Although I don't doubt the result, the methodology matters a lot in a survey like this. Especially the sampling methods. If groups were sampled from different regions or socio-economic groups, that could have an effect.
The first study of Muslim Americans was fielded via telephone to Muslim Americans and a nationally representative sample of adults in the U.S. aged 18 and older. The Muslim-American sample was selected from self-identified Muslim Americans who agreed to be recontacted after participating in the Gallup nightly polling. The general population sample included landline as well as cell phone-only respondents. The survey was administered from Feb. 10, 2010-March 11, 2010, and featured a five-call design. Because of the low number of Jewish American respondents in the sample of U.S. adults, an oversample of the Jewish population was performed using recontacts from the Gallup nightly polling. The data were weighted to correct for disproportionalities in probabilities of selection and response propensities. The data were then weighted to targets for age, gender, region, race, ethnicity, and education from the U.S. Census Bureau. Final weights were applied based on self-identified religious affiliation using targets from the Gallup nightly polling. The response rate for the study was 21%.
Why do they stick us agnostics with the atheists? That's almost like putting us with the Catholics. ;P Seriously, though, I'd like to see if there is a difference between the types of non-believers.