{"id":"W3180016080","doi":"10.2196/30730","title":"Strategies for the Identification and Prevention of Survey Fraud: Data Analysis of a Web-Based Survey","year":2021,"lang":"en","type":"article","venue":"JMIR Cancer","topic":"Survey Methodology and Nonresponse","field":"Social Sciences","cited_by":98,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Patient-Centered Outcomes Research Institute","keywords":"CAPTCHA; Identification (biology); Data quality; Survey data collection; Social media; Incentive; Computer science; Web application; Internet privacy; Psychology; Computer security; World Wide Web; Statistics; Business; Mathematics","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":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.03494123,0.00004454172,0.0001844675,0.00007092434,0.0001318765,0.00002620317,0.0002201911,0.00007003562,0.0001698447],"category_scores_gemma":[0.006015469,0.00003634526,0.00005126188,0.000953305,0.0002343863,0.0001235909,0.00003122933,0.00003949341,1.869834e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001341467,"about_ca_system_score_gemma":0.0009199136,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.01730154,"about_ca_topic_score_gemma":0.4854213,"domain_scores_codex":[0.9889797,0.0103451,0.0002311234,0.0001991076,0.0001543013,0.000090662],"domain_scores_gemma":[0.9840582,0.01496965,0.000217017,0.0003682897,0.000368999,0.00001786695],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00268404,0.0001183022,0.9720436,0.00005809442,0.001118375,3.082008e-7,0.002961057,0.0004777072,0.004795804,0.0006808167,0.003041664,0.01202023],"study_design_scores_gemma":[0.0001743794,0.000009992677,0.9953718,0.000006628183,0.0002054194,1.274781e-8,0.001149002,0.001185436,0.0005574402,0.0001871937,0.001113106,0.0000395585],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9713371,0.0012637,0.0245547,0.0003105564,0.0002054789,0.0003034501,0.001967222,0.000007673182,0.0000501373],"genre_scores_gemma":[0.9984525,0.0002693516,0.000358256,0.00002116959,0.00001902804,0.00006091261,0.000414484,0.00000324602,0.0004010411],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4681197,"threshold_uncertainty_score":0.9937311,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5100129480820306,"score_gpt":0.5479787724618534,"score_spread":0.03796582437982288,"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."}}