Linky. HN discussion. My summary with the negatives flipped:

If you want to change your behavior effectively,

  • think of a small step,
  • that doesn't require willpower every time,
  • uses the environment for encouragement,
  • involves a new behavior, not stopping an old one,
  • is relatively easy to do,
  • is prompted by an external trigger,
  • (which shouldn't be just learning some new information),
  • is a concrete behavior rather than an abstract goal,
  • and resolve to adopt it for some fixed period of time, not forever.

Then the behavior change will be easy.

New Comment
10 comments, sorted by Click to highlight new comments since:

Where should I look to be convinced there are good reasons for thinking this is good advice?

I began by looking at the slides, and following the link on the first slide to the Stanford Persuasive Tech Lab where the authors work.

What a curious site to be part of a university. Even if it's all true, I found nothing on the site reporting actual research into determining whether it's true. If it weren't for the affiliation with Stanford, I'd think I was looking at the web site of a marketing consultancy. This is not to cast any particular doubt on their claims, just to note the absence of any presentation of evidence for them. Maybe I didn't follow enough links.

The director of the lab and first author of the slides has his own Wiki page, and he presents his theory of human behaviour here.

I also managed to find this subsection of the SPTL, on the subject of how users assess the credibility of online sources, which does report actual research.

Well, that's my five minutes.

Try it yourself. After all, it is 'relatively easy to do'.


No, it will be easier. Not easy.

ETA: Another thing: Besides setting achievable, measurable goals within a strict timeframe and concentrating primarily on the benefits, one can further increase the odds of success by taking into account the most probable failure modes. So: Intersperse goals and the steps to achieve them with likely problems.

Thanks, this is great info-per-word.

I wonder if anyone else got tempted to try it immediately, like I did.

Ancedote coming up:

At some point I realized I was drinking way too much coke, so I resolved to not drink any when I went back home for a few days (small step, concrete, environment, fixed period). Once I had succeeded doing that, it was easier to carry the new behaviour back to everyday life. The first time I succeeded for a period of 8 months or so then lapsed again, bootstrapped again with the same process about 2 years ago, haven't drank coke since.

This seems congruent with most of these points except "involves a new behavior, not stopping an old one,". I generally find it's easier to stop doing something than to start doing something else. Anyone else have the same experience?

every so often I'll decide to stop biting my nails and I can devote lots of mental energy to stop myself whenever I see it starting up again. On a really stressful day though, I can't devote that energy and I wind up chewing them off again. Usually I stay on this wagon for a few weeks before I can re-dedicate myself to the non-nail biting mental effort. On the whole though, stop biting my nails is not that all that difficult, the problem is to be consistent about it.

It's difficult to start doing things when the path of least resistance still takes a lot of mental energy. Checking lesswrong is easy, reading science papers for class is hard. Having a goal (not failing class the next day) is a big help though.

There are bitter nail polishes to help people stop thumb-sucking or nail-biting. Have you tried that?

think of a small step,

So the interesting conceptual challenge is how to reformulate large, abstract goals in terms of small, concrete goals.

One problem with group efforts on this subject is that people's abstract goals vary enormously, so Alice may not be able to profit very much by imitating Bob's goal-decomposition strategy.