{"id":"W4393098406","doi":"10.2308/isys-2023-023","title":"Measuring Corporate Human Capital Disclosures: Lexicon, Data, Code, and Research Opportunities","year":2024,"lang":"en","type":"article","venue":"Journal of Information Systems","topic":"Financial Reporting and Valuation Research","field":"Business, Management and Accounting","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Lexicon; Code (set theory); Computer science; Accounting; Human capital; Natural language processing; Business; Data science; Programming language; Economics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003742411,0.0003394712,0.0002982151,0.006725632,0.0006030236,0.002732561,0.0004795182,0.000517484,0.005009476],"category_scores_gemma":[0.04168209,0.0002135363,0.0002764598,0.009087093,0.00072249,0.002316755,0.001566537,0.0007215329,0.002012961],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001733957,"about_ca_system_score_gemma":0.002074827,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006044665,"about_ca_topic_score_gemma":0.006281646,"domain_scores_codex":[0.9956487,0.001769111,0.0007461434,0.0004338669,0.001209316,0.0001929259],"domain_scores_gemma":[0.9657611,0.02323776,0.003570928,0.002411351,0.004608001,0.0004108222],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004069129,0.0003956616,0.350657,0.001715514,0.00008491302,0.0005900831,0.004294632,0.01151305,0.005372552,0.04398409,0.2392718,0.3417137],"study_design_scores_gemma":[0.0001619216,0.0002175363,0.3384934,0.001624234,0.000142832,0.00130062,0.007636936,0.1386985,0.01909217,0.05965701,0.4327205,0.0002542884],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.6447782,0.001294273,0.06878348,0.004317172,0.0003407225,0.001210907,0.2122658,0.005420181,0.06158925],"genre_scores_gemma":[0.7951924,0.0007375577,0.07301397,0.0005603706,0.0001257942,0.001248786,0.1239055,0.000515362,0.004700287],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.006725632,"threshold_uncertainty_score":0.01979202,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5076878867179168,"score_gpt":0.3959824706749596,"score_spread":0.1117054160429572,"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."}}