Thursday, January 18, 2007

NYT digs for bad marriage news

The New York Times gleefully celebrated the claim that 51% of women are not living with a husband. What's wrong with this picture? First, go to the numbers themselves. These data come from the American Community Survey for 2005, whose home page is here.
If you go directly to the simplest table, table S1201, you will find, contra the NYT, that 51% of women are married. (Run your eye down the first column to the row which lists "females." Scoot over one column to the column labeled "Now married, (except separated)".) Voila: 51% of women are now married.
Not, as the NYT reported with thinly disguised glee, that 51% of women are NOT married. To get to the conclusion that the unmarried outnumber the married, our intrepid reporter of social trends has to do some digging.
First problem is right in that table: this number is based on all women over the age of 15. Hmm. Let's look at the 15-19 age group: 2.5% of them are married. What a surprise, that 15 year old "women" are unlikely to be married. By including this age group, we artificially lower the percentage of the population who are married. The next age group, from 20-34 years old, has 44.2% married, and all the other age groups are even higher percentages until you get to the very oldest, over 65, which is surely dominated by widows, and driven by the simple fact that even women who are happily married for a lifetime, are likely to outlive their husbands.
If we exclude the 15-19 year olds from the population, what percentage of women over 20 is married? The teenagers constitute about 8% of the total population of women. My back of the envelope calculation suggests that throwing out the teenagers would raise the percentage of women married by about 4% points to about 55%. Looking at only the women in their child-bearing years, that is, throwing out the over 65 population, we find that 58% of the women between 20 and 64 are married. This is the most significant group for social policy concerns about the well-being of children, demographic growth and the like.


Next problem: Remember I said that Table S1201, reports that 51% of women over 15 are married. So how did the NYT come to the conclusion that 51% of women are not living with husbands? He did it by going to Table B12001, which breaks down married people into categories of "Married, spouse present" and "Married, spouse absent." Run your eye down the table to the Females, and you'll see about 5.5 million women were Married, spouse absent. The 3 million who were legally separated were already accounted for in the previous table. The "other" category includes about 2.4 million women who were married, not legally separated, but who were not living with their spouses at the time of the survey. These women could include those whose husbands are on military deployment, an extended business trip, or in prison or another kind of institution. These 2.4 million women comprise 2% of the population.
And 2% is just the number you need to shift the numbers from "51% of women are married, but not legally separated" to "51% of women are not living with husbands."
Only the New York Times could make a headline out of a footnote.

No comments: