{"id":"W4417210052","doi":"10.1093/jssam/smaf023","title":"Analyzing List-Style Open-Ended Questions: Combining Texts from Individual Answer Boxes Improves Classification with Language Models","year":2025,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Classifier (UML); Factorial; Transformer; Language model; Encoder; Question answering","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":[],"consensus_categories":[],"category_scores_codex":[0.003265413,0.0001326114,0.0003804906,0.0003234078,0.0001477071,0.0003990683,0.0008296658,0.0001048346,0.000004211612],"category_scores_gemma":[0.001431896,0.0001029491,0.00002282221,0.0003660262,0.0001467327,0.0006122043,0.0002859642,0.0002800213,3.854261e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002993533,"about_ca_system_score_gemma":0.0001611205,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005419821,"about_ca_topic_score_gemma":0.0003218711,"domain_scores_codex":[0.9978812,0.001024759,0.0005069541,0.0002604841,0.0001597999,0.0001668086],"domain_scores_gemma":[0.9954553,0.00326708,0.0005708897,0.000305486,0.0003418821,0.00005932052],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.0002073606,0.0001884961,0.02567948,0.00002899377,0.0004072819,0.00002992382,0.00215388,0.00007400652,0.01134507,0.5171075,0.005603977,0.437174],"study_design_scores_gemma":[0.001880412,0.00055325,0.6383463,0.0001266866,0.0001443544,0.00002991103,0.00194726,0.02180797,0.002014549,0.3318743,0.0009336692,0.000341257],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02898635,0.001037376,0.9683693,0.000986509,0.0002477211,0.0001085033,0.0001076651,0.00003332702,0.0001232269],"genre_scores_gemma":[0.4171404,0.0002409852,0.5823389,0.0001409335,0.00001172335,0.000004721342,0.00002279206,0.000004357523,0.00009525599],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.6126668,"threshold_uncertainty_score":0.4198142,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1819431608627828,"score_gpt":0.3925421491784883,"score_spread":0.2105989883157055,"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."}}