Ed Yong has an excellent post reviewing new research which suggests that collective religious rituals are more predictive than religious belief as to support for suicide bombings. The novelty and insight from these studies is that they decompose the independent dimensions from which religious phenomena are constructed. Consider for example that religion may consist of: A) Belief in supernatural agents B) Participation in communal rituals C) Regulation of personal behavior under religious law D) A metaphysical system which explains the nature of the universe And so forth. A study like the one above suggests that it is the second, communal rituals, which heighten the ingroup-outgroup biases which often lead to religiously motivated atrocities. By analogy we can compare these to politically and ethnically motivated violence. Religious actors from within the phenomena in question naturally blur the boundaries of these individual components, and so naturally may claim that it is the "Will of God" which demands that they act in the way they do. Similarly, unbelievers whose own first-hand knowledge of religion is slight may naturally take the words of these violent actors on the fact of it. Researchers such as Scott Atran and Robert Pape have shown that in fact variations in belief have little to do with these actions in the name of belief.