{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0120942,0.000915496,0.000964469,0.002177348,0.0006136353,0.002694064,0.0009943836,0.001595803,0.003762226],"category_scores_gemma":[0.07694288,0.0002934509,0.0008439714,0.001486434,0.0006095132,0.003618407,0.001815175,0.001839786,0.002192746],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008149784,"about_ca_system_score_gemma":0.0006686969,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001168375,"about_ca_topic_score_gemma":0.001737827,"domain_scores_codex":[0.9864066,0.009822274,0.0006378514,0.001518034,0.001246841,0.0003684324],"domain_scores_gemma":[0.8514884,0.1298528,0.005020868,0.004725334,0.007590184,0.001322392],"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":[0.004634351,0.002390685,0.1441202,0.00120208,0.0004359494,0.0002293712,0.003519143,0.03126275,0.0465367,0.003623642,0.011938,0.7501072],"study_design_scores_gemma":[0.0001918038,0.001718379,0.0763791,0.0002999448,0.0002875515,0.0001862311,0.002258557,0.8421453,0.0561255,0.01326285,0.006967902,0.0001769475],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6879877,0.0004395265,0.3016256,0.0008992046,0.0001556255,0.0004799786,0.001291706,0.003162557,0.003958044],"genre_scores_gemma":[0.889342,0.00008131956,0.1067159,0.0001974179,0.0000836684,0.0003021002,0.001640053,0.0001638099,0.001473755],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9879058,"threshold_uncertainty_score":0.06396103,"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."}}