The graph below shows the representation of active-duty Army recruits for 2004, 2005, and 2006 according to the median household income of their ZIP code. (Data are not available for the individual recruit's household income.) Each bar represents the ratio of two proportions: the share of total recruits living in ZIP code areas where median household income falls within the specified range over the share of the total youth population living in the ZIP code. A score of more than one means the income range is over-represented. A score of less than one means the range is under-represented. Note that the peak of the bars falls in the range of $30,000 to $59,999. In other words, neighborhoods with low- to middle-median household incomes are over-represented. Neighborhoods with high-median household incomes are under-represented. (The comparable median household income for the U.S. is $47,837.)

Upper-middle and high-income neighborhoods have fallen in representation from 2004. In other words, wealthier neighborhoods are less represented in 2005 and 2006 than they were in 2004.

Notes: Median household income estimates by ZIP code are drawn from 2005 estimates from Claritas, Inc . (These data were purchased by NPP.) Army data were obtained from the Army through a FOIA request submitted by NPP (2004 data were requested by Peacework Magazine). The graph includes data for 65,842 recruits for 2005. Recruits from Puerto Rico and U.S. territories are not included. Recruits which came from ZIP codes with median household incomes of zero, or less than $5,000, were suppressed since that effectively means we have incomplete data on the ZIP codes in which they live. Fewer than 2% of recruits were excluded.