Reply to Haynes and to Wang
Bibliographic record
Abstract
from the GLI upon request, the full dataset from which the GLI Global reference equations are derived is not.Now that the GLI Global reference equations are formally recommended by the ATS, the public unavailability of the full dataset is of heightened concern. The Reasoning behind Key Decisions That Inform the GLI Global Reference Equations Is OpaqueTogether with the GLI's application of advanced modeling techniques, the size and diversity of its dataset position the GLI as a leading authority on lung function.But the reasoning behind key decisions that influence the GLI equations is opaque.Consider the decision by Quanjer and colleagues in 2012 to exclude all 5,476 observations from the Indian subcontinent and 6,137 observations from Iran; the rationale given was that the data from India, Pakistan, and Iran "did not join well" or "could not be fitted into any group" (6).Or consider the decision by Bowerman and colleagues, when creating the GLI Global reference equations in 2022, to apply extreme sample weights to observations, such that the contribution of an African American woman is weighted more than 15 times the contribution of a Venezuelan (or an Algerian or Israeli) woman (2).Although these and other decisions may well be defensible, the GLI's authority and influence are such that there now needs to be an avenue or a mechanism for greater community dialogue and input, ideally before the adoption of new guidance.The importance of transparency is amplified because of the ATS endorsement of the GLI approach.I recognize that these concerns are thorny, with no obvious solution available at present.They in no way detract from the magnitude and importance of the decision by the ATS to move away from race-specific equations.I hope that by raising these concerns, they may highlight the urgent need for working on the next iteration of non-race-specific spirometry reference equations, one that carefully and explicitly considers its intended end users, that is based on publicly available data, and that is arrived at through an open and transparent process.
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
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 | 0.015 | 0.121 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.004 | 0.008 |
| Scholarly communication | 0.009 | 0.016 |
| Open science | 0.004 | 0.005 |
| Research integrity | 0.035 | 0.056 |
| Insufficient payload (model declined to judge) | 0.023 | 0.017 |
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, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
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".