I went to the study looked for the information on what schools where picked. A single paragraph on the bottom of page 3 says...
5- to 12-year-old children (n = 1,151, mean (M) age = 8.29 years, SD = 2.17 years, n=559 females) were recruited from local schools in six countries around the world: Chicago (USA), Toronto (Canada), Cape Town (South Africa), Istan-bul and Izmir (Turkey), Amman (Jordan), and Guangzhou (China)
I could not find any other information on what kind of schools were used in the study. To be fair, I pressed ctrl-f and typed in 'school', so maybe more information is in there. In addition, the locations of the schools is very lopsided. There is 3 from the Africa and none from South America or Europe. I am not accusing the authors of intentionally leaving out data; I simple expected more data on the schools to be presented.
Also, maybe the make-up of each classroom has an effect. Imagine a class of 19 Christians and 1 Muslim in the deep South and a second class of 20 Muslim children in Cairo. I would expect the Muslim child to share less in the former than in the later. I realize that the study says Christian and Muslim children are less altruistic, but that is just an example of how demographics can change the result.
The is one last thing that has been bugging me. They showed that Christian and Muslim children are less altruistic, but said that religious children are less altruistic. 54.6% of the world's population is Christian or Muslim, but 29% of the world's population is affiliated with a different religion.*
(I feel awesome by just typing that)