Tuesday, July 7, 2009

Review: Why We Make Mistakes

I read a truly remarkable book this weekend: Why We Make Mistakes, by Joseph T. Hallinan.

Everyone knows that people make mistakes. But, I hadn't realized how truly and grossly fallible we are. The book contains many easy-to-digest examples, like how doctors' sloppy handwriting kills thousands of people a year (because it leads to mistakes in dispensing drugs).

Here's a good summary quotation, from the conclusion:
There is an emerging consensus among some psychologists human decision making operates on two levels -- one more rational, one more visceral -- and that these two constantly trade off. [...] Many of our mistakes appear to happen when we are operating on one condition but think we are operating in the other. We may think, for instance, that our decision to take out a loan was dictated by financial considerations alone, only to learn that we were influenced by that picture of a pretty woman. [...]

If this emerging view is accurate, it helps explain why some types of error-prone behaviors are so hard to eliminate: we think we're being rational when we're being visceral, and vice versa. When a mistake does happen, we often end up blaming the wrong cause.... We don't learn from experience, because we're not sure which experience to learn from.

To make matters worse, [... often,] when we are right (and let's not forget, we are right a lot of the time), we tend to attribute our rightness to skill in whatever it is we're right about; but when we're wrong -- that we attribute to chance.
Fortunately, the book also offers advice on how you, personally, can reduce your error rate and become more effective at your tasks. Here is a summary:
  • "Think small." Minor changes can have major effects.
  • Calibrate yourself, to reduce overconfidence. Keep track of your successes and your failures, the things you choose to do and the things you chose not to do. Be aware that you are susceptible to hindsight bias -- i.e., you remember being more right that you actually were.
  • "Think negatively." The devil's advocate will help protect you against overconfidence.
  • Have uninvolved third parties check your work. People like you or close to the work may be subject to the same biases that you are and may make the same oversights.
  • Multitasking doesn't work; distraction leads to more errors. Also, multitasking can make each of the multiple tasks go so much slower it would have been faster to simply tackle them sequentially.
  • Use anecdotes with great care: they are extremely persuasive, much more so than systematically aggregated data, but they are not representative, as well-collected and well-analyzed data are. Thus, if anecdotes are not carefully selected (and perhaps even if they are), people will absorb the wrong lessons from your work.
  • "Get some sleep." Insufficient sleep (and a huge fraction of Americans don't get enough sleep), even a modest sleep debt, leads to meaningfully reduced cognitive ability. (See Sleep Thieves, by Stanley Coren.)
  • Be happy. Happy people are more creative and better problem solvers.
  • Financial incentives are ineffective in generating better creative work.
The book reinforced the notion that's been growing for me as I acquire more and more education (22 years of it so far, oof): that however knowledgable or skilled I am, there is far, far, far more about which I am ignorant and incompetent, and it's often unclear where the boundaries lie. Thus, it leads to better work when I recognize and plan for the ignorance and incompetence -- when I work around or change my limits when I know where they are (the known unknowns), and work diligently and open-mindedly to discover the limits I may not yet be aware of (the unknown unknowns).

The book has a website and a blog.

Anyway, I strongly recommend that you go out and read this book, right away. It will make you a better person.

No comments:

Post a Comment