They didn't do a study to find it, they analyzed a bunch of other studies that were done for other reasons to come up with this. Basically, they took studies that had groups of religious and non-religious people and used the data, defined intelligence in some specific way, then related it to the question "which group is more intelligent?"
My question is what is the control group that would bring validity to such a study? It's easy to massage data to make people who don't believe in God feel good. It would be very easy for me to say, "See? I'm smarter!" but then, because I actually am intelligent I spent some time reading how they came up with this and realize it was not a controlled study looking to answer this question, nor was it 53 controlled studies to find the answer to this question... it really was just some people playing with statistics of people who participated in some other studies that fall within their specific definition of intelligence.