{"id":"W2030541374","doi":"10.1017/s0954394512000129","title":"Models, forests, and trees of York English:<i>Was/were</i>variation as a case study for statistical practice","year":2012,"lang":"en","type":"article","venue":"Language Variation and Change","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":753,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta; University of Toronto","funders":"","keywords":"Multicollinearity; Plural; Variation (astronomy); Inference; Causal inference; Random forest; Econometrics; Interpretation (philosophy); Verb; Moderation; Computer science; Statistics; Linguistics; Variable (mathematics); Argument (complex analysis); Natural language processing; Artificial intelligence; Mathematics; Linear regression","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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.03328795,0.000589948,0.001120117,0.003000732,0.00188178,0.003457078,0.002113807,0.001091176,0.003344442],"category_scores_gemma":[0.09108575,0.0005621242,0.00156318,0.003878297,0.005758971,0.003919045,0.00199999,0.001872935,0.0002140106],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003431486,"about_ca_system_score_gemma":0.001435754,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01904288,"about_ca_topic_score_gemma":0.02839235,"domain_scores_codex":[0.9723036,0.02386169,0.0007112056,0.00174754,0.0009748582,0.000401013],"domain_scores_gemma":[0.8346328,0.155384,0.003105149,0.004102754,0.002121961,0.0006532855],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.0001781672,0.00007542218,0.02969888,0.000301976,0.0002436196,0.0007398779,0.007519057,0.09937933,0.0004657242,0.8036219,0.003306953,0.05446912],"study_design_scores_gemma":[0.0000243432,0.00004092409,0.004280597,0.0001150168,0.00003760287,0.0001873369,0.001173877,0.3152542,0.0002450494,0.6746844,0.003906105,0.0000505307],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1669616,0.001545575,0.8234277,0.002941979,0.00007367829,0.0001159184,0.0004877793,0.0004474369,0.003998327],"genre_scores_gemma":[0.7943203,0.0003291882,0.2034704,0.0001365113,0.00005114889,0.0002452631,0.0003710109,0.000164988,0.0009111309],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9667121,"threshold_uncertainty_score":0.1760456,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06490848551197312,"score_gpt":0.3640252859992587,"score_spread":0.2991168004872855,"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."}}