Do we choose to believe in God? Do we choose not to believe? In one sense, beliefs are not under our control. I cannot, strictly speaking, choose to believe something. In the philosophy classes I teach, I often discuss the nature of belief with my students, and point out our lack of direct control over our beliefs.
For example, if I tell you that I will pay you $1,000 if you sincerely believe that a pink elephant is flying outside your window right now, you cannot do it. You can say you believe it, you can even want to believe it because you'd like the extra cash, but you cannot in fact will yourself to believe it. Why not? Because there is no evidence for the claim, and a mountain of evidence against it.
This applies to religious belief and the lack of it in the following way. One's parents, culture, and society may apply various pressures to have particular religious beliefs (or not), but in my view authentic belief is not produced in this way. Authentic belief is influenced by many things, such as environmental factors, but what is most important is our view of the available evidence for or against a belief. In this way, we have indirect control over our beliefs (what philosophers call "indirect doxastic voluntarism"--we like fancy names for things).
This means that while I cannot directly control whether or not I believe in God, I can control it indirectly by taking stock of the best arguments and evidence on each side of the issue. In this way, I can indirectly choose what to believe, insofar as I make a good faith effort at understanding and evaluating the best available evidence. Then, as a rational being, I follow the evidence. Our choice, then, is to do our best to seek out the truth, wherever that leads us.
Truth, however, poses a problem for the naturalist. The notion of true belief is more at home in a theistic world rather than a naturalistic/atheistic one. Here's why. If our origins and our current cognitive abilities are completely explained by Darwinian processes, then we have reason to doubt the reliability of those processes for producing true beliefs. This is because natural selection does not select for true belief, but rather for behavior that is conducive to survival and reproduction. But there are many sets of beliefs and desires that will yield behavior that is conducive to survival and reproduction, and yet many of these possible sets of beliefs are not only irrational but false.
For example, a prehistoric hominid might really want to cuddle with a tiger, and he might believe that the best way to get a tiger to cuddle with you is to run away and hide in a cave whenever you see one. Of course, the beliefs of our prehistoric hominid are false, but they are conducive to behavior that has survival value. We are in the same situation, if naturalism is true, with respect to the trustworthiness of our cognitive faculties. We have reason to mistrust our beliefs, including our belief that naturalism is true, if naturalism is in fact true. This is because our beliefs have been selected for insofar as they are conducive to survival and reproduction, rather than for their truth value.
However, on a theistic worldview, we have reason to trust our cognitive processes. If God is good, and wants us to know the truth about reality, then God would give us reliable cognitive mechanisms that enable us to represent the world as it really is, at least to a significant degree. So, while the debate about faith and reason continues, there is a solid argument that reason is not at home in a naturalistic world. Reason, rather, finds its home most naturally in a theistic worldview.
For more on the argument against naturalism and responses in defense of naturalism, check out this collection of essays. For a good place to start on the issue of religious belief, see God? A Debate between a Christian and an Atheist.