This database entry is based on information provided by staff of the Baghdad
city morgue covering the period April 14th to August 31st 2003.
The morgue is said to record some 90% of "violent, suspicious" deaths in the
city. Currently about 60% and above of these deaths are the result of gunshot
wounds; this compares to approximately 10% pre-war. People killed by coalition
forces amount to an estimated 15-20% of gunshot victims brought to the morgue
according to a Newsweek report [Aug 18 2003],
but most of the violence is Iraqi-on-Iraqi.
The media reports from which x132's numbers are compiled are those which contain
exact totals tied to specific periods. We excluded reports offering only "snapshot" daily
or weekly rates, or any totals apparently based on simplistic extrapolations
without an appreciation that such rates can fluctuate over time (see below).
As in all our work, two media reports were needed to verify numbers for inclusion
in the count, but the measures referred to above ultimately reduced contributing
reports from 20 or more to 7, and the potential number of deaths (before
adjustments) from over 4,000 to 2,846.
This total of 2,846 was then adjusted in two ways.
First, pre-occupation rates for comparable deaths were deducted based on information
for the equivalent month in 2002 and when this was unavailable, as derived
from an average for the entire year. What we publish is therefore an estimate
of additional deaths in the city directly attributable
to the breakdown of law and order following the US occupation of Baghdad.
This adjustment yielded a total of 1,519 deaths in Baghdad, April 14th
to August 31st, over and above the normal "background level" of violent
deaths in the city.
Second, as in all our database entries, confirmed and
potential overlaps with other, already published incidents in Baghdad during
the relevant period have been subtracted from the current entry. For an example
of how IBC does this, please consult the Notes accompanying
x073. This final reduction merely reflects the fact that IBC has already published
some of these deaths, and therefore cannot now add the 1,519 figure in full
to its database: it does not affect this morgue-derived total except in terms
of its possible distribution among existing IBC entries.
The detailed reports show that the death totals have been rising month by month.
Future data entries for September 2003 and beyond should provide a better
overview of these trends. The latest reports suggest that having doubled
between May and August, the death rate is still rising.
A discussion of some related issues can be found in our 23rd September press
release.