1/n A replication experimental survey dataset I downloaded produces some interesting (but likely underpowered, due to ideological stratification) findings.
2/n First, White liberals are sig. (p=0.03) more likely to endorse sacrificing 12 US soldiers to save white (Macedonian) Muslims than christians. They're also more likely to support a similar sacrifice to save black muslims vs. black christians, but the diff fails to achieve sig 
3/n Second, white moderates show a similar bias but in the other direction: they're more likely to sacrifice 12 US soldiers to save white and black christians than muslims. They're also more reluctant to save black (African) vs. white (Macedonian)
4/n Finally, white conservatives are more likely to make this sacrifice for christians vs. muslims, but the difference is only [marginally] significant (p=0.0526) between the Black christian vs. White muslim conditions.
5/n Example of how these conditions were constructed
6/n
Might interest you @RichardHanania
@RichardHanania 7/n If we move to Liberal subjects overall, we find that white/Macedonian Christians are least likely to be saved at the cost of 12 US casualties: White muslims vs. christians (p < 0.001); Black muslims vs. white christians (p=0.07), Black christians vs. white christians (p=0.03)
@RichardHanania 8/n Among moderates overall, the pro-christian bias remains
@RichardHanania 9/n Among conservatives overall, more significantly endorse sacrificing 12 US soldiers to save black Christians than white and black muslims (probably because a lot of blacks identify as conservative and are christian)
@RichardHanania 10/n In sum, Liberals are biased towards incurring 12 casualties to save Muslim over (especially white) Christian civilians, while moderates and conservatives tend to be biased, to varying degrees, in the other direction (Christians > Muslims).
@RichardHanania 11/n So what to make of these patterns? The pro-christian bias is, relatively speaking, not hard to understand: if you identify with a religion in the US, you're a) unlikely to identify as liberal,. and b) likely to identify as Xian. So what you have is likely an in-group bias.
@RichardHanania 12/n The pattern for liberals is less explicable...until you consider that they might perceive Muslims as part of their [pro-social justice] ingroup. In other words, it could be that libs are applying the stereotypes..
@RichardHanania 13/n ..they have of Muslims and Christians in the US context to their non-US counterparts. What say you?
