Cornell sociologists Cristobal Young and Erin Cumberworth apply 'multiverse analysis'—running all reasonable analytic choices—to disputed social‑science papers. Many famous effects shrink or vanish under this audit, but the piece argues the Regnerus same‑sex parenting study remains robust across specifications. Requiring robustness maps could deter cherry‑picking and clarify where findings are genuinely stable.
— Making multiverse audits a norm would depoliticize contested research by forcing transparent accounting of researcher degrees of freedom before claims enter policy and media.
D. Paul Sullins
2025.08.20
100% relevant
Young and Cumberworth’s multiverse chapter reportedly reexamining Regnerus 2012 and other disputed findings.
Michael Inzlicht
2025.07.30
76% relevant
Heine’s review applies multiple bias-reduction tools (p-curve, z-curve, WAAP‑WLS, selection models, PET‑PEESE) and shows conclusions swing with analytic choices, echoing the call for multiverse-style audits before drawing strong claims in contested literatures.
Lee Jussim
2025.07.01
70% relevant
The piece critiques single‑model testing, unvalidated measures, and lack of adversarial collaboration in a flagship DEI paper, exemplifying the need for robustness audits and multi‑specification checks before claims enter policy and media.
Lee Jussim
2025.06.27
80% relevant
By preregistering and running an adversarial collaboration RRR that flips a celebrated result, the piece exemplifies robust auditing methods for politicized research claims before they shape policy.