History News Network puts current events into historical perspective. Subscribe to our newsletter for new perspectives on the ways history continues to resonate in the present. Explore our archive of thousands of original op-eds and curated stories from around the web. Join us to learn more about the past, now.
Anniversaries: NPR, Hostages in Iran, Internment in America
by Cliopatria
First, if you haven't read it already, I offer you a chance to read the Daily Mirror's post-election editorial,"God Help America." Yes, it's an anti-Bush screed: I am particularly recommending that our Republican readers take a look at it. Great Britain, our"staunchest ally" in the perpetual war, also has people who think that George Bush is
The Yellow Rogue of Texas.
A self-serving, dim-witted, draft-dodging, gung-ho little rich boy, whose idea of courage is to yell:"I feel good," as he unleashes an awesome fury which slaughters 100,000 innocents for no other reason than greed and vanity.
A dangerous chameleon, his charming exterior provides cover for a power-crazed clique of Doctor Strangeloves whose goal is to increase America's grip on the world's economies and natural resources.
And in foolishly backing him, Americans have given the go-ahead for more unilateral pre-emptive strikes, more world instability and most probably another 9/11.
Why else do you think bin Laden was so happy to scare them to the polls, then made no attempt to scupper the outcome?
There's only one headline in town today, folks:"It Was Osama Wot Won It."
Yes, it's vitriolic; doesn't mean it's wrong, though. It's a blazingly good look at how the vast majority of the world seems to view the US, and how the Republican victory looks to a lot of us here at home. We're Americans, too, and we want to be proud of that fact, and we're very, very afraid that we won't be. My brother in London has already gotten some anti-Bush/American comments....
Now, on to the celebrations....
It is the 25th anniversary of the founding of National Public Radio ("the only news source clinically proven to produce informed citizens"), as well as the anniversary of the Iranian Hostage Crisis. In honor of that, Morning Edition has an interview with former hostage Bruce Laingen, currently head of the American Academy of Diplomacy and husband of the woman who created the modern Yellow Ribbon movement, as well as clips from interviews with Jimmy Carter (who sounds more Bush-like than Kerryesque) and another former hostage. In spite of his experience in Iran -- which makes him quite disturbed at the ongoing hostage-taking in Iraq -- he bemoans the state of our relations with Iran:
I continue to believe that we should be in dialogue with the government of Iran. It is ridiculous that twenty-five years later practically to the day today, we have not talked officially with the government of Iran. Today we are face to face militarily with Iranians on their eastern border with Afghanistan, on their western border with Iraq. To the west is Israel, with nuclear weapons; to the north is Russia with nuclear weapons capability; to the east is Pakistan and India, both nuclear weapons states. You have to look at Iran in the context of how they fit: given our respective shared interests in the Middle East, that makes no sense.
It is also the sixtieth anniversary of several Japanese American Internment civil rights cases, and Eric Muller is among the legal, historical and cultural experts attending a conference on that history this week. Though it's not on the formal agenda, it's safe to say that a certain recent book will be discussed.....
Actually, the example I was asking for was another case where research findings were presented with a 95% confidence interval size nearly twice that of the estimate itself. I'm truly curious to hear of such an example (from anyone), since I've never seen such an instance in my life. In fact, today I'm e-mailing an old college buddy of mine who runs a research lab at NIH. He has a PhD in addition to his MD, and so knows more than a thing or two about experimental design, tests of hyptheses, etc. It should prove interesting to hear his reaction.
Jonathan Dresner -
11/8/2004
My point was that the small sample size (and difficulty of maintaining a 'random' sample) is a direct function of the subject which they were studying. But part of the point of scholarship is to study things that are important and useful, and that usually presents new and difficult challenges.
And almost any military action ever taken falls into the category of the example you so deeply crave.
Richard Henry Morgan -
11/8/2004
As for your last question, one answer might be precisely because we can't assign probabilities or confidence intervals to intel.
I don't understand your point about "similarly appropriate methods". The extraordinarily large confidence interval is a function of the small sample size relative to the large population size. If you drew a similarly small sample size of African-American incomes from a similarly-sized population of African-American incomes, you would end up with a similarly fuzzy estimate (though the distribution of incomes in the population might be quite different from the distribution of excess deaths in Iraq). I don't know for a fact that traditional methods are more suspect, but I'm willing to listen to evidence on that. I will, of course, wait patiently for the similar example I asked for. In any case, I wish the authors luck in their future careers.
Jonathan Dresner -
11/8/2004
Your hypothetical rests on an assumption that I believe to be terribly ill-founded: that a survey of African American households taken by similarly appropriate methods would produce a median figure out of line with other estimates. There's no basis for that.
And in this case the "traditionally accepted figures" are even more methodologically suspect than the Lancet figure, whereas that isn't the case with the income distribution.
And as for your concern about policy based on high-confidence information, remind me why we're in Iraq in the first place?
Richard Henry Morgan -
11/8/2004
Well, they're certainly going to get the most possible attention. I can only say, be careful what you wish for, as you just might get it.
I kept citing the bottom end because, as in the hypothetical African-American per capita case, the traditionally accepted figures are considerably lower than the new estimate. I reiterate: I doubt anyone would would make a similar policy decision on discontinuing affirmative action based on similar figures. Nor do I think the NAACP would be complaining to me that I keep pointing out the low end of the confidence interval, and insisting that per capita African-American income could be $200,000.
The reason we can't discuss "why 100K is likely to be an underestimate", is because that introduces the notion of chances, or probability, which implicates the frequency theory of probability -- which poses a problem inasmuch as you can't assign a probabilty given there is only one population of Iraq. But never mind such technical objections. I'd be interested to see just one other example of a critical scale-sensitive policy decision arrived at on the basis of an estimate, the scale of whose 95% confidence interval is nearly twice as large as the estimate itself. Any takers?
Jonathan Dresner -
11/7/2004
You can't fault the authors for wanting their research to get the most possible attention, and if it had come out after the election it would have been the merest blip. It still went through peer-review, and represents an honest attempt at data-based policy evaluation.
And you keep citing the bottom end of the confidence interval... the high end of the confidence interval was 200,000. Why don't we talk for a while about why 100k is likely to be an underestimate? Any takers?
Richard Henry Morgan -
11/7/2004
The gossip wasn't offered as evidence contradicting the study, but as a cause of cynicism -- I could have gone further and added that one of the authors has been a passionate opponent of the war (which is equally not evidence). But the hilariously large confidence interval is more than sufficient to suggest that the number generated shouldn't be used as a basis for policy evaluation.
Ralph E. Luker -
11/7/2004
Gossip, Richard, is only gossip. You insist on higher standards of evidence, except when the point lends credence to what you already believe.
Richard Henry Morgan -
11/7/2004
The criticism of criticism is, of course, a valid point. The numbers themselves, however, speak for themselves as a basis for evaluating actions and developing policy.
Consider another example. Someone offers a study claiming that the per capita income of African-Americans was $100,000, suggesting that this might be relevant for evaluating the success of affirmative action, or even whether to continue that policy. I suspect that in such a context more than one person might be motivated to point out that the bottom end of the 95% confidence interval was a mere $8,000.
The nice thing about this study is that The Lancet, like just about all journals in the "hard sciences", insists on the inclusion of confidence intervals in articles that have statistical estimates. I suggested some time back that history journals and all other journals in the humanities and social sciences do likewise.
Cynicism, in the present case, is stimulated by the gossip that the authors submitted it to The Lancet on the condition that it be published before the election.
Jonathan Dresner -
11/7/2004
For what it's worth, we should note that "criticism" does not necessarily mean "valid criticism."
And there is value in doing this research that is not necessarily partisan: the important thing is that we should attempt to evaluate the effects of our actions and policies, even when they are disasters....
Richard Henry Morgan -
11/7/2004
From what I can tell, the 100,000 figure lies at the center of an estimated range, the downside of whose 95% confidence interval is only 8,000. In any other context, such an estimate would be considered worthless. There were, needless to say, other criticisms. One is that the sample was not randomized. Where they felt they couldn't do the research in a randomly selected area, they chose "similar" areas as proxies (which rather begs the question). Lastly, there is a criticism of the baseline figure from which they calculated excess deaths: it is said to undersate the figure by at least a third. The important thing, of course, is not whether the "study" reflects reality, but whether it can shape reality.
Sharon Howard -
11/7/2004
And even if it were wrong, the main point here is that this is what many people around the world think. If GWB doesn't need to be reality-based, why should anyone else?
Jonathan Dresner -
11/7/2004
That figure has been pretty widely discussed the last week or so: it's from an article in Great Britain's premier medical journal The Lancet. The 100,000 figure is at the high end of their range, but it is reasonably well founded on surveys and previous research on war-zone demographics; it also falls within the margin of error of several other surveys with somewhat lower median figures. You can find discussion of the paper and methodology in some detail at the Volokh Conspiracy (I'm pretty sure, anyway; do your own research) and from the tireless social scientists at Crooked Timber.
That figure, to be fair, is not "deaths as a result of US military action" but is "excess deaths over natural rates" and therefore includes deaths as a result of insurgency (for which we bear some responsibilty as the providers of security), disease (for which we also bear some responsibility), and other causes.
mark safranski -
11/7/2004
Err aside from the fact that the Iraq war had little to do with vanity - what methodology was used to arrive at the " slaughtered 100,000 innocents " ? Would saying 10,000 or 1 million be equally as accurate ?
It's vitriolic, it's wrong and willfully blind as well.
