One week to go before plunging into the slippery world of statistics and the bizarre syntax of R (previous post). Loyal reader Karen was kind enough to send me “Are Blue States Better at Exemplifying Red State Values? The Data Might Surprise You.” (Big Think), a topical look at how statistics can be misleading.
You get one answer if you look at data aggregated by state:
The National Campaign to Prevent Teen Pregnancy data shows that teenagers in Mississippi, the reddest of red states, are more sexually active than the teens of any other state. While the teenagers of liberal New York are the least active. Similar data can be found for teen pregnancy, with nine of the ten worst states for teen pregnancy rates being “red” states.
The same can be found for the divorce rate. Of the states with the ten highest divorce rates, eight of them are solid red, while the state with the lowest divorce rate is bright blue Massachusetts. This study shows the five states with the highest rates of divorce for women all being red, with four of the five lowest rates found in the bluest states.
(I’m not sure what it means to talk about “the highest rates of divorce for women“; it is more common to for a wife to sue a husband than vice versa (data from Massachusetts, for example), but both a man and a woman will be equally “divorced” after the process of litigation is complete.))
Then you get a different answer if you aggregate by county:
when you zoom in to the county level the data reverses, with red counties across the country having more stable marriages and fewer divorces. The worst statistics, given this analysis, are to be found from blue voters in red states.
[Separately, the article reinforces the “white privilege” stereotype:
And, of course, the teenagers who are most likely to have a stable family in the United States are the children of well off, intact, families, while poor, non-white teenagers are the least likely to be in a home with married parents. There is more to the story than mere ideology.
Compared to Asian-Americans, whites are frequent flyers in American family courts and are much more likely to divorce, have an out-of-wedlock child and harvest the child support, etc. Yet the article implies that whites are some sort of paragon of stability and putting interests of kids first.]
This is a standard and well-known statistical “paradox”, but that doesn’t prevent reporters from falling for it over and over again because either they are stupid or they are dishonest and expect their readers to be stupid.
It’s the same reason people in crime-ridden cities care more about “law and order”, not at all surprising if you are capable of grasping the concept that the behavior of a small proportion of people can affect the attitudes of a large proportion of people.
Perhaps the life expectancy of divorced men takes such a hit that divorcees are under-represented amongst surviving men. Or once divorced, the men flee as far away as they can, as one acquaintance who fled from California to Massachusetts to be as far from his ex-wife as possible.
Both effects would skew men vs women divorce statistics.
> I’m not sure what it means to talk about “the highest rates of divorce for women“
A minority of men can marry all women through consecutive marriages. In such a case the divorce rate for men will be higher (presuming only people who have ever been married are considered).
This scenario may not be true, but it would mirror fecundity (fathers have more children than mothers, on average).