What you see depends on where you're looking and how you look at it: publication bias and outcome reporting bias

Abstract

Study publication bias is the decision to publish or not publish a study based on its results. Compared to unpublished work, published studies are more likely to have positive or statistically significant findings. Outcome reporting bias is opting to publish only a subset of the original variables recorded for a study, such that the inclusion of the variables in the published work is selectively based on the results. Statistically significant results have a higher likelihood of being fully reported compared to nonsignificant results, and a significant proportion of published articles describe outcome variables or data analyses that differ from the pre-specified trial protocol as originally conceived. Recognition that publication bias and outcome reporting bias contribute to a distorted perception of drug effects-inflated estimates of efficacy and underreporting of adverse events-has led to the development and expansion of publicly accessible databases that contain transparent information about clinical trials and their results.