Skip to main content

Data documentation

Click on each dataset’s name to see its description, license, citation, and link to its original source. Code used to create these data can be found here. The list covers every data family: a table from the competition, nominal, conjoint, rankings or simulated family names its family under the dataset name, with the source argument the packages need to fetch or cite it.

Every dataset here has been modified. IRW does not redistribute source files: each one is reshaped into the IRW data standard — long format, one row per person-item response, with columns renamed and non-response material dropped. The processing script for a given dataset is the record of exactly what was changed, and is linked above. The license shown beside each description is the license IRW redistributes its own extract under, which is not always the source’s. The two are recorded separately and differ on about one row in nine, nearly always where the source deposit named a permission or named nothing and IRW’s copy goes out under CC BY 4.0; the source’s own license is the Original_License field, which travels with the record to irw_meta.biblio. A restriction does carry forward: where the source is non-commercial or share-alike, IRW’s copy keeps the same clause and reuse of that dataset stays bound by it.

Start typing to search across all datasets by name, description, or reference.