{"id":"W4405971183","doi":"10.1038/s44159-024-00392-z","title":"Using natural language processing to analyse text data in behavioural science","year":2025,"lang":"en","type":"article","venue":"Nature Reviews Psychology","topic":"Mental Health via Writing","field":"Psychology","cited_by":106,"is_retracted":false,"has_abstract":false,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Rigour; Computer science; Popularity; Interpretability; Artificial intelligence; Natural language processing; Social media; Computational sociology; Data science; Pipeline (software); Psychology; World Wide Web","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01599892,0.001325281,0.001069305,0.007745159,0.001289659,0.006302608,0.001285297,0.001437556,0.006718469],"category_scores_gemma":[0.07417971,0.0006248335,0.001619125,0.004556739,0.001929387,0.00554685,0.002360774,0.003177929,0.005163149],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001020214,"about_ca_system_score_gemma":0.002650955,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001701787,"about_ca_topic_score_gemma":0.002049878,"domain_scores_codex":[0.980973,0.01097979,0.003271681,0.002473985,0.002043787,0.0002578222],"domain_scores_gemma":[0.7917477,0.1842596,0.008200056,0.006021121,0.008833704,0.000937696],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000760532,0.0006143252,0.009548639,0.00859058,0.0005735208,0.001287923,0.0100549,0.00415699,0.04995698,0.02650652,0.05105044,0.8368987],"study_design_scores_gemma":[0.0004583504,0.000933814,0.03704641,0.003189943,0.0007407115,0.002941079,0.01484196,0.2203768,0.07387409,0.3760442,0.2687932,0.0007594468],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04930838,0.003807644,0.8992699,0.006287753,0.00166813,0.00184084,0.01868782,0.01152372,0.007605792],"genre_scores_gemma":[0.1457901,0.001769925,0.8226167,0.001189968,0.001040888,0.002100872,0.02133697,0.001078494,0.003076195],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01599892,"threshold_uncertainty_score":0.08461136,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1702667086601565,"score_gpt":0.5825904396932856,"score_spread":0.4123237310331291,"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."}}