When the British medical journal The Lancet reported last week that more than 650,000 Iraqis may have died because of the U.S. invasion of their country, President Bush quickly jumped in to say that number was "not credible," adding that the researchers' "methodology has been pretty well discredited."

When Gen. Peter Schoomaker, the Army's chief of staff, said last week that Army had contingency plans to maintain its current force level in Iraq through 2010 if necessary, his boss, Defense Secretary Donald Rumsfeld, quickly jumped in to say, "Gen. Schoomaker and the Army does (sic) not set force levels in Iraq."

When former Rep. Lee H. Hamilton, D-Ind., co-chair of the congressionally mandated Iraq Study Group, said last week that, "I still have real questions about the capacity, the will, of the Iraqi government to move," his co-chair, former Secretary of States James A. Baker III, quickly jumped in to say that doesn't necessarily mean the panel has decided to recommend the U.S. change courses in Iraq.

All of this backing-and-forthing on Iraq can be attributed to the fact that the midterm elections are just three weeks away. The depth of the tragedy there, and the absence of any viable exit strategy, are key electoral issues. The irony is that it could take an election in the United States to do what elections in Iraq couldn't: Point the way out.

President Bush remains in his accustomed state of denial. Here is one of the world's most respected medical journals publishing a peer-reviewed study by epidemiologists backed by Johns Hopkins University's School of Public Health, part of one of the world's most respected medical schools. And Bush, having, as usual, found someone to tell him what he wants to hear, pronounces its credibility and methodology to be flawed.

The researchers relied on "cluster studies," interviewing a scientifically selected sample of 2,000 Iraqi families about mortality rates before and after the war started. From this they extrapolated that the mortality rate in Iraq was 5.5 deaths per 1,000 before the war and 13.3 per 1,000 since. They obtained death certificates for 92 percent of the reported victims. They did the math and estimated that there was a 95 percent level of confidence that the war resulted, directly or indirectly, in between 392,979 and 942,636 additional deaths.

The midpoint of the range translates to 601,000 direct victims of the war, and 54,000 who died because the health-care system had been devastated. Iraq has a population of 26 million; a similar mortality rate in the United States would yield 7.5 million deaths.

Gen. Schoomaker's "mistake" was admitting, just weeks before the election, that he's doing what a good staff officer should do: advance planning for worst-case scenarios. Rumsfeld, whose aversion to realistic planning is one reason the United States finds itself in a worst-case scenario in Iraq, should have given him a medal instead of upbraiding him.

As for Baker, Hamilton and the Iraq Study Group: They say they won't release any recommendations until after the election, but it's becoming clear that "stay the course" isn't on their short list. Whatever their recommendation, we trust Bush won't have trouble with their methodology.