Just in Time!

Expectations high for panel targeting teen pregnancy;
Critics fear emphasis on condoms
Byline: Cheryl Wetzstein; The Washington Times

Leaders of a new national campaign against teen pregnancy say they plan to reduce the teen-pregnancy rate by one-third by 2005, although they're not sure how.

Well, they are never quite sure how; but intentions are all that matter.  It's all about feeling good about yourself. 

And just in time, too: 

US teen birth rate at all-time low, economy cited
-- headline, Lehigh Valley Express-Times  (et al.) 

Okay.  I know pregnancy rate > birth rate.  The difference is largely in the abortion rate, though there are also miscarriages and things of that sort.  Even so, the target of my wrath is this tidbit:

The U.S. teen birth rate in 2009 fell to its lowest point in almost 70 years of record-keeping — a decline that stunned experts who believe it's partly due to the recession.

Due to the recession?  Due to the recession??  The recession!!!?  A wet towel, please, until the anguish passes. 

Don't these people look at the data?  The article in the local paper included a graph and reprinted the wire story as if the graph did not exist.  I was unable to find the graph in on-line versions of the story in the Express-Times, the Washington Post, and a couple other spot checks; so I recreated it myself from the Statistical Abstracts of the United States.  There is this caveat: the birth rates for any given year sometimes change in the tables from later years.  I did not have time to resolve that point and in most of the cases I glanced at, the adjustment was not very large. 

OK.  Here it is:



The trend line is a 50% Lowess smoother, with 2 iterations to de-emphasize outliers. 

Evidently, the experts were stunned because they had not been looking at the freaking data!!!

The tremendous drop that stunned them is the drop from the penultimate point in the graph to the ultimate point in the graph; one so minuscule as to be near invisible.  This is the speck in the eye of the PR release and wire service reporter.  The beam, of course, is that the birth rate has been decreasing very nearly monotonically since the grand maximum back in the halcyon innocent days of Beaver Cleaver and his evidently round-heeled sister back in the 1950s. 

IOW, the answer to the question why 2009 was the lowest teen birth rate on record is....  (drum roll)  ...because the birth rates have been getting steadily lower (more or less) since 1961 and this is simply the latest point in that ongoing trend. 

The recession?!!!  (Excuse me.  Wet towel, please.  Whap!  Thanks, I needed that.)  Basic rule in quality engineering: a recession that began a couple years ago cannot explain a trend that began half a freaking century ago.  That's like claiming the growth of women in the labor force was due to the Women's Lib movement; or that the increase in foreign auto sales in the 70s-90s was due to the Arab Oil Embargo.  Effects do not precede their causes. 

Any quality engineer worth his pittance would tell you that what actually needs explaining is not the trivial variation from 2008 to 2009, but

• the steady decrease since 1962, and
• the "hump" in the 1990s. 

And don't suppose it's necessarily the first thing to pop into your head.  Like the fact that the age of consent was lower back then.  There is still the 1940s back there, and that can't be entirely due to the 18-19 year old men being off to war.  Do not neglect to consider the quality of the data itself: were all states reporting?  Were the figures adjusted consistently?  How was a "live birth" defined?  A chart of the moving range shows erratic year-to-year swings at the beginning, but settling down to a very consistent first difference after about 1970.  Range problems are often due to measurement system problems. 

But trust me.  It's ain't the recession.