Wolfgang, the reason it is not possible to use impirical data to determine whether or not any particular religions are and have been more prone to certain tendencies than others is because in order to do that, the data needs to be collected over the entire history of each religion, as well as in contexts in which data does not normally get collected and, in some cases, where the incidents of violence are deliberately kept from being reported.

One example of this would be violence committed against the Dalits in India by members of higher castes which, more often than not, does not get reported. Another example of this would be sexual molestation of children by adults working within the context of churches, temples, mosques, religios schools, etc. And also there is the problem of history being written by the victors. We can't call someone's version of "history" empirical data, because we both know that people are quite prone to manipulating historical narratives to suit their idea of posterity and what it ought to be. Plus, we have no way of counting the numbers of people who have belonged to all of the different religions throughout their histories, so a per capita equation is not possible. There is too much data that would have to be left out of any kind of examination of that sort, and for that reason, there is no way to know for sure that the results would be accurate.

Then there is the problem of defining the terms. If one would say "most prone to violence", how do we determine how to measure that? If one country wages a war against another for religious reasons, and that country kills 20 million people using a nuclear weapon, but only fifteen people were involved in making the decision to use the weapon and to launch it (and none on their side get killed), how do we compare that to a country that wages a religiously motivated war on another country in which only one million people are killed, but ten thousand people are involved in doing the killing (and five thousand of the people on their own side get killed)? I say it's not possible to measure these kinds of things without making some very arbitrary distinctions.

And then, of course, there is the problem of violence that is not really religious in nature, for instance when people commit suicide bombings in the context of a struggle for independence, but the acts are attributed to the person's religion instead. There is no way to determine how much of the reportage of this kind of violence really correctly addresses the real reason for the violent act. It would be like trying to suggest that there is a racially determined factor in the fact that in the US, a disproportionate number of prison inmates are Black, instead of suggesting (more correctly), that the majority of prison inmates in the US come from impoverished backgrounds, and that because of a history of institutionalized racism in the US, a disproportionate number of Blacks in the US come from impoverished backgrounds.