Search protocol, raw API responses and results for an evidence map of industrial ignorance-making: OpenAlex official run of 1 November 2026

This package archives the literature search behind the article "The Other Face of Industrial Ignorance-Making: Did Society's Response Change After Exposure?" (Saadat, 2026; Persian original; English version in preparation). Search: OpenAlex works API, fields title and abstract, publication years 1990-2026, protocol version 7. 175 queries, each with at most 5 boolean operators, at most 1000 records kept per query (sorted by citation count). 24 of the 175 queries reached the cap. 34,709 unique works after merging. Recall checks (details in the article, Section 1.3): 13 of 14 seed sources were found (these were used to tune the queries); 8 of 15 further known sources that played no part in tuning were found (checked by DOI and title on 2 November 2026). The keyword screen included here is not a relevance measure: it kept only 3 of the 13 found seed sources in its top tiers. results.csv contains no relevance judgement. Works without an abstract in OpenAlex can be found only if their title matches the query terms. Contents: protocol.json (all parameters and queries), results.csv, recall_check.csv, recall_known_sources.csv, log.txt, offline screening output, every raw API response (cache_raw_responses/), and the two scripts with the API key removed (scripts/). MANIFEST.txt lists SHA-256 checksums of all files. README.txt explains the structure. (The sentence in README.txt saying that works without an abstract cannot be found is superseded by the sentence above.) An earlier run with a cap of 200 records per query was superseded, and runs under earlier protocol versions were discarded; neither is included. OpenAlex data are released under CC0. The search protocol and scripts were developed with AI assistance (Claude, Anthropic); the author is responsible for the content. Related: version 2 of this record (https://doi.org/10.5281/zenodo.23882692) holds a supplement with post hoc keyword screens for experiments testing industry messages (E4), their candidate lists, the final coding, and the pre-specified decision rule. The files of this version are unchanged.