{"id":"W2981744747","doi":"10.1503/cmaj.190085","title":"Testing group differences for confounder selection in nonrandomized studies: flawed practice","year":2019,"lang":"en","type":"article","venue":"Canadian Medical Association Journal","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":17,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Observational study; Confounding; Random assignment; Selection (genetic algorithm); Medicine; Key (lock); Selection bias; Intervention (counseling); Computer science; Family medicine; Pathology; Machine learning; Computer security; Nursing","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.7227249,0.003260078,0.007877702,0.007580434,0.003313002,0.008058622,0.01012975,0.009683926,0.00722278],"category_scores_gemma":[0.915151,0.003283373,0.007263276,0.00677488,0.03202036,0.01241074,0.008195825,0.01526996,0.001126688],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004865201,"about_ca_system_score_gemma":0.007017233,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002395484,"about_ca_topic_score_gemma":0.001537767,"domain_scores_codex":[0.1634507,0.6949196,0.04772253,0.04815224,0.04440368,0.001351239],"domain_scores_gemma":[0.02779511,0.8972758,0.01935314,0.04733611,0.007500984,0.0007389567],"domain_codex":"methods","domain_gemma":"methods","domain_candidate":"methods","domain_consensus":"methods","study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.005364241,0.0006829579,0.08375614,0.01879791,0.02653855,0.003255739,0.01260259,0.01832529,0.002133709,0.4299316,0.02468902,0.3739222],"study_design_scores_gemma":[0.001815839,0.00120985,0.01029971,0.006021515,0.003155122,0.001445804,0.0008892523,0.04726123,0.003145587,0.9108894,0.01361669,0.0002499408],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01365983,0.005474526,0.9571088,0.01302412,0.001822942,0.004037015,0.0005049901,0.0007348403,0.003632861],"genre_scores_gemma":[0.3445635,0.001399864,0.63557,0.008873169,0.001354275,0.006748285,0.000290778,0.0002565571,0.0009435095],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.2772751,"threshold_uncertainty_score":0.3419296,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1467618169955665,"score_gpt":0.4167907857041335,"score_spread":0.270028968708567,"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."}}