{"id":"W2189001323","doi":"","title":"Combining Synthetic Data with Subsampling to Create Public Use Microdata Files for Large Scale Surveys","year":2012,"lang":"en","type":"article","venue":"Survey methodology","topic":"demographic modeling and climate adaptation","field":"Decision Sciences","cited_by":11,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Microdata (statistics); Confidentiality; Computer science; Scale (ratio); Sample (material); Imputation (statistics); Statistics; Sample size determination; Statistical inference; Data file; Data science; Data mining; Econometrics; Database; Missing data; Mathematics; Census; Computer security; Geography; Machine learning","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.04143951,0.0005934833,0.0006766114,0.004275368,0.001134295,0.002817899,0.002189558,0.0009204802,0.008054144],"category_scores_gemma":[0.1915117,0.000832618,0.0009538629,0.006516495,0.001059343,0.002843526,0.003048739,0.001546755,0.001888673],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001369223,"about_ca_system_score_gemma":0.002171108,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002827126,"about_ca_topic_score_gemma":0.002878382,"domain_scores_codex":[0.9728947,0.01984567,0.001774591,0.001754535,0.003415622,0.000314939],"domain_scores_gemma":[0.8168905,0.1091797,0.009549204,0.05133159,0.01176994,0.001279157],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008638821,0.000800517,0.05732182,0.0008906985,0.000508087,0.0005189597,0.006268628,0.07450963,0.003124625,0.1565785,0.0990027,0.599612],"study_design_scores_gemma":[0.0006833273,0.000646238,0.03528041,0.0005218647,0.0001944931,0.0008238824,0.002786688,0.3948414,0.01124592,0.3220346,0.2306135,0.0003277262],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01759798,0.00006599702,0.9696099,0.0008145997,0.0001888141,0.001276023,0.004851934,0.002485863,0.003108855],"genre_scores_gemma":[0.1181851,0.0001080265,0.8653452,0.0003451751,0.0001639992,0.004985014,0.009051907,0.0006463478,0.001169165],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.04143951,"threshold_uncertainty_score":0.2191557,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.7970682118326551,"score_gpt":0.5130019730252127,"score_spread":0.2840662388074424,"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."}}