Monday, March 03, 2008
Joe Mazzafro

My original plan for this month's Intelligence Community(IC) color commentary was do a traditional book review on A.J. Rossmiller's recently published STILL BROKEN: A Recruit's Inside Account of Intelligence Failure from Baghdad to the Pentagon. Given Mr. Rosesmiller's brief experience with DIA, before even reading a word my reaction was that the title was presumptive at best and suggested an agenda. Two hundred and twenty six pages later of mostly anecdotal pap he proved me right. It seems DIA's new hire analysts were consistently producing highly useable intelligence about Iraq only to be thwarted by managers who found their reporting to be unusable because it was too pessimistic to be taken seriously by policy makers. The idea that because something is new to him does not ipso facto make it new to the IC appears to be a condition Rossmiller has not contemplated.

Troubling me though was my own experience where I saw the system not wanting to take seriously analysis that did not fit an established community position. The Iraqi Weapons of Mass Destruction (WMD) Special National Intelligence Estimate (SNIE) will come to mind immediately as an example of IC group think gone bad, but I also just read David Halberstam's THE COLDEST WINTER this past January so the Korean War was on my mind. Here the conventional wisdom of General McArthur and his J2 Charles Willoughby, supported by Washington because who would know better than the commander in the field, was the Chinese Army would not be sent to roll back “United Nations” forces from the Chinese border. Reports from tactical units that they had sighted or captured Chinese troops suggesting that the People's Liberation Army (PLA) had crossed the Yalu River before the bugles sounded were dismissed by Willoughby and accepted by McArthur (if he actually saw them) as either wrong or only episodic evidence of scouting units.

Clearly there is a systemic IC issue regarding how well established analytical positions can filter out information or analysis not in synch with in vogue views. I was wrestling with this natural almost generational conflict when the following joke came through my Outlook Inbox and diverted my attention.

A biker is riding by the zoo, when he sees a little girl leaning into the lion's cage. Suddenly, the lion grabs her by the cuff of her jacket and tries to pull her inside to slaughter her, under the eyes of her screaming parents. The biker jumps off his bike, runs to the cage and hits the lion square on the nose with a powerful punch. Whimpering from the pain the lion jumps back letting go of the girl, and the biker brings her to her terrified parents, who thank him endlessly.

A reporter has seen the whole scene, and addressing the biker, says - Sir, this was the most gallant and brave thing I saw a man do in my whole life.

- Why, it was nothing, really, the lion was behind bars. I just saw this little kid in danger, and acted as I felt right.

- Well, I'll make sure this won't go unnoticed. I'm a journalist, you know, and tomorrow's papers will have this on the first page. What motorcycle do you ride?

- A Harley Davidson. The journalist leaves.

The following morning the biker buys the paper to see if it indeed brings news of his actions, and reads, on first page:


Surely there is nothing inaccurate in this news summary, but it's not what happened! After a good laugh, the importance of context finally occurred to me. Fairly early on intelligence officers responsible for producing or reporting intelligence analysis are advised rather sternly about the importance of providing a context so their audience can understand what is being presented and why its worthy of their attention. A.J Rossmiller seems to have missed this lesson. I don't doubt the accuracy of the author's observations, but STILL BROKEN doesn't give me enough context to know if what he reports is what really happened or more importantly if another DIA new hire would have the same experience he did.

This takes me to the soft underbelly of information sharing in the IC ------ the need for context, which I see as a value added commodity done best by people with broad understanding of an issue vice time on the target. In today's push to use Information Technology to enable intelligence sharing, fusion, correlation, and analysis are often used incorrectly as interchangeable terms. Fusion is the connection of two or more data points without a specific reason for doing so. Correlation provides a rational for why two data points should or should not be linked together (e.g. time coincidence, geography, sequencing, etc.). Analysis provides and explanation for what correlated information means (e.g. this piece of SIGINT tied to this radar tract indicates a cruise missile headed towards x x x).

Unlike fusion and correlation, analysis needs a context for it to be useful. Where could a cruise missile in that location have been launched from? Are cruise missile launchers known to be in that region? What would be the motivation for launching a cruise missile on this trajectory? Are there alternative explanations for the events being observed? Why is one explanation of the observed events more plausible than the other explanations? This is why I see the ODNI’s Analytical Space (“A” Space) initiative as so important ------ at a level above fusion and correlation “A” Space will allow analysts to explore their hypothesis with their peers and enable t others from across the IC interested in a particular topic to help a intelligence analysts to find or build the appropriate context for the shared data at hand.

That's what I think; what do you think?