Lisa Blaydes and Drew Linzer have a “very interesting paper”:http://www.stanford.edu/~blaydes/AA_ISA.pdf. As they note, public debate in the US revolves around two quasi-theories – that Muslims (as the cliche goes) ‘hate us for our freedoms’ (or as Blaydes and Linzer put it more prosaicly, ‘the nature of Islamist anti-Americanism is cultural rather than military or political’), or, alternatively, that many Muslims are anti-American because of what the US does in the world. Both theories presuppose that ‘individuals form their opinions about the United States primarily as a _direct_ reaction to what the U.S. is or does.’ But, as Blaydes and Linzer point out, political elites likely play an important role in shaping public opinion. Blaydes and Linzer
bq. submit that observed levels of anti-Americanism among Muslims in a given country depend primarily on the intensity of anti-American messages being voiced by prominent political elites within that country. Simply put, the reason many Muslims tell public opinion researchers that they hold an unfavorable opinion of the United States is because trusted political leaders tell them so. But what is especially important about this association is that it is predominantly _domestic_ forces that determine the strength with which elites press anti-American claims. In particular, when competition between a country’s Islamist and secular-national political factions is great, political leaders from both sides have strong incentives to use anti-American rhetoric to boost mass support. Less intense conflict between these two groups dampens those incentives, leading to more balanced elite attitudes towards the United States, and thus less anti-American sentiment in the minds of individual Muslims. Since Islamists tend to be more anti-American than their secular counterparts, this logic explains why|seemingly paradoxically|while religious Muslims are more anti-American than their less pious compatriots, anti-American attitudes are most prevalent in more _secular_ countries where the political division between religious and non-religious individuals is the greatest.
They find that:
bq. The variable measuring a country’s level of reformer-Islamist struggle can, in particular, explain variation of fity percentage points or more in the probability of holding an unfavorable attitude towards America (depending on values of the other two variables). Among Muslims, anti-American attitudes are most prevalent when reformers and Islamists are most competitive, in countries that are predominantly Muslim, and where the level of imports from the United States on a per capita basis is large.
Not my area – so there may be nuances that I am missing here – but this seems to me like an especially nice and publicly salient paper.