Loading...
Loading...
Contributors:
Dataset URI:
doi:10.7910/DVN/FJCLVHThis dataset contains the core analysis code and supplementary materials for the study “Beyond Agreement: Handling Measurement Error in LLM-Assisted Content Analysis”. The code implements every estimator, test, and certification rule of the article and the two simulation studies of the supplementary material (S1 and S4). It consists of plain NumPy functions. A user supplies the data loading and the loop over settings, and pastes these functions in. Files: readme.txt: Requirements (Python 3.10 or later, NumPy), a table mapping each function to the section of the article and the table it produces, and a ten-line example. calibration.py: Fits the calibration line, computes the curvature statistic, the regression and inverse estimates, the article-level bootstrap intervals, the equivalence test on the raw bias, and the certificate (Parts I and III of the article). routes.py: Corrects a corpus mean (Route A), a difference between two strata with the invariance test (Route B), and a class proportion (Route C), and runs the compatibility check across instruments (Part II; Tables A2–A4). bridge.py: Builds article values from unit labels, fits the article calibration line, and applies the slope condition of the unit bridge (Part II; Table A5). simulate.py: Generates labeled corpora with a set slope, intercept, curvature, mix, and within-article dependence, and runs one cell of the simulation studies (S1 and S4; Tables A1 and A7). Supplementary Material.pdf: Simulation study (S1), the two empirical studies on NewsMTSC and PerSenT (S2–S3), the simulation behind the slope floor of the weak-instrument rule (S4), the reporting template with a worked example (S5), and the prompts (S6), with Tables A1–A8. The two datasets used in the empirical studies are public (NewsMTSC; PerSenT) and are not redistributed here.
Keyword(s):
MeSH Term(s):
Grants and Funding:
No Grants or Funding Data Available
Associated Publication(s):
No Associated Publication Data Available
Supplementary Data:
No Supplementary Data Available
Related Dataset(s):
No Related Dataset Data Available
The National Library of Medicine (NLM) provides information about biomedical datasets. Inclusion on an NLM Web page does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: NLM Web Policies (nih.gov)