A new study claims that religion may be on the way out in some parts of Europe, largely because it isn't as useful to adherents as it once was. 

The study was authored by Daniel Abrams and Haley Yale of Northwestern University's engineering and applied mathematics department and Richard Wiener, a physics professor at the University of Arizona. Looking at census data in nine different countries, the study applies a mathematical model of the dynamics between groups. The model appears to show religion eventually disappearing from those countries.

The team chose Australia, Austria, Canada, the Czech Republic, Finland, Ireland, New Zealand, Switzerland, and the Netherlands because those countries gathered census information on religious affiliations, some of it dating back more than 100 years. (In the U.S., such data is spottier). The survey data showed more people in those nations were saying they had no religious affiliation at all. The largest figure was 60 percent for the Czech Republic and 40 percent for the Netherlands.

One thing that caught the researchers' attention was the speed of change. For example, surveys taken in the Schwyz Canton in Switzerland, about 20 miles southeast of Zurich, show about 5.5 percent of respondents saying they were unaffiliated in a survey taken a decade ago. In 1950 the number was near zero. It seems to have nearly doubled every time a survey was taken since then.

The model uses a concept called relative utility. It's a measure of how useful it is to be a member of some religious group. A relative utility score of 0.5 means it is just as useful to be religious as not to be, and people will simply join the larger group (other things being equal). A relative utility of zero would mean that it is a near-absolute advantage to be a member of a religious group (this might apply in a country where apostasy is punishable by law).

Abrams said he assumed the relative utility remained constant, at least over the last century, and when he ran the data through the model he found a surprising result: the realtive utility score was greater than 0.5 over that period.

The across-the-board increase in the numbers of non-religious might be for several reasons. One, the researchers say, is that bigger groups have an easier time attracting new members (this dynamic also applies to social networking Web sites). As unaffiliated groups get larger more people see a social benefit to joining.

Another is changes in the way people live. Churches once filled a lot of the social functions in a community. That is less true in a modern society, where there are jobs, schools, and other organizations that aren't necessarily church-related. That alters the relative utility score. But in this case,

Abrams applied a similar model in 2003 to show why some languages, especially smaller regional ones, were declining. It is more socially useful to speak English, for example, in the U.K. than it is to speak Welsh, or French in France rather than Provençal.

The study also says this theory might apply to other social systems, such as encouraging people to quit smoking. As the social benefit of smoking decreases (because there are more laws restricting it) more people will be led to quit, at the same time losing the draw for new smokers.

Not everybody is convinced that the model is all that robust. Paul PZ Myers, a professor of evolutionary biology at the University of Minnesota (and prominent atheist) wrote on his blog that the perceived utility metric might be a bit circular. That's because it is derived from the data and not measured directly, and then used to calculate future trends. Also, the perceived utility is not necessarily a constant. I'd like to see more of them pick up on this mode of analysis, and then I'll trust it more, he wrote.