“Black-Box” Epidemiology
Bibliographic record
Abstract
To the Editor: We consider the article by Greenland et al1 and the authors’ response2 to the related commentaries to be a very important contribution. However, we think that the following points, not raised in the commentaries,3–6 should be addressed. To propose that research in risk-factor epidemiology should not concern itself with a theoretical component, and thus become only an empiric part of a current or future theory, does not permit one to make a distinction between scientific investigation and ordinary knowledge. To have epidemiologic articles concerned only with the description of methods, analysis, and data would not permit a differentiation between the knowledge generated by such persons as reporters, detectives, and so on, and the knowledge designated as “scientific” by Greenland et al.1 This would mean that the knowledge generated in epidemiologic studies would be considered as lacking scientific content. In the example that was used, to speak of a unifying theory could fall into a certain contradiction. Although the value of the articles with null associations is clearly shown, the example that is cited is based solely on positive associations. This does not take into account that, from the point of view of the falsificationist approach, the evidence against the theory is the most relevant. Although it was not possible to develop a theory on the process of transformation from health to illness, placing replication and precision on the same plane as validity and novelty begs the following question: When would it be pertinent to end the replication of a fact about a topic of investigation in the absence of a theory that explains and predicts a disease? Given that the value of replication and precision is evident, how many studies are necessary and what degree of precision is acceptable to end research on a topic if a theory is never proposed? Finally, to propose that risk-factor epidemiology need not show its use for public health creates the problem of distinguishing between epidemiologic and basic research. Today, various agencies that fund research require that proposed projects include a description of the usefulness of, and benefits to be derived from, the findings. To lose sight of the use of public health studies would mean that one of the most important functions of all epidemiologic investigation—the only discipline of research capable of establishing the relation between environmental factors and the health of human populations—would be lost. Juan Manuel Mejía-Aranguré Epidemiologia Clinica, Hospital de Pediatria, Centro Medico Nacional “Siglo XXI”, Atizapan, Mexico, [email protected] Arturo Fajardo-Gutiérrez Manuel Ortega-Alvarez Clinical Epidemiology, Hospital de Pediatria, Centro Medico Nacional SXXI, Atizapan, Mexico
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 | no category Domain: not available · Genre: Commentary About the Canadian research system: no · About a Canadian topic: no | Not applicable | low |
| gpt | no category 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 | 0.026 | 0.151 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.003 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.003 | 0.008 |
| Scholarly communication | 0.008 | 0.010 |
| Open science | 0.008 | 0.003 |
| Research integrity | 0.032 | 0.045 |
| Insufficient payload (model declined to judge) | 0.012 | 0.007 |
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.
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".