Bibliographic record
Abstract
Missing data are the bane of all clinical research. With the possible exception of the CAPRIE trial,1 in which the investigators went to extraordinary lengths that enabled them to followup 99.8% of their 19 000 participants over two years, it is highly unusual for a study to end with complete data on all subjects. There are many reasons for this: a person may omit an item on a questionnaire or refuse to complete it entirely; a vial of blood may be dropped or the analyser fail to function one day; or a participant may not appear for his or her appointment. Longitudinal studies (those that follow participants over time) can be subject to all of these mishaps, but now the problem is magnified in that they could happen at each of the assessment sessions; in addition to which, participants may drop out of the study entirely before all the data are collected. Furthermore, the more sophisticated, multivariable statistical techniques that use two or more variables in the same analysis, such as multiple regression or factor analysis, make the problem even worse, in that most of them require complete data for all of the subjects. If a person is missing one variable out of the, say, 10 that are being analysed, then that subject is dropped entirely from the analysis. Simulations have shown that if as little as 10% of the data is missing, as many as 60% of the subjects could be eliminated.2 One consequence of missing data is a loss of statistical power: differences between groups that would be statistically significant may no longer be so because of the reduced sample size. It may appear at first glance that this is a relatively simple problem to remedy—begin with more people to allow for missing data, or recruit more …
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 | 0.401 | 0.772 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.005 | 0.003 |
| Bibliometrics | 0.009 | 0.012 |
| Science and technology studies | 0.004 | 0.019 |
| Scholarly communication | 0.007 | 0.018 |
| Open science | 0.012 | 0.012 |
| Research integrity | 0.008 | 0.011 |
| Insufficient payload (model declined to judge) | 0.031 | 0.005 |
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; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
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".