Dissociative identity disorder across languages, regions and time: coded dataset, codebook and analysis registers
Description
De-identified derived data and analysis code for a reliability-audited synthesis of the multilingual DID case literature. Three independently developed large language models applied one frozen coding instrument to 232 full-text sources in 19 languages, yielding 358 analytic records spanning 1882 to 2026, audited by thirteen human coder units across 31 variables. The deposit contains the frozen consensus value for every record, the coding instrument, the source-level covariates and the full-text screening record, the inter-model and human-model reliability results, the registered descriptive and inferential results including all 851 primary and sensitivity rows, the frozen analysis registers, and the analysis code with the environment of the run of record. Source documents are not redistributed, because of copyright and the risk of reproducing identifiable clinical narratives.
Files
Institutions
- Kyoto Institute of TechnologyKyoto, Kyoto