{"id":"W3158739288","doi":"10.1371/journal.pone.0264270","title":"Avoiding bias when inferring race using name-based approaches","year":2022,"lang":"en","type":"article","venue":"PLoS ONE","topic":"Names, Identity, and Discrimination Research","field":"Social Sciences","cited_by":54,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"Fonds National de la Recherche Luxembourg; Canada Research Chairs","keywords":"Inference; Race (biology); Imputation (statistics); Computer science; Data science; Racial bias; Econometrics; Artificial intelligence; Machine learning; Mathematics; Sociology; Missing data","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.002443268,0.00007122612,0.0001360366,0.0002039413,0.002924214,0.0002251987,0.0003179506,0.00003452244,0.00137636],"category_scores_gemma":[0.001459435,0.00008529867,0.00005290015,0.0003528815,0.0001360906,0.0002465906,0.0001865195,0.0003610918,0.00002480213],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003619035,"about_ca_system_score_gemma":0.0003491872,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003044803,"about_ca_topic_score_gemma":0.0005194689,"domain_scores_codex":[0.9971963,0.0008261462,0.0001505131,0.0002118988,0.00128595,0.0003292496],"domain_scores_gemma":[0.9993085,0.0002795245,0.00007421071,0.0001386425,0.00008914304,0.0001099835],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"qualitative","study_design_scores_codex":[0.0001256839,0.00809605,0.7773073,0.0007804884,0.0003897986,0.00007605825,0.1098644,0.005069181,0.01253204,0.07497501,0.001670452,0.009113609],"study_design_scores_gemma":[0.008170347,0.0007566771,0.03334814,0.00102349,0.001366971,0.000006618535,0.4565305,0.3305382,0.06372008,0.0686195,0.03122958,0.004689963],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9762029,0.0002738242,0.0005915731,0.001887106,0.00007225478,0.000310552,0.000006986099,0.00009794715,0.02055688],"genre_scores_gemma":[0.9949356,0.0000161413,0.001798103,0.0001045036,0.0001530976,0.00004958364,0.000007373276,0.00001380042,0.002921775],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7439591,"threshold_uncertainty_score":0.9995365,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4989682414370745,"score_gpt":0.3578808874921893,"score_spread":0.1410873539448851,"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."}}