101.4 US Soldiers Killed in Iraq in July!

Eighty U.S. soldiers were killed in Iraq in July. If you're a glass half full guy, you note that that's the lowest number in 2007. If you're a glass half empty guy, you note that July is typically a low casualty month (the Iraqi legislature is apparently not the only group that takes a break during Iraq's dog days) so the number is actually larger than it appears. What the GHE guys are proposing, implicitly, is a seasonal adjustment procedure where you account in some way for the typical monthly pattern. A simple procedure would be to subtract the monthly mean from the raw total and then add back in the annual mean. Government statistical agencies routinely seasonally adjust all sorts of economic data using more sophisticated methods such as the Census Bureau's X-12-ARIMA method. The inflation or employment figures that you read in the paper, for example, are typically not the raw figures but the seasonally adjusted ones. If you seasonally adjust US combat deaths using X-12-ARIMA, this is what you get:

The blue line is the raw (not seasonally adjusted) data and the red line is the seasonally adjusted data. July's total of 101.4 is down from May's 115.1 and June's 119.6, but higher than April's 94.8. The low for 2007 was actually January (78.6), right before the surge.