Adjustments to the signed likelihood root and analysis of an embedded experiment in a survey
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
This thesis consists of two projects. The first project is to develop an adjustment to the signed likelihood root (r) so that the normal approximation to the distribution of the adjusted r is improved. By using Taylor series expansions, we have developed an additive adjustment to r, which leads to a second-order approximation to its distribution. The theory is developed, simulations are recorded to indicate repetition accuracy, real data is analyzed, and connections to alternatives are discussed. The second project is dedicated to the analysis of an embedded experiment in a survey. We derive the Horvitz-Thompson estimator of the average treatment effect and its variance for a general design. Five estimators of corresponding variance are proposed and examined under a design combination of simple random sampling without replacement and completely randomized design. In the presence of auxiliary information, a new model-assisted estimator for the average treatment effect is developed and the variance of the estimator is derived. We show that the new estimator is approximately design-unbiased when a general model is employed to incorporate the auxiliary information. Moreover, it doesn't require auxiliary variable information at the population level and is relatively easy to implement and compute. Simulations carried out indicate that the new estimator gains in efficiency and its relative bias is negligible. Reliable variance estimators based on simulation experiments are suggested. The method proposed is applied to a synthetic data provided by Statistics Canada with multiple treatments under a design combination of stratified random sampling and randomized block design.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| 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 it