Speaking of Hackery

by Kieran Healy on May 31, 2006

Jesus wept. “This nonsense again.”:http://www.newsmax.com/archives/ic/2006/5/29/132706.shtml?s=ic

bq. Despite media coverage purporting to show that escalating violence in Iraq has the country spiraling out of control, civilian death statistics complied by Rep. Steve King, R-IA, indicate that Iraq actually has a lower civilian violent death rate than Washington, D.C. … Using Pentagon statistics cross-checked with independent research, King said he came up with an annualized Iraqi civilian death rate of 27.51 per 100,000. While that number sounds high – astonishingly, the Iowa Republican discovered that it’s significantly lower than a number of major American cities, including the nation’s capital. “It’s 45 violent deaths per 100,000 in Washington, D.C.,” King told Crowley. Other American cities with higher violent civilian death rates than Iraq include: Detroit – 41.8 per 100,000. Baltimore – 37.7 per 100,000. … The American city with the highest civilian death rate was New Orleans before Katrina – with a staggering 53.1 deaths per 100,000 – almost twice the death rate in Iraq.

Astonishing, indeed, until you recall that Washington D.C. is a city, and Iraq is a country. As it happens, only last week we were looking at OECD data on “national rates of death due to assault”:https://crookedtimber.org/2006/05/26/incarceration-again/. The U.S. rate is “exceptionally high”:http://www.kieranhealy.org/files/misc/assault-deaths-99.png compared to peer nations — around 6 per 100,000. But this is still some distance behind Iraq, I think. In the meantime, I have an offer for Rep. King. He should pay my expenses for a vacation to DC, including a flight to the city, a taxi to a local hotel, a few dinners out at restaurants. Maybe some tickets to some museums and local sights, perhaps a concert or a game. At the same time, he could take a parallel trip to Baghdad and do the same things — commercial flight in, local taxi, wander out for dinner, etc. We’ll both bring camcorders and see how it works out. If DC is so much more dangerous than Iraq I’m sure something like this would really show up people who say the situation in Iraq is terrible.

{ 1 trackback }

06.01.06 at 7:13 am



rilkefan 05.31.06 at 11:57 pm

I’m slowed by a headache and don’t get the city/country thing. If it’s some density issue then uhh isn’t Iraq largely represented by Baghdad which would make for a high rate naturally? I’m guessing the real clunker is the quoted death rate in Iraq – for one thing, 27.51 has too many significant figures. Scaling up by 260 gives 7k/year, which isn’t anywhere near the Lancet figures…


Guest 05.31.06 at 11:57 pm

As far as I got with that one was http://www.newsmax.com… ‘Nuff said.


Functional 06.01.06 at 12:07 am

Astonishing, indeed, until you recall that Washington D.C. is a city, and Iraq is a country.

Why is this relevant?

It *would* be relevant if he were comparing *raw numbers of deaths.* But isn’t he comparing the death rate per 100,000? If so, then your rebuttal is akin to saying, “The New York City murder rate nay be higher than the murder rate in Charlotte, but it’s a bigger city.”

As it happens, only last week we were looking at OECD data on national rates of death due to assault. The U.S. rate is exceptionally high compared to peer nations—around 6 per 100,000. But this is still some distance behind Iraq, I thnk.

Um, right. You’ve seen crime statistics before, surely. You should know how to tell the difference between a nationwide average, and a rate that pertains to a single city. If the proposition is that New Orleans had a death rate of 53 per 100,000, that is not remotely refuted by pointing to a nationwide average of 6 per 100,000. All you’ve shown is that there are a lot of American locales that are much safer than New Orleans, and they all average out to 6. Again, so what?


"Q" the Enchanter 06.01.06 at 12:13 am

Perhaps commenter 3 above should have signed “dysfunctional.”


John Quiggin 06.01.06 at 12:23 am

Rilke is right. The implied number of deaths nationally for Iraq is out by a factor of around 10. This is illustrated by the obvious absurdity of the suggestion that DC is more dangerous than Baghdad (or any other place in Iraq, except maybe some parts of the Kurdish area and perhaps the Green Zone itself).


Betamax Guillotine 06.01.06 at 12:30 am

Forest for the trees, folks. What mendacity.

Kieran’s modest proposal is the best way to defuse this extremely disingenuous nonsense. In addition, the last time I checked, we didn’t have the U.S. Army sending combat patrols into D.C. and no one shows up at D.C. hospitals with arms and legs blown off from IEDs.

File Rep. King’s efforts under “The Thinking…It Hurts Too Much! Please Do It For Me!”.


felix 06.01.06 at 12:36 am

Why is this relevant?

I bet I can find a neighborhood in a city in Iraq that has experienced greater per capita death rates than Berlin during World War II. Would you accept this as evidence that Berlin during World War II was a safe place to live and that concerns about the dangers of living there during that time were absurd overreactions (quite probably the work of liberals working to undermine the acting government)?

If so, you should be placed on the “pay no mind” list due to stunningly poor arguments. If not, you also should be placed on that very same list for wasting our time by asking what the relevance was in the first place.

The message to take away from the quoted passage and the response is, of course, that Republican politicians lie reflexively, and that there is quite a large constituency in the US that either enjoys being lied to, or can’t tell when it is being lied to.


anno-nymous 06.01.06 at 1:51 am

From Iraqbodycount, these are the numbers for “civilians reported killed”:

“6,331 from 1st May 2003 to the first anniversary of the invasion, 19th March 2004

11,312 from 20th March 2004 to 19th March 2005

12,617 from 20th March 2005 to 1st March 2006”

12,617 deaths in 346 days in a population of 26,075,000 corresponds to 51.04 per 100,000 per year. For what it’s worth. (I figured I’d stick to the 2 decimal places standard).


Thomas Palm 06.01.06 at 3:15 am

Since the Iraqbodycount only counts deaths reported in the media the numbers are way too low. One of the creators has guessed they get about half the real number of deaths, and that may be optimistic. Then there are all the non-civilian deaths in Iraq that has no counterpart in Washington DC.

I think the suggestion of sending King on a vacation to Baghdad is a good one. It might give him some perspective, or maybe it’ll just give him a hole in the head.


Charles S 06.01.06 at 5:13 am

Another major factor is that Iraq Body Count (and, presumably, the Rep’s sources as well) is only counting civilian deaths due to the US military or insurgents. It is not counting murders or deaths from violence over all. So the category of deaths that lead to 40+ deaths per 100,000 in violent US cities (civilian-civilian violence) is simply not included in either the Iraq Body Count number or Rep. King’s number.

While I can’t recall seeing a murder rate (seperate from military and para-military violence) for Iraq or for Bahgdad, every more annecdotal description I have seen has reported extremely high rates of rape and kidnapping (unrelated to paramilitary and military actions), and I would be surprised if the rate of murder were not also at least at the levels of major US cities.

So even if the rate of deaths from US or insurgent actions were lower than the total violent death rate in some US cities, that does not demonstrate that the total rate of violent deaths in Iraq is less than the total rate in US cities. Given that Iraq Body Count works out to 50 per 100,000, with an estimated under-reporting factor of 2, and assuming that Iraq has only the civilian-civilian murder rate of New Orleans, that would mean that Iraq would have a total violent death rate of 150 per 100,000. And even that rate, as thomas palm points out, would ignore non-civilian violent deaths. If you or your brother or son or father (with the exception of US troops, armed combatants in Iraq seem to be overwhelmingly male) were killed because he took up arms to kill Americans, or to drive out the Sunnis in the neighborhood, or to try to keep the Sunnis from driving your family out of the neighborhood, or to join the Iraqi military or police, or to try to ensure your tribe’s dominance of the local political scene, would that make you feel safer than if he had just been killed while buying groceries?

It certainly wouldn’t provide much comfort to me.


KCinDC 06.01.06 at 7:01 am

Along with the various reasons King’s number is just plain wrong, the objection of comparing a city to a country seems valid as well. The people in Baghdad, for example, are not living in an ideal Iraq in which violence is uniformly distributed across the country. They’re living in Baghdad, which presumably has a higher death rate than Iraq as a whole, so the Baghdad death rate would be much more relevant in assessing their situation.


nick s 06.01.06 at 7:02 am

Remember, Rep. King is from Iowa. He sees more brown and black folks in the chamber of the House of Representatives than he does in his home district. No wonder he instinctively thinks D.C. is such a scary place.


Donald Johnson 06.01.06 at 11:27 am

Actually, Charles S, Iraq Body Count counts all civiian deaths caused by violence (criminal AND war-related) if they are reported in the media and pass their tests (I think two sources, but am not sure). But as someone else pointed out, they admit their numbers might be low by a factor of two. I think they probably miss a huge proportion of civilian deaths caused by coalition forces, because the Brookings Institute’s Michael O’Hanlon estimates there are roughly 1000 insurgents killed each month, and I doubt the US could kill so many without killing a comparable number of civilians, unless the insurgents conveniently place themselves in areas far from innocent people. In most months Iraq Body Count shows only dozens of civilians killed by the US.

Robert Fisk visited the Baghdad morgue last summer and found there had been 1000 murders in one month. Baghdad has a population of 5 million. Show me the American city with a murder rate that high.


MattXIV 06.01.06 at 2:54 pm

“The implied number of deaths nationally for Iraq is out by a factor of around 10. ”

Which is why Kieran’s post doesn’t really make sense, since it focuses on the city vs country argument. If King’s number was accurate, he’d have a good case that Iraq wasn’t as bad as it seems; but to reach King’s number, Iraq conditions in Iraq would have to be a lot better than they are now and we’d expect that would reflected in the other indicators. I’d like to know what methodology was used to come up 25.7, since it will inevitably be used again to cook future estimates.


JR 06.01.06 at 3:12 pm

King is rank and obvious liar.

He first claimed that he got his data from Pentagon reports “cross-checked with independent research.” That is a lie. The Pentagon does not collect data on civilian deaths in Iraq – or if it does, the information is classified.

King later admitted that he had no Pentagon data and was relying entirely on the Iraq Coalition Casualty Count website, http://www.opinionjournal.com/best/?id=110008402

That site, http://www.icasualties.org, does not claim to know the actual number of civilian deaths and only collects newspaper reports. Its total death toll for 2005 (with no data at all for two months) is 9658. Iraq Body Count, which collects yearly data from March to March, shows 12,617 for 3/05 to 3/06, without data for about one month, and says that it believes that the figures are understated.

But the biggest error is that both RBC and ICCC are counting violent deaths from the war – not ordinary violent death (murders, car accidents, suicides, etc.) of the sort that make up the Washington, DC figures. So unless you believe that ordinary crime and violence no longer exist in Iraq, you have to add the war-related violent death rate to an underlying non-war-related violent death rate- which is likely to be higher than usual due to the absence of ordinary public order.


y81 06.01.06 at 6:01 pm

Sad to say, it’s not at all clear to me whether it is more dangerous to be an average Iraqi male (randomly distributed across the country) or an average black urban male in an American city like Washington or New Orleans. No one who writes for or comments on this blog knows either, but their moral fervor is undimmed by their factual ignorance.


Charles S 06.01.06 at 8:23 pm

Donald Johnson,

You’re right, I was paying attention to the goal, and not the methodology.

I do suspect that civilian-civilian murder is probably under-reported in the press compared to murders by the military/militias/insurgents, although I can’t provide any evidence to support that claim.


your claim for the biggest source of the error is the same as mine, and, as Donald pointed out, we are both wrong. Check out the second paragraph of IBC’s explanation of their project, and you will see that they include all violent civilian deaths that make the papers. How many violent civilian deaths that are not caused by military/militia/insurgents make the papers is an open question (certainly answerable, but not trivial).


KCinDC 06.01.06 at 10:42 pm

And Y81 continues comparing apples with oranges. If the Iraqi male is randomly distributed across the country, while the American male must be in a large city (where crime, along with anything else that involves interaction between people, good or bad, is more frequent), then you’re stacking the deck to support your “clap harder” argument. And we won’t even get into why you need to make the American male black.


Donald Johnson 06.01.06 at 10:43 pm

There’s some evidence to support the notion that US-caused deaths are under-reported. I know I’ve seen at least one or two accounts (quite apart from the Haditha massacre) where US forces claim to have killed a certain number of insurgents and when reporters go to the scene, they find that the locals give a very different account. How typical is this? Well, I’ve only seen one or two cases, but I suspect that reporters rarely get to check up on the accounts the military gives of their engagements. The estimate I’ve seen of insurgent deaths (roughly 1000 per month) is awfully high and I doubt very much that reporters are able to check out more than a small fraction of these killings to see if the locals on the scene agree with the US military version of who got killed. But that said, this is just conjecture on my part. Reasonable conjecture, I think. It’s well known that Vietnam bodycounts often included civilians counted as VC–maybe the military has changed its ways, but I doubt it. Even apart from deliberate massacres, which are presumably rare, I’d just expect a bureaucracy to cover up its mistakes when possible and there’s no incentive for the US government to admit that large numbers of civilians are dying in its counterinsurgency operations if such a thing is indeed happening, except in those cases where someone catches the incident on tape.


JR 06.01.06 at 11:03 pm

Charles S- IBC doesn’t claim to include all civilian violent deaths, only “excess civilian deaths caused by criminal action resulting from the breakdown in law and order.” How they can tell an excess death from a normal criminal death is beyond me, but if you look at the incident-by-incident reports, very few of the reported deaths could be anything other than war-related. The normal average everyday violent death is a husband murdering his wife or a truck driver running a light and wiping out a couple of pedestrians- and deaths like that don’t show up on the IBC list.


roger 06.01.06 at 11:25 pm

I’m impressed by the peaceableness of Iraq. I think Representative King could do us all a favor and invite the Badr brigade and Muqtada al Sadr’s militia over to patrol these here states — starting with Sioux Falls, Iowa. Nothing like those militia to bring those murder numbers down. Oh, and we probably need to bring our own troops home now, too – obviously their job is finished in Iraq.
I feel a lot better about the President’s excellent adventure, now.


y81 06.02.06 at 2:18 pm

kcindc, I guess it’s the same reason a previous commentator had to accuse Rep. King of being a racist. What I don’t know is why you didn’t find that gratuitous racial reference troubling.

Seriously, it seems to me that the above statistical discussion does show that the overall rate of violent death in Iraq is of the same order of magnitude as that of urban black men in the U.S. Sort of interesting. No doubt, being an insurgent in Baghdad is yet more dangerous, as is being a member of the Crips. But neither is anything like being a European 18-year-old male in 1915.


John 06.03.06 at 4:57 pm

If you want to assess the risk that a randomly selected person will die of inury vs. that of a radomly selected person in the United States dying from injury, you should use all injury. In 2003, the latest year for which data are available at the CDC site http://webapp.cdc.gov/sasweb/ncipc/mortrate10_sy.html, the crude rate in the United States was 56.39 deaths caused by injury per 100,000 population. At that rate, we’d expect Iraq (population about 26,783,383, see http://www.cia.gov/cia/publications/factbook/geos/iz.html )to have about 15,103 deaths by injury per year…or about 41 per day.

If I had to bet I’d bet Iraq has a little more than that…but not that much. I do not think the risk of death by injury in Iraq is as high as the impression one might form from watching the news. But I do think the insurgency is accomplishing quite a bit in terms of psychological effect. That’s evidenced by this discussion.

Comments on this entry are closed.