{"id":"W3086958902","doi":"10.18148/srm/2020.v14i3.7639","title":"Automatic Coding of Open-ended Questions into Multiple Classes: Whether and How to Use Double Coded Data","year":2019,"lang":"en","type":"article","venue":"DOAJ (DOAJ: Directory of Open Access Journals)","topic":"Survey Methodology and Nonresponse","field":"Social Sciences","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Coding (social sciences); Computer science; Training set; Artificial intelligence; Statistical learning; Data mining; Machine learning; Natural language processing; Statistics; Mathematics","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":["metaresearch"],"category_scores_codex":[0.2681754,0.002904511,0.002931352,0.00915946,0.003261794,0.006818983,0.005076758,0.003112736,0.00476164],"category_scores_gemma":[0.6047383,0.002385417,0.002043462,0.007461981,0.004883721,0.01168271,0.008805081,0.005224057,0.003911543],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005240328,"about_ca_system_score_gemma":0.009481668,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003980557,"about_ca_topic_score_gemma":0.006599309,"domain_scores_codex":[0.6262692,0.2987103,0.02426196,0.02303721,0.02422761,0.00349372],"domain_scores_gemma":[0.1839466,0.5830063,0.0605123,0.08332319,0.08566258,0.003549036],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.003200277,0.001631705,0.04733761,0.004027523,0.0004990761,0.0002469833,0.04174072,0.009123275,0.02593115,0.01407075,0.01660879,0.8355821],"study_design_scores_gemma":[0.003172406,0.004182281,0.1088194,0.01029213,0.0008131704,0.001911582,0.05487275,0.4665065,0.124168,0.1454425,0.07695494,0.002864447],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.143884,0.0004073363,0.8271336,0.003423954,0.0007106681,0.01300215,0.0008831634,0.004891802,0.005663225],"genre_scores_gemma":[0.2445579,0.0001864413,0.7397083,0.0008127115,0.0001374639,0.01110903,0.001057004,0.0007595875,0.001671599],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7318246,"threshold_uncertainty_score":0.90247,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.6833831604247071,"score_gpt":0.6409189352124031,"score_spread":0.04246422521230409,"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."}}