Replication package for: Misreading Weibo: How Automated Sentiment Tools and Machine Translation Represent Chinese Online Discourse on the Gaza War

Data and code for the article "Misreading Weibo: How Automated Sentiment Tools and Machine Translation Represent Chinese Online Discourse on the Gaza War." The package contains the collection log for 163,168 Weibo posts retrieved with eight conflict-related search terms (1 October 2023 to 31 January 2024), the deduplicated and cleaned corpora (86,052; 81,692; 55,102 posts, with a flag for the 48,832 posts that contain their search term), a 5,000-post stratified sample with English machine translations, both human coders' completed files for a nested 1,200-post sample, the coding protocol, all collection, cleaning, scoring, corpus-linguistic and validity scripts, and every output, table value, and figure reported in the article. Sentiment tools tested: SnowNLP and a HowNet lexicon (Chinese originals), and VADER and TextBlob (English translations). The package is split into two zip files (unzip both into the same folder). Running four scripts in code/08_revision reproduces all results exactly; see README.md. The raw HTML pages saved during collection are available from the authors on request. This version supersedes the anonymized release prepared for peer review. An earlier, unrelated deposit for a previous version of the project (10.5281/zenodo.18005944) was withdrawn and should not be used.