The Marriage Gap

Gallup takes a look at the "marriage gap" in American politics. I'm always waiting for analysts to dig a bit deeper into this issue. Marriage would seem to be a trait that's correlated with a lot of other politically important elements of personal identity. For example, a white person is more likely to be married than is a black person and there is, of course, a large "race gap" in voting patterns. Similarly, one assumes that more religious people are more likely to be married than are less religious people. Twentysomething secular college-educated urbanites tend to be Democrats, and tend to be unmarried, but my strong sense is that married twentysomething secular college-educated urbanites are Democrats just like those of us who are unmarried.

What's really needed, in short, is a bigger sample and some multivariate regression to try to see if there's any reason to think that marriage is exercising a large independent influence on American political behavior.

Comments

I agree. Add to the list of independent variables - just in case someone is planning on doing this - levels of home ownership and the percentage that are parents.

Posted by Curtis | February 20, 2008 2:23 PM

Perhaps I am an outlier; it took me half a decade of marriage but I eventually converted my spouse to a registered Democrat. My in-laws are still wingnuts every one.

Marriage in no way made me less likely to be a Democrat.

Posted by moxie | February 20, 2008 2:23 PM

What's the third category? Those don't add up to 100%. Married and not married should, theoretically, add up to 100%.

Posted by Mo | February 20, 2008 2:25 PM

All my college-educated married friends are Democrats and all my college-educated single friends are Democrats. I would also tend to believe that other factors are more important than marriage in terms of political affiliation.

Posted by drjimcooper | February 20, 2008 2:26 PM

My totally unscientific intuition about this is that marriage is an effect rather than a cause of political inclinations, i.e. liberal progressive types are somewhat less likely to get married than are traditional conservative types.

Posted by Virginia | February 20, 2008 2:32 PM

What's the third category? Those don't add up to 100%. Married and not married should, theoretically, add up to 100%.

No: The three options on political IDs should add up to 100%. So, for married people: 32% D + 32% I + 35% R= 99%. For not marrieds, D+I+R=98%. This chart doesn't say that 32% of Dems are married and 41% are not married; it says 32% of married people are Dems and 41% of not-marrieds are Dems.

Posted by flippantangel | February 20, 2008 2:40 PM

Steve Sailer's way ahead of you on this one. See, for example, his recent post: affordability of family formation. The correlation is to high to ignore here.

Posted by Juan | February 20, 2008 2:41 PM

People in large urban areas tend to get married later. People in large urban areas tend to be Democrats. So, all this chart shows is that...people in large urban areas tend to be Democrats.

Posted by Peter Bautista | February 20, 2008 2:41 PM

flippant,

Ahh, that makes more sense. Thanks.

Posted by Mo | February 20, 2008 2:42 PM

Also less likely to be married: women over the age of 65, a Democratic voting bloc. And then there are gays and lesbians, who I assume lean Democratic although not to an extreme degree.

Posted by ChuckE | February 20, 2008 2:44 PM

I'm always waiting for analysts to dig a bit deeper into this issue.

What's to dig? The analysis is very simple: get married, people!

Posted by Al | February 20, 2008 2:45 PM

+1 (err, 2 including my husband) twentysomething, secular, college-educated, urban Republicans, for what that's worth (little).

FWIW, marriage did make me somewhat more Republican, in that I see the GOP as more likely to get rid of the ridiculous marriage penalty (the tax brackets that are less than double that for single people, the fact that I pay his marginal on my first dollar of income, the IRA contribution limits that are less than double that for singles, etc.) that discourages work among high-income people, not to mention discouraging marriage among lower-income folks.

Posted by Amber | February 20, 2008 2:49 PM

It is hard to hold everything else constant and just look at one item.

Smart people are more likely to be married than dumb people.
Rich people are more likely to be married than poor people.
Old people, to a point, are more likely to be married than very young people.
Heathly people are more likely to be married than sick people.

Then you have the other side
Religious people are more likely to be poorer and not as smart as non religous people
People who came of age during the Bush years are more likely to be Democrats than older people.

Then you get into the really politicaly correct positions
Blacks are not as smart as white who are not as smart as East Asains.
Smokers are not as smart as non smokers.
Gays are rarely allowed to be married.

Mix all of these points together and you end up without enough information to form a valid conclusion.

Did you know that children with bigger feet can spell better than children with little feet?

Posted by neil wilson | February 20, 2008 2:55 PM

Less likely to get married: younger, urban, and liberal.

I can count multiple couples who aren't married because their gay friends can't and ergo it's bullshit.

Posted by dbt | February 20, 2008 2:55 PM

FWIW, marriage did make me somewhat more Republican, in that I see the GOP as more likely to get rid of the ridiculous marriage penalty (the tax brackets that are less than double that for single people, the fact that I pay his marginal on my first dollar of income, the IRA contribution limits that are less than double that for singles, etc.) that discourages work among high-income people, not to mention discouraging marriage among lower-income folks.

Why don't you just file seperately? Then it's like you're not married at all. If one partner makes a lot more than the other, marriage is strictly better, financially. If they make the same it as least as good. What is there to complain about?

Posted by mpowell | February 20, 2008 2:56 PM

It makes sense that marrieds trend conservative.

They spend the whole day:

- sublimating self-interest to a shared good
- pooling and allocating resources
- in many cases, sacrificing money and time to take care of poor, powerless little people just because somebody's got to do it.

After all that they need a break.

Posted by ao | February 20, 2008 2:57 PM

", in that I see the GOP as more likely to get rid of the ridiculous marriage penalty"

FWIW my wife and I pay less taxes now that we're married versus when we were single. We also would pay more total taxes if we filed "married filing separately" which has the same tax brackets as single. I just don't see where this "marriage penalty" talk comes from.
It is true that there is some encouragement for one person to not work in high-income families, but there are tax advantages to it. Filing married/jointly instead of head of household will reduce the overall tax rate. Anyways, I have a feeling in plenty of families where the husband makes a high income the wife wants to stay at home regardless of the tax rate. Even a 0% tax rate on that extra income may not be enough to coax them into working.

Secondly, tax cuts without reductions in spending are meaningless (unless you plan on dying soon). It just means we'll be paying for the spending at some point, just not right now

Posted by JordanT | February 20, 2008 2:58 PM

Its good to see interest in demographics. Still, lets say the obvious...

Income

Affluent people get married, have stable lives, children, better health care, etc. (Think millionaires not billionaires please.)

Outside of NYC and DC there are millions of people who settle down for a few decades and rake in their share of the pie and get quite comfortable doing so.

That electoral college is here to stay urbanites.

Posted by Kim | February 20, 2008 3:00 PM

I agree with the pointing to Mr. Steve Sailer's analysis.

I think one should also see how age correlates with party identification.

I believe the younger generations (i.e., Xs and especially Ys) are emphatically not Republican.

Posted by bartkid | February 20, 2008 3:00 PM

Steve Sailer has a recent column correlating the cost of raising a family with states voting for Bush vs. voting for Kerry, finding a very strong correlation. Basically, red states are those in which it is inexpensive to raise a family, and blue states are those in which it is expensive. Red state women end up spending a greater portion of their (rather pathetic) lives actually raising children than blue state women do.

Posted by Jim W | February 20, 2008 3:01 PM

I'm a married twentysomething (earlier than most - married at 23) quasi-religious college-educated urbanite and I am and always have been a Democrat. But that's just it, I always have been a Democrat. I was before I got married, and getting married was not formative event in my political identity. I doubt that it is for anybody.

Posted by Aaron | February 20, 2008 3:10 PM

Age gap.

Young people tend to be liberal and typically aren't married yet.

Posted by TH | February 20, 2008 3:15 PM

in that I see the GOP as more likely to get rid of the ridiculous marriage penalty

Translation: I want the government to subsidize my life-style choices more than they already do.

Posted by mpowell | February 20, 2008 3:18 PM

We discussed this in my campaign politics class here at Ohio State. The professor concluded that the perceived marriage gap in voting patterns among women may have been a wage gap, as single women, particularly single mothers, are more likely to have lower income and therefore more likely to vote Democratic.

Posted by Matthew Struhar | February 20, 2008 3:25 PM

It also seems problematic to look at "not married" people as a single group: You've got your not-married-never-had-kids folks, you've got single parents, you've got divorced people, you've got tons of old women who simply outlived their husbands. Why would you expect the very different ways in which those people are "not married" to lead to similar political views diffrent from "marrieds"?

Posted by flippantangel | February 20, 2008 3:26 PM

Good point re: the variety of "not marrieds." OTOH it seems to me that if there's something about being married that matters, then it shouldn't matter that there's diversity among non-marrieds. So for example there are lots of races that are not Caucasian, but there's something about Caucasians that leads them to be more Republicans, so it's still useful to know what that thing is. Or: there are lots of ways to not be super-rich, but super-rich people tend to be Republicans, so we can look for something about being super-rich that matters.

Posted by Ted | February 20, 2008 4:03 PM

Re: What's really needed, in short, is a bigger sample and some multivariate regression

Mr. Yglesias,

I believe you mean 'multiple' regression, not 'multivariate'. 'Multiple' regression means having more than one independent variable, 'multivariate' means more than one dependent variable.

Posted by Hector | February 20, 2008 4:25 PM

I believe you mean 'multiple' regression, not 'multivariate'. 'Multiple' regression means having more than one independent variable, 'multivariate' means more than one dependent variable.

Nope.

Multivariate regression: y = x'b + e where x is a vector that can include multiple right-hand side variables. One dependent and multiple "independent."

Multiple regression: a phrase used by people who don't really know anything about empirical analysis.

And marriage would be endogenous as a right-hand side variable, so who cares about actual estimating anything.

Posted by SavageView | February 20, 2008 4:38 PM

Married filing separately is not the same as single. For instance, the 25 percent tax bracket is (roughly) 30-70k for singles, 60-120k for married filing jointly, and 30-60k for married filing separately. The brackets for married filing separately are half those for married filing jointly, NOT equal to those for single. (There are lots of other disincentives to file separately, including that you can't make Roth IRA contributions if you make more than 10k.)

It's not that I want the government to subsidize my lifestyle choice, honestly. It's that I want tax policy to promote work and productivity over not work. If you raised marginal rates across the board to cover the elimination of the marriage penalty, I'd be fine with that.

Last year, when my husband and I weren't married, we lived in the same house, had our same jobs, same mortgage, etc. Then, we both paid 0/15/25 percent income tax. This year, he got a raise, so he'll pay 0/15/25/28 percent, while I'll pay 28 percent on the first and every dollar. Our lifestyle didn't change at all--only our tax rates did. Now I work because I don't have anything better to do, and honestly, I'll probably always work, because I enjoy it. But it's pretty ludicrous to say that women either want to work or want to stay home, and that tax rates/other incentives don't affect their decisions. It's the marginal effects that matter.

Posted by Amber | February 20, 2008 4:39 PM

http://www.amconmag.com/2008/2008_02_11/article.html

Steve Sailer's wonderful recent full length piece on this subject is on line here.

Posted by anotherScott | February 20, 2008 4:47 PM

Someone commented that some liberal heteros don't get married out of solidarity with gays. Is this a widespread sentiment, or a pretty eccentric one? Never heard of it before.

Posted by fromMars | February 20, 2008 4:53 PM

"It also seems problematic to look at "not married" people as a single group"

This data is not too helpful; need to know age breaks of the respondents among other variables. This is just the filler soundbites that the 24/7 cable news like to throw out there. Not useful for planning political campaigns.

Posted by CParis | February 20, 2008 4:58 PM

From my new article "Value Voters" in The American Conservative, here's the answer to Matt's question:

"Democratic pollster Stanley Greenberg’s multiple regression analysis of the 2004 exit polls revealed:

"'The marriage gap is one of the most important cleavages in electoral politics. This is true even when controlling for other demographic and behavioral factors such as gender, age, race, gun ownership, union household membership, _party identification_, education, income, and church attendance. Controlling for all these other variables, the odds of voting for Kerry were 1.56 times greater if the voter was unmarried than if the voter was married. In contrast, once other demographic and behavioral factors were controlled for, a voter’s gender had no significant effect on their likelihood to vote for the Democrat. [Italics mine]

http://www.amconmag.com/2008/2008_02_11/article.html

Posted by Steve Sailer | February 20, 2008 5:02 PM

Also from my "Value Voters" article:

"The impact of marriage on the Red-Blue divide among states was long difficult to quantify graphically because the government only provides [summary] data on “getting married,” yet it’s “being married” that drives voters toward the GOP. Many white people get married in Nevada, for example, but the state is only purplish-red because they also get divorced frequently. Next door Utah, however, is the most rock-ribbed Red State (Bush won 72 percent) because the locals get married and stay married.

"Consequently, I devised a measure called “Years Married” (modeled on Total Fertility) that estimated how many years a woman could expect to be married during her childbearing years of 18-44.

"For example, white women in Utah lead the nation by being married an average of 17.0 years during those 27 years from age 18 through 44. In contrast, in liberal Washington D.C., the average white woman is married only 7.4 years. In Massachusetts, where Bush won merely 37 percent, years married average just 12.2.

"Applied to white women, this new measure proved to be the single-best predictor imaginable of Bush’s share of the vote by state in the last two elections. Bush carried the top 25 states, while Kerry won 16 of the lowest 19.

"The 2004 correlation coefficient was a stratospheric 0.91, accounting for an astonishing r-squared equal to 83 percent of total variation in voting by state. This has to be one of the highest correlations for an unexpected factor ever seen in political science."

http://www.amconmag.com/2008/2008_02_11/article.html

Posted by Steve Sailer | February 20, 2008 5:09 PM

Many astute observations of our political scene are contained in "Simpsons" cartoon. When the citizens of Springfield debate such issues like bear attacks, taxes, aliens, one lady always pipes in "and what about the children?"

I think that many people, especially women, adopt a more narrow outlook after marriage. Perhaps their news sources also change, as well as the quantity and quality of their news exposure.

Finally, GOP harps at their "pro-family agenda" and that somehow convinces many people. Perhaps even the singles, but with the opposite effect.

Posted by piotr | February 20, 2008 5:33 PM

Yeah, I have to chime in and note that marriage is a highly correlated variable --- married individuals tend to be a little bit whiter, a bit older, a bit higher income and asset accumulation, and a hell of a lot straighter, or at least more accomplished at having a wide stance than the entire voting age population. All of those correlating factors tend to produce a little bit more Republican voter than the general population.

Posted by fester | February 20, 2008 5:48 PM

Marriage --> kids --> scared witless --> Republican.

Posted by EpicureanQuaker | February 20, 2008 6:09 PM

"Red state women end up spending a greater portion of their (rather pathetic) lives actually raising children than blue state women do."

What's really pathetic is when superannuated blue state women belatedly wish they gotten married and had kids like the red state women they mock. See, for example, Lori Gottlieb's essay about settling in the current Atlantic, Marry Him! (Matt, shouldn't you have plugged this yourself?).

Posted by Fred | February 20, 2008 6:57 PM

That's the main theme of Maureen Dowd's writing: how her conservative Catholic family kept telling her that if she lived out all this feminist ideology but how women should be just as competitive as men, she'd end up alone, childless, and embittered. And it drives her crazy that they were right.

Maureen could have settled a long time ago for John Tierney, but she had to keep going for the out-of-her-league Michael Douglas types. But Michael married Catherine Zeta-Jones instead of her.

Posted by Steve Sailer | February 20, 2008 7:02 PM

SavageView,

Nope. Multiple regression means more than 1 predictor variables. Multivariate regression means more than 1 response variables.

http://www.gseis.ucla.edu/courses/ed231a1/notes2/mvreg.html

"Multivariate multiple regression is a logical extension of the multiple regression concept to allow for multiple response (dependent) variables. Multivariate regression estimates the same coefficients and standard errors as one would obtain using separate OLS regressions. In addition, multivariate regression, being a joint estimator, also estimates the between-equation covariances. This means that it is possible to test coefficient across equations.

"The matrix formula for multivariate regression is virtually identical to the OLS formula with the only change being that Y is a matrix response variables and not a vector."

http://www2.chass.ncsu.edu/garson/PA765/regress.htm

"Multiple regression, a time-honored technique going back to Pearson's 1908 use of it, is employed to account for (predict) the variance in an interval dependent, based on linear combinations of interval, dichotomous, or dummy independent variables. Multiple regression can establish that a set of independent variables explains a proportion of the variance in a dependent variable at a significant level (through a significance test of R2), and can establish the relative predictive importance of the independent variables (by comparing beta weights)."

Posted by Hector | February 20, 2008 7:15 PM

Mr. Sailer,

"The 2004 correlation coefficient was a stratospheric 0.91, accounting for an astonishing r-squared equal to 83 percent of total variation in voting by state. This has to be one of the highest correlations for an unexpected factor ever seen in political science."

If this finding is so astonishing, how come you publiched it in a journal of opinion, and not in a professional social-science journal?

Come to think of it, why don't you try publishing some of your race theories in some professional journals, if they're so good?

Posted by Hector | February 20, 2008 7:24 PM

"If this finding is so astonishing, how come you publiched it in a journal of opinion, and not in a professional social-science journal?"

Because, among other reasons, The American Conservative pays me money for articles. I like money. You can exchange it for food and shelter.

Posted by Steve Sailer | February 20, 2008 8:13 PM

By the way, an earlier finding of mine that I published in The American Conservative in December 2004 -- that Bush's share of the vote by state in 2004 was very highly correlated (r-squared of 74%) with the Total Fertility Rate in each state among non-Hispanic whites -- was subsequently replicated by professional social scientists.

In a 2006 paper entitled “The ‘Second Demographic Transition’ in the US,” demographers Ron J. Lesthaeghe and Lisa Neidert of the University of Michigan confirmed the findings that I first published in my “Baby Gap” article in The American Conservative in 2004: the white total fertility rate correlates extraordinarily well with whether a state voted for Bush or Kerry. They note that this provides “to our knowledge one of the highest spatial correlations between demographic and voting behavior on record.”

Yet the Baby Gap appears to be somewhat less important than the Marriage Gap.

Posted by Steve Sailer | February 20, 2008 8:25 PM

Steve,

Is there big money in articles about black people being stupid, the oncoming brown horde from the South, and Jews running our foreign policy, you know, hypothetically?

You must live and eat like a king.

Posted by SoCalJustice | February 20, 2008 8:26 PM

SoCalJustice, doesn't it just eat you up alive inside to know that Sailer gets taken seriously by people like Columbia University's Andrew Gelman, among others? So what have you accomplished in life?
http://www.stat.columbia.edu/%7Ecook/movabletype/archives/2008/02/red_states_blue.html

There's a discussion of Sailer's affordable family formation theory at that link at Gelman's site.

Posted by scottynx | February 20, 2008 9:00 PM

No, it doesn't eat me up.

Steve got quoted on some professor's blog. Is that really the best his p.r. sycophant can do?

Good for him.

So what have you accomplished in life?

Amongst other things, I'm not publicly on record claiming Black people are genetically and intellectually inferior to White people.

How about you, Chappie?

Posted by SoCalJustice | February 20, 2008 9:18 PM

scottnyx,

You know what is really funny, though?

When the professor realizes just who are the type of people that read Steve's blog, then feels the need to comment on it, but is too polite to be honest about what he just witnessed.

Maybe Steve is onto something about "political correctness."

That was a classic moment, thanks for the link!

Posted by SoCalJustice | February 20, 2008 9:29 PM

Sailer's piece contains some interesting info, but unfortunately it ends up in tedious immigrant-bashing territory the author calls home. Also, Sailer fails to note that GOP policies, nationally and often at state and local level, work against "affordable family formation", and if people are voting GOP for that reason then they are idiots voting to screw themselves and their children.

Posted by Jonf | February 20, 2008 9:58 PM

SoCalJustice,

If Sailer's main goal were making money, he'd be politically correct. There's more money to be made writing falsehoods that make people like you feel good about yourself than there is in writing honestly about people as they are. Sailer has managed to make a living without compromising his integrity. Like the Montgomery Gentry song goes, "that's something to be proud of".

Posted by Fred | February 20, 2008 10:49 PM

No, Fred, Sailer has a career because there's a small but steady demand for respectable-sounding racism. If he were non-racist (or, in your phrase, "politically correct") he'd be competing against every other writer out there. Instead, he's the smartest, best-informed racist writer in America. Congratulations, Sailer!

Posted by too many steves | February 20, 2008 11:05 PM

Fred,

Who ever said his "main goal" was making money?

Sailer has managed to make a living without compromising his integrity.

You should bust that out the next time there's an open mic night at your local comedy club. The crowd might not get it, likely having no idea who he is, but it's pretty hilarious nonetheless.

Posted by SoCalJustice | February 20, 2008 11:18 PM

"Someone commented that some liberal heteros don't get married out of solidarity with gays. Is this a widespread sentiment, or a pretty eccentric one? Never heard of it before."

This is the kind of thing that its cute to say if you're an urban liberal hetero. Once the moral barriers against living together without marriage come down, people are going to latch on to any fine-sounding nonsense that justifies sex and companionship without religious or legal commitment. In the '60s they would have talked about free love. In the '80s they would have talked about taxes. People afraid of divorce and unsure of whether they've chosen a permanent partner will say anything to avoid the marriage talk. I'm not other-bashing here; in fact, I'm in an unmarried cohabitating relationship myself. At least I'm not so egocentric and faux-idealistic as to pretend I'm making the difference for gay rights by living with a girl and not marrying her.

Also, marriage penalty complaints are BS (Amber). The "penalty" is insignificant, especially if Bush's tax cuts are made permanent. I don't know how much it was before, but at this point the married brackets are exactly double the single brackets at low levels, and only slightly less than double at high levels. If you think two rich people in a legal relationship aren't able to shoulder slightly more of a tax burden than their single fellow-citizens, they you should get a divorce to save your $800 or whatever it comes to. You just told us that your husband makes over $128,000/year and you have your own job, and you're whining about a few percentage points here and there?

Posted by cgaros | February 21, 2008 12:00 AM

SoCalJustice,

Why do you feel the need to try to bait Sailer? If you disagree with one of his points, argue your side. Resorting to ad hominem petulance seems pathetic.

Posted by Fred | February 21, 2008 7:45 AM

Re: Because, among other reasons, The American Conservative pays me money for articles. I like money. You can exchange it for food and shelter.

Weak, Sailer, very weak. If you think your ideas are so good, then why don't you find an evolutonary psychologist at a university to collaborate with you on an article, and try to get your race theories published in a professional journal. Or even in some semi-popular journal somewhere that is read by both scientists in the field and the general public.

What's the real reason you stay away from the professional journals and the professional scientists, Sailer?

Posted by Hector | February 21, 2008 8:52 AM

Re: Sailer has managed to make a living without compromising his integrity.

Oh, I'm sure. Seems to me though that if one's 'integrity' involves theories of racial superiority, then give me a good old hypocrite any day.

Too many steves,

Re: Instead, he's the smartest, best-informed racist writer in America.

This was bloody hilarious. I almost choked on my coffee laughing.

Posted by Hector | February 21, 2008 8:59 AM

Fred,

Why do you feel the need to regulate free speech?

I thought you "conservatives" were against that kind of thing?

Steve Sailer is many things, including smart and funny.

The fact that he uses his talents, such that they are, to race bait and pander to the worst instincts of (thankfully, only some) White Americans is a tremendous stain on him and any outlet that publishes him - and should be open to mockery and, at times, scorn.

Plus, it's fun to "bait" race baiters.

Does that answer your question, Fred?

Here's a question for you: why do you feel he needs your protection? He's an adult, he can, and should, defend himself.

Posted by SoCalJustice | February 21, 2008 9:53 AM

Weak, Sailer, very weak. If you think your ideas are so good, then why don't you find an evolutonary psychologist at a university to collaborate with you on an article, and try to get your race theories published in a professional journal.

You need to bone up on modern philosophy - and sociology - of science. Kuhn and Feyerabend and so forth. If only science were the disintrested search for truth! The fact is that scientists have their biases and their cultures that exclude the views of outsiders.

Posted by Justin | February 21, 2008 9:57 AM

Who's regulating speech, SoCal? I just asked why you felt the need to resort to ad hominem attacks in the hopes that it might prompt you to engage in a discussion of ideas instead. That you chose to attack the man instead of critiquing his ideas indicates that you are either unwilling or unable to engage in such a discussion.

Posted by Fred | February 21, 2008 1:46 PM

There's an economics journal article that had some evidence that states with no-fault divorce laws had bigger "gender gaps" in partisan preference. In addition, married women are more likely to vote Republican than unmarried women. The basic theory behind the article is that women gravitate to the Democrats if they believe that they cannot reliably get economic security based on a man's financial support, but will vote Republican if they economic support from a husband is something they can reasonably depend on (e.g., if they are married, if they live in states with stricter divorce laws).

Posted by jonp72 | February 21, 2008 2:30 PM