Suppose that you're a social scientist, studying ethnic tensions in the Balkans.
And suppose that you're focusing on an area where the ethnic make-up of the population is about 90% Croatian & about 10% Serbian.
And suppose that your thesis is that, in their endless, pointless feuding with one another, Croats & Serbs are just about equally guilty: i.e., that Croats hate Serbs just as much as Serbs hate Croats, and that Croats are just as likely as Serbs to express their hatred through acts of violence. And vice versa.
Given all that, what would you expect, in your area, when it come to acts of violence motivated by ethnic hatred?
Would you expect more attacks by Croats against Serbs? Or would you expect more attacks by Serbs against Croats?
On the one hand, potential Croat victims out-number potential Serb victims by about ten to one. Which might lead you to expect a big majority of Serb perpetrators and Croat victims.
On the other hand, potential Croat perpetrators out-number potential Serb perpetrators by about ten to one. Which might lead you to expect exactly the opposite - i.e., a big majority of Croat perpetrators and Serb victims.
Should this all just work out to a wash? Or what?
My own initial, and very weakly held, theory is that, in this situation, one should expect approximately equal numbers of victims and perpetrators on each side. 'Cause even though Croats out-number Serbs by about ten to one, the number of Croat *encounters* with Serbs *necessarily* exactly equals the number of Serb encounters with Croats.
But I am *very* ready to be instructed otherwise, if I'm missing something.
Am I missing something?