{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":25,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":25,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"478ad34d6f7d","filters":{"venue":"Journal of Survey Statistics and Methodology"}},"results":[{"id":"W4378232119","doi":"10.1093/jssam/smad015","title":"Automated Classification for Open-Ended Questions with BERT","year":2023,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Topic Modeling","field":"Computer Science","cited_by":25,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo; Western University","funders":"Social Sciences and Humanities Research Council; Social Sciences and Humanities Research Council of Canada; Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Coding (social sciences); Artificial intelligence; Boosting (machine learning); Natural language processing; Machine learning; Language model; Training set; Statistics","authors":[{"name":"Hyukjun Gweon","is_ca":true},{"name":"Matthias Schonlau","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4858752837440197,"gpt":0.4643715025409455,"spread":0.02150378120307417,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009821454,0.0009717103,0.0007556367,0.002442108,0.0004476257,0.001281019,0.001521868,0.001466596,0.004296798],"category_scores_gemma":[0.05542549,0.0003372126,0.0006503614,0.001565665,0.0006401632,0.002610463,0.001649653,0.002226998,0.003621974],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00120911,"about_ca_system_score_gemma":0.0009440864,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002051675,"about_ca_topic_score_gemma":0.002885373,"domain_scores_codex":[0.9930335,0.004431301,0.0003845817,0.0009350243,0.0008485885,0.0003670184],"domain_scores_gemma":[0.9385139,0.04560056,0.00287909,0.006273583,0.005735077,0.0009978196],"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":[0.001980311,0.001055633,0.09056731,0.0009987942,0.0001489321,0.000266471,0.002277793,0.06424219,0.01685926,0.009624635,0.04375131,0.7682275],"study_design_scores_gemma":[0.00006830709,0.0003063315,0.02997775,0.0001200293,0.00003128884,0.0001675511,0.0007839292,0.9298141,0.01379093,0.01450488,0.01037114,0.00006373871],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5554429,0.0009808594,0.4126046,0.001689665,0.0003507421,0.0008010605,0.005922379,0.01389659,0.008311239],"genre_scores_gemma":[0.8951675,0.00009474348,0.09531221,0.0002027122,0.00007845092,0.0003758139,0.006134529,0.0001856511,0.002448325],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.009821454,"threshold_uncertainty_score":0.05194145,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3095502049","doi":"10.1093/jssam/smaa023","title":"Machine Learning for Occupation Coding—A Comparison Study","year":2020,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Data-Driven Disease Surveillance","field":"Medicine","cited_by":24,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Koch Institute for Integrative Cancer Research, Massachusetts Institute of Technology; Robert Koch Institut; Universität Mannheim; Social Sciences and Humanities Research Council of Canada; Deutsche Forschungsgemeinschaft; Institut für Arbeitsmarkt- und Berufsforschung","keywords":"Computer science; Coding (social sciences); Machine learning; Artificial intelligence; Multinomial distribution; Data mining; Statistics; Mathematics","authors":[{"name":"Malte Schierholz","is_ca":false},{"name":"Matthias Schonlau","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.400401439441831,"gpt":0.4705134066249507,"spread":0.07011196718311963,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01694215,0.0008501569,0.0008303429,0.002942241,0.0008997315,0.00199656,0.001856336,0.001448836,0.003745319],"category_scores_gemma":[0.06901777,0.0002847299,0.001225649,0.00298517,0.00106234,0.002800574,0.001308104,0.002236977,0.001336861],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002513106,"about_ca_system_score_gemma":0.00160088,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01115459,"about_ca_topic_score_gemma":0.006810096,"domain_scores_codex":[0.9881981,0.008676227,0.0004607773,0.001105959,0.001182372,0.0003766715],"domain_scores_gemma":[0.9058338,0.07850874,0.002803256,0.004851162,0.007031919,0.0009711168],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.006119942,0.0044772,0.3391311,0.001813921,0.001562427,0.0002549437,0.002252256,0.08193183,0.001026108,0.01291286,0.02636205,0.5221554],"study_design_scores_gemma":[0.0006985183,0.002724193,0.2539698,0.0007449757,0.0006943414,0.000338304,0.004961906,0.7021742,0.00270902,0.01490342,0.01589185,0.0001895232],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.964805,0.002601005,0.02262511,0.001633907,0.0002978493,0.0003159638,0.001286929,0.0003572337,0.006076986],"genre_scores_gemma":[0.9811615,0.0003964672,0.01479587,0.000157368,0.0001023903,0.0001827638,0.002279029,0.00006386267,0.0008607627],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9830579,"threshold_uncertainty_score":0.08959979,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3132157226","doi":"10.1093/jssam/smaa046","title":"Who Counts? Measuring Disability Cross-Nationally in Census Data","year":2021,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Health disparities and outcomes","field":"Social Sciences","cited_by":21,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta; University of Toronto","funders":"","keywords":"Microdata (statistics); Standardization; Terminology; Census; Harmonization; International Classification of Functioning, Disability and Health; Medical model of disability; Psychology; Gerontology; Actuarial science; Medicine; Political science; Environmental health; Business; Population; Psychiatry; Physical therapy","authors":[{"name":"David Pettinicchio","is_ca":true},{"name":"Michelle Maroto","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6228852929621991,"gpt":0.5393534200347042,"spread":0.08353187292749487,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0139088,0.0003162318,0.0005018019,0.004399329,0.0004311169,0.001773804,0.0008564432,0.0004053759,0.002516471],"category_scores_gemma":[0.05957206,0.0002758037,0.0005270666,0.008681746,0.0005788217,0.002841516,0.001993696,0.001050684,0.0006422422],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001249474,"about_ca_system_score_gemma":0.001776659,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02908897,"about_ca_topic_score_gemma":0.02784789,"domain_scores_codex":[0.9836369,0.00959858,0.002817473,0.001145175,0.002285092,0.0005167509],"domain_scores_gemma":[0.9799446,0.006917753,0.006155436,0.002661923,0.003890886,0.0004294735],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00004407005,0.00009140057,0.7735083,0.001191517,0.0006368778,0.00005920103,0.007670299,0.001027802,0.0002526786,0.01861768,0.06743923,0.1294608],"study_design_scores_gemma":[0.00004052572,0.0001056779,0.8171711,0.003901259,0.0002506983,0.000304025,0.01640746,0.004018336,0.001219167,0.01631135,0.1401305,0.0001398223],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.633811,0.01197901,0.07043922,0.02605697,0.001447808,0.001754384,0.1918856,0.0004597449,0.06216632],"genre_scores_gemma":[0.8992199,0.004317359,0.04791232,0.001923788,0.0003068533,0.0022417,0.04257979,0.0001282674,0.001369954],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02908897,"threshold_uncertainty_score":0.07355767,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2102082048","doi":"10.1093/jssam/smu011","title":"Small Area Prediction of Proportions with Applications to the Canadian Labour Force Survey","year":2014,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":18,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Estimator; Statistics; Small area estimation; Econometrics; Mathematics; Mean squared error; Multinomial distribution; Census; Estimation; Current Population Survey; Standard error; Table (database); Population; Benchmarking; Computer science; Demography; Economics; Data mining","authors":[{"name":"Emily Berg","is_ca":false},{"name":"Wayne A. Fuller","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3842642301409165,"gpt":0.4131436403482028,"spread":0.02887941020728629,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01106618,0.0007956003,0.001095859,0.002390618,0.001185123,0.001202959,0.002148845,0.0006350814,0.008678866],"category_scores_gemma":[0.07054941,0.0005132234,0.001145129,0.004916168,0.001052401,0.001013265,0.001750049,0.002143414,0.001324596],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003258241,"about_ca_system_score_gemma":0.007476464,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.3221852,"about_ca_topic_score_gemma":0.3055909,"domain_scores_codex":[0.9936578,0.003647719,0.0001926199,0.0008814487,0.001407348,0.000213103],"domain_scores_gemma":[0.9769297,0.01616316,0.0009167995,0.002648212,0.003116947,0.0002253172],"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":[0.0002430165,0.0001296868,0.03525268,0.0002495291,0.0002349437,0.0003085697,0.0008919213,0.2918066,0.001083773,0.1847474,0.01341949,0.4716323],"study_design_scores_gemma":[0.00004926106,0.00007527692,0.01229983,0.00007297634,0.0000392161,0.00009110229,0.0002006718,0.9040748,0.0009538796,0.06959777,0.01247875,0.00006642962],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.006805361,0.00009573777,0.9902565,0.0001295018,0.00006391813,0.0002446175,0.0005632947,0.0005937765,0.00124718],"genre_scores_gemma":[0.153108,0.0003184107,0.8375044,0.00009362943,0.00007464361,0.001185048,0.001872277,0.0002551051,0.00558846],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6778148,"threshold_uncertainty_score":0.6406198,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2599342551","doi":"10.1093/jssam/smw035","title":"Adaptive and Network Sampling for Inference and Interventions in Changing Populations","year":2016,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"HIV, Drug Use, Sexual Risk","field":"Medicine","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Sampling (signal processing); Sampling design; Inference; Computer science; Population; Adaptive sampling; Sample (material); Smoothing; Data mining; Simple random sample; Tracing; Machine learning; Statistics; Artificial intelligence; Mathematics","authors":[{"name":"Steven K. Thompson","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6918727978663437,"gpt":0.5395960375672009,"spread":0.1522767602991428,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.08948983,0.001207841,0.001797483,0.002637914,0.001336659,0.002298268,0.003234647,0.002343399,0.005411882],"category_scores_gemma":[0.2693467,0.001037459,0.002118969,0.00239683,0.007183936,0.004288729,0.003731557,0.004987875,0.0003153515],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002691573,"about_ca_system_score_gemma":0.002511389,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004168263,"about_ca_topic_score_gemma":0.003293781,"domain_scores_codex":[0.8744895,0.115962,0.001538378,0.003681037,0.003895711,0.000433335],"domain_scores_gemma":[0.6631449,0.3087742,0.007659683,0.01519088,0.004365028,0.0008653486],"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":[0.0001757855,0.000114946,0.004069144,0.0004774224,0.0004529489,0.0001036234,0.0006251532,0.09926884,0.0003093452,0.8261264,0.001595694,0.06668071],"study_design_scores_gemma":[0.0001242411,0.0001668126,0.0008906093,0.0001762333,0.00006603564,0.00004492421,0.00009847936,0.2842137,0.0003568563,0.709453,0.004374321,0.00003479674],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002656997,0.0003018832,0.9950383,0.0005754548,0.00008374419,0.0002241664,0.00004847172,0.00005671963,0.001014229],"genre_scores_gemma":[0.1319576,0.0009371111,0.8613431,0.0005748388,0.0003079689,0.003551678,0.000165903,0.00005812535,0.001103515],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.08948983,"threshold_uncertainty_score":0.4732731,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2464696916","doi":"10.1093/jssam/smaa042","title":"Estimating the Size and Distribution of Networked Populations with Snowball Sampling","year":2020,"lang":"en","type":"preprint","venue":"Journal of Survey Statistics and Methodology","topic":"Census and Population Estimation","field":"Mathematics","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Snowball sampling; Sample size determination; Inference; Computer science; Population; Sampling (signal processing); Population size; Graph; Sample (material); Selection (genetic algorithm); Statistics; Sampling design; Econometrics; Mathematics; Theoretical computer science; Machine learning; Artificial intelligence; Demography; Telecommunications","authors":[{"name":"Kyle Vincent","is_ca":true},{"name":"Steve Thompson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5549742675041466,"gpt":0.474974966433552,"spread":0.07999930107059461,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006386574,0.0004639976,0.0008241511,0.001612707,0.0003728932,0.001117215,0.001561804,0.0005779675,0.00184584],"category_scores_gemma":[0.02821905,0.0005676949,0.0005765164,0.001129748,0.001195438,0.001514068,0.001802433,0.0006840465,0.0002169834],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007255494,"about_ca_system_score_gemma":0.0005880147,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004733108,"about_ca_topic_score_gemma":0.00429237,"domain_scores_codex":[0.9976997,0.001455007,0.00006993666,0.0003957729,0.0002863569,0.00009320144],"domain_scores_gemma":[0.9868217,0.009924386,0.001083132,0.001325903,0.0006320939,0.0002128684],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004972031,0.0001651759,0.05230464,0.0001503424,0.0003062573,0.0003576714,0.0005109439,0.7062225,0.00517732,0.135491,0.00210127,0.09671567],"study_design_scores_gemma":[0.00003369175,0.00005024381,0.00364651,0.0000261995,0.00002027601,0.00005387224,0.00009212627,0.9325207,0.00108661,0.06171593,0.0007416775,0.00001214029],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.116794,0.00008477025,0.8816352,0.0001068236,0.00002043729,0.0001133121,0.0001957032,0.0001010349,0.0009487265],"genre_scores_gemma":[0.7697065,0.0001815996,0.227808,0.00007008424,0.00004121894,0.0003025161,0.0006531052,0.00003841405,0.001198553],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006386574,"threshold_uncertainty_score":0.03377587,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2196855157","doi":"10.1093/jssam/smv022","title":"Clarifying Some Aspects of Variance Estimation in Two-Phase Sampling","year":2015,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Survey Sampling and Estimation Techniques","field":"Mathematics","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Estimator; Jackknife resampling; Variance (accounting); Mathematics; Statistics; Sampling (signal processing); Bias of an estimator; Minimum-variance unbiased estimator; Computer science; Applied mathematics","authors":[{"name":"Jean‐François Beaumont","is_ca":true},{"name":"Audrey Béliveau","is_ca":false},{"name":"David Haziza","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6371680919234616,"gpt":0.5436168220745937,"spread":0.09355126984886786,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.07277423,0.00114091,0.002048965,0.001979416,0.0009978085,0.002871408,0.002752796,0.003344018,0.003860806],"category_scores_gemma":[0.233593,0.0009895831,0.00184646,0.003289538,0.006331606,0.006191041,0.003440426,0.005256986,0.0005206838],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001802336,"about_ca_system_score_gemma":0.002243176,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001568803,"about_ca_topic_score_gemma":0.0008983193,"domain_scores_codex":[0.943688,0.04558609,0.001834367,0.003253603,0.004759948,0.0008779557],"domain_scores_gemma":[0.8031482,0.1687876,0.008373334,0.0125046,0.006645516,0.0005407486],"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":[0.00004216302,0.00004151801,0.001476384,0.0002789361,0.00006777218,0.0001556481,0.000551512,0.01234997,0.0006658935,0.9547664,0.00117599,0.02842768],"study_design_scores_gemma":[0.00006776582,0.000138031,0.001311825,0.000154792,0.00004113189,0.000139722,0.0001392871,0.07387768,0.0009372903,0.9166383,0.006509857,0.00004432649],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004781648,0.0006666347,0.991342,0.001232965,0.0001083568,0.00007146948,0.00003466063,0.00003433436,0.001728075],"genre_scores_gemma":[0.2378473,0.002011553,0.7530146,0.002133699,0.001110062,0.0009356132,0.0001566692,0.0001201498,0.002670449],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.07277423,"threshold_uncertainty_score":0.3848715,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4392953941","doi":"10.1093/jssam/smae003","title":"Text Messages to Facilitate the Transition to Web-First Sequential Mixed-Mode Designs in Longitudinal Surveys","year":2024,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Survey Methodology and Nonresponse","field":"Social Sciences","cited_by":8,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Economic and Social Research Council; University of Essex","keywords":"Web survey; Sample (material); World Wide Web; Mode (computer interface); Computer science; Quarter (Canadian coin); Mixed mode; Psychology; Geography; Human–computer interaction","authors":[{"name":"Pablo Cabrera‐Álvarez","is_ca":false},{"name":"Peter Lynn","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6092991703284136,"gpt":0.5016061607140164,"spread":0.1076930096143972,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.05317186,0.00103794,0.0007950296,0.001148391,0.0009213054,0.001818269,0.001356468,0.001584064,0.0140379],"category_scores_gemma":[0.1448588,0.001073826,0.0007942292,0.001020564,0.0009533288,0.00189288,0.002156841,0.001804348,0.002053606],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007008924,"about_ca_system_score_gemma":0.001035374,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003702733,"about_ca_topic_score_gemma":0.0007308,"domain_scores_codex":[0.9372129,0.05589284,0.002550471,0.001728461,0.001987636,0.0006276486],"domain_scores_gemma":[0.7653698,0.1906823,0.01378292,0.02088692,0.00696677,0.002311167],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.03442573,0.02755745,0.03687182,0.004972509,0.0006429399,0.0002838606,0.01993622,0.0133889,0.04625816,0.02175564,0.01139524,0.7825116],"study_design_scores_gemma":[0.05032706,0.260322,0.2307256,0.00366429,0.002155701,0.0006536182,0.00663055,0.1158866,0.1265178,0.07742117,0.1243854,0.001310123],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5378194,0.0003836742,0.4097843,0.001668116,0.0008184823,0.0345525,0.001504277,0.003237159,0.010232],"genre_scores_gemma":[0.4327792,0.0001569886,0.4975649,0.001082918,0.0003199969,0.06448951,0.0003757004,0.0002612524,0.002969597],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9468281,"threshold_uncertainty_score":0.281203,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3190138084","doi":"10.1093/jssam/smab022","title":"A Model-Assisted Approach for Finding Coding Errors in Manual Coding of Open-Ended Questions","year":2021,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Reliability and Agreement in Measurement","field":"Decision Sciences","cited_by":7,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Actua; University of Waterloo; Roche (Canada)","funders":"","keywords":"Coding (social sciences); Computer science; Statistics; Natural language processing; Artificial intelligence; Mathematics","authors":[{"name":"Zhoushanyue He","is_ca":true},{"name":"Matthias Schonlau","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7870611956524755,"gpt":0.5441244124235828,"spread":0.2429367832288927,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.05606314,0.002557892,0.002239944,0.006392811,0.001906153,0.003326988,0.004967253,0.003189472,0.002808691],"category_scores_gemma":[0.1579796,0.001601642,0.002124382,0.003226386,0.001789765,0.003241448,0.004391115,0.004175576,0.001588715],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003281877,"about_ca_system_score_gemma":0.005248129,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009178651,"about_ca_topic_score_gemma":0.01451252,"domain_scores_codex":[0.9337932,0.0496453,0.002749562,0.006656197,0.006266864,0.0008888423],"domain_scores_gemma":[0.765063,0.1815021,0.01477125,0.0189258,0.01823824,0.001499638],"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":[0.001521416,0.001233367,0.02988261,0.0007020172,0.0009402582,0.0003612837,0.004892926,0.1611781,0.00865052,0.0118042,0.01096145,0.7678719],"study_design_scores_gemma":[0.00007303025,0.0001427766,0.002178024,0.00007859848,0.00004606645,0.000118862,0.0003501175,0.9820195,0.002986599,0.01077024,0.001165308,0.00007074501],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01791999,0.0000631522,0.9778147,0.0003239305,0.00004360281,0.0003998852,0.0001374609,0.002852733,0.0004444862],"genre_scores_gemma":[0.2254403,0.00004146875,0.7714089,0.0002798273,0.00003612025,0.001233656,0.0005281527,0.000209075,0.0008225068],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9439369,"threshold_uncertainty_score":0.2964937,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4226069045","doi":"10.1093/jssam/smac023","title":"Augmenting Survey Data with Digital Trace Data: Is There a Threat to Panel Retention?","year":2022,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Survey Methodology and Nonresponse","field":"Social Sciences","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Deutsche Forschungsgemeinschaft","keywords":"Panel data; Panel survey; Panel discussion; German; Quarter (Canadian coin); Survey data collection; Data retention; TRACE (psycholinguistics); Worry; Computer science; Psychology; Econometrics; Computer security; Advertising; Business; Statistics; Mathematics; Economics; Demographic economics; Geography","authors":[{"name":"Mark Trappmann","is_ca":false},{"name":"Georg‐Christoph Haas","is_ca":false},{"name":"Sonja Malich","is_ca":false},{"name":"Florian Keusch","is_ca":false},{"name":"Sebastian Bähr","is_ca":false},{"name":"Frauke Kreuter","is_ca":false},{"name":"Štefan Schwarz","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7842892671678372,"gpt":0.5220393702231063,"spread":0.2622498969447309,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.2272346,0.0004616347,0.0008555807,0.001260664,0.002927787,0.004779514,0.003327643,0.002279633,0.01078063],"category_scores_gemma":[0.5621638,0.0008352515,0.001125471,0.002753598,0.003410114,0.005741832,0.004859527,0.003911303,0.002462034],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001986068,"about_ca_system_score_gemma":0.004441191,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006010947,"about_ca_topic_score_gemma":0.005636422,"domain_scores_codex":[0.6944813,0.2628323,0.006975172,0.007051365,0.02442397,0.004235863],"domain_scores_gemma":[0.2611,0.5606416,0.05487818,0.0836899,0.03570382,0.003986632],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.002546496,0.002058869,0.4630425,0.00183574,0.001181369,0.0003432152,0.02280509,0.003051209,0.00362783,0.01561939,0.02806612,0.4558221],"study_design_scores_gemma":[0.0009037677,0.01077032,0.6075813,0.008343349,0.001612345,0.001668528,0.05480074,0.03690199,0.02633104,0.06337669,0.1871917,0.0005183768],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7513102,0.001777768,0.1319218,0.06167452,0.001389219,0.002859164,0.002171566,0.001104126,0.04579164],"genre_scores_gemma":[0.9593022,0.0002980485,0.02495184,0.009016387,0.0003626865,0.001864937,0.0005720988,0.0001224037,0.003509242],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7727654,"threshold_uncertainty_score":0.9529573,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4309622736","doi":"10.1093/jssam/smac027","title":"Jackknife Bias-Corrected Generalized Regression Estimator in Survey Sampling","year":2022,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Survey Sampling and Estimation Techniques","field":"Mathematics","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Statistics Canada","funders":"","keywords":"Jackknife resampling; Estimator; Statistics; Mathematics; Mean squared error; Bias of an estimator; Population; Minimum-variance unbiased estimator; Sample size determination; Sample (material); Variance (accounting); Econometrics","authors":[{"name":"Marius Stefan","is_ca":false},{"name":"M. Hidiroglou","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7039901809482423,"gpt":0.5070651107249131,"spread":0.1969250702233292,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02514324,0.0008661597,0.001862215,0.001852541,0.000537216,0.0008227293,0.002828906,0.001548239,0.001866535],"category_scores_gemma":[0.1021934,0.0006057534,0.0009696434,0.002518096,0.001267672,0.001715594,0.00169246,0.001551796,0.0006654788],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009220726,"about_ca_system_score_gemma":0.001451418,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004361883,"about_ca_topic_score_gemma":0.003593879,"domain_scores_codex":[0.9709872,0.0241124,0.0005299925,0.001663481,0.00226527,0.000441623],"domain_scores_gemma":[0.9586712,0.02783542,0.003977719,0.005190094,0.003984939,0.0003406338],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0004557665,0.0001771087,0.02678973,0.0007083539,0.0008397843,0.0003889695,0.0006080325,0.4988889,0.003696787,0.1664106,0.006494174,0.2945418],"study_design_scores_gemma":[0.00009805144,0.0002380934,0.005544463,0.0001976913,0.0001197937,0.0003675528,0.0001260768,0.9083502,0.003586112,0.07462406,0.006663633,0.00008412466],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.008054975,0.0002530308,0.9908744,0.00006606098,0.00002974741,0.00005153217,0.00004990075,0.0002097161,0.000410721],"genre_scores_gemma":[0.2757016,0.0003929843,0.7210756,0.0002184083,0.00007814714,0.000384884,0.0003069294,0.0001651136,0.001676181],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02514324,"threshold_uncertainty_score":0.1329717,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3186908110","doi":"10.1093/jssam/smab007","title":"Lack of Replication or Generalization? Cultural Values Explain a Question Wording Effect","year":2021,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Cultural Differences and Values","field":"Psychology","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Saskatchewan","funders":"Economic and Social Research Council","keywords":"Replication (statistics); Generalization; Epistemology; Econometrics; Mathematical economics; Computer science; Psychology; Positive economics; Mathematics; Statistics; Economics; Philosophy","authors":[{"name":"Henning Silber","is_ca":false},{"name":"Endre Tvinnereim","is_ca":false},{"name":"Tobias Stark","is_ca":false},{"name":"Annelies G. Blom","is_ca":false},{"name":"Jon A. Krosnick","is_ca":false},{"name":"Michael Bošnjak","is_ca":false},{"name":"Sanne Lund Clement","is_ca":false},{"name":"Anne Cornilleau","is_ca":false},{"name":"Anne-Sophie Cousteaux","is_ca":false},{"name":"Melvin John","is_ca":false},{"name":"Guðbjörg Andrea Jónsdóttir","is_ca":false},{"name":"Karen Lawson","is_ca":true},{"name":"Peter Lynn","is_ca":false},{"name":"Johan Martinsson","is_ca":false},{"name":"Ditte Shamshiri-Petersen","is_ca":false},{"name":"Su‐Hao Tu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5870137882448033,"gpt":0.5534318512773253,"spread":0.03358193696747802,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.3319109,0.001392047,0.002126631,0.00289963,0.001846997,0.003540235,0.005230397,0.003026291,0.01035802],"category_scores_gemma":[0.6974455,0.001492479,0.002753541,0.002489725,0.01532635,0.009144391,0.006254712,0.004667397,0.001342292],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002924085,"about_ca_system_score_gemma":0.002593202,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005206643,"about_ca_topic_score_gemma":0.002854274,"domain_scores_codex":[0.6465013,0.2544641,0.02206345,0.03325057,0.03997155,0.003748963],"domain_scores_gemma":[0.1254609,0.705332,0.03559264,0.1150976,0.0172268,0.001289995],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.005493981,0.001950037,0.5647778,0.006893303,0.00515579,0.00179996,0.06844357,0.003531537,0.01247731,0.1243528,0.01864447,0.1864794],"study_design_scores_gemma":[0.001711594,0.002842512,0.569925,0.003687858,0.002858769,0.001770439,0.02455313,0.02133107,0.02330338,0.309432,0.03799847,0.0005857981],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6773633,0.006479692,0.2069645,0.02676508,0.003461801,0.006093627,0.001680452,0.001369143,0.06982235],"genre_scores_gemma":[0.981834,0.0002518524,0.01196956,0.002929011,0.0002872793,0.001303643,0.0002267202,0.0002194712,0.000978493],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6680891,"threshold_uncertainty_score":0.8238728,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3018909183","doi":"10.1093/jssam/smaa004","title":"Multiply Robust Bootstrap Variance Estimation in the Presence of Singly Imputed Survey Data","year":2020,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Winnipeg; Université de Montréal","funders":"National Institute of General Medical Sciences; Natural Sciences and Engineering Research Council of Canada; National Institutes of Health","keywords":"Estimator; Quantile; Imputation (statistics); Statistics; Robustness (evolution); Variance (accounting); Point estimation; Econometrics; Mathematics; Robust statistics; Population; Computer science; Missing data","authors":[{"name":"Sixia Chen","is_ca":false},{"name":"David Haziza","is_ca":true},{"name":"Zeinab Mashreghi","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6810327651611238,"gpt":0.4918226381347552,"spread":0.1892101270263686,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01627766,0.0008134197,0.001535705,0.001962506,0.0005729265,0.001088384,0.002615853,0.001927646,0.001609984],"category_scores_gemma":[0.09556707,0.0006758928,0.00148506,0.001898563,0.001411404,0.001546324,0.002530872,0.001773666,0.0006660212],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004324937,"about_ca_system_score_gemma":0.0008587586,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007181971,"about_ca_topic_score_gemma":0.0007506695,"domain_scores_codex":[0.9813107,0.01443601,0.0005640582,0.001134369,0.002280458,0.0002744175],"domain_scores_gemma":[0.9557781,0.03187373,0.003046924,0.006941973,0.002128867,0.0002305338],"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":[0.0005335619,0.0002156358,0.01445979,0.0007042721,0.001078251,0.0005529055,0.0007916652,0.1111894,0.008747135,0.1759443,0.003711478,0.6820714],"study_design_scores_gemma":[0.0001286178,0.0004862366,0.008841,0.0002299221,0.0002756073,0.001006909,0.0001520387,0.7462594,0.01277897,0.2196075,0.01009333,0.0001405505],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006676668,0.0002687072,0.9924636,0.00007617315,0.00001729587,0.00003748964,0.00002667651,0.000179211,0.0002541506],"genre_scores_gemma":[0.190259,0.0004430836,0.8077152,0.0001175665,0.0001193145,0.0003417161,0.000244309,0.0001226235,0.0006371939],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9837223,"threshold_uncertainty_score":0.0860855,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4396570341","doi":"10.1093/jssam/smae021","title":"Should We Offer Web, Paper, or Both? A Comparison of Single- and Mixed-Response Mode Treatments in a Mail Survey","year":2024,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Survey Methodology and Nonresponse","field":"Social Sciences","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Center for Science and Engineering Statistics; U.S. Census Bureau; Washington State University; American Institutes for Research; McMaster University; U.S. Department of Education; National Science Foundation","keywords":"Mixed mode; The Internet; Web survey; Mode (computer interface); Consistency (knowledge bases); Sample (material); Response time; Incentive; Response bias; Psychology; Computer science; Advertising; World Wide Web; Social psychology; Business; Artificial intelligence; Economics; Human–computer interaction","authors":[{"name":"Sarah K Heimel","is_ca":false},{"name":"Rebecca L. Medway","is_ca":false},{"name":"Rachel Horwitz","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6823083142602526,"gpt":0.5520341180469343,"spread":0.1302741962133184,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.05662949,0.001236464,0.001619808,0.001418255,0.001528024,0.002990016,0.00241673,0.003983324,0.01260879],"category_scores_gemma":[0.1593973,0.001344299,0.001628072,0.000904031,0.00224013,0.003222919,0.001787206,0.003384129,0.002957463],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001926146,"about_ca_system_score_gemma":0.002016296,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008475673,"about_ca_topic_score_gemma":0.0007073649,"domain_scores_codex":[0.9079624,0.07728196,0.003764908,0.003554426,0.005622519,0.001813822],"domain_scores_gemma":[0.7960584,0.1641828,0.01893594,0.01363278,0.003499316,0.00369065],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"nonrandomized_trial","study_design_scores_codex":[0.3360558,0.1719557,0.05243,0.003575611,0.002576306,0.0002484213,0.01123423,0.008618061,0.01341523,0.01417943,0.007583295,0.3781279],"study_design_scores_gemma":[0.1304123,0.510956,0.1884192,0.001828778,0.00363297,0.0002424926,0.009338362,0.06658991,0.02752429,0.03333931,0.02667935,0.001037025],"study_design_candidate":"nonrandomized_trial","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.971797,0.0002488841,0.009720054,0.001413295,0.0005010347,0.008322372,0.000469103,0.0003479683,0.007180299],"genre_scores_gemma":[0.9466997,0.0001966993,0.02033045,0.001906539,0.0003811428,0.0272695,0.0002926526,0.00006143917,0.002861897],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9433705,"threshold_uncertainty_score":0.299489,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3097049086","doi":"10.1093/jssam/smaa016","title":"Targeting Key Survey Variables at the Unit Nonresponse Treatment Stage","year":2020,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Survey Sampling and Estimation Techniques","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"","keywords":"Estimator; Non-response bias; Weighting; Missing data; Statistics; Propensity score matching; Computer science; Econometrics; Inverse probability weighting; Set (abstract data type); Mathematics; Medicine","authors":[{"name":"David Haziza","is_ca":true},{"name":"Sixia Chen","is_ca":false},{"name":"Yimeng Gao","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6424708043159415,"gpt":0.4784392389837912,"spread":0.1640315653321503,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.03461761,0.0007210404,0.001526347,0.001352248,0.001067968,0.001482753,0.00251941,0.00181469,0.007528339],"category_scores_gemma":[0.09535882,0.0004901005,0.001491673,0.003042921,0.001376234,0.001647961,0.002639391,0.002558958,0.001525198],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001140402,"about_ca_system_score_gemma":0.002711464,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001668719,"about_ca_topic_score_gemma":0.001826222,"domain_scores_codex":[0.9643867,0.02989766,0.00103097,0.001954475,0.002050562,0.000679477],"domain_scores_gemma":[0.9640807,0.02043003,0.004475085,0.007896169,0.002735659,0.0003824139],"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":[0.000710257,0.001169337,0.05910905,0.001537811,0.0008001334,0.000267492,0.001683906,0.03711893,0.004227994,0.253103,0.01307106,0.627201],"study_design_scores_gemma":[0.0008067029,0.001900024,0.06569655,0.00103821,0.0008919468,0.0003101975,0.001517994,0.4578405,0.02619729,0.3909906,0.05262213,0.0001878503],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03045611,0.0003371737,0.9629795,0.001054,0.0002284123,0.00166686,0.0003884945,0.00021234,0.00267708],"genre_scores_gemma":[0.5502732,0.0005201657,0.4358749,0.00111065,0.0002414395,0.005997603,0.0007264179,0.0001014323,0.005154131],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9653824,"threshold_uncertainty_score":0.1830776,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2521098594","doi":"10.1093/jssam/smw013","title":"The Pseudo-EBLUP Estimator for a Weighted Average with an Application to the Canadian Survey of Employment, Payrolls and Hours","year":2016,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Spatial and Panel Data Analysis","field":"Economics, Econometrics and Finance","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Mathematics; Estimator; Mean squared error; Statistics; Heteroscedasticity; Small area estimation; Bias of an estimator; Best linear unbiased prediction; Mixed model; Monte Carlo method; Minimum-variance unbiased estimator; Computer science","authors":[{"name":"Susana Rubin-Bleuer","is_ca":false},{"name":"Leon Jang","is_ca":false},{"name":"Serge Godbout","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1877846169135838,"gpt":0.3363112003694582,"spread":0.1485265834558744,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0129994,0.0005717196,0.00101201,0.001641329,0.0006564235,0.001172336,0.002486935,0.0008050224,0.003510683],"category_scores_gemma":[0.06711335,0.0004314415,0.001082519,0.003309861,0.001134056,0.001826353,0.001412159,0.001747057,0.0005660167],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002090217,"about_ca_system_score_gemma":0.006964821,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.07717991,"about_ca_topic_score_gemma":0.0846097,"domain_scores_codex":[0.9908922,0.005041022,0.0003200283,0.00110184,0.00238755,0.0002573791],"domain_scores_gemma":[0.9779772,0.01234645,0.001259531,0.003922235,0.004281663,0.0002129088],"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":[0.0002010783,0.0001280188,0.03994922,0.0007497065,0.000835317,0.0002469672,0.0006341988,0.1644709,0.002188415,0.2219605,0.01507514,0.5535605],"study_design_scores_gemma":[0.000112742,0.000255981,0.04493628,0.0003546074,0.0002420852,0.000371308,0.0002982204,0.7406816,0.00360158,0.1684882,0.04051367,0.0001435859],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.009093446,0.000480683,0.9881953,0.000215874,0.0000609453,0.0001231536,0.0004783734,0.0002158958,0.001136372],"genre_scores_gemma":[0.2231642,0.0009157456,0.768112,0.0003843725,0.0001126463,0.000843414,0.001689058,0.0002117104,0.004566879],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9979098,"threshold_uncertainty_score":0.1534613,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W7133121762","doi":"10.1093/jssam/smaf036","title":"Smoothed pseudo-population bootstrap methods with applications to finite population quantiles","year":2025,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Survey Sampling and Estimation Techniques","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada","keywords":"Estimator; Resampling; Confidence interval; Bootstrapping (finance); Quantile; Jackknife resampling; Poisson sampling; Population; CDF-based nonparametric confidence interval; Sampling (signal processing)","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.4880688058564743,"gpt":0.5424089670887016,"spread":0.05434016123222729,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00960835,0.0005820905,0.0009761308,0.002101296,0.0004790851,0.001099885,0.001967717,0.0009794052,0.004324331],"category_scores_gemma":[0.05820232,0.0005416679,0.001026825,0.002021916,0.001454034,0.001521853,0.001906395,0.002023974,0.0009450864],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006293656,"about_ca_system_score_gemma":0.0009509239,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001516155,"about_ca_topic_score_gemma":0.001238669,"domain_scores_codex":[0.9941885,0.004428512,0.0001419238,0.0003119178,0.0008161473,0.0001129853],"domain_scores_gemma":[0.9657571,0.02732205,0.00126733,0.002932251,0.002451593,0.0002696301],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002537075,0.0001352016,0.004621275,0.0003786369,0.0002415447,0.0003524182,0.0005176216,0.3588019,0.003215226,0.3979915,0.003401915,0.2300891],"study_design_scores_gemma":[0.00004247344,0.00007740106,0.001146612,0.00005480048,0.00002526197,0.0000924636,0.00004631059,0.8518042,0.001336269,0.1407801,0.004567918,0.00002610253],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002828886,0.0001161177,0.9964844,0.00005174743,0.00001896853,0.00002373184,0.00002063115,0.0001930827,0.0002624532],"genre_scores_gemma":[0.2742997,0.0004954411,0.722311,0.0002014182,0.000135286,0.0005747176,0.0002841214,0.0002930734,0.001405316],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00960835,"threshold_uncertainty_score":0.05081439,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4392712152","doi":"10.1093/jssam/smae002","title":"Estimation of a Population Total Under Nonresponse Using Follow-up","year":2024,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Survey Sampling and Estimation Techniques","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Statistics Canada","funders":"","keywords":"Estimation; Statistics; Econometrics; Computer science; Population; Mathematics; Medicine; Environmental health; Economics","authors":[{"name":"Marius Stefan","is_ca":false},{"name":"M. Hidiroglou","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.554927493029936,"gpt":0.5121732745467844,"spread":0.04275421848315164,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03570823,0.0005681533,0.001386053,0.002914332,0.0004891516,0.001057165,0.001994652,0.0009670279,0.002417361],"category_scores_gemma":[0.1098353,0.0004771174,0.0009482721,0.002510694,0.000865261,0.001965709,0.001918406,0.0009622602,0.0003372134],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008023783,"about_ca_system_score_gemma":0.001162498,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001325238,"about_ca_topic_score_gemma":0.001447126,"domain_scores_codex":[0.9753203,0.02011395,0.0006477819,0.001601391,0.002034655,0.0002819707],"domain_scores_gemma":[0.9425818,0.04159984,0.005178969,0.007306805,0.002989335,0.0003431979],"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":[0.0004420873,0.0006067135,0.08236342,0.0007391815,0.0009976077,0.0001608681,0.001653826,0.08726244,0.002300905,0.1057591,0.002610684,0.7151032],"study_design_scores_gemma":[0.0003977906,0.002084716,0.08719757,0.0004487725,0.0007677904,0.0003693101,0.0009956377,0.6739724,0.01263434,0.2043847,0.01659394,0.0001529985],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06233973,0.0001992264,0.9353634,0.0002263887,0.00003049097,0.000328934,0.0001421191,0.0001571246,0.001212561],"genre_scores_gemma":[0.6120263,0.0003423649,0.3830637,0.0002677017,0.00007309697,0.001398921,0.0006162398,0.00004724075,0.002164411],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.03570823,"threshold_uncertainty_score":0.1888455,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4206410105","doi":"10.1093/jssam/smab057","title":"Neighborhood Bootstrap for Respondent-Driven Sampling","year":2021,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"HIV, Drug Use, Sexual Risk","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Resampling; Estimator; Leverage (statistics); Sampling (signal processing); Statistics; Consistency (knowledge bases); Computer science; Variance (accounting); Econometrics; Respondent; Tree (set theory); Bootstrap aggregating; Sample (material); Mathematics; Artificial intelligence","authors":[{"name":"Mamadou Yauck","is_ca":true},{"name":"Erica E. M. Moodie","is_ca":false},{"name":"Herak Apelian","is_ca":false},{"name":"Alain Fourmigue","is_ca":false},{"name":"Daniel Grace","is_ca":false},{"name":"Trevor Hart","is_ca":false},{"name":"Gilles Lambert","is_ca":false},{"name":"Joseph Cox","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5731953672614928,"gpt":0.5175135239374811,"spread":0.05568184332401171,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.03910657,0.0007968874,0.001674482,0.002066131,0.001314755,0.001480154,0.003823871,0.002106114,0.005331551],"category_scores_gemma":[0.1642949,0.0006845673,0.001655116,0.002181126,0.002619683,0.001912339,0.003017966,0.002768849,0.001055583],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001374217,"about_ca_system_score_gemma":0.001689248,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005719395,"about_ca_topic_score_gemma":0.003291787,"domain_scores_codex":[0.9723287,0.02307852,0.0004881966,0.001464478,0.002275915,0.0003641994],"domain_scores_gemma":[0.8550141,0.1192398,0.004464358,0.01288946,0.007300717,0.001091502],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006470848,0.000230059,0.01720008,0.0005212743,0.0004283392,0.0006089783,0.0008569509,0.2852244,0.001361621,0.5858139,0.006418095,0.1006892],"study_design_scores_gemma":[0.00006655122,0.00008866259,0.001146384,0.0001072497,0.00003305753,0.0001053212,0.0000806938,0.8854811,0.000562674,0.1091079,0.003193175,0.00002727363],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.009167859,0.0002288083,0.9885039,0.0002578003,0.00006941634,0.0001944901,0.0001533844,0.0001935645,0.00123071],"genre_scores_gemma":[0.3985195,0.0004272586,0.5955581,0.0004812927,0.000269937,0.00143418,0.0008556653,0.0001657818,0.002288371],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9608935,"threshold_uncertainty_score":0.2068178,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3193938275","doi":"10.1093/jssam/smab029","title":"Bootstrap Estimation of the Conditional Bias for Measuring Influence in Complex Surveys","year":2021,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Statistics Canada","funders":"","keywords":"Estimator; Statistics; Sampling (signal processing); Variance (accounting); Sampling design; Conditional variance; Sample (material); Econometrics; Mathematics; Estimation; Population; Computer science; Autoregressive conditional heteroskedasticity","authors":[{"name":"Jean‐François Beaumont","is_ca":true},{"name":"Cynthia Bocci","is_ca":true},{"name":"Michel St-Louis","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7007546710550773,"gpt":0.5016242371275665,"spread":0.1991304339275107,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.05431816,0.0008607756,0.001588373,0.004048616,0.0008987564,0.001687674,0.002356936,0.001758286,0.003384317],"category_scores_gemma":[0.3355904,0.0006348335,0.001727736,0.004081226,0.003437231,0.002843384,0.003249537,0.002384265,0.0006209873],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001380768,"about_ca_system_score_gemma":0.001185478,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002853516,"about_ca_topic_score_gemma":0.001819096,"domain_scores_codex":[0.9492401,0.04064842,0.001524843,0.002259632,0.005821417,0.0005055041],"domain_scores_gemma":[0.6821668,0.2675654,0.01525008,0.02296295,0.01112958,0.0009252589],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005819841,0.0002413702,0.09318882,0.0008635066,0.001437035,0.0005838199,0.001438145,0.1771745,0.003736717,0.4577108,0.0049423,0.2581009],"study_design_scores_gemma":[0.000108314,0.0003053213,0.02608529,0.0005375631,0.0002754661,0.0004495257,0.0003029062,0.6222063,0.005314244,0.3365316,0.007760551,0.0001228408],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01776364,0.0003468381,0.9798472,0.0001533309,0.0000544484,0.0001293161,0.0001006283,0.0001749072,0.001429645],"genre_scores_gemma":[0.5966239,0.0005599454,0.3998951,0.0003158354,0.0002034551,0.0009868597,0.0005149719,0.0002110574,0.0006889548],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.05431816,"threshold_uncertainty_score":0.2872653,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4402591820","doi":"10.1093/jssam/smae032","title":"Model-Based Prediction for Small Domains Using Covariates: A Comparison of Four Methods","year":2024,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Institute for Work & Health; McGill University; McGill University Health Centre","funders":"Institut de Valorisation des Données","keywords":"Covariate; Computer science; Statistics; Econometrics; Artificial intelligence; Data mining; Machine learning; Mathematics","authors":[{"name":"Victoire Michal","is_ca":true},{"name":"Jonathan Wakefield","is_ca":false},{"name":"Alexandra M. Schmidt","is_ca":true},{"name":"Alicia Cavanaugh","is_ca":true},{"name":"Brian E. Robinson","is_ca":true},{"name":"Jill Baumgartner","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7662515417175665,"gpt":0.5679552565310851,"spread":0.1982962851864813,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01844433,0.00116597,0.002236554,0.001510521,0.0005274764,0.001571272,0.002652668,0.001553729,0.001360968],"category_scores_gemma":[0.04027356,0.0006304921,0.001625828,0.001545272,0.0008757801,0.002719437,0.00225085,0.0028482,0.000336546],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007306555,"about_ca_system_score_gemma":0.001341537,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004808705,"about_ca_topic_score_gemma":0.00254553,"domain_scores_codex":[0.9934047,0.005028183,0.0002238014,0.0005706307,0.0006461492,0.0001265408],"domain_scores_gemma":[0.9579965,0.03525061,0.001699902,0.002583517,0.001989909,0.0004796731],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006515424,0.0003632248,0.01526618,0.0003370093,0.0008226472,0.0001282224,0.0003371258,0.7252387,0.0009646729,0.02054769,0.001968152,0.2333748],"study_design_scores_gemma":[0.00002991937,0.00005670328,0.0008638537,0.00003483142,0.00002779928,0.00002020555,0.00002295262,0.9928418,0.0001665419,0.005563091,0.0003575002,0.00001476214],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.033331,0.001710352,0.9630207,0.0005040528,0.00005882774,0.0001142531,0.0001123212,0.0004587895,0.0006896218],"genre_scores_gemma":[0.5508233,0.001988514,0.4440655,0.0003524855,0.0002431465,0.0004380876,0.0007521375,0.000280139,0.001056732],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01844433,"threshold_uncertainty_score":0.09754407,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3041151490","doi":"10.1093/jssam/smab004","title":"Imputation Procedures in Surveys Using Nonparametric and Machine Learning Methods: An Empirical Comparison","year":2021,"lang":"en","type":"preprint","venue":"Journal of Survey Statistics and Methodology","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"","keywords":"Imputation (statistics); Nonparametric statistics; Computer science; Machine learning; Artificial intelligence; Data mining; Data set; Econometrics; Missing data; Mathematics","authors":[{"name":"Mehdi Dagdoug","is_ca":false},{"name":"Camélia Goga","is_ca":false},{"name":"David Haziza","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5742043220523099,"gpt":0.5636730961020205,"spread":0.01053122595028932,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.1388914,0.0005960376,0.00134219,0.003456406,0.0007205585,0.002412758,0.001953996,0.001758024,0.003077113],"category_scores_gemma":[0.3922183,0.0004911525,0.001722077,0.006594833,0.002534029,0.004810079,0.002519685,0.002149339,0.0004633788],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001120763,"about_ca_system_score_gemma":0.001582771,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001399302,"about_ca_topic_score_gemma":0.001115916,"domain_scores_codex":[0.8072861,0.1783908,0.002652236,0.002510478,0.008475523,0.0006849077],"domain_scores_gemma":[0.3414253,0.6034805,0.01871352,0.02414095,0.01141555,0.0008241487],"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":[0.002449241,0.001171407,0.1539497,0.002534267,0.003943099,0.0002143285,0.002762455,0.1226231,0.0004608427,0.1834532,0.009550793,0.5168876],"study_design_scores_gemma":[0.0007252959,0.002420154,0.1894022,0.002608278,0.001273051,0.001099207,0.003297981,0.5156788,0.001814469,0.2623169,0.01902459,0.0003392057],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.3292924,0.01711799,0.6415478,0.002442937,0.0002213683,0.0005806891,0.0007020584,0.0003601213,0.007734559],"genre_scores_gemma":[0.8537108,0.002898018,0.1408037,0.0003379005,0.0001874051,0.0004079714,0.0007185184,0.0001320023,0.0008036966],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1388914,"threshold_uncertainty_score":0.7345369,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"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,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Classifier (UML); Factorial; Transformer; Language model; Encoder; Question answering","authors":[{"name":"Ruben L. Bach","is_ca":false},{"name":"Matthias Schonlau","is_ca":true},{"name":"Katharina Meitinger","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1819431608627828,"gpt":0.3925421491784883,"spread":0.2105989883157055,"validation_status":"score_only:v0-immature-baseline"},"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"},"labels":[],"label_agreement":null},{"id":"W7119808922","doi":"10.1093/jssam/smaf043","title":"Correcting Selection Bias in Non-Probability Two-Phase Payment Survey","year":2025,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Bank of Canada","funders":"","keywords":"Selection (genetic algorithm); Selection bias; Variance (accounting); Calibration; Payment; Estimation","authors":[{"name":"Heng Chen","is_ca":true},{"name":"John Tsang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4738508169731649,"gpt":0.5186313866550564,"spread":0.04478056968189142,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0488295,0.0006711688,0.001213799,0.001482544,0.0006055344,0.001789892,0.002622339,0.001678163,0.004204954],"category_scores_gemma":[0.250248,0.0008565342,0.001044871,0.00245824,0.001540636,0.003672973,0.002484631,0.001834564,0.000574153],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009561467,"about_ca_system_score_gemma":0.001845787,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00183914,"about_ca_topic_score_gemma":0.001482813,"domain_scores_codex":[0.9550615,0.03838867,0.0009245457,0.001992747,0.003066132,0.0005662702],"domain_scores_gemma":[0.8129236,0.1545085,0.01030949,0.01544299,0.006303846,0.0005115237],"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":[0.0003738172,0.0003741662,0.03889026,0.0005784099,0.000297427,0.0003948653,0.000946632,0.1179039,0.00159833,0.5983975,0.004685098,0.2355598],"study_design_scores_gemma":[0.0001629969,0.0003014765,0.0101351,0.0001262406,0.0000814122,0.0002166956,0.0001647249,0.663299,0.002045761,0.3173563,0.0060531,0.00005716707],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0225213,0.00008502519,0.9755367,0.0003346114,0.00004494828,0.0002286141,0.00009272679,0.0001477042,0.001008474],"genre_scores_gemma":[0.5543257,0.0002528156,0.4413976,0.0004304781,0.0001190559,0.0009232375,0.0003470583,0.00006841293,0.002135566],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0488295,"threshold_uncertainty_score":0.2582381,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4248615235","doi":"10.1093/jssam/smab047","title":"A Rescaling Bootstrap Approach For Imputed Survey Data","year":2021,"lang":"en","type":"article","venue":"Journal of Survey Statistics and Methodology","topic":"Survey Sampling and Estimation Techniques","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Winnipeg","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Imputation (statistics); Estimator; Statistics; Mathematics; Econometrics; Variance (accounting); Computer science; Missing data","authors":[{"name":"Zeinab Mashreghi","is_ca":true},{"name":"Huiqi Deng","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.8566941241856907,"gpt":0.5528146051570141,"spread":0.3038795190286766,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009453639,0.0006697754,0.001343869,0.002342267,0.0005600858,0.0007783429,0.002704028,0.001020844,0.002515689],"category_scores_gemma":[0.04220826,0.0004576498,0.001039461,0.003021076,0.001038366,0.001312166,0.00182362,0.001760561,0.0007151063],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004708489,"about_ca_system_score_gemma":0.0006441947,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001068551,"about_ca_topic_score_gemma":0.0008008559,"domain_scores_codex":[0.9890282,0.007897723,0.0004143242,0.0009141612,0.001522435,0.000223095],"domain_scores_gemma":[0.987044,0.00775983,0.0007746919,0.00268583,0.00155455,0.0001809659],"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":[0.0002989801,0.0002190541,0.008560794,0.0005754995,0.000511213,0.0008226168,0.0007146913,0.1641323,0.008979416,0.1914281,0.00609569,0.6176615],"study_design_scores_gemma":[0.00004976302,0.0001203384,0.00257371,0.00008341052,0.0000686132,0.0003031907,0.00008599492,0.9018056,0.003361879,0.08369855,0.007789481,0.00005944334],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002481619,0.0001382105,0.9968208,0.00005513504,0.00002689644,0.00002779846,0.00004013723,0.0001742342,0.0002351972],"genre_scores_gemma":[0.1884731,0.0004660548,0.8088207,0.0001892153,0.0001804431,0.0004274081,0.0004323984,0.0001802096,0.0008303854],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.009453639,"threshold_uncertainty_score":0.0499962,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}