More Calculus Please

I've long been sympathetic to the argument, advanced over the years in various forms by members of the Greeberg Quinlan Rosner team, that the "marriage gap" is an under-recognized feature of American politics and that one of liberalism's most promising growth areas is simply in finding better ways to engage and mobilize unmarried women who are a large and quite progressive bloc of the population with low voter turnout rates. You can see the latest form of the argument in this report and, as I say, I find it convincing.

I do, however, keep being disappointed by the relative lack of statistical sophistication you see here. After all, unmarried people are demographically quite different from married people in a number of ways including age, race, sexual orientation and religious affiliation — all characteristics that are plausibly big driver's of voting behavior. They do a decent job of showing that the "marriage gap" holds up even when you look at the major sub-samples of the population (it's not, in short, just driven by the different marriage rates of blacks, whites, and Latinos) but this is still a pretty crude way of looking at the interplay of factors. What would really be nice would be some regression analysis that could help us try to estimate the impact of marriage independent of other demographic factors.

Relatedly, it's always worth saying that proposals to "target" this or that slice of the electorate sometimes seem to me to involve underestimating the heterogeneity of the group. It's true, for example, that one would expect a 25 year-old unmarried white woman who graduated from Wellesley, took an entry-level job at a DC think tank, and is now enrolled at Georgetown Law School and a 25 year-old unmarried African-American mother of two who dropped out of high school to both be loyal Democrats but it's not at all clear that there's a common "single woman" or even "single 25 year-old woman" characteristic that's driving this common voting behavior, even though they're both common archetypes in major American cities. A Republican strategist looking to make inroads with these voters, for example, would probably adopt different strategies depending on which woman they were trying to court.

Comments

"I do, however, keep being disappointed by the relative lack of statistical sophistication you see here. "

Worth noting that Greenberg Quinlan Rosner is presumably in the tank for one of the Democratic Presidential candidates, and that their work at this particular moment is likely designed with that in mind...

Posted by Petey | November 2, 2007 5:37 PM

Ann Coulter agrees with you - in her own way. Of course, that's why she wishes women couldn't vote, but that only goes to reinforce your point about how important unmarried women are to the Dems.

Posted by Shawn | November 2, 2007 5:54 PM

Single women without kids are a lost cause for the GOP. Their main issue tends to be unfettered access to abortion, which the GOP can't advocate without pissing off its pro-lifers.

The GOP could plausibly poach a few votes from inner city single women with kids by pushing school choice, but the handful of votes Republicans would gain wouldn't make it worth it, especially if it frightened affluent suburban parents who "chose" their good public schools by moving to an affluent area.

Posted by Fred | November 2, 2007 6:15 PM

GQR is Chris Dodd's pollster. I tend to think this sort of work is designed with the party in mind in general, as Dodd is more or less generic democrat in the sense that his personality or policies don't bring anyone new into the fold.

Posted by Nicholas Beaudrot | November 2, 2007 6:18 PM

Quite right on the stistics point. Anyone who's worked on an electoral campaign (Petey?) knows that even the biggest-name polling firms give you a packet consisting fo a written summary plus "toplines" and "crosstabs". A few simple regressions would be vastly more informative, but you never get them. Or these days, they could just give you the underlying data and let you play with it yourself. But again, doesn't happen -- polling really doesn't seem to have advanced since the 1970s.

There's a business opportunity there for someone, methinks...

Posted by lemuel pitkin | November 2, 2007 6:23 PM

"GQR is Chris Dodd's pollster."

Huh. I learned something new every day. I was assuming they were in bed with Clinton owing to previous ties, but I guess they're too liberal for Mark Penn's taste.

"I tend to think this sort of work is designed with the party in mind in general, as Dodd is more or less generic democrat in the sense that his personality or policies don't bring anyone new into the fold."

Yup.

Posted by Petey | November 2, 2007 6:30 PM

Its not calculus you want more of -- its regression coefficients. And I agree, more sophistication in these reports would be appreciated.

Posted by KC | November 2, 2007 6:37 PM

Matt,

Greenberg _did_ do a multiple regression analysis on the marriage gap and I wrote about it in early 2005:

"Greenberg built a multiple regression model and found:

"Marital status was a statistically significant predictor of likelihood to vote for Kerry…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 are Greenberg's.]

That 1.56 times greater likelihood of voting for Kerry if you are single is a remarkably large effect when you consider that Greenberg is statistically controlling for party identification along with a laundry list of other famous demographic factors.

http://www.vdare.com/Sailer/050123_vindicated.htm

Posted by Steve Sailer | November 2, 2007 6:50 PM

In December 2004, I explained how the Marriage Gap paints the electoral map red or blue:

The more years of their young adulthoods that the white people in a state spend in wedlock on average, the more Republican the state is overall.

I figured out how to estimate by state the expected number of years between the ages of 18 and 44 that a woman will be married (to be precise, married with her husband present).

For example, white women in Utah, where Bush had his best showing with 71 percent of the total vote, led the nation by being married an average of 17.0 years during those 27 years from age 18 through 44.

In contrast, in Washington D.C., where Bush only took 9 percent, the average white woman is married only 7.4 years. In Massachusetts, where Bush won merely 37 percent, her years married average just 12.2.

California is next at 12.5. That coincides with my observations: four of my seven best friends from my Los Angeles high school's class of 1976 did not get married for the first time until this millennium, when they were in their forties.

Marriage in L.A. is increasingly reverting to what it was in Jane Austen's novels: a luxury that many cannot currently—and some may never—afford.

Indeed, there is much more diversity among states in years of marriage than in years of education—a factor that is frequently discussed although it is less politically significant.

Leaving anomalous Washington D.C. aside, the range between the state with the most-educated whites (Hawaii with 14.2 years of schooling on average) and the worst (West Virginia with 12.2) is only 42% as large as the Utah-Massachusetts gap in years married.

Overall, Bush carried the top 25 states ranked on years married for white women. The correlation coefficient with Bush's share of the vote is 0.91, or 83 percent of the variation "explained." That's extremely high. Years married also correlates with the 2000 election results at the 0.89 level (80 percent). So it's no fluke.

The r-squared when years married and fertility are combined in a multiple regression model is improves to 88 percent. (Small-sounding change, perhaps, but actually an important (30%) reduction in the unaccounted variation - from 17 percent to 12 percent.)

In other words, both years married and fertility play statistically significant roles, with years married somewhat more important.

http://www.vdare.com/Sailer/050123_vindicated.htm

Posted by Steve Sailer | November 2, 2007 6:56 PM

If you want to understand the full picture of why some states vote Republican and some Democrat, you have to look at my over-arching theory of "Affordable Family Formation."

In parts of the country where it is economical to buy a house with a yard in a neighborhood with a decent public school, you’ll generally find more Republicans.

You'll find fewer in regions where it’s expensive.

It’s a stereotype that a mortgage, marriage, and babies tend to make people more conservative.

But it’s a true stereotype.

The arrow of causality points in both directions. Some family-oriented people move to family-friendly states, but the cost of forming a family also affects how many families are formed overall.

http://blog.vdare.com/archives/2005/05/08/affordable-family-formation-the-neglected-key-to-gops-future/

Posted by Steve Sailer | November 2, 2007 7:01 PM

Um, OT, but this is incredible.

Posted by NCProsecutor | November 2, 2007 8:14 PM

one of liberalism's most promising growth areas is simply in finding better ways to engage [...] unmarried women

We need about 10 million diamond rings, and a whole shitload of flowers.

Posted by The Confidence Man | November 2, 2007 8:43 PM

re: figured out how to estimate by state the expected number of years between the ages of 18 and 44 that a woman will be married (to be precise, married with her husband present).

Why did you exclude older people from your numbers? That's seems pretty arbitrary, especially since you're going to pick up a lot of really long term marriages if you count the whole population. By counting only the young you are skewing the data set toward a handfdul of strict religious sect (like the LDS) which pressure the young into marrying early and strongly frown on divorce.

Re: figured out how to estimate by state the expected number of years between the ages of 18 and 44 that a woman will be married (to be precise, married with her husband present).

Um, most people married in 1800. Why would you think otherwise? Of course the poor did not go in for lavish church weddings, and option that certainly exists today. Marriage itself is not unaffordable if you don't insist on an expensive wedding. In many ways it's cheaper (if a bit more of a strain) to live with someone than alone. What's increasingly unaffordable is not getting married, it's having and raising children.

Re: It’s a stereotype that a mortgage, marriage, and babies tend to make people more conservative.

Then they shouldn't be voting Republican these days because the GOP is no longer any kind of conservative party: it's a flat out radical party, and its program is deeply inimical to anyone save the truly wealthy.

Posted by Jonf | November 2, 2007 9:20 PM

Re: Single women without kids are a lost cause for the GOP. Their main issue tends to be unfettered access to abortion, which the GOP can't advocate without pissing off its pro-lifers

Spectacularly wrong. The youth of America tend to be more opposed to abortion than their elders. The most strongly pro-life demographic is young men aged 18-24, and the next most is young women aged 18-24.

Not that this should be surprising. The young are always more idealistic then their elders. And much as many liberal Democrats like to deny it, the plain fact is that the pro-life position is the more idealistic, and the pro-choice position the more cynical. Idealistic young people prefer to see the unborn baby as a child of God rather than a disposable piece of meat. On the issue of abortion, the Democrats have betrayed every principle they ever claimed to stand for.

I agree with JonF about economic conservatism btw.

- Hector (pro-life, young, Democrat)

Posted by Hector Dauphin-Gloire | November 2, 2007 9:36 PM

My measure of Years Married is modeled on the well-known Total Fertility Rate statistic, which likewise is focused on women below age 45.

My interest is in how current and recent conditions affect family formation, which is obviously the province of the pre-menopausal.

Gregory Clark's "A Farewell to Alms" shows statistically that marriage was by no means universal in Western Europe in 1800, and poorer people tended to marry later and have fewer children, just as reading of Jane Austen novels would suggest.

Posted by Steve Sailer | November 2, 2007 9:41 PM

My analysis was largely anticipated by Ben Franklin in 1751:

"For People increase in Proportion to the Number of Marriages, and that is greater in Proportion to the Ease and Convenience of supporting a Family. When Families can be easily supported, more Persons marry, and earlier in Life."

A quarter of a millennium ago, Franklin explained the virtuous cycle connecting low land prices, high wages, marriage, and children:

"Europe is generally full settled with Husbandmen, Manufacturers, &c. and therefore cannot now much increase in People… Land being thus plenty in America, and so cheap as that a labouring Man, that understands Husbandry, can in a short Time save Money enough to purchase a Piece of new Land sufficient for a Plantation, whereon he may subsist a Family; such are not afraid to marry;… Hence Marriages in America are more general, and more generally early, than in Europe."

So, what can Republican government do to help preserve the traditional American patrimony of high wages and affordable land prices (and, in turn, help itself by creating new family values voters?) Franklin offered a sensible answer, which is even more logical now. Restrict immigration. As old Ben asked in his 1751 essay:

"[W]hy should the Palatine Boors [Germans] be suffered to swarm into our Settlements, and by herding together establish their Language and Manners to the Exclusion of ours?"

Good question.

http://www.vdare.com/sailer/060305_patriarchy.htm

Posted by Steve Sailer | November 2, 2007 9:44 PM

Not to beat a dead horse, but it's a little surprising to see a post calling for more sophisticated statistical analysis headlined "More Calculus Please". Does the Atlantic have someone write your blog heds now?

Posted by Doh | November 2, 2007 10:21 PM

If memory serves "More Calculas Please" is one of Michael Ledeen's slogans

Posted by Comment | November 2, 2007 10:36 PM

If memory serves "More Calculus Please" is one of Michael Ledeen's slogans

Posted by Comment | November 2, 2007 10:37 PM

sorry your wizards had to be the first ones to find out how good the new celtics are...

ps: I hate reading wizards fans complain about Arenas. you guys should feel lucky to have him. imagine how we warriors fans feel. you guys took our three best players of that era - jamison, arenas, and hughes while our management resigned crap guys like dunleavy and troy murphy and the team languished in last place for more years while you guys were going to the playoffs.

Posted by Warriors In 2k-never | November 2, 2007 11:05 PM

Re: Gregory Clark's "A Farewell to Alms" shows statistically that marriage was by no means universal in Western Europe in 1800, and poorer people tended to marry later and have fewer children, just as reading of Jane Austen novels would suggest.

That makes a bit more sense than saying that the poor did not marry at all. Also, you should remember that in the pre-modern era common law marriages were recognized by law and considered to be real marriages too, except in the sacramental/religious sense of the term. Among the upper classes women were fairly uselwss exccept as wives. The three careers available to the aristocrats (the military, government and the clergy) were closed to women. So daughters were a burden valuable only for their marriage prospects. Lower down the socioeconomic ladder women could and did labor productively, whether independently or beside their husbands and fathers, so there was no great rush to get daughters to the altar.

Re: So, what can Republican government do to help preserve the traditional American patrimony of high wages and affordable land prices (and, in turn, help itself by creating new family values voters?

Resign from office, repent in sackcloth and ashes, and give up politics until they are ready to truly serve the American people, not the elite.

Re: Why should the Palatine Boors [Germans] be suffered to swarm into our Settlements, and by herding together establish their Language and Manners to the Exclusion of ours?"

Um, some of them folk are my ancestors you are talking about. Are you going to accuse me of being less than a real American because I had some great-great-great grandparents who hailed from Thuringia and Westphalia? Besides, without the Germans American beer would be even worse than it is, if that's possible.

Posted by JonF | November 3, 2007 7:37 AM

JonF,

That makes a lot of sense. It is true however that a fair number of people never did marry at all, particularly in the 19th and early 20th centuries; spinsterhood was not uncommon, and then of course the Church was, at least in the Catholic countries, an alternative to marriage.

It is interesting that this doesn't seem to have characterized the rest of the world outside western Europe (other than perhaps a few countries like Tibet which also had a celibate clergy, severe land shortages, etc. It's apparently so unique that there's a special name for it, the "European marriage pattern", and it goes back centuries. The extreme example is I think Ireland in the late 19th century where men where typically in their mid-30s when they married, and a significant number never married at all. It would be interesting to speculate on the reasons why late marriage only ever characterized western Europe.

Posted by Hector Dauphin-Gloire | November 3, 2007 10:15 AM

Re: It is true however that a fair number of people never did marry at all, particularly in the 19th and early 20th centuries; spinsterhood was not uncommon, and then of course the Church was, at least in the Catholic countries, an alternative to marriage.

This has been true from day one and it is still true today and will probably be true 5000 years from now assuming we humans do not use genetic engineering to transform ourselves into something unrecognizable. Also, a lot of those "confirmed bachelors" and "maiden ladies" were really gay people passing under the radar during an era that did not deem it proper to peep into other people's bedroom door keyholes. Even in cultures where mariages were forced on everyone, many of those marriages existed in name only, without offspring or any other bond save the purely legal one.

Re: It is interesting that this doesn't seem to have characterized the rest of the world outside western Europe

Europeans (from Rome onward) had more freedom about marriage. Even women did in principle, since the Church required consent to be freely given. Yes, that was sometimes ignored when Lord Rich-and-Mighty browbeat his son or daughter into an unwilling marriage for the good of the family, but at lower levels there was less likely to be that sort of pressure applied, and Church authorities were less likely to cooperate when that did happen. By contrast many other civilizations faetured arranged and complusory marriages. And sometimes these worked out well when the kinfolk were wise and discerning in their choices, and sometimes of course they were disastrous matches (and in extreme cases undesirable spouses would fall victim to domestic murder, freeing a now mature widow or widower to look elsewhere on their own). By the way, I think this applies to civilizations only: simpler cultures seem to have allowed a great deal of latitude in marriage. The pattern you mention for Europe (and it came to apply in the Eastern countries as well, despite the persistence of older Greek customs at Byzantium for a time*) was reinforced by Christianity, but I don't think it can be wholly written off to Christianity: the pattern was Rome's as well, before Christianity came on the scene. Perhaps it represents a survival of looser marriage customs from Europe's pre-civilzied past, lasting just long enough for Christianity to reanimate them and make them the norm.

* Religious celibacy was very popular in the East for many centuries and this, I think, broke down the stricter, family-dictated approach to marriage. Byzantine hagiography is full of tales of St. So-and-So telling his or her parents to take a hike when marriage was pressed. Also, Roman law did apply in the East and in time won out over local customs.

Posted by Jonf | November 3, 2007 12:20 PM

Matt,

Something any decent marketer will tell you is that "targeting" people on the basis of demographics is a terrible means of gaining new customers -- or voters. You target on the basis of NEEDS. Needs may be correlated with demographics, but they are not the same thing. So, for instance, people with kids are more likely to care about the quality of the public schools. But there are people who care about public education who don't have kids and people with kids who don't care about public education (e.g. their kids go private). So the fact that unmarried women who vote tend to vote democratic doesn't mean _all_ unmarried women will have this tendency. If you increased the turnout of unmarried women you might find that the ones who weren't voting before are actually more republican. A better strategy is to motivate people who agree with you on the issues to go vote.

Posted by Jim MacLean | November 3, 2007 9:34 PM