Recovering the raw data behind a non-parametric survival curve
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
BACKGROUND: Researchers often wish to carry out additional calculations or analyses using the survival data from one or more studies of other authors. When it is not possible to obtain the raw data directly, reconstruction techniques provide a valuable alternative. Several authors have proposed methods/tools for extracting data from such curves using a digitizing software. Instead of using a digitizer to read in the coordinates from a raster image, we propose directly reading in the lines of the PostScript file of a vector image. METHODS: Using examples, and a formal error analysis, we illustrate the extent to which, with what accuracy and precision, and in what circumstances, this information can be recovered from the various electronic formats in which such curves are published. We focus on the additional precision, and elimination of observer variation, achieved by using vector-based formats rendered by PostScript, rather than the lower resolution image-based formats that have been analyzed up to now. We provide some R code to process these. RESULTS: If the raster-based images are available, one can reliably recover much of the original information that seems to be 'hidden' beneath published survival curves. If the original images can be obtained as a PostScript file, the data recovered from it can then be either input into these tools or processed directly. We found that the PostScript used by Stata discloses considerably more of the data hidden behind survival curves than that generated by other statistical packages. CONCLUSIONS: When it is not possible to obtain the raw data from the authors, reconstruction techniques are a valuable alternative. Compared with previous approaches, one advantage of ours is that there is no observer variation: there is no need to repeat the digitization process, since the extraction is completely replicable.
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.
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 | 0.055 | 0.169 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.011 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.005 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.003 |
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 it