Comments

1. #1 sod

   January 21, 2008

   Some of the field surveyors were employed by Moqtada Sadr’s Ministry of Health.

   pretty funny. the IFHS poll, with its results CHEERED at by the same people, was COMPLETELY performed by people under control of the health ministry.

   I said: ‘do a bigger survey to improve the accuracy, and I will make sure it gets the proper attention in the news media.’

   and i like this part. if the US government (or David Kane and Co) was interested in any sort of “truth” on this subject, all they need to do would be:

   approach some neutral polling agency and speak the same words: “do a bigger survey…”

   the costs of such an operation would be immensely low and is well within reach of Pentagon or most right wing groups.

2. #2 David

   January 21, 2008

   Speaking of “dishonesty,” what would you call this claim by Tirman from the article:

   The deaths-by-violence in that latter survey [IFHS] remained the same from year-to-year, however, which is not plausible-all observers agree that violent deaths were rising sharply in 2005 and 2006.

   Who are these “all observers” that Tirman is talking about? Besides the IFHS authors themselves, the IBC folks would disagree. Their data is largely consistent with the sort-of-steady death rate. They don’t count for Tirman? And, although I am not nearly as well-informed as the folks behind IFHS and IBC, I am an “observer” who finds the IFHS/IBC rates perfectly plausible. The key reason, perhaps, that death tolls in 2005/2006 were not that much different/higher than 2004 is because 2004 included two major battles for Falluja as well as other large scale operations.

   As always, IFHS/IBC/me could be wrong, but Tirman does his own reputation for honesty no good when he makes false claims about “all observers.”

   And, by the way, sod’s suggestion is an excellent one. Which “neutral” organization would he suggest? Lancet supporters should be careful what they wish for . . .

3. #3 Thom

   January 21, 2008

   Wow. I didn’t realize that Munro was an agitator for Iraqi invasion. That article he wrote for National Review is a bit on the scary side.

4. #4 sod

   January 21, 2008

   The key reason, perhaps, that death tolls in 2005/2006 were not that much different/higher than 2004 is because 2004 included two major battles for Falluja as well as other large scale operations.

   the Samarra bombing led to a MASSIVE increase in violence, but the number of dead was rising BEFORE it already.

   i have not seen anyone argue that the 2004/5 deathrate was higher than 05/06 before.

   why don t you start looking for sources for such a wild claim?

   remember, violence increased in a way, that even the US military had to aknowledge it.

5. #5 Ken

   January 21, 2008

   Seems like a lot of people are happy to criticize surveys that give results that they don’t like but don’t often go asking that more thorough, comprehensive surveys be carried out. Of course, if future surveys give results that critics, mostly from the safety of their comfortable homes, find unacceptable, they will be considered “unacceptable”.

   I can’t see that there can be long term benefits to ignorance of the real extent of excess deaths, health effects, refugee migrations, infrastructure damage or other hard consequences of decisions to go to war. Certainly it will not win the hearts and minds of the Iraqi people or their neighbors to demonstrate a willful blindness to those consequences. But it also won’t provide future military and policy planners with clear insights to have grossly wrong estimates of damage imbedded in the historical record of previous military interventions.

6. #6 guthrie

   January 21, 2008

   Ken- what makes you think that anyone in power is interested in what future military and policy planners think?

7. #7 John Tirman

   January 21, 2008

   Pentagon Cites Spike In Violence in Iraq
   Averting Civil War Called Main Goal

   By Ann Scott Tyson
   Washington Post Staff Writer
   Saturday, September 2, 2006; Page A01

   Rising sectarian bloodshed has pushed violence in Iraq to its highest level in more than two years, and preventing civil war is now the most urgent mission of the growing contingent of 140,000 U.S. troops in the country, according to a new Pentagon report released yesterday.

   Executions, kidnappings and other sectarian attacks targeting Iraqi civilians have soared over the past three months, contributing to a 51 percent rise in casualties among the population and Iraqi security forces, the report said. More than 3,000 Iraqis are killed or wounded each month, and by July, 2,000 of the casualties were the result of sectarian incidents, it showed.

   The Pentagon report, though consistent with what news media have reported for months, is significant because it represents an official acknowledgment of trends that are widely believed to be driving the country toward full-scale civil war….

   Overall, the number of weekly attacks in Iraq escalated to nearly 800, the highest level since the Pentagon began gathering the statistics in April 2004, the report said. Statistics collected by defense experts, however, indicate that the level is higher than at any time since the U.S.-led invasion overthrew the Saddam Hussein government in April 2003. [end]

   ——-
   The report the WaPo article refers to has tables on pp 31-32 that show a rather unambiguous rise in violence over the period I was describing. http://media.washingtonpost.com/wp-srv/nation/documents/Security-Stabilty-ReportAug29r1.pdf

   IBC last month declared it was switching from one media report from two media reports to confirm a civilian death. Whatever their utility, one cannot rely on their figures for much of anything. “De nihilo nihil.”

8. #8 David Kane

   January 21, 2008

   sod,

   Have you read the IFHS paper? It claims (table 4) that both its data and IBC show a lower death rate for 2004/2005 than for 2003/2004. And IFHS shows a lower rate for 2005/2006 then for 2003/2004. Anyway that you look at the IBC data, you can’t see anything like the explosive growth in death rates that L2 reports. (That doesn’t make L2 wrong, it just means that Tirman is not to be trusted when he makes claims about “all observers.”)

9. #9 David Kane

   January 21, 2008

   I appreciate John Tirman’s response and citation. But, really! If the phrase “all observers” (when used in conjunction with a discussion of civilian mortality in Iraq) does not include IBC, then I am just going to have to take every other factual claim he makes with a big chunk of salt.

   Also, there is a difference between “unambiguous rise” and the doubling (03/04 — 04/05) and then redoubling (04/05 — 05/06) which L2 describes. For all I know, Iraq was 4 times more violent in 05/06 than 03/04, but that is not at all what IBC or IFHS data show.

10. #10 Lee

    January 21, 2008

    David Cane attacks this statement as untrue:
    >blockquote>The deaths-by-violence in that latter survey [IFHS] remained the same from year-to-year, however, which is not plausible-all observers agree that violent deaths were rising sharply in 2005 and 2006.

    Later he says:

    Anyway that you look at the IBC data, you can’t see anything like the explosive growth in death rates that L2 reports.

    Note that Kane is moving the goalpost – the Tirman article claimed violetn deaths are rising sharply, Kane shifts that to whether IBC shows the same growth as L2. In any case, Kane is maintaining his claim that IBC does not support Tirman.

    Kane, have you looked at the fricking IBC data or analysis? This is a March 18, 2007 press release, currently on the IBC site.

    Year Four: Simply the worst

    March 18th 2007
    Summary

    On every available indicator the year just ended (March 2006 – March 2007) has been by far the worst year for violence against civilians in Iraq since the invasion:

    * almost half (44%) of all violent civilian deaths after the initial invasion phase occurred in the just-ended fourth year of the conflict

    * mortar attacks that kill civilians have quadrupled in the last year (from 73 to 289)

    * massive bomb blasts that kill more than 50 people have nearly doubled in the last year (from 9 to 17)

    * fatal suicide bombs, car bombs, and roadside bombing attacks have doubled in the last year (from 712 to 1476)

    * one in 160 of Baghdad’s 6.5 million population has been violently killed since the beginning of the war, representing 64% of deaths recorded so far

    These are the stark headlines derived from Iraq Body Count’s ongoing compilation and analysis of media reports of civilian casualties in the Iraq conflict, which has documented 65,000 violent deaths to date.
    Trends since 2003.

    Following the six week “Shock and Awe” invasion phase (March 19 – May 1, 2003), which alone caused the deaths of some 7,400 civilians, the violent death toll has steadily risen year-on-year. There were 6,332 reported civilian deaths in the 10.5 months following the initial invasion in year one, or 20 per day; 11,312 in year two, 55% up on year one’s daily rate; 14,910 in year three (32% up on year two); and a staggering 26,540 in year four (78% up on year three, and averaging 74 per day). Not counting the 7,400 invasion-phase deaths, four times as many people were killed in the last year as in the first. And from the invasion to the present, at least 110,000 civilians have been wounded, 38,000 of them during year four.

    Trends in the last year.

    Even within the last year, there has been a marked upward trend in violence. This trend is reflected in IBC’s monthly figures, which peaked in July at nearly 3,000 and have since remained elevated at around 2,500 or higher throughout the second half of the year. These IBC trends are broadly in line with the Pentagon’s latest assessment of trends in the security situation (which however include attacks on US and Iraqi troops as well as civilians). In the data collected by the Pentagon most casualties were Iraqis, despite that 68% of the attacks targeted US-led coalition troops.

    David, you choose to focus on one sentence in Tirman;s article, in which you claim that he was wrong (in essence because you think he left out a necessary modifier, “most, for example) and therefore you argue that one cant trust his honesty. That is a serious claim, especially from someone like you who so blatantly misrepresents the basic data like this. IBC reports a sharp upturn if violence in the year ending in March 2007, and increasing deaths in every year over the year before, as Tirman said.

    Kane, do you enjoy discrediting yourself like this?

11. #11 Lee

    January 21, 2008

    I left out the link:
    http://www.iraqbodycount.org/analysis/numbers/year-four/

    And re Kane’s talk about IBC not supporting the L2 “4x” rate, note this choice quote from that press release:

    There were 6,332 reported civilian deaths in the 10.5 months following the initial invasion in year one, or 20 per day; 11,312 in year two, 55% up on year one’s daily rate; 14,910 in year three (32% up on year two); and a staggering 26,540 in year four (78% up on year three, and averaging 74 per day). Not counting the 7,400 invasion-phase deaths, four times as many people were killed in the last year as in the first.

12. #12 Tim Lambert

    January 22, 2008

    Since Kane asked about “dishonesty”, let’s look. John Tirman:

    >The deaths-by-violence in that latter survey [IFHS] remained the same from year-to-year, however, which is not plausible-all observers agree that violent deaths were rising sharply in 2005 and 2006.

    David Kane:

    >Have you read the IFHS paper? It claims (table 4) that both its data and IBC show a lower death rate for 2004/2005 than for 2003/2004. And IFHS shows a lower rate for 2005/2006 then for 2003/2004.

    from [IFHS Table 4](http://content.nejm.org/cgi/content/full/NEJMsa0707782v1/T4):

    IBC Burnham IFHS
    2004/2005 0.44 6.6 1.56
    2005/2006 0.72 12.0 1.67

    Both IBC and Lancet2 show sharp increases for 2005/2006. But Kane cherry picked from table 4 to argue that it showed that IBC didn’t increase. Kane is not to be trusted.

13. #13 sod

    January 22, 2008

    sod, Have you read the IFHS paper? It claims (table 4) that both its data and IBC show a lower death rate for 2004/2005 than for 2003/2004. And IFHS shows a lower rate for 2005/2006 then for 2003/2004. Anyway that you look at the IBC data, you can’t see anything like the explosive growth in death rates that L2 reports. (That doesn’t make L2 wrong, it just means that Tirman is not to be trusted when he makes claims about “all observers.”)

    the downtick that IBC and IFHS are shwing from the 03/04 to 04/05 numbers can be easily explained by their methodology:

    reporting on death toll was huge during the initial invasion. so IBC got big numbers.

    reporting death by US bombardment to the Sadr health ministry feels like the right thing to do.

    reporting ethnic cleansing death by shiite death squads in 06 might not have felt like such a good idea…

    both IBC and Lancet2 show about a doubling in number from 04/05 to 05/06. if you have sources that contradict a sharp increase in violence during that time, the bring them on!

14. #14 Boris

    January 22, 2008

    I am just going to have to take every other factual claim [Kane] makes with a big chunk of salt.

15. #15 David Kane

    January 22, 2008

    Talk about cherry picking! I cite Table 4 and then Tim, instead of giving us all of Table 4, just gives us 2/3′s of it. Here is what Tim gave us plus the extra line.

    IBC Burnham IFHS
    2003/2004 0.60 3.2 1.77 *Left out by Tim!
    2004/2005 0.44 6.6 1.56
    2005/2006 0.72 12.0 1.67

    Why did Tim not give us the line for 2003/2004? After all, Tirman is making a claim about the increases in 2005 and 2006, and there is no way to consider an “increase” for 2005 unless you look at the data for 2004.

    Now, admittedly, we are comparing apples and oranges (a bit) here since Tirman’s quote just discusses (calendar) years 2005 and 2006 while the table gives ranges that cover half of each year. But the fundamental point is that Tirman dishonestly pretends that “all observers” agree with him when it is clear that the IBC data is much more consistent with IFHS than with L2. (And you can be sure that this is what IBC folks like JoshD believe.) Does anyone really believe that redoing the analysis for calendar years would change this result? I would take that bet.

    For both IBC and IFHS, the death rate goes down and then back up. For L2 it doubles and then redoubles. One of these three is not like the others.

    But the best part here is Tim’s obfuscation. He quotes me correctly. But every claim in that quote is true! He then quotes accurately a portion of the table. But nothing in that part of the table contradicts what I said! He then implies that this, somehow, shows that I am not to be trusted. His less careful readers, without bothering to check each detail, think that he has somehow caught me out.

    As always, I am a fan of Tim’s work. He is careful and smart. But this sort of trickery does nothing for his reputation.

16. #16 sod

    January 22, 2008

    David, let us hear those sources that claim that violence did not go up in 05/06!

    But the fundamental point is that Tirman dishonestly pretends that “all observers” agree with him when it is clear that the IBC data is much more consistent with IFHS than with L2. (And you can be sure that this is what IBC folks like JoshD believe.)

    what was that? a wild attempt to misinform again?

    But the deaths-by-violence in that latter survey remained the same from year to year, which is not plausible — all observers agree that violent deaths were rising sharply in 2005 and 2006.

    so the IBC guys will NOT agree with Tirman, that violence increased in 05/06?
    what are you willing to bet? i m holding!

17. #17 Tim Lambert

    January 22, 2008

    David Kane, I posted the relevant part of table 4 and linked to the entire table. You were caught cherry picking.

    The [graph on the front page](http://www.iraqbodycount.org/) of the Iraq Body Count shows violent deaths rising sharply in 2005 and 2006. But David Kane claims that the IBC folks would deny this.

18. #18 David Kane

    January 22, 2008

    I sometime wonder why I bother to argue these points. Tirman’s claim is:

    The deaths-by-violence in that latter survey [IFHS] remained the same from year-to-year, however, which is not plausible-all observers agree that violent deaths were rising sharply in 2005 and 2006.

    Now, the main (only?) place where any “year-to-year” comparisons are made by IFHS is in Table 4, so this is what Tirman must be referring to. Here is the key data.

               
               IBC  Burnham IFHS
    2003/2004  0.60    3.2  1.77 *Left out by Tim! 
    2004/2005  0.44    6.6  1.56
    2005/2006  0.72   12.0  1.67
    

    Note that I have added in the first row which, for some unknown reason, Tim failed to tell us about.

    It is obvious looking at this data that L2 (Burnham) tells a radically different story than IBC or IFHS. In this context, it is reasonable for Tirman to summarize the IFHS data as “remained the same.” Of course, it did not stay exactly the same. It was down and then up. But, in the context of L2, remain the same is a fair summary. And the same would apply to IBC. It did not stay exactly the same. It too was down and then up. And, sure, you can note that IBC and IFHS differ in that IBC shows a 20% increase from 2003/2004 to 2005/2006 while IFHS shows a 6% decrease. But, in comparison with L2, this different is irrelevant since L2 shows a 400% increase.

    And Tirman admits as much when he writes in comment #7.

    IBC last month declared it was switching from one media report from two media reports to confirm a civilian death. Whatever their utility, one cannot rely on their figures for much of anything. “De nihilo nihil.”

    So I am correct in claiming that Tirman does not consider IBC to be included in the set of “all observers.” And so we have the spectacle of Tim defending a claim that he thinks that Tirman is making about IBC data when, in fact, Tirman is making no such claim. Tirman does not claim that IBC data refutes IFHS. He argues that IBC data cannot be relied on “for much of anything.”

19. #19 Lee

    January 22, 2008

    David Kane,

    This is from the IBC press release I posted and linked above – which you seem to think no one will read, so you can pretend it isn’t there.

    IBC reports year-over-year increases in deaths in EVERY YEAR after the initial invasion:

    Following the six week “Shock and Awe” invasion phase (March 19 – May 1, 2003), which alone caused the deaths of some 7,400 civilians, the violent death toll has steadily risen year-on-year. There were 6,332 reported civilian deaths in the 10.5 months following the initial invasion in year one, or 20 per day; 11,312 in year two, 55% up on year one’s daily rate; 14,910 in year three (32% up on year two); and a staggering 26,540 in year four (78% up on year three, and averaging 74 per day). Not counting the 7,400 invasion-phase deaths, four times as many people were killed in the last year as in the first.

    Also, David, you keep changing your argument. The Tirman quote you attacked was this:

    The deaths-by-violence in that latter survey [IFHS] remained the same from year-to-year, however, which is not plausible-all observers agree that violent deaths were rising sharply in 2005 and 2006.

    It is clear that all the cited sources, including the IBC which yo said contradicted this,a ctually support it. There WAS a sharp rise in 2005 and 2006, and that rise shows in the numbers as well – they explicitly point that out themselves in this press release, a press release that had the purpose of pointing out exactly that, so they WOULD – they DID – agree with Tirman that there was increasing violence in that period.

    Also, David you seem to be changing your argument to ‘more consistent with IFHS than with L2″ In count of deaths, sure. But that is not what the Tirman said – he said there were increases in violence, agreed to by all observers. IBC agrees – they said so.

20. #20 Tim Lambert

    January 22, 2008

    Kane claims that the 64% increase in IBC in 2005/6 in table 4 can be fairly described as “remains the same”…

21. #21 David Kane

    January 22, 2008

    Yeah, Tim! That’s exactly what I claim . . . Anyway, you ought to e-mail John Tirman. He says that we “cannot rely” on IBC data and yet you, clever fellow, have shown that IBC data proves his point perfectly.

    I now more fully understand why so many knowledgeable people think that spending time discussing things with you is a waste of effort.

    But I am tougher than they are, or perhaps more of a glutton for punishment. I’ll have more updates on related topics in the next week or so that, with luck, you will find worthwhile enough to bring to your reader’s attention.

    And, by the way, it sure would be nice to devote a new thread to this topic.

22. #22 sod

    January 22, 2008

    But the fundamental point is that Tirman dishonestly pretends that “all observers” agree with him when it is clear that the IBC data is much more consistent with IFHS than with L2. (And you can be sure that this is what IBC folks like JoshD believe.)

    David, this claim of yours is simply a false citation. shouldn t you at least excuse?

    Anyway, you ought to e-mail John Tirman. He says that we “cannot rely” on IBC data and yet you, clever fellow, have shown that IBC data proves his point perfectly.

    what nonsense is this?
    IBC data shows a big increase from the 04/05 to the 05/06 number. as does Lancet2.
    you can notice this, even if you think that the IBC numbers in general are much too small….

23. #23 Lee

    January 22, 2008

    I note that Kane is steadfastly failing to respond to the IBC press release pointing out that their own data shows exactly the pattern that Tirman claims.

    Kane, those 2003-2004 number in the IBC column in the IHFS paper include the deaths during the initial 6 week shock and awe phase of the war. As IBC themselves point out, remove the invasion-phase deaths from the count, so that we are looking at deaths during our attempt to stabilize and protect Iraq, and yo9 see a steady pattern of sharp year-on-year increases in violence. Which is what Tirman claimed.

    IOW, IBC has already agreed with the Tirman claim that Kane is disputing – that, indeed, Kane has said is not supported by IBC data, and that IBC would disagree with.

24. #24 dhogaza

    January 22, 2008

    Everything David posts is a lie. It’s amazing.

    How does this man live with himself?

    And Tirman admits as much when he writes in comment #7.

    IBC last month declared it was switching from one media report from two media reports to confirm a civilian death. Whatever their utility, one cannot rely on their figures for much of anything. “De nihilo nihil.”

    So I am correct in claiming that Tirman does not consider IBC to be included in the set of “all observers.”

    Tirman explicitly tags IBC as belong to the set of “observers whose figures are not worth much of anything”.

    This set is clearly a subset of “all observers”.

    Therefore, David is lying when he claims that Tirman’s statement supports his claim that Tirman doesn’t consider IBC to be included in the set of “all observers”.

25. #25 Will McLean

    January 22, 2008

    Some defenders of Burnham et al are acting as though the trendline in the IFHS survey of violent death in Iraq is wildly out of whack with other reports. I think they are imputing much greater precision to the study than the authors claim.

    If you graph the IBC tally for the same three periods post-invasion, you get a V shape: high at the beginning because that period includes the invasion proper, lowest in the middle period and higher at the end as internal violence grows.

    The “best estimate” figures for IFHS are a shallower V. However, if you multiply the IBC figures by three, they still fall between the highest and lowest IFHS estimate for each period.

    I wouldn’t expect the trendlines to match exactly. Aside from sampling error, not every death date would be remembered correctly, which would probably tend to smooth the curve.

    On the other hand, Burnham et al is a worse fit to the IBC trendline. You need to multiply the IBC rate by 12 to get the Burnham estimate of violent deaths, which seems like an improbable amount of underreporting by the media. And if you do that, the IBC derived figure for March 2003- April 2004 is a lot higher than the high end of the range of Burnham estimates for the same period. Instead of a V, the Burnham estimates graph as a relentless upward slope.

26. #26 sod

    January 22, 2008

    sorry Will,

    but i have NEVER EVER heard or read anything about a “V-shape” in the trend of violent dead in Iraq, up until right now.

    i have heard a LOT about a constant increase. and i ve heard about the MASSIVE increase in 06.

    again:

    if you have any evidence for this development of violent dead in Iraq, bring it on!

27. #27 Will McLean

    January 22, 2008

    sod writes:

    “sorry Will,

    but i have NEVER EVER heard or read anything about a “V-shape” in the trend of violent dead in Iraq, up until right now.”

    Only because you haven’t been paying attention.

    Look at post #18. IBC recorded a higher death rate 2003/2004 than 2004/2005, because 2003 includes the invasion phase. If you don’t believe me, go to the IBC site and look at the graph on the homepage. See the huge spike on the far left?

    Iraqis that were killed during the invasion phase are just as dead as those that died later.

28. #28 trrll

    January 22, 2008

    Yeah, Tim! That’s exactly what I claim . . . Anyway, you ought to e-mail John Tirman. He says that we “cannot rely” on IBC data and yet you, clever fellow, have shown that IBC data proves his point perfectly.

    So what? He said “all all observers agree that violent deaths were rising sharply in 2005 and 2006.” He didn’t say anything about reliability, and his argument certainly does not turn on whether or not he regards the IBC numbers as reliable.

    Moreover, a measure may be more reliable in some respects than in others. Everybody acknowledges that IBC’s methodology omits some deaths, so their numbers are at best a lower limit, although there is disagreement as to how large the downward bias is. But even biased numbers can provide an estimate of change from year to year so long as the degree of bias remains reasonably constant, because a constant bias will divide out.

29. #29 Lee

    January 22, 2008

    Will, that isn’t a ‘v-shape’ It is a steadily rising tide of violence, with a 6-week period of high death rate for entirely different causes, the invasion, grafted onto the far left end.

    This is why the IBC press release removes the initial invasion phase deaths when comparing years – the point is to isolate the internal violence effects from the invasion effects.

    Tirman climbed sharply rising violence – the IBC data supports that claim.

30. #30 Will Mclean

    January 22, 2008

    Lee writes:

    “Will, that isn’t a ‘v-shape’ It is a steadily rising tide of violence, with a 6-week period of high death rate for entirely different causes, the invasion, grafted onto the far left end.”

    The causes of the violent death rate during the invasion phase were not relevant to the IFHS results. It measured violent deaths during the relevant period. Period. Neither IFHS nor Burnham et al tried to remove invasion deaths from their results.

    If the IFHS survey, or Burnham et al, asked Iraqis if household members had been violently killed, but told that any that died in the six weeks following the invasion didn’t count because they died fom an entirely different cause, that would be entirely different.

31. #31 Lee

    January 22, 2008

    Will, that is true – dead people dont care how they died. But it is irrelevant to the point under discussion.

    Kane made a charge about what Tirman said. Contrary to what Kane charged, and in accordance with Tirman’s statement, violence WAS increasing sharply – including in the IBC data.

    Looking at the data on an annual granularity makes it look like 2003/2004 IBC data show high violence, and dispute Tirman’s claim – but they don’t. Violence immediately after the invasion was relatively low and it kept increasing – just as Tirman said, and just as IBC confirms in their press release.

32. #32 Kevin Donoghue

    January 23, 2008

    Will McLean: Neither IFHS nor Burnham et al tried to remove invasion deaths from their results.

    While the Roberts (2004) and Burnham (2006) studies didn’t actually aim to exclude soldiers killed in the invasion phase, their definition of a household is very likely to have had that effect – so that they are missing some deaths on the left-hand side of the V curve. Part of the difference between those two studies may be that soldiers’ deaths were more likely to be included due to a form of recall bias; after a few years, respondents would be more likely to forget the fact that a soldier wasn’t a member of the household at the time of his detah. As for IBC, deaths (especially civilian deaths) were obviously a much hotter topic for reporters during the invasion phase; that’s why IBC came in for so much criticism from war bloggers in the early days.

33. #33 Donald Johnson

    January 23, 2008

    “As for IBC, deaths (especially civilian deaths) were obviously a much hotter topic for reporters during the invasion phase; that’s why IBC came in for so much criticism from war bloggers in the early days.”

    I think the bulk of the invasion deaths were actually counted in the months immediately following the invasion–some news organisations went around hospitals and asked for statistics. It was easier for reporters to move around then because the insurgency was at a low level initially. Whether the hospital and morgue figures were complete I don’t know. But anyway, IBC counted about 7000 civilian deaths in that phase, which is still (by their count) the most violent one or two months of the war.

    One other factor I’ve seen mentioned elsewhere (or maybe here) is that IFHS relied to some extent on IBC data to do their estimates for the most violent areas that they couldn’t survey themselves. So it’s not two independent analyses.

34. #34 z

    January 25, 2008

    If only they hadn’t banned DDT, the death toll in Iraq would be negative. Damn you Rachel Carson!!