Chapomatic

Photon Courier took the ball (a short post here) and ran with it, with his curiosity piqued about the Gilovich study.

There was no substantive difference between the scenarios given to the two groups (unless you believe that the mode of refugee transportation is somehow substantive)…yet those who heard the WWII-reminiscent version were more likely to conclude that aggression must be met with force, while those who heard the Vietnam-reminiscent version were more likely to recommend a hands-off policy. Again, there was nothing in the scenarios to make anyone conclude as a matter of logic that the first version was more similar to WWII and the second version was more similar to Vietnam.

I agree–this is potentially a powerful model in analyzing several things, including information flow in an information warfare environment, how the pre-Iraqi invasion’s public debate self-assembled, and other very interesting things. Here’s a little more raw data related to the article.

The term anchoring effect seems to be a term of art already. A quick Google shows lots of applications, including an alternate term hindsight bias. I can see this as a tool in describing not just strategic decisions but also tactical decisions–if you’ve fixated on the one target out there you might just miss the other one in your blind spot. Confirmation bias is similar, and both reinforce each other.

The original authors of the HBR article, Giovanni Gavetti and Jan Rivkin, have written something similar back in 2002/2003 aimed at how to teach thinking by analogy (that at the moment I’m too cheap to purchase). The article in question has a summary that I find a bit overcompressed and unhelpful except as an abstract.

Dr. Rivkin gave an interview to Computerworld magazine that is a “read the whole thing” good summary.

Why is analogical reasoning so useful in a field like IT? Analogies are most powerful in settings where there’s not enough clarity to use deductive reasoning nor so much ambiguity that you have to go for trial and error. Many pockets of IT have this middle ground that’s familiar enough to make links to more familiar settings but not clear enough to identify cause and effect. In that middle ground, analogies may be the only options we’ve got.

Give me an example. Intel for many years resisted entering the low end of the market. Then [Harvard Business School professor] Clayton Christensen introduced them to an example in the steel industry. U.S. Steel had let minimills take over the low end with cheap concrete reinforcing bars called rebars. He pointed out that this was the beginning of the troubles for the U.S. steel business. Once the minimills got a beachhead at the low end, they moved up. At Intel, this really struck a chord. Andy Grove feared if they ceded the low end of the market, the high end might follow. He even began to refer to low-end PCs as “digital rebar,” and soon thereafter Intel introduced the Celeron processor to fight it out on the low end and prevent other companies from getting a beachhead.

In this case, the analogy wasn’t about learning from someone’s success but trying to prevent a repeat of someone’s failure. It was about what they thought U.S. Steel should have done.

Tell me about some of the drawbacks to analogical thinking. The core pitfall is choosing a source based on superficial similarities to the target. When Ford was looking at redesigning its supply chain, it turned for guidance to Dell’s key principle of virtual integration. There is good reason to look at Dell. Some aspects of what it does look like what Ford does. They both take fairly standardized components and assemble them into a vast variety of models.

But other things are quite different. A large portion of Dell’s cost advantage comes from the fact that virtual integration enables it to buy inputs late. A PC that arrives from Dell has a microprocessor bought later than the microprocessor bought for another supplier. In a setting where the price of microprocessors declines dramatically over a short period, that difference translates into a large cost advantage for Dell. But prices in the auto process are not coming down so rapidly, so the power of virtual integration and less inventory is not nearly as great. The good news is that Ford didn’t fall into that trap.

Another potential problem is the anchoring effect. Can you explain? People get attached intellectually and emotionally to their analogies, and it’s very hard to shake. If you look at Sun, Scott McNealy often uses analogies drawn from the auto business. He argues that buyers should be interested in the whole package, not the components, because when they buy a car, they care about the whole car, not where the carburetor comes from. But you have to question how dispassionately he can assess that analogy. His father worked for years in the auto business, and his sons are named for auto models: Maverick, Scout, Colt and Dakota.

Also in the “interesting but tangential” department is this chat with Gilovich discussing basketball and something called the “hot hand”.

This comment may be connected to the Gilovich paper cited in the HBR article:

From: Leigh Thompson

Date: 04 May 2000

Time: 09:33 AM

Comments

I asked Tom to comment on his article and this is what he said:

“I’m delighted to learn that the first article I ever published is still “in circulation.” The story of the origin of that research is easy to tell. Bob Abelson was a visitor at Stanford during my first year of grad school, right after his “Scripts, Plans, Goals, and Understanding” book with Roger Schank had come out and interest in scripts, schemas, etc. (what Abelson referred to as “things that go bump in the mind”) was at its peak. Anything that emphasizes the role of top-down processing in the top-down/bottom-up mix highlights the possibility of the MISAPPLICATION of various mental representations in the effort to perceive the world. Although that idea was often talked about as an obvious bi-product of schematic processing, there was to my knowledge no truly compelling demonstration of it and so I sought to generate one. I was immediately drawn to the Vietnam/Munich context because this pair of “dueling metaphors” was inevitably employed by politicians whenever the prospect of US military involvement in foreign lands was discussed. The sports studies were conducted because it was the domain in which I could most readily see how to make the priming (although we didn’t call it that back then) of the different schemas based on something that was irrefutably irrelevant to the judgment at hand.

I have not read that article in quite a number of years, and I’m afraid to discover what kind of writer I was then. At any rate, I hope you enjoy it.”

This quote, if it refers to the article cited, is exactly what came to mind to me when I read about the study.