BIOL 210 Lecture Notes - Lecture 9: Confirmation Bias, Publication Bias, Statistical Hypothesis Testing
4 Oct 2017
School
Department
Course
Professor
Document Summary
Publication lag + review period (date received, accepted, and published) In some disciplines, older papers are not cited after a while. The cited half-life of journals in these disciplines is short. In some papers, the half-life is lo(cid:374)g as the field does(cid:374)(cid:859)t (cid:272)ha(cid:374)ge (cid:448)ery fast. Right now only worry about authors, affiliations, acknowledgement, reviewers, disclosure of conflicts of interest. Possible biases in science : confirmation bias, publication bias, funding source bias, linguistic and country bias, institutional bias, authority bias, gender bias. Info by doctors first biased by sales reps, then by publications: compare hospital records to publication records (a lot of studies were done but never published -> you get publication bias) Indefinite conclusions even less published than negative results. Infer studies that have negative results statistically: correlation between brain size and iq literature is biased: studies with small samples are reporting large effect and get published.
