I can’t add much to this excellent report by Doug Chapin at the Program for Excellence in Election Administration.

I will out myself as Geek #1, but I suspect my friends in the community know this already.  The demographics on the military were as of 2008, and were obtained from a slide presentation from the Defense Manpower Data Center.

The comments about income and region were drawn from a project I worked on with the Triangle Institute for Security Studies that compared demographic and attitudinal profiles of the rank and file, officers, and civilians, the results of which appeared in a number of publications including Soldiers and Civilians: The Civil-Military Gap and American National Security. The surveys were administered from 1998-9, but I have not heard anything to indicate that the recruitment and volunteer patterns have changed substantially since.

Finally, the comment about the relationship between serving in a hierarchical organization which stresses the importance of voting (as well as providing voting assistance officers)  and the possibility that members of the military might participate at higher rates than their demographic and income profiles would predict is admittedly speculative.  I based my comments on extensive research on military sociologists, such as the late Charles Moskos, as well as a preliminary but intriguing pattern I found in vote validation studies conducted by the National Election Studies.  While the total number of military respondents in the survey is small, there is a statistically significantly greater propensity to over report being registered (10% difference) and to over report turnout  (17% difference).

What this means is, to take one example, in the 2008 Cooperative Congressional Election Study, 89% of the general public who said they voted were validated as having voted, compared to 73% of the military.  These findings are preliminary because there were only 51 military respondents in the CCES, and I suspect (but do not know) that all were stationed in the US.  (The CCES survey is publicly available here.)