Bayes factor (BF₁₀)

The Bayes factor compares how well two assumptions predict the data: H₁ (there is a difference) against H₀ (there is none). BF₁₀ = 8 means: the data are about 8 times more likely under H₁ than under H₀. Unlike the p value it can also favor H₀ (BF₀₁ = 1/BF₁₀).

Example: BF₁₀ = 8.4: moderate evidence for a difference between the groups. BF₀₁ = 4.8: the data are about 5 times more likely under H₀ (no difference) than under H₁ - moderate evidence against a difference.

Caution: The Bayes factor depends on the prior (here: Cauchy with r = √2/2 as in JASP) and is not a probability of H₁. A value near 1 means "inconclusive", not "no effect". The classification (anecdotal to extreme) according to Lee and Wagenmakers (2013) is a rule of thumb.