{"id":"W4400218600","doi":"10.2196/59641","title":"Large Language Models Can Enable Inductive Thematic Analysis of a Social Media Corpus in a Single Prompt: Human Validation Study","year":2024,"lang":"en","type":"article","venue":"JMIR Infodemiology","topic":"Misinformation and Its Impacts","field":"Social Sciences","cited_by":35,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Institute on Drug Abuse; National Institutes of Health; National Cancer Institute; Research to Prevent Blindness","keywords":"Social media; Computer science; Thematic analysis; Thematic map; Natural language processing; Psychology; Linguistics; Artificial intelligence; Sociology; World Wide Web; Geography; Qualitative research; Cartography","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001599744,0.00009363137,0.0003772448,0.0005751911,0.0001322837,0.00004140499,0.0001594455,0.0001453867,0.0002234006],"category_scores_gemma":[0.0004218456,0.0000817339,0.00008257377,0.001269714,0.0001054488,0.0003432583,0.00005222352,0.0001708005,0.000006332595],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001819332,"about_ca_system_score_gemma":0.0001446766,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001521751,"about_ca_topic_score_gemma":0.01543106,"domain_scores_codex":[0.9984309,0.0004424974,0.0004655969,0.0001490619,0.0002346424,0.0002773131],"domain_scores_gemma":[0.9993144,0.0002709206,0.0001745205,0.0001147256,0.00007123517,0.00005422856],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.000007351431,0.0002063121,0.002911046,0.00002355434,0.0001325613,0.000002557433,0.9389111,0.0003950862,0.0002224136,0.05657697,0.0001323198,0.00047874],"study_design_scores_gemma":[0.001054516,0.0002608974,0.05181595,0.00006730851,0.0004427664,8.284747e-7,0.8923174,0.02222824,0.0002133613,0.0309096,0.0002850564,0.0004040482],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9873239,0.00002233338,0.0001481993,0.0002156641,0.00007147279,0.0005239933,0.00002763644,0.00006382597,0.01160293],"genre_scores_gemma":[0.9996188,0.000001863203,0.00003217872,0.00007559291,0.00005401696,0.00003800745,0.00004413212,0.000005437861,0.0001300198],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04890491,"threshold_uncertainty_score":0.8610894,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08916595273790014,"score_gpt":0.405704456337404,"score_spread":0.3165385035995039,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}