From Econlog:

Last week I stumbled upon a little gem outside of Larry Iannaccone's office: a chapter by Rodney Stark and Alan Miller on the religious gender gap. Long story short: Women are more religious than men by virtually every measure in virtually every culture. ... In fact, the gender gap is smallest in the most traditional societies, and largest in the least traditional societies! In societies that approve of single motherhood, with a high abortion rate, low fertility, and high female labor force participation, the religiosity gap between women and men is especially large.

I think this points to a needs-oriented solution. Women have children. They feel the need to provide for those children. Collectivization through religious affiliation provides additional help to their children.

But why do they feel they need additional help? For married women, men are not a constant (mortality, divorce, etc.), so religion is a hedge. Single women compare what they can provide to what a married couple can provide, and feel poorer for the comparison, so they augment by collectivizing.

Addendum on 2006-12-11
Thinking about this further, I realized I needed to explain why females on average want to help their children more than males do. I think this can be explained genetically. Because the mother gives birth to the child, she knows that it's the child has her genes. The guy doesn't have the physical proof, so barring a paternity test, he must have a lower level of confidence that the child is carrying his genes. All things being equal, females on average should care more about their children's welfare than males on the basis of this information disparity.

So, if this reasoning dominates other variables in the great big Bell Curve of Life, then social charities should attract females more than men. Because religions often (citation needed) are/were the dominant social charity, females are more religious than men. A test of this would be to measure social charity participation in "traditional" and "non-traditional" societies to see if more females than men were involved.

The upshot of this is that social charities can be seen as reverse insurance policies, i.e. you pay for them after you've used them. When you were a suckling idiot, you depended on your parents and the social charity to which they belonged. Now that you are all grown up, you need to pay into that same social charity to pay back your debt.

This can be an explanation for why people who grew up in a household but then disavowed the religion would be given a hard time. Since they made it to adulthood, they have used the social insurance, but because they are opting out of the social charity, they are essentially welching on their social debt.

Since most people in the US are born into a religion (citation needed), Atheists are people who have welched on their social debt and turned their backs on their family's social charity. Since public office is ostensibly about helping those around you, voters would have a hard time believing an Atheist would make a good representative because they have freely admitted their shirking of social debt. Note that changing your religious affiliation is fine, because you're still paying into some socially recognized charity.

In order to change their electability, Atheists need to create their own social organization with an embedded social charity. Until there's a recognizable large Atheist organization (with lots of women), the electability of Atheists will continue to be impaired.

Correlation does not imply causation. Respectfully, I think your theories about women and religion are far-fetched. - Don Spidell
Hey Don...just out of curiosity do you have a hypothesis as to why women are more religious than men? I don't think Patrick is saying that he is right, I think he is taking a stab as to why that is. While I think he reasoning sounds intuitive, and a lot like evolutionary psychology, I would be open to hearing a different POV. Maybe something I am not thinking of. My addition to the conversation would be that I think women are more compassionate than men, and since a lot of churches concern themselves with doing good works for their community women may feel compelled to join up so they can also have an impact on their community. I might add that men are (typically) more driven by logic and faith is difficult to reconcile with logic (typically). I would guess though there are a lot of factors as to why women are more religious than men, and it would be interesting to think about it some more. Nathan.
Yeah, they are far-fetched ;) , but they are predictive, so we can test the predictions. It's glossing over a lot (this theory : what actually happens :: Galilean Mechanics : Quantum Mechanics ;), but I think it's a working theory. What would this theory predict that seems wrong? -- Patrick.
With regards to whether males or females are more logical, I'd have a hard time making that case, as there are many games where being the "irrational" player makes you the winner. For example, the game of Chicken is won by the first player to rip off his steering wheel and throw it out the window. So our ideas of what is rational or irrational are a bit fluid. Given this, I'd have a hard time saying that someone who throws fits and gets what they want (President Andrew Jackson) was an irrational person, given the evidence that they were highly successful. Given that we don't really yet know what actually is happening in our heads, and what I take to be waffling over what is rational and irrational, I'd not make the claim that the sexes are different in that regard. The nice thing about the above theory is that it doesn't need to prove that men and females are differently abled. -- Patrick.
I know, it was me who made that claim :) Again, it's just an idea...not something I have done any research on. :) - Nathan
I suppose it would help to define "religious" first. Does that mean church attendance? Church membership? Church participation? I'd like to see the stats on women being more religious than men before I'd offer a guess as to why they are. (I'm glad I got some discussion going with my bold statement :-) ). -- Don Spidell
Don, sure that is fair. Though, we don't have to do hard science here, I think we can just posit things and see where they lead us. I plan to read the article Patrick is referring to and see if we can tell what the criteria they used was.
Patrick, Having been a soldier in the Human Services Army for nearly a decade, I can tell you it's a female dominated field. I think this gets back to compassion. I also think it has something to do with the pathetic wages that are can make a lot more money doing much more stimulating things. I included a break there, but I am guessing Patrick's software will eat it. - Nathan

Well, I think we don't actually need to measure "religiosity". That article is just what got me thinking about this stuff. I think we can get by with just counting how often people go to church, and what percentage of income they tithe. I.e. A prediction this makes is that non-churchgoing females will probably also belong to some social charity. So, the propensity to belong to social charities ( Red Cross, Medecins Sans Frontieres, etc.) will be higher in females with low church attendance records than females who have higher church attendance (as long as their church has an embedded social charity). The assumption here is that attendance and percent of income tithed is a good proxy for social insurance payments.