This time in statistics. Some days ago Atrios posted a link to Rush Limbaugh's website where this picture was prominently displayed (it isn't there now).
The point of the picture is to make the Iraq fatality numbers look tiny. The error in the picture? That the risk of death is not related to the population at risk. There are far fewer Americans in Iraq than in this country, yet Limbaugh ignores the difference in these base numbers. To make the terms comparable, we'd need to make the bases comparable. If the American military dies in Iraq at the rate of 2,300 out of the, let's say, 130,000 currently there, then this rate applied to the general American population would give us about 5.3 million war dead.
Limbaugh lies with statistics, a common hobby for some writers on the Iraq war.
If you still doubt me, consider this: That the number of people who have died trying to climb Mt. Everest is not in the thousands must mean that mountain climbing is every bit as safe as just staying in bed all day long.
A similar attempt at using absolute numbers to make arguments about rates of war deaths is taking place today in the right blogosphere. It's not too hard to see what is wrong with it if you remember to check what the base is.