Is the replication crisis worst in psychology or medicine?

Written by AI. Help improve this answer by adding to the sources section. When the sources section is updated this article will regenerate.

The phrase “replication crisis” is used to describe the growing realization that many published scientific results cannot be reproduced when other researchers try to follow the original methods. The crisis has been discussed most often in two broad domains—psychology and biomedicine/clinical medicine—and the evidence so far suggests that psychology has the larger proportion of non-replicable findings, although medicine has the most high-stakes failures and fraud scandals.

Replication rates and headline numbers

In 2015 the Open Science Collaboration re-ran 100 prominent psychology studies and obtained statistically significant results for only 36 % of them, with effect sizes roughly half of those originally reported [1].
A 2023 review of the psychology literature summarized dozens of large-scale replication projects and estimated that roughly 75 % of headline claims fail to replicate—leading its author to declare that “three out of four claims are false” [2].
Comparable, systematic across-the-board numbers for clinical medicine are scarcer. What exists tends to focus on particular subfields rather than the whole of medicine. Fraud investigations such as the long-running Alzheimer’s β-amyloid case demonstrate that some influential biomedical papers can rest on manipulated data [3], but they do not show the overall base rate of failure.
The lack of large-scale replication exercises in medicine makes a direct numerical comparison difficult. However, where replications have been attempted (e.g., pre-clinical drug studies), success rates have often been well below 50 %, implying that reproducibility problems extend to medicine even if their exact magnitude is still uncertain [5].

Patterns and mechanisms Psychology – Heavy reliance on small samples and flexible analytic choices (“p-hacking”) mean that chance findings can make it into print easily [1][4]. – Replications are comparatively cheap, so large open projects have accumulated precise estimates of failure rates [1].

Medicine – Clinical trials are expensive and involve regulatory oversight, so fewer public replication attempts exist; nevertheless, high-impact fraud or unreplicable biomarker work (e.g., in Alzheimer research) shows that the problem can have direct consequences for patients [3][5]. – Medicine’s incentive structure can put enormous financial value on spectacular but fragile early findings, encouraging selective reporting [5].

Do the sources agree?

Both the 2015 Science consortium and the 2023 Substack essay argue that psychology’s replicability is particularly poor [1][2].
The New York Times opinion piece stresses that biomedical fraud—while not necessarily common—can derail entire therapeutic areas and waste billions [3].
Vox’s overview of scientific fraud treats psychology and medicine as jointly troubled fields, calling for stronger sanctions for data fabrication across the board [5].

Thus, all authors concede that neither field is immune; they differ on whether psychology’s lack of replication (high false-discovery rate) is worse than medicine’s mixture of smaller-scale non-replication plus occasional blockbuster fraud.

Public-discourse timeline (selected milestones) 2015 – Science publishes the Open Science Collaboration’s mass replication of psychology studies, igniting mainstream discussion of the crisis [1]. 2020-2022 – Internal critiques of celebrated psychology effects such as stereotype threat appear, arguing that many results are smaller or nonexistent upon re-analysis [4]. 2023 – Popular outlets summarise the replication literature; one Substack article frames the issue as 75 % false claims in psychology [2]; Vox calls for criminal penalties for repeat fraudsters across scientific fields [5]. 2025 – A New York Times op-ed links a decades-long Alzheimer drug failure largely to fraudulent microscopy images, sparking debate over oversight in biomedical research [3].

Conclusion Based on the best quantitative evidence now available, psychology exhibits the lower replication rate. Medicine, however, remains vulnerable to high-profile frauds that can dominate an entire therapeutic area. Because medicine lacks field-wide replication audits on the scale already carried out in psychology, the true gap may narrow—or even reverse—once equally systematic studies are completed.

Sources

Question

Is the replication crisis worst in psychology or medicine?