A very interesting 2008 Heritage Foundation report, based on 2006-07 data. Here’s one finding: Enlisted recruits are somewhat more likely to come from middle-class and wealthier neighborhoods than from poorer ones:
Individual or family income data on enlistees do not exist. The Defense Department does not maintain records on the household income of recruits or officers. Examining the earnings of most recruits before they joined the military is not possible because, for most of them, their first full-time job is in the military.
Instead, we approximated the recruits’ household incomes by assigning each recruit the median household income of the census tract in which they lived….
Using the median household incomes in their census tracts, the average household income for all 2006 recruits was $54,834 per year (in 2008 dollars). The average enlisted recruit in 2007 had a household income of $54,768. This is modestly above the national average of $50,428….
[L]ow-income families are underrepresented in the military, and high-income families are overrepresented. Individuals from the bottom household income quintile make up 20.0 percent of the population of those who are 18-24 years old but only 10.6 percent of the 2006 recruits and 10.7 percent of the 2007 recruits. Individuals in the top two quintiles make up 40.0 percent of the population, but 49.3 percent of the recruits in both years.
Enlisted recruits are also vastly less likely than their age peers to lack a high school diploma (and though “[t]he military requires at least 90 percent of enlisted recruits to have high school diplomas,” the actual percentage of recruits with high school diplomas is greater than 98%). There’s more in the study itself, including data by race and geography. Thanks to InstaPundit for the pointer.
UPDATE: I initially neglected to note that this was a 2008 report, but I just learned of it, and I thought it was still worth passing along; naturally, if anyone has more recent data, I’d love to see it. I also erroneously reported the results of the Armed Forces Qualifying Tests, which, as a commenter pointed out, are hard to interpret given that most of the public doesn’t take them; I’ve therefore deleted the reference to the AFQTs — sorry for the error, and thanks to the commenter for noting it.