POL SCI 137A Chapter Notes - Chapter 7: Statistical Power, Publication Bias, Grey Literature
Document Summary
Publication bias in the social sciences: unlocking the file drawer. Study conducted on publication bias: analyzing known population of conducted studies (221 total) in tess (hence all exceed a substantial quality threshold) Strong results 40% more likely to be published than null results. Strong results 60% more likely to be written up. Stage at which publication bias (pb) occurs - authors do not write up. Pb: when publication of study results is based on the direction or. Ex. greater likelihood of statistically significant results being and submit null findings significance of the findings published than statistically insignificant results. File drawer problem - selective reporting of scientific findings. Pb documented in the biomedical and social sciences. Pb detection: (common) replicate metanalysis with and without unpublished literature. Examine published literature and rely on assumptions about the distribution of unpublished reserach. Compare published studies to gray literature (dissertations, working papers, conference papers etc. )