« previous | MUCK HOME | next »
Petraeus' Methodology is Better than... What, Exactly?
Gen. Petraeus and Amb. Crocker are before the Senate Foreign Relations Committee this morning, and are still on opening statements. As they reiterated their statements, I noticed something in Petraeus' description of the credibility of his methodology for sectarian attacks and civilian casualties. Here's how Petraeus describd his "rigorous, consistent data collection and analysis:
Two US intelligence agencies recently reviewed our methodology, and they concluded that the data we produce is the most accurate and authoritative in Iraq.
As Josh noted yesterday, it would be nice to know which agencies these are, as CIA and DIA reportedly have qualms about MNF-I's methods. But looking at that statement closely, it may be possible to square the methodological circle. Notice that Petraeus didn't say that those agencies blessed MNF-I's methods as the "most accurate and authoritative," full-stop. He said that they found it's the most accurate and authoritative in Iraq. The alternative collection and analysis on Iraq is conducted by different agencies of the Iraqi government, which release sometimes-conflicting data. And needless to say, the Iraqi government has a huge incentive to downplay both civilian and especially sectarian casualties.
It would be hard for the professionals at MNF-I to have a worse methodology than the Iraqis. But that doesn't mean MNF-I has a better method of tallying both figures than other elements of the U.S. government. Perhaps the Senators today will get some clarification from Petraeus.
Comments (15)
zennurse wrote on September 11, 2007 10:48 AM:Petraeus just tried to clarify the difference between his data and the GAO data by saying their cut off was 5 weeks earlier. I am pretty sure the GAO report included a trip to Iraq as recently as 8/30 and that data had been provided to the GAO as recently as a day or two before the hearing.
Is this nonsense?
Thanks for all your work, Spencer, very helpful
OCPatriot wrote on September 11, 2007 11:02 AM:Doesn't anyone find it disgusting that Petaeus won't tell what his methodology is and who the agencies were that vetted these numbers? And, worse yet, that he doesn't seem to be bothered by other groups which produced different, more damning numbers, and were open about their source? Also doesn't anyone care about the fact that the Iraq police are dysfunctional by EVERYONE'S standards and can't be fixed without disbanding them and starting over?
chabuka wrote on September 11, 2007 11:13 AM:Its time to impeach Bush...he received a four-star B.J. from Gen. Betray-Us
chabuka wrote on September 11, 2007 11:13 AM:Its time to impeach Bush...he received a four-star B.J. from Gen. Betray-Us
chabuka wrote on September 11, 2007 11:14 AM:Its time to impeach Bush...he received a four-star B.J. from Gen. Betray-Us
EthanS wrote on September 11, 2007 11:18 AM:How can Petraeus have monthly counts of Civilian Dead that are LESS than the IBC number? IBC is an absolute LOWER bound - it relies on published, confirmed, media reports!
The real number must be higher; probably by an order of magnitude.
Related: What happened with the AP monthly numbers in November 2006? The monthly average more than doubles, even though IBC's numbers don't change much. From Jan 06 to Oct 06, IBC numbers are 250% - 350% higher than AP. But starting in November, IBC numbers are only 25%-50% higher. AP must have changed their reporting system.
And again, any value less than IBC is simply NOT credible.
Michael Lafferty wrote on September 11, 2007 11:18 AM:NOTE: I am not claiming that IBC is a true estimate of Iraqi civilian dead. I think the Lancet studies are the most accurate estimate we have. My points are only that (1) IBC can point to EVERY death in their count and it's verified. It is a hard floor - a starting point for any estimate. (2) AP's 2006 numbers are NOT consistent before and after November 2006. Something changed.
Come on: figure it out. What he is really saying is that the command has US Army and Marine Corps military intelligence personnel at two different staff levels collecting, reviewing and reporting this information.
Don't believe for a moment that there is any 'independent' intelligence collection or analysis outside the theater by non-Department of Defense personnel. He cannot control the information outside of his command, so he has elected to keep it all in house where he can exert complete control. This is also a convenient excuse for the application of classification standards: as the commander, as he is the final classification authority and it's information collected in a war zone, blah, blah, blah…
This is all such a crock. And, I say this as a former military intelligence officer in both the active and reserve components of the US Army.
joejoejoe wrote on September 11, 2007 11:21 AM:Gen. Petraeus is probably talking about the Army Intelligence agency and the Marine Intelligence agency. They are listed on the list of the 'Intelligence Community' on intelligence.gov.
http://www.intelligence.gov/1-members.shtml
EthanS wrote on September 11, 2007 11:22 AM:M. Lafferty, that's what I was thinking.
The "18 agencies" include the 'independent' Army Intelligence and Marine Intelligence agencies, which are NOT part of DIA. Those two probably approved the methodology.
==
zennurse,
some dude named steevo wrote on September 11, 2007 11:35 AM:Of course, GAO ASKED for the August data for their report. DoD refuse to provide it! That's chutzpah!
Having served in the military, I wouldn't be surprised if they had just made up these numbers out of thin air. Generals are used to seeing the world through a filter and rarely see things objectively. Worse, they aren't even aware that they do it...
jeffgee wrote on September 11, 2007 11:44 AM:Wouldn't recent levels of Iraq violence be lower in August because it's 115 degrees in the shade and there are fewer people out and about because of it?
TheraP wrote on September 11, 2007 11:54 AM:two thoughts:
1) he says the "data we produce" - now data usually refers to the "raw data." And statistics is not usually referred to as "data." So someone reviewed methodology but only commented on "data."
2) let's consider the statistics of the number of intelligence agencies and the percentage that agrees with their "data." Ummm... it's a pretty tiny percentage. So what's the overall confidence level here? Ummmm.... not too good!
Keith wrote on September 11, 2007 11:55 AM:If the data used by the GAO and Petraeus is the same, then they should track almost exactly (point for point) up until the GAO cut-off. In other words, the GAO data should back-up/confirm his data. I wonder if a Senator will follow-up by comparing his charts with GAO's.
jak1 wrote on September 11, 2007 1:48 PM:>>Two US intelligence agencies recently reviewed our methodology, and they concluded that the data we produce is the most accurate and authoritative in Iraq.<<
This reminds me of the movie Revenge of the Nerds!
You know the part where the Nerds go before the board. And the board assures them they will get a fair hearing by the board, to hear their complaint against the Alpha Omegas.
To bad for them the board was made up of the Alphas.
JEP wrote on September 11, 2007 2:38 PM:War profits galore!
Just how do they divvy op the profits from $12 Billion a week?
Bush's exit strategy?
...January 20, 2009.
That is the only exit he's watching right now.
ANYONE ELSE NOTICE THE R'S ARE ALL CO-OPTING JULY, 2008 AS IF THAT REPRESENTS "BRINGING THEM HOME?"
If the entire surge strategy hinged on July 2008 since the inception of the surge, what else can that date represent but a political timeline, not a military one.
Petreaus mentioned it more than once, One of those times, he actually backpedalled to diminish the emphasis he had already on that date as the "end of the surge", he's also tossing them a bone in December, another in March and then bringing us back to where we were before the surge.