Dan Kahan points me to this news article by Mike Alberti:
Every generation born in the United States has lived longer than the one before, and so it has been easy to assume that the upward trend would continue. But while life expectancy in the country as a whole has continued to rise steadily for the last twenty years, a new study shows that life expectancy for women has actually declined during this period in 313 U.S. counties, most in the Southeast, the Southern Midwest, and Appalachia. Life expectancy for men, by contrast, only declined in six counties (although the study did find that men in these same areas tended to have worse outcomes than men elsewhere). In what some experts have dubbed a public health crisis, these findings mean that children born today in many parts of the United States can expect to live shorter lives than their parents.
Here’s the map:
Changes in years of life expectancy for women in U.S. counties, 1987-2007
Source:Institute for Health Metrics and Evaluation at the University of Washington
My main thought is I’d like to see where these changes are coming from.
“Life expectancy” is calculated based on death rates at each age at different time points (see here). If life expectancy is going down, then death rates at some age groups are going up. I’d like to see where this is happening.
This would be more helpful, I think, then speculating based on averages. I mean, sure, averages are fine. But if you have the underlying data (which they must have, otherwise they couldn’t have computed teh life expectancies), then let’s take a look!
This sort of thing comes up a lot with politically-charged statistics: you get some numbers, or maybe a graph or map (graphs and maps are good!), then some discussion, but not enough backtracking to how the original underlying numbers could directly address some of the points of discussion. I’m not trying to criticize the researchers or the reporter here, just suggesting that more could be done when looking at summary measures, whether they be college rankings, quality-of-life measures, economic freedom scores, life expectancies, or whatever.