Sunday, October 31, 2004

More On The Lancet Report

Updating my earlier post after I read the actual report:

One thing no one else pointed out is what they write about the possible exceptionality of their Fallujah sample: they write they are not convinced it is an outlier, for they didn't visit other destroyed towns like Najaf or Ramadi, and they hit a low-death-rate outlier in also destroyed Sadr City.

As for the numbers - first deaths counted, then the death rates per thousand people ('/t/y') per year I calculated (from their figure of 110,538 person-months before and 138,439 person-months after), then the difference, then the post-war total from the difference (using 17.8 months and 24.4 million people):

  • Accidents: 4 before 13 after, 0.434 /t/y before 1.127 /t/y after, increase 0.693 /t/y or 25,082;
  • Chronic disorders (incl. heart failure): 22 before 29 after, 2.388 /t/y before 2.534 /t/y after, increase 0.146 /t/y or 5,284;
  • Infectious diseases: 1 before 5 after, 0.109 /t/y before 0.433 /t/y after, increase 0.324 /t/y or 11,727;
  • Infant mortality: 6 before 10 after, 0.651 /t/y before 0.867 after, increase 0.216 /t/y or 7,818;
  • Violence: 1 before 73/21 (without Fallujah) after, 0.109 /t/y before 6.328/1.820 /t/y after, increase 6.219/1.711 /t/y or 225,086/58,374.

Without Fallujah, these add up to 108,285 deaths, but this analysis was done without probability calculations - difference of mean and median, I suspect. So here are figures corrected for the 98,000 statistical maximum and rounded:

  • Accidents: 22,700;
  • Chronic disorders (incl. heart failure): 4,800;
  • Infectious diseases: 10,600;
  • Infant mortality: 7,100;
  • Violence: 203,700/52,800.

Note that the figure for violence, the one that can be compared to the IBC figure also includes killed soldiers and insurgents, unlike the IBC figure.

Further breakdown of violent deaths: the 1 before the war was by the regime, 2 were by unknown perpetrators, 7 were by criminals, 2 by insurgents, 3 by US infantry soldiers, and 58 by aerial bombardement. The breakdown of Fallujah victims is not given, but if all were due to aerial bombardement, 6 remain for bombing victims outside Fallujah. More than 50% of the foreign soldier-killed were certainly civilians (28 children, 4 women, 1 old man out of 61) - following the authors' logic about larger exposure of men, I guess 7 civilian men aged 15-60, thus two-thirds civilian. Converted to absolute numbers with the same multipliers as before:

  • Killed by the Regime pre-invasion: 2,700/year
  • Killed by unknown perpetrators: 5,600
  • Killed by criminals: 19,500
  • Killed by insurgents: 5,600
  • Killed by US ground forces: 8,400
  • Killed by aerial bombardement: 161,800/15,000
  • Iraqi soldiers/guerillas killed by foreign forces: 56,700/7,800
  • Civilians directly killed by foreign forces: 113,500/15,600.


At 4:17 PM, Anonymous Anonymous said...

I have long felt it to be a moral certainty that the Iraq Body Count figures are too low. Fallujah and the other bombed cities cannot properly be left out. Thanks for this fine piece of work.

At 2:43 AM, Anonymous Anonymous said...

Hi DoDo- thank for your analyses here and at lenin's blog. Your criticisms of the Lancet paper as presenting transformed data in the form of rates per time are well founded, but partly necessitated by the cluster analysis used.

I wrote to ask if have you been able to extract any estimate from the Lancet article of total dead in Fallujah (till 2004)? Your help would be appreciated.

Best, ion

At 5:30 AM, Anonymous Anonymous said...

I'm extrapolating using Table 2 in the Lancet pdf for Fallujah deaths. There were 142 reported deaths, 52 of these in the Fallujah cluster. The body-count excluding Fallujah would thus comprise 90 deaths extrapolated to the 98,000 excess mortality. Thus, the 52 violent deaths in Fallujah would estimate out to 56,600 excess deaths.

I've emailed the authors to try to confirm.

Regards, ion

At 6:37 PM, Blogger DoDo said...

If I read the article correctly, the Fallujah clusters were meant to give an estimate for all of Anbar Province, not just Fallujah, and somewhere towards the end of the article they say the Fallujah figures would mean 200,000 dead there.

However, since the Fallujah clusters are a much smaller sample than the total, the error margin of such a sub-estimate is much much higher - it is much less reliable than the press reports.

I'd say the Fallujah death toll is an unknown.

Just for the April invasion, we have two reports of 800 dead and one of 1200 dead, and these were the centrally collected data. Dead rotting under the rubble or buried by their families in promptu were not counted. Fallujah is a town of reported 200,000 to 300,000 people, given the scale of the destruction, 10,000 dead would be not far off IMO.

Fallujah had hundreds of casualties in the year before, and my rough guess for the bombings since September is another one thousand dead.

At 3:42 AM, Anonymous Anonymous said...

Hey, excellent website. A great Iraq resource is Deaths in Iraq. It breaks all of the casualties down by age, race, branch of the military, country, etc.


Post a Comment

<< Home