{"id":"W3112544537","doi":"10.12758/mda.2020.10","title":"Coding Text Answers to Open-ended Questions: Human Coders and Statistical Learning Algorithms Make Similar Mistakes","year":2021,"lang":"en","type":"article","venue":"Social Science Open Access Repository (GESIS – Leibniz Institute for the Social Sciences)","topic":"Computational and Text Analysis Methods","field":"Social Sciences","cited_by":11,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Statistical model; Coding (social sciences); Computer science; Statistical learning; Natural language processing; Artificial intelligence; Statistical analysis; Correlation; Speech recognition; Machine learning; Statistics; Mathematics","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":["metaepi_narrow","sts","scholarly_communication","open_science"],"consensus_categories":["sts"],"category_scores_codex":[0.007989625,0.0002915738,0.0005309152,0.0002142557,0.0495997,0.0156352,0.005818289,0.0001321333,0.00008476751],"category_scores_gemma":[0.002313321,0.0002556333,0.000170784,0.004522327,0.007641832,0.004070735,0.002008686,0.0002863742,0.000004760421],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005832238,"about_ca_system_score_gemma":0.003359754,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01406156,"about_ca_topic_score_gemma":0.00933588,"domain_scores_codex":[0.9944677,0.0009087381,0.000605252,0.001266734,0.001764022,0.0009875959],"domain_scores_gemma":[0.9972093,0.0009134946,0.0003960308,0.0002059408,0.0009115475,0.0003636779],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00006153836,0.0002263611,0.003390973,0.00003009375,0.0001357553,0.00003334605,0.03547216,0.0008894805,0.004703922,0.8245245,0.003640845,0.1268911],"study_design_scores_gemma":[0.002987371,0.0003893364,0.1094559,0.0002794527,0.000915307,0.00003207381,0.2534027,0.006984742,0.003258807,0.08912863,0.5295445,0.00362121],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.08058678,0.0005469745,0.1844043,0.06694654,0.006785936,0.006863231,0.0002382462,0.0004216247,0.6532063],"genre_scores_gemma":[0.97359,0.00004762382,0.01898111,0.001162323,0.001071979,0.0002665928,0.00002224503,0.00002194229,0.004836222],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8930032,"threshold_uncertainty_score":0.9999896,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1583214395006781,"score_gpt":0.5040170616981421,"score_spread":0.345695622197464,"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."}}