Retractions in medicine: the tip of the iceberg
Bibliographic record
Abstract
In 1983, in the aftermath of what was then considered one of the most significant cases of scientific fraud ever, The New York Times reported that 82 papers by John Darsee, formerly of Harvard and Emory, had been retracted (available at https://www.nytimes.com/1983/06/14/science/notorious-darsee-case-shakes-assumptions-about-science.html). That idea persisted: ∼30 years later, Nature said that >80 of Darsee’s papers had been withdrawn.1 In truth, just 17 papers by Darsee have ever been retracted (available at: http://retractiondatabase.org/RetractionSearch.aspx#?auth%3dDarsee%252c%2bJohn%2bR). That may seem surprising, given how high-profile the case was, but we have learned in the decades since that thousands—or even tens of thousands—of papers that should have been retracted have not been. Last year, there were >2300 retractions, up from just 38 in the year 2000 (Figure 1).2 Even accounting for the growth in papers published, the rate has increased dramatically. There are far more eyeballs on papers today, including the eyeballs of sleuths who find image manipulation, plagiarism, duplication suggestive of paper mills, statistical anomalies, and other issues (available at: https://retractionwatch.com/2018/06/17/meet-the-scientific-sleuths-ten-whove-had-an-impact-on-the-scientific-literature/). Take the example of John Carlisle, an anaesthetist whose work spotting data too good to be true, and randomization issues, has led to scores of retractions, including one in the New England Journal of Medicine (available at: https://www.npr.org/sections/health-shots/2018/06/13/619619302/errors-trigger-retraction-of-study-on-mediterranean-diets-heart-benefits).
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Direct model labels (unvalidated)
Per-model category and study-design labels from the labeling rounds. They are machine output, unvalidated, and the disagreement between models ships as data. No study design here is MEDLINE-validated yet.
| Model arm | Categories | Study design | Confidence |
|---|---|---|---|
| gemma | MetaresearchResearch integrity Domain: Evaluation · Genre: Commentary About the Canadian research system: no · About a Canadian topic: no | Not applicable | low |
| gpt | Research integrity Domain: not available · Genre: Commentary About the Canadian research system: no · About a Canadian topic: no | Not applicable | low |
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.076 | 0.389 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.005 | 0.004 |
| Science and technology studies | 0.008 | 0.022 |
| Scholarly communication | 0.019 | 0.022 |
| Open science | 0.004 | 0.008 |
| Research integrity | 0.026 | 0.039 |
| Insufficient payload (model declined to judge) | 0.022 | 0.009 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedLabeled directly by 2 models reading the full record.
The models disagree on parts of this classification; every voice is preserved in the section at the end of the page.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".