At Wizbang, Jay Tea analyzes the breathless reports that deployed troops are donating to Obama six times more often than McCain and provides much needed (especially for lefty bloggers who seem to struggle with understanding statistical relevance) perspective:
Let's take those numbers and play with them a little.
Minimum donation to be counted by Open Secrets: $200
Average donation to Obama: $455 and change.
Average donation to McCain: $410 and change
Total donations to Democrats: $63,882
Total donations to Republicans: $76,027
Average donation to Democrats: $456 and change
Average donation to Republicans: $500 and change
Total donations: 292
Total donations to Democrats: 140
Total donations to Republicans: 152
Total number of deployed troops donating to Obama and McCain: 160
Total number of Americn troops deployed abroad: about 370,000
Total number of deployed troops in Iraq: about 170,000
Total number of troops deployed in Afghanistan: about 20,000
Percentage of US troops deployed overseas donating at least $200 to presidential campaigns this year: 0.079%. Or, in simpler terms, less than one in ten thousand. Or, roughly, one in 12,500.
So yeah, Obama has outraised McCain from US troops stationed overseas, by overwhelming ratios in both number and amount. But the actual numbers are so damned small as to be statistically irrelevant.
And I think I kinda like that 99.9% of our troops aren't spending at least $200 on presidential campaigns.
The only credible conclusion that one can draw from the data is that among a tiny, tiny subset of US troops serving overseas (probably made up of those with a highly paritsan political bent), Obama has raised more money than McCain at this point of the election cycle. As our own Saint Paul (who knows a thing or two about such statistical matters) put it in an e-mail:
"That story is funny for a lot reasons. One being that the population of soldiers overseas who have donated to political campaign cannot in anyway be generalized to all soldiers or even all soldiers overseas. Those willing to give their money are almost certainly a small subset more inclined to be caught up in emotionalism and who probably would exhibit past voting behavior for Democrats. That small population size is supporting evidence of this (ain't that many soldiers who are prone to emotionalism or to be Democratic voters)."