Scientific American, in "Getting a Rational Grip on Religion -- Is religion a fit subject for scientific scrutiny," poses a legitimate, in my humble opinion, question. I think the short answer is, yes religion is a fit subject for scientific inquiry. There is no reason not to use the scientific method to study religion or religious phenomena. The problem, when there is one, comes when scientists fail in the proper use of their own methods.
There is a natural tendency to do so. After all, the scientist is typically self-limited to the use of strictly materialistic hypotheses about observations that may relate to non-material phenomena. No scientific method has ever been successfully developed that was able to incorporate extra-physical hypotheses. Thus it is easiest to take a reductionist approach to phenomena that are not purely physical in nature. The most obvious example is attempts to explain away free will: either to claim (as Stephen Hawking does) that free will is an illusion or to claim (as John Polkinghorne does) that free will is an epiphenomenon of non-linear dynamics in the human brain. Hawking errs in claiming that free will does not actually exist, though his conclusion follows quite logically from his major premise (that all phenomena are physical). Polkinghorne errs, not in the conclusion that free will exists, but in his explanation of it, which rests on a serious misunderstanding of non-linear dynamical systems (the so-called butterfly effect). Other examples include all the attempts to explain away the miracles of Jesus. (He didn't walk on water, he walked on rocks in the water. He didn't multiply loaves and fish, they just got shared among stingy people who already had them. The man born blind was only a fake. None of these miracles ever really happened, these were all just stories made up by people who were mesmerized by Jesus, etc.) All such explanations rely on certain assumptions about the lack of veracity or the lack of mental competence or wholeness of the witnesses. Although such explanations tend to the lame, they have accumulated a vast currency in our modern secular culture.Thus, any explanatory theories that arise from such an investigation will be material in nature, and therefore will serve primarily to explain why people develop various kinds of superstitions. Since there are so-called religious phenomena that typically have a material explanation, such as hearing voices or seeing a Madonna statue appear to weep, this is usually neither a difficult enterprise nor a particularly interesting one. Most mature believers have long ago divested themselves of superstitions and a tendency to find supernatural explanations for natural phenomena.
The Scientific American article is a review of a new book by Daniel C. Dennett, Breaking the Spell: Religion as a Natural Phenomenon. Although the title suggests Professor Bennett is in the tradition of the highly logical Hawkings of the world who simply say that a phenomenon cannot be what it seems to be, it turns out that he is more in the spirit of the Polkinghornes who latch onto a trendy idea that seems to explain something that's difficult to explain. In the particular book in question, that idea seems to be that all ghost sightings (and, ultimately, all supernatural experiences) are mental projections in the imagination, a type of schizoid psychotic phenomenon. According to Benet, these schizoid phenomena get commemorated in cultural traditions which get passed on from one generation to another. The explanatory power of his theory, however, doesn't lie in how these traditions get generated, but in how they are passed on. His theory builds on recent work in sociology that studies cultural units that get passed on by emulation, an extra-genetic method of inheritance. In that theory, such a unit is referred to as a "meme." (Please note, meme rhymes with gene.) What's new here is not so much theory or observation as it is terminology.The result would be a cacophony of superstitions--memes vying with memes--some more likely to proliferate than others. In a world where agriculture was drawing people to aggregate in larger and larger settlements, it would be beneficial to believe you had been commanded by a stern god to honor and protect your neighbors, those who share your beliefs instead of your DNA. Casting this god as a father figure also seems like a natural. Parents have a genetic stake in giving their children advice that improves their odds for survival. You'd have less reason to put your trust in a Flying Spaghetti Monster.
Thus belief in a stern father-god is more adaptive than belief in a flying spaghetti monster, and is therefore more likely to survive in a culture. In other words, Bennett's study contributes only to the already vast literature on the cultural endurance of superstition, and it evidently does so only by introducing new terminology to represent old phenomena. Where it fails is in its attempt to explain religion in general as simply another example of superstition. As such, it joins a cultural unit (a "meme") with a great deal of staying power, the myth that all religion is mere superstition.