Editore"s Note
Tilting at Windmills

Email Newsletter icon, E-mail Newsletter icon, Email List icon, E-mail List icon Sign up for Free News & Updates

August 31, 2007
By: Kevin Drum

UP IS DOWN AND VICE-VERSA....Juan Cole is pissed:

I saw on CNN this smarmy Bush administration official come and and say that US troop deaths had fallen because of the surge, which is why we should support it. Just read the following chart bottom to top and compare 2006 month by month to 2007. US troop deaths haven't fallen. They are way up.

....How brain dead do the Bushies think we are, peddling this horse manure that US troop deaths have fallen? (There are always seasonal variations because in the summer it is 120 F. in the shade and guerrillas are too heat-exhausted to fight; but the summer 2007 numbers are much greater than those for summer 2006; that isn't progress.) And why does our corporate media keep repeating this Goebbels-like propaganda? Do we really live in an Orwellian state?

I'll leave that last question to Atrios, who, if memory serves, is pretty good at answering questions like that. My part in this is simpler. Prof. Cole wants someone to turn his troop fatality numbers into a graph to make it clear exactly what he's talking about, and obviously I'm your man for that kind of grunt work. So here it is. Pictorial evidence that troop fatalities in Iraq are down1 this year, just like surge proponents are saying. Pay attention, Congress.

1See post title for clarification of commonly accepted surge terminology conventions.

Kevin Drum 11:17 AM Permalink | Trackbacks | Comments (64)

Bookmark and Share
 
Comments

once you understand that up is down it all becomes clear.

Posted by: supersaurus on August 31, 2007 at 11:25 AM | PERMALINK

Kevin-
Why do you hate America? Don't you know your use of so called "mathematical facts" and "graphical evidence" only emboldens the terrorists?

You probably think it's none of our damn business where Craig puts his winkie either?
Godless liberal.

Posted by: cboas on August 31, 2007 at 11:28 AM | PERMALINK

Once again I have a units issue.
The grapgh should be Deaths per Amount of troups.
(Or something similar)

Posted by: Robodruidhatesspam on August 31, 2007 at 11:29 AM | PERMALINK

As Craig's arresting officer said, Embarrassing, embarrassing. No wonder why we're going down the tubes.

It ought to be the tagline for the Republican Party.

Posted by: frankly0 on August 31, 2007 at 11:29 AM | PERMALINK

As Craig's arresting officer said, Embarrassing, embarrassing. No wonder why we're going down the tubes.

It ought to be the tagline for the Republican Party.

Not bad, but shouldn't they also explain that by "tubes" they don't mean dump trucks?

Posted by: shortstop on August 31, 2007 at 11:31 AM | PERMALINK

Hmmm... Since you are the man for this kind of grunt work, it would be really interesting to add data for previous years to the chart.

Posted by: tom on August 31, 2007 at 11:31 AM | PERMALINK

The Republican public support for the war is still pretty strong, or so say the radio pundits. I wonder if the politicians feel comforted by that, knowing the '08 elections are over a year away and things can only get worse.

Now is the time for them to accept Reid's olive branch. Now is the time to cut funding and get out of Iraq.

Posted by: MarkH on August 31, 2007 at 11:34 AM | PERMALINK

Some wingnut with a blog can now say "Even liberal Kevin Drum says 'troop fatalities in Iraq are down this year'".

Do you think he'll bother explaining your sophist inversion to his readers?

Posted by: floppin' pauper on August 31, 2007 at 11:36 AM | PERMALINK

Prof. Cole wants someone to turn his troop fatality numbers into a graph to make it clear exactly what he's talking about, and obviously I'm your man for that kind of grunt work

And how are the death numbers any worse than the numbers for Bill Clinton? As ALICIA COLON of the New York Sun pointed out, "What is never compared is the number of military deaths during the Clinton administration: 1,245 in 1993; 1,109 in 1994; 1,055 in 1995; 1,008 in 1996. That's 4,417 deaths in peacetime but, of course, who's counting?"

Posted by: Al on August 31, 2007 at 11:37 AM | PERMALINK

Come on, Al. That was poor even by your pathetic standards.

Posted by: exasperated on August 31, 2007 at 11:41 AM | PERMALINK

Ahhh kevin,I so admire your self definition. For those interested in other numbers use the machinery on "Iraq Coalition Casualities" where you can search for deaths by nation, month, age, sexuality, etc.

http://icasualties.org/oif/

Posted by: shrink in sf on August 31, 2007 at 11:43 AM | PERMALINK

Yes! Al- you are the MAN!
Coming in to save the day with the illogical douche-baggery.
"The guy before was WAY worse than me, therefore I'm not so bad."

Posted by: cboas on August 31, 2007 at 11:43 AM | PERMALINK

Forget about arguing over numbers. The surge is working, because they say it is.

I think the Dems would be better served by focusing on the ability of the military to sustain the surge, not to mention the 2-3-4 deployments that are tearing apart the military.

Posted by: trainwreck on August 31, 2007 at 11:46 AM | PERMALINK

Al, when's the last time I mentioned you're an idiot? You seem to need some reminding.

Those numbers for the Clinton administration include all deaths due to illness and accidents, for example.

It's apples and oranges.

But, again, you're neither. You're an idiot. Maybe write it down on your sleeve?

Posted by: frankly0 on August 31, 2007 at 11:46 AM | PERMALINK

I agree that information on the deaths per total force might be informative.

Posted by: kevin. on August 31, 2007 at 11:51 AM | PERMALINK

Ah Kevin,

ONe glance at your graph makes evident the fact that you left off September, October, NOvember, and DECember. CAught cherry pickiing the months!!

Posted by: EGbert on August 31, 2007 at 11:57 AM | PERMALINK

The Republican scum will say anything and the media will just follow along. But on the (real) question of what the higher death toll means, is it because the troops are actually actively doing more fighting (including higher troop numbers) or is it because the other side has gotten more effective, or (no doubt) some combination of the two? The raw death numbers don't tell you much.

Posted by: abc on August 31, 2007 at 11:58 AM | PERMALINK


I agree that information on the deaths per total force might be informative.

That's great. We should send in 2 Million soldiers, so even if 2000 die per month, it will only be 0.1%, which is smaller than the percentage of the people who die from Restless Leg Syndrome.

Posted by: gregor on August 31, 2007 at 11:58 AM | PERMALINK

Facts serve at the pleasure of the president. Acknowledging that Up is now Down means that we're makin' progress.

Posted by: Nemo on August 31, 2007 at 12:01 PM | PERMALINK

So "Al", where are the numbers for the Reagan-Bush years and the last four years of Clinton ?

If you`re gonna play the game of "The Numbers" at least be balanced.

"At Least the War on the Environment is Going Well" - bumper sticker

Posted by: daCasacadian on August 31, 2007 at 12:02 PM | PERMALINK

Fallen from 120/month to 80/month since May.

You people are idiots.

Posted by: am on August 31, 2007 at 12:02 PM | PERMALINK

Fallen from 120/month to 80/month since May.

It's called summer, cretin.

Heard of it?

Posted by: frankly0 on August 31, 2007 at 12:05 PM | PERMALINK

The funny thing is that it does look like the numbers for Aug 2007 -- although higher than Aug 2006 -- are lower than might be expected given the, ehem, "surge" in deaths for Jan though July 2007 as compared to 2006, but the given the chance to make their argument with truth or lies, the GWB admin will always opt for the later.

Posted by: Disputo on August 31, 2007 at 12:05 PM | PERMALINK

I'm as much of a liberal as the next guy (at least in this crowd), but at this point troop deaths are lower in August than they were in January (Feb 1 being more-or-less the surge start date), both in an absolute sense, and as a percent of 2006 deaths.

In the interest of being anti-surge, I really don't think that this provides any evidence of the surge working. Saying that August happens to have a lower increase in troop deaths (over 2006) than January does is a single data point trend (ie useless). My disagreement with Juan Cole is that it's not a clear indictment of the surge either.

Posted by: prankster on August 31, 2007 at 12:11 PM | PERMALINK

Once again I have a units issue.
The grapgh should be Deaths per Amount of troups.
(Or something similar)

As Kevin said, that would be informative, but would not address the claim the actual death numbers are down. Deaths per 1000 (or whatever) would be the rate. They are not claiming the death rate is down (so far as I know) but that the raw numbers are down.

And while we're looking for numbers, what's the casualty numbers/rate? We need to know the level of wounded, too.

And then there are all those pesky dead Iraqi's who we are destroying to protect.

Posted by: martin on August 31, 2007 at 12:12 PM | PERMALINK

It's interesting to me that overall, Kevin's commenters are top-notch, but his trolls are dumber than dirt.

Posted by: SteveK on August 31, 2007 at 12:12 PM | PERMALINK

The funny thing is that it does look like the numbers for Aug 2007 -- although higher than Aug 2006 -- are lower than might be expected

The numbers here for a single month are so small and subject to random events (e.g. maybe it was a hotter August in 2007?) that it would be absurd to base a case on so slender a reed.

And, beyond that, for what it's worth, the number for August 2007 did not include the last several days of August, rendering the apparent visual significance of the chart more misleading.

Posted by: frankly0 on August 31, 2007 at 12:15 PM | PERMALINK

If the Bush regime is so concerned about troop fatalities they should consider that our troop fatalities in Iraq from March 1992 to February 2003 were running at zero a year. They seem to have spiked significantly once we attacked and invaded Iraq....

Posted by: Stefan on August 31, 2007 at 12:17 PM | PERMALINK

I'm as much of a liberal as the next guy (at least in this crowd), but at this point troop deaths are lower in August than they were in January (Feb 1 being more-or-less the surge start date), both in an absolute sense, and as a percent of 2006 deaths.

Again, it's the heat, not the insurgency. The rebels and soldiers go out less when it's 120 degrees, so deaths in summer will tend to be less than deaths in winter and spring given that the overall rate of activity due to the extreme heat is much lower.

Posted by: Stefan on August 31, 2007 at 12:19 PM | PERMALINK

All Orwell all the time. Why not? Being monomaniacal about sports has worked for ESPN.

Posted by: JeffII on August 31, 2007 at 12:20 PM | PERMALINK

Yes, indeed - If only more Repug's tubes could be tied.

From the Australian of August 31,'07, Dennis Shanahan, in Baghdad writes with the lede of:

Surge Working: top US General

In the article he quotes Petraman as saying:
"General Petraeus told the Australian, during a face to face interview at his Baghdad headquarters, that there had been a 75 per cent reduction in religious and ethnic killings since last year, a doubling in the seizure of insurgent weapon caches between January and August, a rise in the number of al-Quada "Kills and captures", and a fall in the number of coalition deaths from roadside bombings.

Petraeus added, "We say we have achieved progress and we are obviously going to do everything we can to build on the progress and we believe that al-Quada is off balance at the very least"

The General went on to add that the surge strategy involving the extra 20,000 troops would continue for a few months before the troop level in the country was phased down. But the objective was to hold the gains that had been made so far."

And, this is going to change in the next two weeks, how??? The spin is in full force.

Posted by: thethirdPaul on August 31, 2007 at 12:20 PM | PERMALINK

No one is as dumb as Kevin's trolls. That's because the Washington Monthly couldn't afford real trolls so they just have some cheap perl scripts doing an XSLT translation from neopet chats mashed up with V1AGRA spam.

You can read more about this at SIGXML this year. I think it's quite clever, economical, and achieves 95% of what real trolls achieve, which is of course hits and more ad impressions.

Posted by: jerry on August 31, 2007 at 12:23 PM | PERMALINK

Kevin,

Kudos for the pickup on the chart - Yes, the good Professor did say that he was engaged in a conference and asked if anyone could create a chart.

Posted by: thethirdPaul on August 31, 2007 at 12:25 PM | PERMALINK

I'm as much of a liberal as the next guy (at least in this crowd), but at this point troop deaths are lower in August than they were in January (Feb 1 being more-or-less the surge start date), both in an absolute sense, and as a percent of 2006 deaths.

When you have year-to-year data, why are you looking at troop deaths in January 2007 vs. August 2007 rather than at January 2006 vs. January 2007 and August 2006 vs. August 2007? You're looking at an apple and demanding that we acknowledge that it's not an orange.

Posted by: Mnemosyne on August 31, 2007 at 12:49 PM | PERMALINK
And how are the death numbers any worse than the numbers for Bill Clinton?

Since the stats are not for global troops deaths, but for "troop fatalities in Iraq", I'm going to go out on a limb and guess that the comparable numbers under Bill Clinton were pretty close to zero. I don't recall any reports of NFZ or Desert Fox fatalities, though there may have been a few. But, if you think the comparable numbers were worse under Clinton, please, present some evidence of worse comparable numbers, not global deaths of active-duty service members to compare to the death toll in Iraq.

Posted by: cmdicely on August 31, 2007 at 12:52 PM | PERMALINK

With the billions that Halliburton shareholders, like Richard Bruce Cheney, are making, a few hundred dead American grunts is no big deal. Come on - war profiteering is hard work. Bush himself comes from a long line of war profiteers and he couldn't give a horse's patootie whether a few poor, dumb American kids get blown apart, as long as the dividend checks keep rollin' in....

Posted by: The Conservative Deflator on August 31, 2007 at 1:11 PM | PERMALINK

Al - you're a fucking genius, your comments just keep getting better and better.

Posted by: blowback on August 31, 2007 at 1:19 PM | PERMALINK

Here's a graph showing how the Pentagon changes the numbers
The only rationale is for political spin. When a government is so dishonest in everything it does, it's a shitty government.

Posted by: Mike on August 31, 2007 at 1:24 PM | PERMALINK

The depressing thing is not that proof of failure on the Adminstations part is not reason for withdrawl.
Proof of failure is reason for sending more troops. And proof that liberals and media are undermining Bush's efforts.
Bush Losing does not equal Democrats right.
Bush Losing prooves Democrats more wrong.

Posted by: cboas on August 31, 2007 at 1:28 PM | PERMALINK

I saw on CNN this smarmy Bush administration official come and and say that US troop deaths had fallen because of the surge, which is why we should support it.

Logically, troop deaths should be up because there are more targets to shoot at, and they are adopting new tactics that get them out from their protected enclaves and in and among the Iraqis.

Posted by: Swaggeering Jingoistic RSM on August 31, 2007 at 1:36 PM | PERMALINK

I would be interested to see 2005 and 2004 in there as well.

Posted by: jefff on August 31, 2007 at 1:40 PM | PERMALINK

Logically, troop deaths should be up because there are more targets to shoot at, and they are adopting new tactics that get them out from their protected enclaves and in and among the Iraqis.

Which is one reason why measuring The Surge success by lower US fatalities is absurd. Another reason is that leaving Iraq would be even more successful than The Surge... oh, wait....

Posted by: Disputo on August 31, 2007 at 1:41 PM | PERMALINK

Kevin --- All of the seasonality of Coalition fatalities charts that your heart could desire are over at the Newshoggers as I put them together at the start of August to look at July casualty figures in expectation. Here is the short story for 2007

"all seven months of 2007 have seen higher total coalition fatalities than their corresponding month in 2006. Five of the seven months have seen their war to date maximums in 2007."

If you include August in this conclusion, seasonally adjusted August 2007 was slightly better than we would expect compared to July, but August 07 Coalition fatalities were higher than August 06, and as of 12:00pm EST on 8/31 two fewer coalition deaths in August 07 than the war to date high of August 2005.

Posted by: fester on August 31, 2007 at 1:51 PM | PERMALINK

"Statistics: The only science that enables different experts using the same figures to draw different conclusions."
Evan Esar
Can we quit the pissing match an decide that this means nothing to the moron running the show right now?

Posted by: cboas on August 31, 2007 at 2:02 PM | PERMALINK

Proof of failure is reason for sending more troops.

When are you joining up?

Posted by: thersites on August 31, 2007 at 2:12 PM | PERMALINK

you obviously don't understand my point.

Posted by: cboas on August 31, 2007 at 2:23 PM | PERMALINK
It's called summer, cretin.

Didn't they open up for Fountains of Wayne?

Posted by: kenga on August 31, 2007 at 2:39 PM | PERMALINK

"All of the seasonality of Coalition fatalities charts that your heart could desire are over at the Newshoggers..."

Ahh yea, that chart of monthly deaths for every year of the war was exactly what I wanted. The seasonal pattern is easily discernible to the eye in every year of the war. Even 2003 matches up pretty well, which I did not particularly expect.

Posted by: jefff on August 31, 2007 at 3:17 PM | PERMALINK

Seasonality also explains why the Petraeus/WH report was scheduled for Sept.

Posted by: Disputo on August 31, 2007 at 3:59 PM | PERMALINK

I've been following the casualty figures posted at icasualties.org for some time and making my own graphs based on them. (The latest one is here, if you're interested; in a few days I'll be updating to include August.) The variations over 2-3 months can be huge and are worthless in assessing trends. but when you look at all the data, it's clear that US troop deaths have been distinctly higher since the beginning of last fall. The jump in military deaths came before the surge started, and the surge has not led to a reduction.

Posted by: Dr. Drang on August 31, 2007 at 4:37 PM | PERMALINK

I agree with Juan Cole that the short term drop in US combat deaths is a weak argument in favor of showing that the surge is working.

However, I am unimpressed with Cole's general approach to argument, which seems to be:
Respond only to the weakest argument put forth by the US government then compares the US government to a Nazi.

No wonder Yale turned him down.

Posted by: ex-liberal on August 31, 2007 at 4:41 PM | PERMALINK

How long do y'all bet it is before they start saying "But the fatality growth rate is decreasing sharply! Hallelujah!"

Posted by: Minivet on August 31, 2007 at 4:49 PM | PERMALINK

Speaking of data..According to this source, troop levels are as follows (I dug up the figures for July and August from different sources). The fatality figures are those provided by Juan Cole:

Jan 2006 155,000 (fatalities = 62)
Jan 2007 132000 (fatalities = 83)

Feb 2006 138,000 (fatalities = 55)
Feb 2007 137000 (fatalities = 81)

March 2006 138000 (fatalities = 31)
March 2007 142000 (fatalities = 81)

April 2006 138000 (fatalities = 76)
April 2007 155000 (fatalities = 104)

May 2006 133000 (fatalities = 69)
May 2007 155000 (fatalities = 126)

June 2006 127000 (fatalities = 61)
June 2007 162000 (fatalities = 101)

July 2006 127,000 (fatalities = 43)
July 2007 154000 (fatalities = 79)

August 2006 138,000 (fatalities = 65)
August 2007 160,000 (fatalities = 77)

A couple of statistical bits: The average deaths for 2006 was 58 per month (standard deviation = 15); the average for 2007 was 92 (standard deviation 17). That's a significant difference (P=.0009; t = 4.2).

Upthread, someone wonders if the difference is simply the result of more boots on the ground? Kudos for considering base rates! Here's the data:

Number of troop deaths per 10,000 soldiers:
Jan 2006 = 4.8
Jan 2007 = 6.3

Feb 2006 = 4.0
Feb 2007 = 5.9

March 2006 = 2.3
March 2007 = 5.7

April 2006 = 5.5
April 2007 = 6.7

May 2006 = 5.2
May 2007 = 8.1

June 2006 = 4.8
June 2007 = 6.2

July 2006 = 3.4
July 2007 = 5.1

August 2006 = 4.7
August 2007 = 4.8

Average deaths per 10,000 soldiers per month in 2006 = 4.2
Average deaths per 10,000 soldiers per month in 2007 = 6.1

So no, the increase in absolute deaths is not the result of an increase in the numbers of soldiers. The incidence of deaths per 10K soldiers is significantly higher in all months of 2007, except August (P = .0044; t = .519)


Posted by: PTate in FR on August 31, 2007 at 4:50 PM | PERMALINK

Kevin -- What is the source for your numbers? They don't line up with icasualties.org, whatever criteria is used. E.g., for Aug-07, you show about 78(?); icasualties.org shows total (84) vs. hostile only (58), or total (84) vs. US only (64).

Posted by: has407 on August 31, 2007 at 4:51 PM | PERMALINK

No wonder Yale turned him down.

Yale turned Juan Cole down for the same reason DePaul fired Norm Finkelstein -- insufficient fealty to the Zionist mafia. It's the new McCarthyism.

Posted by: Disputo on August 31, 2007 at 4:54 PM | PERMALINK

never mind... I see Cole is using US-only deaths from icasualties.org; the numbers for Aug-07 have changed up from 77 to 80.

Posted by: has407 on August 31, 2007 at 4:56 PM | PERMALINK

US Active Duty Military Deaths 1980 - 2004

Not up to date, but shows uptick in 'hostile action' deaths due to Iraq war. Lowest total deaths were in period from 1997 to 2001.

http://www.murdoconline.net/pics/Death_Rates.pdf

Posted by: ApostasyNow on August 31, 2007 at 5:17 PM | PERMALINK

If August deaths are 80, rather than 77, my data should read:
Aug 2006 4.7 deaths per 10,000
Aug 2007 5 deaths per 10,000

(& I copied the wrong value for the t value above. For those who care, the values, adjusting for 80 deaths in August, should read-- P is .0035, and the t = 3.49.)

Posted by: PTate in FR on August 31, 2007 at 5:23 PM | PERMALINK

thethirdPaul From the Australian... "General Petraeus told the Australian, during a face to face interview at his Baghdad headquarters, that there had been a 75 per cent reduction in religious and ethnic killings since last year...

More a bald faced interview. Based on the DoD's own report (Jun-07), the year-to-date total for 2007 Jan-Apr ~3400 vs. all of 2006 Jan-Dec ~11600 supports a claim of a "75 per cent reduction". However, comparing the same period Jan-Apr shows a doubling from 2006 ~1700 to 2007 ~3400.

Posted by: has407 on August 31, 2007 at 5:24 PM | PERMALINK

Are those tubes we're going down as a nation the same ones that Senator Stevens was talking about?

Posted by: angryspittle on August 31, 2007 at 5:30 PM | PERMALINK

p.s. To put that in perspective--according to the DoD's own numbers--a 75% reduction for 2007 Jan-Aug will require a net negative ~2400 deaths for May-Aug '07, or massive revision downward of the Jan-Apr '07 numbers shown in the DoD's Jun-07 report.

Posted by: has407 on August 31, 2007 at 5:33 PM | PERMALINK

But, but, If Bush couldn't lie than he wouldn't have anything to say.

Yeah, Murdock helps Bush lie too. Fox isn't conservative - talks about liberals news might as well be talk about half-way honest media to Murdocks lies.

Posted by: Me_again on August 31, 2007 at 10:23 PM | PERMALINK

What a disgusting gaggle of nitpickers you all are! Every number on that chart and every number each of you has cited in your criticisms represents a once living and breathing human being - soldier, marine, man, or woman. Up or down raw numbers, rate of growth, comparisons to previous years - none of it matters. Bush is going to continue the status quo until he is out of office, no matter what information this post compiles/graphs or what Gen. Petraius et al report. The real issue is why are we in Iraq, what are we doing to an overexposed military, what the impact will be for our survivors, and what horrors have we wreaked upon that country? If the goal (always shifting) was to have a friendly nation (read compliant) in the middle east, then we have failed no matter how long we stay. All of us need to google Mark Twain's "The War Prayer" . . . it will curl your soul.

Posted by: mary on September 2, 2007 at 11:37 AM | PERMALINK

Figures lie and liars figure. Using the same data set that Kevin used, the most violent months since the war started, by month, were January 2005, February 2007, March 2007, April 2004, May 2007, June 2007, July 2007, August 2005, September 2004, October 2006, November 2004, and December 2006. Put another way, over the past year, 7 of the months have been the deadliest of the war. August 2007 (83) was, for all intents and purposes, the same as August, 2005 (85). Clearly, we're winning.

For the past 12 months, every month's deaths exceeded the average number of deaths for that month except for November, 2006.

Posted by: Conman on September 4, 2007 at 12:52 PM | PERMALINK




 

 

Read Jonathan Rowe remembrance and articles
Email Newsletter icon, E-mail Newsletter icon, Email List icon, E-mail List icon Sign up for Free News & Updates

Advertise in WM



buy from Amazon and
support the Monthly