What difference do retractions make? An estimate of the epistemic impact of retractions on recent meta-analyses
Bibliographic record
Abstract
ABSTRACT Every year, several hundred publications are retracted due to fabrication and falsification of data or plagiarism and other breeches of research integrity and ethics. Despite considerable research on this phenomenon, the extent to which a retraction requires revising previous scientific estimates and beliefs – which we define as the epistemic impact - is unknown. We collected a representative sample of recently retracted studies that had been included in recent meta-analyses, and compared the summary effect size of these meta analyses with and without the refracted studies. On average, the retractions had occurred about six years prior to the publication of the corresponding meta-analyses. Our results suggest that retractions have varying impacts depending on their causes. In particular, removing from an analysis a study retracted because of issues with data, methods or results, led to a statistically significant reduction of the estimated effect size. Assuming that the results of these retracted studies are completely false, then the meta-analyses that had included them had overestimated the summary effect sizes by, averaging across effect size metrics, 30% (median, 13%). However, retractions due to plagiarism or other issues not related to data, methods or results had no impact on the conclusions of meta-analyses. Since retractions due to plagiarism or other non-data related issues typically constitute over 75% of total retractions, our results suggest that the epistemic impact of most retractions is likely to be null. However, our results also suggest that retractions due to issues with data, methods or results should be accompanied by a revision of relevant meta-analyses, and by extension a downwards revision of prior scientific beliefs.
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: Empirical About the Canadian research system: no · About a Canadian topic: no | Observational | low |
| gpt | MetaresearchResearch integrity Domain: Evaluation · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Other design | low |
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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".