{"id":"W4280577454","doi":"10.1098/rspb.2021.2780","title":"Slow improvement to the archiving quality of open datasets shared by researchers in ecology and evolution","year":2022,"lang":"en","type":"article","venue":"Proceedings of the Royal Society B Biological Sciences","topic":"Research Data Management Practices","field":"Computer Science","cited_by":37,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University; Université de Montréal; Carleton University","funders":"Horizon 2020 Framework Programme","keywords":"Data quality; Computer science; Data sharing; Mandate; Quality (philosophy); Data curation; Reusability; Data science; Seniority; Linked data; World Wide Web; Medicine; Semantic Web; Political science; Engineering; Software","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":["open_science"],"consensus_categories":["open_science"],"category_scores_codex":[0.014917,0.00007816606,0.0001529168,0.00003349052,0.000683756,0.0006360667,0.008846902,0.00002323551,0.00001124946],"category_scores_gemma":[0.001944011,0.00004065408,0.00004826649,0.001108305,0.0006545015,0.00209922,0.02195634,0.0002676354,3.971514e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001068922,"about_ca_system_score_gemma":0.00006589854,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00180959,"about_ca_topic_score_gemma":0.00002592928,"domain_scores_codex":[0.9978595,0.0002055674,0.000287957,0.0005402727,0.0007354511,0.0003712783],"domain_scores_gemma":[0.998874,0.000533776,0.0002647351,0.0002136952,0.00005072403,0.00006305719],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001114157,0.0005249932,0.7459012,0.0001000434,0.00005381983,2.423736e-7,0.002287749,0.0002681187,0.05260374,0.1358657,0.05345201,0.008830918],"study_design_scores_gemma":[0.0003961707,0.001456102,0.9515121,0.00002131618,0.00000376697,6.230291e-7,0.00626911,0.02217522,0.001171838,0.009421799,0.007375369,0.0001965757],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9750171,0.00008115762,0.0004911309,0.02258818,0.00004496103,0.0009408382,0.0001449225,0.00001028762,0.0006814814],"genre_scores_gemma":[0.9955293,0.00005027371,0.003786529,0.0003914254,0.000006487204,0.0001205282,0.000004101142,0.000001275502,0.0001101136],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2056109,"threshold_uncertainty_score":0.9965157,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1941139242982793,"score_gpt":0.4060641246116984,"score_spread":0.2119502003134191,"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."}}