Daily Archives: September 17, 2010

Heritage hack job

Blogging has definite advantages over peer review. Like publishing whatever you want, instantly, under the institutional legitimacy conferred by the brand of the institution you work for. Unfortunately, those benefits can be used for ill as well as for good.

This guy at the Heritage Foundation, who claims a master’s degree in political science, is advertised as the “intellectual godfather” of welfare reform. Their blog has a breathless post about the new poverty announcement, which manages to blame both Obama and the “collapse of marriage” for the spike in poverty reported by the Census Bureau.

The recession is irrelevant to the post (in which — after a token link to the Census report — every single link, 5 in a row, is to the same report, published on the same day). A cyclical spike in poverty of course undermines the premise and conclusion: “Child poverty is an ongoing national concern, but few are aware of its principal cause: the absence of married fathers in the home.” Somehow, neither “recession” nor “unemployment” appear in either the blog post or the “full paper.”

Anyway, the reason for this post is this: I’m no statistician, but I think that if p is the proportion of something, then 1-p is the inverse of the proportion. Let’s go to the evidence.

The Heritage report tell us that “throughout most of the 20th century, marital childbearing was the overwhelming norm in the United States.” And gives us this graph:

That is pretty compelling (made all the more dramatic by a y-axis that starts at 50%, giving the appearance of married childbearing careening toward zero). But wait:

I recommend a careful study of the graphs. Since bandwidth isn’t yet completely free, there must be a reason to include both of them, besides the need to justify the salary of a “senior research fellow.”

He writes:

“The flip side of the decline in marriage is the growth in the out-of-wedlock childbearing birth rate, meaning the percentage of births that occur to women who are not married when the child is born. As Chart 3 shows, throughout most of U.S. history, out-of-wedlock childbearing was rare.”

Whew. To help you sort through all this evidence, there is also a clarifying footnote: “In each year, the marital birth rate in Chart 1 and the out-of-wedlock birth rate in Chart 2 will sum together to equal 100 percent of all births.”

Peer review or not, on that I think I can agree* (that is, my preconceptions are confirmed).

And, hopefully we can all agree that filling up a report with graphs and footnotes and links is an important part of seeming like an expert.


* Except for the references to “Chart 1” and “Chart 2,” which should be “Chart 2” and “Chart 3.”


Filed under In the news, Politics