Anthony Higney

Post-doc Economist

less than 1 minute read

I was reading O’Rourke’s history of meta-analysis paper and he mentions what is possibly the first paper to systematically look at publication bias based on significance. This is Sterling’s 1959 paper “Publication Decisions and Their Possible Effects on Inferences Drawn from Tests of Significance–Or Vice Versa”.

The main table is striking:

Table 32 from Sterling (1959). It has 4 columns. First column is a list of journals with "total at the bottom". Second is percent of artciles that test a hypothesis. Third is percent that reject H0. Fourth is percent that do not. All journals have ober 95% of artciles rejecting H0. Total is 97.28% of artciles that test reject H0

97% rejected the null!

You May Also Enjoy

True P Values

2 minute read

layout: single title: “Taleb Paper on P-Hacking” date: 2024-12-12 use_math: true published: false categories: blog tags: Publication Bias Meta-Analys...