= 27d567bc2d48d9f40e9ccc20ea370b15)
mapnoterritory comments on The Power of Reinforcement - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (467)
Daniel Kahneman in Thinking, Fast and Slow:
There reason for that lies in regression to the mean when training (example of flight instructors in the israel airforce):
Since positive reinforcement is so counterintuitive: don't forget to reward yourself for rewarding somebody for good behaviour! :)
Speaking of regression to the mean, that seems to be one topic that wasn't really covered in the sequences that really should have been.
So you (or at least Kahneman) implicitly admit that punishment is effective at changing behavior.
Yes, I think so and apparently so does Kahneman. I don't think this is particularly controversial. Kahneman does say that positive reinforcement is more efficient (both in animals and humans).
Everyone who's looked at the data thinks that punishment can change behavior. The question is whether punishment makes the changes you want- and people dramatically overestimate the usefulness of punishment and dramatically underestimate the usefulness of positive reinforcement.
Also it depends on the definition of what you "want" -- for example if you punish someone for bad behavior, what exactly is your goal?
All three goals are pleasant, though only the first one is officially desirable. The punishment works in all directions. Perhaps this is the reason why behavior change by punishment is popular more than it deserves; and why people rationalize its usefulness even when the first goal visibly fails.
Agreed. Hopefully, instructors care most about the first- but in general human interaction, the others can easily rise to prominence.
Depends, the current "everyone is special, everyone deserves an A for trying" culture almost certainly overvalues positive reinforcement.
I see a difference between 'niceness' and 'positive reinforcement'. The "everyone deserves an A for trying" approach is 'nice' but it generally isn't skillful positive reinforcement; I think a major problem with it is underestimating how much it rewards behaviors that look like trying but aren't trying.
There's also a basic value question- if you're trying to build self-esteem, it's not clear that an "A for trying" approach overvalues positive reinforcement, though if you're trying to build understanding, it clearly would be a misapplication of positive reinforcement.
Everyone getting an A isn't reinforcement. Reinforcement has to be conditional on something. If you give everyone who writes a long paper an A, that's reinforcing writing long papers. If you give everyone who writes a well-written paper an A, that's reinforcing well-written papers (and probably more what you want to do).
But if you just give everyone an A, that may be positive, but it simply isn't reinforcement.