{"id":"W4392356050","doi":"10.29115/sp-2023-0022","title":"Keep the noise down: On the performance of automatic speech recognition of voice-recordings in web surveys","year":2024,"lang":"en","type":"article","venue":"Survey Practice","topic":"Survey Methodology and Nonresponse","field":"Social Sciences","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Speech recognition; Noise (video); Computer science; Voice activity detection; Speech processing; Artificial intelligence","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":["metaresearch"],"category_scores_codex":[0.2970477,0.001160187,0.001065266,0.002409568,0.001308986,0.005554759,0.00262004,0.003361241,0.001863656],"category_scores_gemma":[0.7233706,0.001057607,0.00130273,0.003712778,0.004356337,0.006108145,0.003754359,0.001714722,0.002115219],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002335376,"about_ca_system_score_gemma":0.00225331,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006783809,"about_ca_topic_score_gemma":0.005720876,"domain_scores_codex":[0.5305219,0.3982349,0.01930615,0.01570804,0.03312989,0.003099072],"domain_scores_gemma":[0.1297363,0.7848721,0.02589876,0.02793117,0.03052468,0.00103695],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.004036489,0.0004841921,0.6224042,0.001835293,0.001097856,0.0002652752,0.05120629,0.005219049,0.004222371,0.00217903,0.004842184,0.3022078],"study_design_scores_gemma":[0.0002062066,0.00337791,0.9151629,0.00219596,0.000666812,0.000553628,0.01486446,0.02937181,0.01477349,0.004833679,0.01352836,0.0004648013],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.907164,0.002996077,0.06526038,0.004094738,0.000491164,0.001060065,0.0008433666,0.0005398607,0.0175503],"genre_scores_gemma":[0.9736191,0.0004684955,0.02251493,0.0009490469,0.000228654,0.0005111918,0.0005023908,0.0002870793,0.0009191322],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7029523,"threshold_uncertainty_score":0.8668653,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2129304017765198,"score_gpt":0.426281586716043,"score_spread":0.2133511849395232,"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."}}