{"id":"W4388542568","doi":"10.2139/ssrn.4610561","title":"On Using Harmonized Data in Statistical Analysis: A Note of Caution","year":2023,"lang":"en","type":"article","venue":"SSRN Electronic Journal","topic":"Advanced Statistical Methods and Models","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Statistical analysis; Econometrics; Statistics; Computer science; Data mining; Mathematics","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":[],"consensus_categories":[],"category_scores_codex":[0.003432002,0.000112982,0.0003790503,0.0003488557,0.00006393665,0.00001325586,0.0002450435,0.00005602214,0.00002099838],"category_scores_gemma":[0.002324471,0.00009628663,0.00005899275,0.0008138141,0.00004155314,0.0001023531,0.00007242933,0.0009231708,0.00000424838],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00036932,"about_ca_system_score_gemma":0.000591716,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005324941,"about_ca_topic_score_gemma":0.0004298565,"domain_scores_codex":[0.9976349,0.0002511454,0.000472054,0.0002315924,0.0003200198,0.001090286],"domain_scores_gemma":[0.998161,0.001199759,0.0001852729,0.0003402393,0.00005319195,0.00006055615],"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":[0.0001306467,0.00008045776,0.00008115368,0.00001676957,0.0002339355,0.00001846153,0.00007807581,0.001743624,0.0007399577,0.9837689,0.00001696041,0.01309106],"study_design_scores_gemma":[0.0004572723,0.00009317375,0.0001840758,0.00002412573,0.0002514985,0.00001477152,0.0001468466,0.1966531,0.00003440491,0.8020549,0.000007156607,0.00007866435],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07911938,0.00006357324,0.9204748,0.0000710258,0.00004095892,0.00008428626,0.0000885816,0.00001820409,0.00003917869],"genre_scores_gemma":[0.7890138,0.0002926476,0.2105135,0.00001339663,0.00003632156,0.000001893792,0.00003168475,0.00001973262,0.0000770245],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7099613,"threshold_uncertainty_score":0.4010769,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1833010166755009,"score_gpt":0.486145505220095,"score_spread":0.3028444885445941,"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."}}