Kevin Lewis, the author of the Boston Globe "Uncommon Knowledge" derived from recent publications in the social science journals, touts yet another study of the effects of prayer. In this case it's asserting of "how prayer prevents drinking".

A recent study supports an interesting approach to curbing alcohol consumption: regular prayer. In surveys, people who reported praying more often also reported less alcohol consumption and fewer alcohol-related problems, and more prayer was associated with less consumption and fewer problems over the next several months. Of course, people who pray a lot may be less prone to drink anyway, so the researchers randomly assigned people to regular prayer or nonprayer tasks and then asked them to report their alcohol consumption after four weeks. Those who were assigned to pray drank significantly less than those who weren’t.

Now, I am religious, I suspect, but don't know, that thinking and meditating on ethics, morality, closely examining my conduct for its morality, as well as its rationality, are good and effective. I've got nothing at all against prayer, so long as it's not of the kind that Mark Twain cites in one of his better known short pieces. But prayer cannot be the subject of scientific study.

In the abstract of this study, published in Psychology of Addictive Behaviors it says:

Participants assigned to pray every day (either an undirected prayer or a prayer for a relationship partner) for 4 weeks drank about half as much alcohol at the conclusion of the study as control participants.

"Undirected prayer" starts out being too undefined and undefinable. As I pointed out here before, you cannot even control whether or not the activity you're calling "prayer" is the same action even twice in a row, not even in the same person. The same person very well might be doing different "things" depending on the impossible to define "state of mind" they are in from minute to minute. And "prayer", even as defined in the vast and changing ideas of various religions and religious people, is of wildly different formal definition. Is the rote mumbling of the 23rd Psalm the same kind of thing as the intense concentration of mindfulness meditation?

If you can't define a behavior in order to isolate that behavior, if you can't detect its actual presence except by self reporting*, if you can't reliably exclude it from your control group, what do your results mean? How can you know what they mean?