Frauke Kreuter
 Social Data Science and AI

Bolei Ma

→ Group Frauke Kreuter
 Social Data Science and AI

Frauke Kreuter

Prof. Dr.

Core PI

Social Data Science and AI

Stefan Feuerriegel

Prof. Dr.

Core PI

Artificial Intelligence in Management

Abstract

Reproducibility in the social and behavioral sciences is typically evaluated by independent researchers who reanalyze the original data to assess whether the published findings can be recovered. However, such approaches are resource-intensive and difficult to scale. Here, we show that large language models (LLMs) can automate reproducibility assessments. Using N=76 published studies with predefined claims from the behavioral and social sciences, we compare LLM-generated analysis with the original findings and human reanalysis. For 7 studies, the LLM could not produce a viable effect size estimate. For the remaining studies, our LLM pipeline recovered the original effect sizes in 41% of studies using a +/-0.05 tolerance in Cohen's d. Further, our LLM pipeline reached the same qualitative conclusion as the original study in 96% of cases, where conclusions indicate whether the reanalysis supports the original claim. For comparison, human reanalysts recovered the original effect sizes in 34% of studies and reached the same qualitative conclusion in 74% of cases. Together, these results show that LLMs can serve as a scalable tool for automated reproducibility assessment and provide a foundation for systematic auditing of empirical results in the social and behavioral sciences.

misc HMS+26

Preprint

Authors

Links

Research Areas

 A1 | Statistical Foundations & Explainability

 C4 | Computational Social Sciences