Saturday, October 08, 2011

Marriage durability by state

In taking a fresh look at how what Steve Sailer deems the marriage gap held up in the 2008 election, I came across an interesting US Census report on marriage rates and things related with data from 2009. The marriage gap still existed in '08, with McCain's share of a state's vote and the median duration of first marriages (age adjusted and for all races) in that state correlating at .85 (p = .00). Steve found a correlation of .91 (p = .00), though he looked exclusively at white women between the ages of 18 and 44. Those are extremely strong relationships for anything in the social sciences.

Perhaps counterintuitively at first, divorce rates also correlate strongly with voting Republican at the state level (r = .53, p = .00). That's surely not something that red state "values voters" are proud of, is it? Well, the confounding factor is that divorce rates are measured among the population at large, not just among those who are (or were) married. So in states where relatively few people get married, it's not surprising that divorce rates are going to tend to be lower there than in states where more people tend to get married. Indeed, marriage and divorce rates correlate with one another at .39 (p = .01) at the state level.

How about an attempt at measuring the durability of marriages by looking at the divorce rate in the context of the marriage rate? The following table ranks states in this way by taking the number of marriages per 1,000 people aged 15 and older that occurred in 2009 and dividing it by the number of divorces per 1,000 people aged 15 and older that occurred over the same period of time. So it's not measuring the healthiness of 'the institution of' marriage per se*, but instead how committed those who actually get married tend to be:

1. Idaho
2. North Dakota
3. Hawaii
4. Wyoming
5. Utah
6. District of Columbia
7. Delaware
8. Connecticut
9. New York
10. New Jersey
11. California
12. Virginia
13. Illinois
14. South Carolina
15. Nebraska
16. Texas
17. Washington
18. Iowa
19. Kansas
20. Alaska
21. Maryland
22. Wisconsin
23. Minnesota
24. North Carolina
25. West Virginia
26. Montana
27. Colorado
28. Massachusetts
29. Pennsylvania
30. Florida
30. New Mexico
32. Missouri
33. Arkansas
34. Georgia
35. Nevada
36. Arizona
37. Louisiana
38. Oklahoma
39. South Dakota
40. Oregon
41. Indiana
42. Michigan
43. Ohio
44. Kentucky
45. Mississippi
46. Vermont
47. Tennessee
48. New Hampshire
49. Rhode Island
50. Alabama
51. Maine

There is a bit of a generational issue here, as we're gauging to some extent how eager relatively young people are to get married with how likely older married people are to get divorced. Theoretically, younger people in a given state who are getting married today could be entering into much stronger (or weaker) marriages than their divorcing parents did when they were married compared to what is taking place in another state could make states where marriage is suddenly getting much stronger or much weaker stand out less with the old strength (or weakness) negating the new weakness (or strength). Or, the marriage rate in a state may be undergoing a major shift, with relatively few (or more) of its younger people getting married compared to what has taken place in the state's past. It's difficult to account for these differences, although states undergoing the most demographic changes are the most likely to be effected.

There aren't any clear correlations with things like political leanings or IQ that emerge (both are statistically insignificant, with only slight positive relationships existing with higher IQ and voting Democrat), as this accompanying map illustrates. Mountain states tend to do the best, while the upper Northeast and South fare the worst.

I suspected that among states where marriage rates are higher matrimony would be more superfluous than in states where fewer people get married in the first place, as if the relatively few people who actually do get married in states like Maine and Rhode Island actually mean it, whereas in states like Wyoming where everybody ties the knot, there are plenty of marginal people with lots of slack in their strings. Not so, however--the correlation between the marriage rate and marriage durability at the state level is a healthy .52 (p = .00). Marriages tend to be more durable in states where more people get married in the first place.

* In DC, for example, the marriage rate is on the low end among all the other states, but the divorce rate is the even further down the list, second from the bottom. Consequently, marriage durability in the nation's capital is actually quite high.


Steve Sailer said...

"The marriage gap still existed in '08, with McCain's share of a state's vote and the median duration of first marriages (age adjusted and for all races) in that state correlating at .85 (p = .00)."

Is this .85 using this new Census report's data or my old numbers applied to 2008?

Steve Sailer said...

Okay, I come up with r = 0.64 for this Census Bureau's measure of women's first marriage durability when correlated with McCain's 2008 percentage, although quite a bit of that is driven by D.C. being such an outlier. But this isn't as good of a measure as years married among white women 18-44 of what I'm interested in since women who never married aren't counted and it's heavily driven by older people from an era when everybody got married and it's driven somewhat by racial differences between states (although that's not as important). But 0.64 still isn't bad.

Audacious Epigone said...


Appendix table 3 (p24) in the Census report noted in the post is what I used. Wonder why we're getting a variance. I'm showing a correlation of .85 with DC included, and .80 without DC. That is almost as strong as your more applicable and restrictive measure for the '04 election.

Steve Sailer said...

You're right, I get 0.87 for p. 24 correlated with McCain's share of the vote.

Audacious Epigone said...


The slight variance probably comes from my use of the age-adjusted numbers.

Somewhat relatedly, I wonder if there is an intra-state correlation in fertility rates among whites, blacks, and Hispanics. That is, in states where whites have more kids than whites in other states, do blacks and Hispanics also have more kids than blacks and Hispanics in other states? The lab of the states website you originally used looks to be defunct now. Any idea where we might find that data? I haven't looked, but I will.

bbartlog said...

Divorce laws may result in some people filing in a different state than they married in, for strategic reasons. This shouldn't have much effect in most cases but for small-pop states like Maine and Idaho it could be a confounding factor.

Nietzsche said...

There's countless statistics that prove the younger the couple the more likely they are going to get married. The rise in the average age for marriage and decline in total marriages would explain the general drop in divorce rates.

It's still an abysmal failure in general, when half of every marriage from Vegas shotgun marriages to pseudo-arrange marriages kick the bucket.

Apparently this study by UCSD concludes that teen/early 20s women are 3 times more likely to end up in poverty and 66.6% more to divorce within 15 years if they have offspring from the union.

It would make sense for young divorced teen mothers of any ethnicity to vote democrat in order to enhance their welfare benefits and have the government as a surrogate father.

Steve Sailer said...

Yes, that's why I used "years married" and the Census Bureau uses "duration of first marriage" -- we're trying to look at differences in successful long term marriages.

Harper said...

Just read Sailer's article where he quotes you. I completely agree with the marriage line.

However, I disagree with this point in the Sailer article:

"since the GOP is inevitably the white party, you want marginally white people from places like Latin America and South Asia to identify as white."

This has many problems, not least of which is the fact that most of the Hispanics coming to the US are not white.

As various bloggers recently pointed out:

"According to the CIA World Fact Book, Mexico is:
60% mestizo
30% Amerindian
Less than 10% European (mostly Spaniard)

Examining genetic ancestral markers, Rubén Lisker has found lower-income mestizos in Mexico City to be:

59% Amerindian
34% European [mostly Spaniard]
and 6% black

These people clearly are not white."

Since these people clearly aren't white (i.e. Diaspora Europeans), I don't see what advantage can be gained by lying. In Mexico they don't lie about this -- they call them mestizos, which may make them try to be white even more (e.g. bleaching their skin). But by telling them they're white, why should they even try? And by pretending they're white, it will only cheapen what it means to be white in the US (i.e. a Diaspora European). You don't make a commodity more valuable by cheapening it.

Anonymous said...

NAM is now at #1 at urban dictionary.

Audacious Epigone said...


I think the idea is not so much to have them identify as white as it is to have them not identify as non-white. I'm not sure how effective that is capable of being though, especially when there is no shortage of groups prodding them to identify as non-white.