{"id":"W4403115895","doi":"10.2139/ssrn.4943579","title":"Learning Source Biases: Multisource Misspecifications and Their Impact on Predictions","year":2024,"lang":"en","type":"preprint","venue":"SSRN Electronic Journal","topic":"Media Influence and Politics","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Econometrics; Data science; Economics","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":"codex-gemma-dda1882f352a","candidate_categories":["research_integrity"],"consensus_categories":[],"category_scores_codex":[0.002182772,0.0002837012,0.0002667462,0.0002811934,0.00121335,0.0004290803,0.0003880764,0.0003174697,0.00008991205],"category_scores_gemma":[0.0008474898,0.0002130606,0.0002641701,0.0002061004,0.0002736007,0.00007127474,0.000143232,0.008095034,0.00009833207],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00158837,"about_ca_system_score_gemma":0.00632804,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001795718,"about_ca_topic_score_gemma":0.001381473,"domain_scores_codex":[0.9966046,0.0004063552,0.0003291936,0.0003381069,0.000354592,0.001967104],"domain_scores_gemma":[0.9985412,0.0005752783,0.0002179228,0.0002239233,0.0001232216,0.0003184191],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001006128,0.0003605226,0.009061807,0.000127837,0.002360865,0.00001815991,0.2597634,0.05902267,0.0001461221,0.4884806,0.004869832,0.1756875],"study_design_scores_gemma":[0.0004194966,0.0007660895,0.001176441,0.001012597,0.0003071614,0.0002300081,0.1605308,0.004531218,0.00002488324,0.6078677,0.2222539,0.0008797667],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9672506,0.01036014,0.002227102,0.00577657,0.001096222,0.0004528009,0.00006368638,0.0002719968,0.01250081],"genre_scores_gemma":[0.9639897,0.02362978,0.00002008622,0.00008425426,0.002333179,0.00002136253,0.00001868687,0.00004059525,0.009862325],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.217384,"threshold_uncertainty_score":0.9993051,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03081311897199608,"score_gpt":0.3396370264904079,"score_spread":0.3088239075184119,"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."}}