Thursday, 9 June 2005

Making decisions: rules of thumb






I'm fascinated by rules of thumb (the fancy name for which, I've now learned, is "cognitive heuristics") based on maths analyses which actually work in making real life decisions.

An article by Robert Matthews in the Financial Times in late April (yes, I'm very behind catching up on stuff!) says that finding rules humans can use is in fact an active area of research and mentions ongoing work by the Adaptive Behaviour and Cognition Group (the ABC Group) at Berlin's Max Planck Institute for Human Development looking particularly at "fast and frugal heuristics" - rules you can apply quickly and effectively, yet which are frugal in terms of how much data you need. Some of the rules have been found to be as effective as complicated sophisticated maths techniques.

Here are the rules of thumb mentioned in the article:
  • "Take the best" (TTB) - to decide quickly between two options, look for a characteristic both of them have which you think correlates best with what you're trying to assess, then use that to make your decision. An example given was if you have to guess which city is bigger, you could figure that the the more famous one is likely to be bigger (i.e. decide that fame correlates with size) and then pick the one which you've heard of, or which has the more well known football team. And if that doesn't work, try another feature which you think should correlate with the "target property" (in this case size) and use that. This rule even works better than mathematical models, in situations where you don't know the options well and don't have much information to go on (e.g. the research included people choosing stocks based on mere familiarity - picking companies they'd heard of - beating the market, while economics experts using more "sophisticated" methods actually lost money!). Apparently considering all the options carefully may not be as effective as you'd think, because the payoff may not be worth the extra effort and you could be misled by meaningless detail into actually making the worse decision. (Another aspect of the "familiarity" rule seemingly used by animals is "Do what you see most others do" when searching for food. I'd venture to suggest that rule is applied, consciously or not, by most people too, and not just in relation to food!) We will probably be most efficient in our decision making when we're pushed for time because that's when we're forced to use the TTB rule, so the author suggests, when trying to choose food in a restaurant, just to say "We're ready to order" and make yourself pick something fast, to end up with a good choice!
  • Don't accept the first promising choice - that's just Colley's Rule, I've covered it in a previous post (which I've just updated to add a link to another interesting item). This article describes it another way, as using the first good option to set the baseline of quality, then take the first one after that which beats it.
  • In weighing up likelihood and payoffs, focus on what if it goes wrong - should you do something if you don't know how likely it is to succeed? Look at what might happen if you do it and it goes wrong, and if the consequences are no worse than the consequences of not doing it, then do it. The example given is, should you take a job offer if you don't know how well it will work out? If, even supposing the job goes all wrong, you'd be no worse off than if you rejected the job, then the rule says take the job.
  • And a rule of thumb about rules of thumb! Rules of thumb themselves just make it more likely you'll make the right choice, but don't guarantee it - so be prepared to get some wrong in return for, as the author puts it, "being right more often than not".

Technorati Tags: , , , , , , , , , , ,

4 comments:

SatishTalim said...

Since you are talking to Technorati, I thought I would comment about a problem I am facing. I am now using Technorati tags to create Categories for my posts (Blogger does not give this facility). Two days back, I created a tag for a category called ABJ2EEArticles. Till today, this tag does not carry my post (I have checked and re-checked the code, no problems there). I created another tag for a category called ABGeneral, sometime back and lo it got created. Is there a problem if I have a digit in the Tag?

Improbulus said...

Hi Satishtalim, I think you probably meant to comment against my pt 2 post?

From other comments I've had on other posts, the problems with Technorati appear to be escalating and I think now merit a separate post, which I am going to put up in a day or two (have been very busy!), so please check back in a couple of days.

Thomas said...

Robert Matthews directly (http://www.robertmatthews.org/)has written again an article in todays (01.06.2005) Financial times with the title " The agonies of too much choice". He states three rules:
1. Take the best
2. Collee's rule
3. You can't win them all (whatever that means).
So he is now down to three rules ...
Thomas

Improbulus said...

Yes, that's interesting isn't it? Though the last "rule" I mentioned isn't so much a rule as a warning not to take it all too slavishly, I think.