Research Choices Can Undermine Social Science Credibility

A fresh book argues that seemingly minor decisions made during the research process—such as how data is analyzed or which questions are asked—can significantly erode the trustworthiness of social science findings. Cornell political scientist, Robb Willer, and Yale statistician, Molly Roberts, detail these concerns in their forthcoming book, according to reporting from the Cornell Chronicle.

The authors contend that a lack of methodological rigor in social science research contributes to a broader “credibility crisis” facing the field. This crisis, they argue, stems not from intentional fraud, but from practices that, even as common, introduce bias and reduce the reliability of results. The book specifically calls for social scientists to adopt a “multiverse” analysis, a robust approach to data analysis that acknowledges the multitude of analytical choices researchers make and assesses the impact of those choices on the conclusions reached.

This call for increased methodological scrutiny arrives as broader questions about the reliability of scientific research gain traction. A recent article in asteriskmag.com questioned whether social science can be trusted, highlighting concerns about reproducibility and the influence of researcher bias. The article points to a growing awareness of the need for greater transparency and accountability within the scientific community.

The “multiverse” approach advocated by Willer and Roberts involves systematically exploring a range of analytical options and reporting the results of each. This contrasts with the traditional practice of selecting a single analytical path and presenting only those findings. By acknowledging the inherent subjectivity in research decisions, the authors believe social scientists can build greater confidence in their conclusions.

The need for evidence-based social science is also underscored by a recent publication in Nature, which examines the “why, what, and future implications” of the approach. This article emphasizes the importance of grounding social science research in empirical evidence and utilizing rigorous methodologies to ensure the validity of findings.

Concerns about trust in science extend beyond social science, impacting fields like conservation science. A review published by Frontiers highlights the complex interplay between credibility and trust, noting that public perception of scientific authority is influenced by factors beyond the scientific merit of the research itself. The review suggests that effective communication and engagement with stakeholders are crucial for building and maintaining trust in science.

As of March 22, 2026, neither the American Psychological Association nor the National Science Foundation have issued statements responding to the book’s arguments or outlining plans to promote “multiverse” analysis among researchers.

Willer and Roberts’ book is scheduled for release in May 2026.