A growing number of studies have found that large brains may help animals survive by avoiding predation. These studies provide an alternative explanation for existing correlative evidence for one of the dominant hypotheses regarding the evolution of brain size in animals, the social brain hypothesis (SBH). The SBH proposes that social complexity is a major evolutionary driver of large brains. However, if predation both directly selects for large brains and higher levels of sociality, correlations between sociality and brain size may be spurious. We argue that tests of the SBH should take direct effects of predation into account, either by explicitly including them in comparative analyses or by pin‐pointing the brain‐behavior‐fitness pathway through which the SBH operates. Existing data and theory on social behavior can then be used to identify precise candidate mechanisms and formulate new testable predictions.