« Free Riders | Main | Unions and Corporations »

November 07, 2007

One Cheer for the Washington Post

This is the headline and first paragraph of an article in The Washington Post:

Giuliani Is Still Standing By Questionable Figures

The former New York mayor would have us believe that he was off by one percentage point at most in calculating his chances of surviving prostate cancer in Britain. In fact, he was spectacularly wrong the first time and equally wrong the second time. Epidemiologists say that his claim rests on a faulty statistical methodology that would not earn a passing grade at top medical schools in the United States.

The first graf of the article is pretty good work. Indeed, the whole article is pretty good work. I'd like to see it on A1 rather than A6, and I'd like to see Giuliani's lies get mentioned in everyday coverage of the candidate, but baby steps.

That said, most people will still only real the headline. And the headline does not accurately convey the article's conclusions. Giuliani is not "standing by questionable figures.' He's continually "lying about American health care." There's a big difference.

The media really has to get comfortable saying the "L" word.

November 7, 2007 | Permalink


I think we're unlikely to see the Post calling Giuliani a liar in a headline. They could at least say "Giuliani Is Still Standing by Incorrect Figures", though.

Posted by: Herschel | Nov 7, 2007 11:53:01 AM

Interesting point. I took a look at the WaPo website to see how priminently (or not) this article was displayed. I found the link in their politics section and found essentially the same, though somewhat longer, article that appeared to be part of a regular feature called The Fact Checker. And it ran under a much better headline than the print version: "Four Pinocchios for Recidivist Rudy"

Posted by: Marlowe | Nov 7, 2007 11:58:54 AM

Swear to god, when I read that headline I thought it was about Guilliani and Kerik.

Posted by: Steve Balboni | Nov 7, 2007 12:11:51 PM

Didn't you see that picture of him next to Pat Robertson this morning? Talk about "Standing By Questionable Figures"!

Posted by: Bill Camarda | Nov 7, 2007 12:24:28 PM

I'm with Steve. Up until the second to last paragraph I kept expecting you to switch to Kerik.

Posted by: eriks | Nov 7, 2007 12:31:55 PM

Mendacity. That's the strongest word media people (including some lefty blogs) will apply to Rudy's lies.

Perhaps the media is applying the narrow legal application of perjury (which I think requires intention for the most part) to political lies.

Or maybe the media just isn't willing to do the work that would be required to document the lies and report them with supporting evidence and testimony.

But that doesn't change the fact that public lies are the method of choice for gaining office and maintaining it: the Lee Atwater legacy. But after 8 years of Bush/Cheney/GOP lies being unchallenged, I guess we shouldn't be surprised that truth/falsity is no longer a media concern.

I read last night in TNR (Michael Crowley) about some wingnut jerk (a Bob Jones Univ. prof who sent a mass email in John McCain's 2000 primary bid in South Carolina) charging that "McCain has chosen to sire children without marriage". He was challenged on CNN, the professor responded "Can you prove that there aren't any". There you have conservative GOP campaign tactics in one sentence. If you can't prove that some charge isn't true, then it must be true, even if made up out of whole cloth. Lies as truth, the way to win elections.

Posted by: JimPortlandOR | Nov 7, 2007 12:52:01 PM

"Epidemiologists say that his claim rests on a faulty statistical methodology that would not earn a passing grade at top medical schools in the United States. "

What a bizarre sentence. Giuliani is neither an epidemiologist nor a student at a top medical school in the US. Why would we expect him to have a grasp of such a statistical methodology. The sentence would make much more sense if it said it would not earn a passing grade at the worst medical schools in the US.

Posted by: Ginger Yellow | Nov 7, 2007 1:00:57 PM

You're not going to get the media to use the L-word. Lie implies intent, and reporters should not be in the business of speculating about what's in someone's head. But they should feel comfortable using the f-word - that is, "false." The figures aren't questionable. They're false. Inaccurate. Incorrect. Wrong.

Posted by: Bloix | Nov 7, 2007 2:01:25 PM

I just checked the paper version and it gives Giuliani "four Pinochios" for this claim. According this section's metric, this stands for "Real whoppers." I think they did a good job. I would say two cheers--for Michael Dobbs at least.

Posted by: d | Nov 7, 2007 2:05:22 PM

Here's the link to the Health Affairs study "Toward Higher-Performance Health Systems: Adults’ Health Care Experiences In Seven Countries, 2007"

It is extremely important reading for anybody interested in health policy. Note that this study is about perceptions and does not provide "objective" performance measures, although it includes data on alleged medical errors. The US responses stand out negatively in almost all aspects of the study. Interesting is that Germans also reported a high level of insatisfaction and lack of confidence despite reporting the lowest incidence of medical errors, easy access to medical care and short waiting times. My guess is that the insatisfaction reflects in part bad-mouthing by politicians and journalists, who constantly conjure up alleged health care crises and "cost explosions" (Germany's health care expense as a share of GDP is pretty stable at 11%). The neoliberal politics of the past 25 years has spawned countless "health care reforms" which all resulted in higher out-of-pocket costs. 20 years ago, everything from doctor's visit to prescription drugs to dental fillings was essentially free. Nowadays Patients are paying every time, much as in the US (if not as much). While these payments may not really cause hardship, they create a perception that you pay more and more, getting nothing in return.


Posted by: piglet | Nov 7, 2007 2:06:54 PM

Using prostate cancer as a proxy for judging the relative strengths of the NHS versus our system introduces two very difficult to gauge variables that make cancer epidemiology particularly vexing:

1) lead time bias: here, as the Post explains, the concept is relatively simple. Massive PSA screening catches the cancer at an earlier stage in its natural disease progression. So if a patient destined to die of his prostate cancer were diagnosed relatively early in the evolution of said cancer, he might survive the 5 years we typically use as a cut-off to measure cancer survival. Because the US has greater screening, we might have improved 5 year (but possibly equivalent 10 year survival) for Prostate Ca.

2) the question of indolence: are the prostate cancers we're detecting destined to kill these men? If they are, one would expect that improved screening, because we're presumably catching the cancer at an early, less malignant stage, might death rates. If they aren't, if they're simply indolent and slow growing neoplasms, screening should NOT improve death rates from Prostate Ca, because these tumors never would've killed these men even if they had gone undetected. Judging by equivalent likelihood of death per 100,000 people, it appears we're detecing largely indolent, non-fatal disease.

A more apt comparison between the UK and the US in terms of cancer treatment (eg how we deal with cancers once diagnosed) might look at lung cancer or brain cancer--diseases for which we have no early detection method to sully the data. My hunch (based on an August article in the Lancet Oncology) is that we'd have improved overall survival in the US and, because there would exist no significant differences in detection methods, one could attribute this to better treatment here.

Posted by: sean | Nov 7, 2007 4:01:14 PM

The good thing is the Giuliani is a liar narrative seems to be building. It takes a couple of points like this to take hold in the media's all powerful 'framework' we seem to be on the way with this one. It's moved from the blogs to leftish columnists and now to A6. We just need one more easy to explain, preferably trivial fib and it will haunt him through the rest of the campaign.

Posted by: AJ | Nov 7, 2007 6:44:57 PM

Sean, I don't follow the logic in your part 2. If the death rate is based on actual patients with tumors we know of - rather than a statistical estimate of what we think the incidence is - then if you find a bunch of wimpy tumors that aren't going to kill anyone, you certainly have lowered the apparent death rate.

Say there are 100 people with prostate neoplasms, and 10 of those are going to be fatal in 5 years (assuming for the moment that there's no treatment at all) and the rest of them will never be fatal. And say we detect all the potentially fatal ones no matter what. If we only notice 10 of the remaining 90 in screening, then the death rate is 50%; if we detect all of them, it drops to 10%.

Posted by: Hob | Nov 7, 2007 7:44:32 PM

I agree Giuliani should have backed off his misuse of 5-year survival when it was pointed out to him. To be fair, though, lots of non-slimeballs who should know better use that statistic for comparative purposes, outside of a randomized trial, with abandon, and equally inappropriately. I think Giuliani's numbers originally came from the Commonwealth Fund, and they do it all the time, and of course their data are constantly being cited by others; I hope they've figured it out by now.

Most of all, I'm glad that the public is finally getting a lesson in statistics, and I think the press overall is doing a good job of covering this one.

Posted by: Emily DeVoto | Nov 7, 2007 9:38:04 PM


Here's the reasoning: the death rate (we can call it incidence) is the number of deaths from Prostate CA per year per 100,000 people(say 28 per 100,000 people). Let's say in the UK and US (to make this really simple), we have 100 and 1000 respectively (say the US has 10x pop of UK) people that develop prostate CA (detected or not). 30 (or the US case 300) of those have very aggressive disease that would present to a physician regardless of widespread screening. Suppose that these people are destined to die, even with treatment. Furthermore, lets say, for the sake of logical clarity, the UK does no screening. The remaining 70 cases are never picked up but it doesn't matter, because these tumors are so indolent (slow growing) that they'll never kill the patient. The US picks these cases up and treats them. Technically the mortality here would be 100% in the UK versus 30% in the US. But the incidence (death from prostate CA per 100,000 people) is still the same. And the US tries to claim better mortality, when the fact of the matter is that 700 of the cases the US claims to have cured would have lived regardless of treatment. That's why we look at incidence (number of prostate CA deaths per 100,00 people). If they're the same, it suggests 1) screening aint helping 2) the aggressive cases (those destined to die), will die regardless of whether they're treated here or in the UK.

Posted by: sean | Nov 7, 2007 10:04:58 PM

"I think Giuliani's numbers originally came from the Commonwealth Fund, and they do it all the time, and of course their data are constantly being cited by others"

You haven't followed the issue have you. We had so many threads about his already I'm getting tired of it. Giulianis figures came from the Manhattan Institute, a right-wing think tank. And no, reputable experts don't use these figures.

Posted by: piglet | Nov 7, 2007 10:45:13 PM

From 1988, another word that means "lie": 'Contradictions'


Posted by: Bloix | Nov 8, 2007 9:31:38 AM

Sean: I'm pretty sure we're saying exactly the same thing and I just misread the direction of your argument. Didn't mean to muddy the waters.

Posted by: Hob | Nov 8, 2007 1:04:23 PM

The comments to this entry are closed.