Talk:Intelligence analysis

From Wikipedia, the free encyclopedia

MILHIST This article is within the scope of the Military history WikiProject. If you would like to participate, please visit the project page, where you can join the project and see lists of open tasks and regional and topical task forces. To use this banner, please see the full instructions.
Start This article has been rated as Start-Class on the quality scale.

This entry is a bit wordy to me. Does anyone agree? Specifically, I want to shorten up the lead in, and make some fairly major edits and cuts to the rest of the article. It just reads a bit like an analysis manual than a encyclopedia article to me. I'll wait a while for comments before editing.Chance.williams 21:02, 11 November 2007 (UTC)

What do you see as problem areas? One of my reasons for going into depth, and putting this as part of a series of articles, is that most intelligence articles were more like spy fiction than an encyclopedic discussion of a complex subject.
It may be appropriate to split off certain parts into separate articles, such as the discussion of cognitive traps in analysis. I would hesitate, however, before completely removing material.
Are there other articles on intelligence that you find to be a better example? I started writing these due to what I saw a lack of coherence to the overall pieces of intelligence. Jumping directly into espionage, for example, with no real criteria for when it is appropriate versus other collection techniques, was one concern. There is also a very real and current political debate about the best way both to provide intelligence to policymakers, and be sure it is both read and not cherry-picked to prove a position already established in the absence of information. Howard C. Berkowitz 21:30, 11 November 2007 (UTC)

Please don't take my suggestions as a criticism, I don't claim to be an expert editor or anything. However, I do recommend the following:

  1. Spread the info in section 1 Intelligence_analysis#Be_Bold_and_Honest or at least re word it. It sort of reads like a pep talk for analysts. Not a bad thing, but not encyclopedic either.
I don't completely disagree. Let me share my thinking. This is, after all, Wikipedia, and it struck me how Frank Watanabe, who wrote a fair number of the suggestions to analysts, sounded almost exactly like the editor's introduction to Wikipedia. I like the compare-and-contrast means of teaching, and it struck me that inserting Wikipedia references into Watanabe's article in Studies in Intelligence might be a valid way to go.
Do look up in the hierarchy at the "Intelligence Analysis Management" article; there are a few things that were a tossup between going in this article and the management article. There's also a section in "Intelligence Collection Management" dealing with initial ratings. When I first encountered ratings, back with the Intelligence Branch of the Mongol Horde, that was considered analysis, but it seems to be more of a consensus that it is the end stage of collection. I mention that here because some people may consider rating to be the start of analysis.
  1. Shorten the lead to two to four paragraphs.
  2. Make short paragraph for, and then separate section 2 Intelligence_analysis#Getting_Started into a different article. This section reads like a "how-to" or process description, which is fine, but for this article I think we should primarily answer "what" intel analysis is.
Hadn't thought about a "how", but I have no major problem with spawning subordinate articles. There are, in fact, some areas of analysis where I would have liked to go farther, as in biographical--which also mixes in with the HUMINT, HUMINT Clandestine Sources, and Counterintelligence articles.
  1. I would remove section 3.2 Intelligence_analysis#Methods_of_Analysis from this article and combine it with the proposed "how" article.
  2. The intelligence cycle is barely mentioned. We should add in info on how analysis fits in, since this is vital to the analysis process.
Have you looked at the overall structure? The cycle is two levels higher in the hierarchy.
Intelligence cycle management (tasking, budgeting, prioritization retained here)
Intelligence collection management
HUMINT (and subordinate things)
MASINT (6 main subarticles)
SIGINT (3 subarticles and more to go)
Intelligence analysis management
Intelligence analysis
Intelligence dissemination management
Intelligence cycle security (subarticles in progress)
Counterintelligence


  1. Add a short "history of" section, and if it gets too long make a separate article if it seems necessary. The info is pretty much already there, but I think a section on this would be a little better.
  2. If time and space permit, I'd like to add an honorable mention of paradox of warning and how it relates to analysis.
Good point. While I didn't use the specific term, the concept appears in several places. Look at "warning" in the "Intelligence dissemination management" article. I made some references to the best exposition of the idea I've seen, which is in Luttwak's Coup d'Etat. It's interesting to see how the concept is evolving, especially with some of the semantic web/ontology, and even taking it back to information theory with Shannon and Wiener.
  1. Section 3.3 can be slightly shortened, and then wikified.
I'll be interested to see what you are thinking. By wikified, I can see internal links, but if you are thinking of other articles now existing, I'm concerned that there are too many intelligence stubs floating around. Now, if you have some thoughts about hooking this into cognitive psychology and the like, that could be very interesting. Howard C. Berkowitz 03:22, 12 November 2007 (UTC)

I'm going to start making some changes. Feel free to revert if you think I've gone too far.Chance.williams 02:09, 12 November 2007 (UTC)

Contents

[edit] Possible approaches

First, I'd encourage you to look at the structure of intelligence related articles that is on my userpage and elsewhere. At the high level, the intelligence cycle is at the very top, with some efforts to unify US and UK doctrine. (makes note to self to see if I can find any fUSSR/Soviet intelligence doctrine in open literature, more recent than IRONBARK).

Note that some of the lower level things, such as clandestine HUMINT, dealt with things I found were cross-functional, as it covers concepts that are viewed differently by [positive] HUMINT and counterintelligence. When I looked back at HUMINT, I realized that an article on Special Reconnaissance, possibly picking up organic and long-range surveillance, was needed to do more rounding out of HUMINT. HUMINT still needs something for stability operation.

You mentioned a how-to that might be subordinate to intelligence analysis. Be sure you've read intelligence analysis management, at least the source rating part of intelligence collection management, and perhaps at least the indications & warning part of intelligence dissemination management. At that point, I'd suggest discussing an outline for how-to. One of my methods is to do a version onto a sandbox userpage, and then replace articles when the new article is in a solid first draft. Howard C. Berkowitz (talk) 00:37, 17 November 2007 (UTC)

[edit] Specialized analysis disciplines

Right now, intelligence analysis deals with cognitive and general approaches. I'm thining of putting more specialized analytical disciplines, such as financial, order of battle, economic (volunteer needed), medical, etc. Thoughts???? Howard C. Berkowitz (talk) 06:48, 27 November 2007 (UTC)

What you are doing is brilliant. The other analytical disciplines would be interesting. Your articles on Counterintelligence and HUMINT tradecraft are also good. Though, I can foresee a wikipedia editor may try to delete the good work you are doing because it is coming close to [being a how-to guide] to the subject and also [quite lengthy]. The part on "Getting Started" sounds like a how-to. Not that I care about it being a how-to, or the length, I think your articles are great. I'm just giving you a heads up that an Wiki-editor may blowtorch some of the articles. Hey, if you really want to go beyond the wikipedia side of things and do an indepth how-to on the subject head over to the Conflict Wiki which covers this sort of thing. I think the guy who runs that wiki would appreciate your input. unsigned comment added by Evud (talk • contribs) 11:06, 27 November 2007
I may just have committed a sin, but SineBot is far too verbose.
Let me digress philosophically a bit. In the broad subject here, for which "Conflict Wiki" may indeed be a good term, there is a great deal of misinformation, and, in some cases, what I can only call comic book input. I shake my head when, on the one hand, I'm working on a [sourced] taxonomy of HUMINT, but then wander over to a page full of fictional double agents or counterterror operations, with moles called double agents and vice versa.
Where I'm going with my concern is that without a strong foundation, which may indeed involve things on the "how to" level, I don't know how to break the vicious circle: we are speaking about a discipline here, not vignettes.
I've already had a couple of blowtorches, which, thankfully, were restored by others, both in intelligence and -- much more so -- in computer networks. When someone tells me that I cannot delete previous material merely because it is wrong, or some tertiary textbook cites it, but what I'm putting is primary source material that I had a hand in developing, I shake my head.
It is my understanding that some CIA safehouse personnel weren't sure if they were going to have to get an ambulance after a KGB defector watched a video of a James Bond movie -- he was laughing so hard that it was interfering with his breathing. Believe me, I'm cutting a lot of introductory material to the bone, from the perspective of understanding what actually is involved. It's interesting that there is actually a good deal of open literature on many of these topics, but I've very rarely seen either classic texts, or things like Studies in Intelligence, or even Military Thought brought to us courtesy of Penkovsky's IRONBARK material, cited.
I suppose I'm worried about where this might go. Do you think that an intelligence (and possibly special operations) project might be a way of managing common expectations and blowtorch application? Howard C. Berkowitz (talk) 15:56, 27 November 2007 (UTC)


[edit] Issues with this article

The definiton given at the beginning of the article gives me some problems. Why is "producing formal descriptions" included into the definiton of intelligence analysis? Producing a description of anything, regardless of how formal, does not imply that analysis was done when producing that description. Analysis implies reasearching and delving into a subject in order to find out more about the topic. Also, intelligence analysis is not done at just the strategic level. I'm sure that intelligence analysis can happen at any level of an intelligence agency. This defintion just seems to be weak and not covering the whole topic.

Within the article, there are examples of poor writing and writing that violates Wikipedia's guidance to article writing. In Wikipedia's words to avoid, [[1]], the word "obviously" is mentioned as a word that doesn't exactly fit within an encyclopedia.

At the bottom of the second paragraph, the article says that an intelligence agencies' opponents may be a monster and of demonic cunning. When I read that I was abhored that it had found its way into Wikipedia. It is poor writing in general, and definately should not be in Wikipedia.

S. Randall (talk) 23:22, 28 November 2007 (UTC)

[edit] Still too verbose

Okay, I understand that maybe my deletions were considered too extreme, fair enough, I bit off more than I could chew as a new editor. However, 646 words and 6 paragraphs before we even get to the table of contents? I just believe that's way too much for this article. We've got an Einstein quote that is really irrelevant, we've got weasel words, and a lot of information that properly belongs in the body of the article. The wikipedia article on layout says:

"The lead is shown above the table of contents (for pages with more than three headings). The appropriate lead length depends on the length of the article, but should be no longer than four paragraphs."

Unlike my last edits I'm not going to change anything until it's discussed, but I'm a firm believer that simpler usually = better.Chance.williams (talk) 20:15, 11 December 2007 (UTC)

I could see moving the second, third, and sixth paragraphs before the TOC. I think the material about incremental information contributing to a failure is important to a lead, and at least a reference to cognitive traps. Whether or not it involves an Einstein quote, I also believe that it is essential to get across that intelligence analysis takes place in a context where some paranoia is needed: this isn't a subtle natural process, but a scenario in which very smart people are trying to hide what they are doing.
That additional material, with an appropriate transition, could go into cognitive traps. As far as weasel words, the reality is that intelligence analyses almost always have to contain some level of uncertainty. It's rarely as clear as having SIGINT cue you to send a U-2 over Cuba and see missiles being emplaced. Indeed, Pat Lang, who was then the National Intelligence Officer for Warning, was figuratively running about DC three days before the Iraqis invaded Kuwait, with information almost as slam-dunk that Iraqi units were being prepared for what (weasel) appeared to be the real thing, not a realistic drill.
It may be that another subarticle is appropriate, but the lack of general knowledge about analytic tradecraft have greatly confused US, and indeed world, reaction to the reports about Iranian weapons programs. Real-world intelligence is not straightforward. I could write some equivalent articles about medical diagnosis in the presence of uncertainty, and probably have appropriate weasel words. Radiologists and pathologists love weaseling; I wouldn't be surprised to see a radiologist report on an arm that had been run over by a bulldozer as "appearing consistent with multiple fractures".
My suggestion would be that you propose alternate text, and we (and others) can discuss that text to see if it suggests more certainty than is implicit to the intelligence process. Howard C. Berkowitz (talk) 21:21, 11 December 2007 (UTC)

[edit] Analysis of Competing Hypotheses

Should the existing article, which has the same source, be moved into the section here, with a redirect at the existing page? The older article does have a list of items not here, but, otherwise, they cover the same material from Dick Heuer's work. Howard C. Berkowitz (talk) 18:46, 18 January 2008 (UTC)