Skip to main content
. 2018 Jul 6;35(2):200–210. doi: 10.1093/bioinformatics/bty565

Table 1.

Different hypotheses tested in various data scenarios

Hypothesis strong weak
Null β = 0 and log (α) = 0 β ≠ 0 and log (α) = 0
Alternative β ≠ 0 and log (α) ≠ 0 β = 0 and log (α) ≠ 0

Note: The BTH model integrates over the mean effect size, β, testing the union of the weak and strong alternative hypotheses against the union of the weak and strong null hypotheses.