Peacewise comments on We Change Our Minds Less Often Than We Think - Less Wrong

39 Post author: Eliezer_Yudkowsky 03 October 2007 06:14PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Article Navigation

Comments (115)

Comments (115)

Sort By: Old

You are viewing a single comment's thread. Show more comments above.

Comment author: gwern 06 January 2012 09:11:51PM * 2 points [-]

You attacked my character as a means of dismissing my discussion, it's a low tactic and one I didn't expect from LW, it is indeed the typical internet trash talk I mentioned....With regards to the intention to be immature, I have no knowledge of your intentions - only the observation that attacking my character without addressing the substance of my argument is an immature act.

If you still think that...

I maintain that the police officer is a better judge of the event than either the driver or you or I, hence it is more likely that it was in fact reckless driving than it wasn't reckless driving.

You wish to defer to the cop's expertise on whether it breaks the law? Excellent! I wish to defer to teens' expertise on what they enjoy. I'm glad we could come to agreement that teens overestimate risk but enjoy risky behavior much more than older people.

It is quite clear that in perceiving their rewards with such a high (subjective/personal) value, many of them have indeed made an error for one third teenager deaths are in car accidents. One should consider if death, both the risk of it and the actual occurrence of it is truly a fair price to pay for driving fast (or under the influence of alcohol).

'fairness' does not enter into it. As a transhumanist, I do not think death is a fair price for much of anything.

That aside, you repeat your 1/3 number as if it means anything in the absence of other information, as explained already. It does not.

We might consider a pre-drive of the route as an expense which made the reward vs investment unfavourable, this would reveal that the perceived reward is not so high as to overcome some (amount) of minutes of the teenagers time. Something to consider, I'd appreciate your input on that line of reasoning.

This is so far your only point worth a damn. I suggest you continue this line of reasoning, sans the fucking anecdotes.

Comment author: Peacewise comment score below threshold [+] (50 children)

Comment author: Peacewise 07 January 2012 03:06:46AM * -4 points [-]

Gwern wrote "You wish to defer to the cop's expertise on whether it breaks the law? Excellent! I wish to defer to teens' expertise on what they enjoy. I'm glad we could come to agreement that teens overestimate risk but enjoy risky behavior much more than older people."

I've been comfortable all along with accepting what teens find enjoyable, I do not agree that teens overestimate risk and frankly I'm surprised you could glean that from what I've written. Let me be clear, the article reveals that teenagers underestimate risk.

Dobbs wrote - "It was the brain scans she took while people took the test. Compared with adults, teens tended to make less use of brain regions that monitor performance, spot errors, plan, and stay focused—areas the adults seemed to bring online automatically. This let the adults use a variety of brain resources and better resist temptation, while the teens used those areas less often and more readily gave in to the impulse to look at the flickering light—just as they're more likely to look away from the road to read a text message."

Estimating risk is about planning, monitoring performance, spotting errors and staying focused, whilst the final comment quoted above provides a suitable anecdote highlighting another situation where teens are more likely to underestimate risk. Further teens more readily give in to impulse - that reveals that teens readily don't estimate risk (at all), do you see that? Estimating risk is in opposition to giving into impulse, an impulse is a sudden urge - estimation isn't something that's done suddenly, estimation is calculated by examining the context.

Gwern continues with "'fairness' does not enter into it. As a transhumanist, I do not think death is a fair price for much of anything."

I used the term "fair" in the context of the gain outweighs the cost, it's a colloquialism you obviously understand since you use it yourself, hence your comment "fairness' does not enter into it" is false. However, you're quite right that death isn't a fair price for much of anything, providing support for the risk of death not being much of a fair price for anything either... again you just keep destroying your own argument.

So on one hand we've got teens who are shown in your quoted article to anecdotally underestimate risk and also shown in research to utilise less brain processes that estimate risk and on the other hand its been shown in other research presented in the same article that teens place a higher value on rewards - all of that leads to the inescapable conclusion that some teens are overconfident in their decision making ability.

It's become apparent to me that this discussion is an example of the disconfirmation bias. Perhaps you'd care to follow the procedure for minimizing/removing disconfirmation bias before you make another post, I have already done so several times.

Comment author: gwern 09 January 2012 09:16:17PM 4 points [-]

Estimating risk is about planning, monitoring performance, spotting errors and staying focused, whilst the final comment quoted above provides a suitable anecdote highlighting another situation where teens are more likely to underestimate risk. Further teens more readily give in to impulse - that reveals that teens readily don't estimate risk (at all), do you see that? Estimating risk is in opposition to giving into impulse, an impulse is a sudden urge - estimation isn't something that's done suddenly, estimation is calculated by examining the context.

This is just missing the point. The mind is what the brain does. If a teenager chooses the higher reward and glances away, then by definition the areas involved in inhibition etc aren't going to be as busy! If they were equally busy in both the glancers and non-glancers, no one would be discussing them in the first place!

However, you're quite right that death isn't a fair price for much of anything, providing support for the risk of death not being much of a fair price for anything either... again you just keep destroying your own argument.

Unfortunately, we pay with death for both action and inaction. Destroying my own argument indeed. If you seriously mean that, then you must mean that no risk should ever be taken, which is not a position many will sympathize with.

So on one hand we've got teens who are shown in your quoted article to anecdotally underestimate risk and also shown in research to utilise less brain processes that estimate risk and on the other hand its been shown in other research presented in the same article that teens place a higher value on rewards - all of that leads to the inescapable conclusion that some teens are overconfident in their decision making ability.

Nothing inescapable about it. What we have is a worthless anecdote you insist on supporting your position, extremely strong evidence against underestimation of risk, brain-imaging results you do not understand, none of which forces the conclusion of overconfidence as opposed to teens intrinsically having higher rewards just as they intrinsically oversleep and all the over changes that go with puberty and being young adults.

It's become apparent to me that this discussion is an example of the disconfirmation bias.

Or, as it is more commonly known, the confirmation bias.

Comment author: Peacewise comment score below threshold [+] (31 children)

Comment author: Peacewise 10 January 2012 08:53:28AM * -4 points [-]

Thanks gwern for returning to the discussion, cheers.

This is just missing the point. The mind is what the brain does. If a teenager chooses the higher reward and glances away, then by definition the areas involved in inhibition etc aren't going to be as busy! If they were equally busy in both the glancers and non-glancers, no one would be discussing them in the first place!

Perhaps I am missing the point. I reason that one point is that if the areas involved in inhibition aren't busy, then the person isn't estimating risk, in context, they are instead being overconfident in their decision making, they've not judged the situation, instead acting in impulse.

Unfortunately, we pay with death for both action and inaction. Destroying my own argument indeed. If you seriously mean that, then you must mean that no risk should ever be taken, which is not a position many will sympathize with.

Well no actually, I do seriously believe that you've destroyed your own argument and I don't mean that no risk should be ever taken, instead I mean that when the risk is death of oneself or others then the risk is so high as to outweigh most, if not all rewards.

Nothing inescapable about it. What we have is a worthless anecdote you insist on supporting your position, extremely strong evidence against underestimation of risk, brain-imaging results you do not understand, none of which forces the conclusion of overconfidence as opposed to teens intrinsically having higher rewards just as they intrinsically oversleep and all the over changes that go with puberty and being young adults.

The evidence isn't extremely strong against underestimation of risk, Dobbs wrote about growing from teenager to adult... "When this development proceeds normally, we get better at balancing impulse, desire, goals, self-interest, rules, ethics, and even altruism, generating behavior that is more complex and, sometimes at least, more sensible. But at times, and especially at first, the brain does this work clumsily. It's hard to get all those new cogs to mesh."

If one gets better at balancing those things during development, ie. growing up, that reveals one has a lack in balancing those things, which are to do with judgement - and poor judgement is one form of overconfidence. http://en.wikipedia.org/wiki/Overconfidence_effect

Dobbs goes on to write, "This let the adults use a variety of brain resources and better resist temptation, while the teens used those areas less often and more readily gave in to the impulse to look at the flickering light..." this reveals that teens give in to impulse, which is about lacking judgement which goes to them being overconfident.

Dobbs continues "If offered an extra reward, however, teens showed they could push those executive regions to work harder, improving their scores."

This draws out a reason why they underestimate risk - because if the reward isn't "extra" or perceived as higher to the teen, then they likely won't push those executive regions to work harder, as is revealed in the articles previous paragraph, which I haven't quoted.

Dobbs continues "Add stakes that the teen cares about, however, and the situation changes. In this case Steinberg added friends: When he brought a teen's friends into the room to watch, the teen would take twice as many risks, trying to gun it through lights he'd stopped for before. The adults, meanwhile, drove no differently with a friend watching."

This can actually be interpreted either way, Steinberg chooses a higher reward, whilst one can just as reasonably choose an underestimation of risk – a rationale for which is alluded to in article is that social feelings/thoughts are more sensitive for the teenagers, hence the brain is more focussed on social cognition than risk estimation (as the risk estimation processes require more effort), hence less brain cycles on risk estimation – hence risk underestimation.

If I may use a real world scenario based upon the research and one that does happen, i.e. not fictional. When a teenagers friends are brought into their car, the teen would take twice (or some other >1 multiplier) as many risks, trying to gun it through lights he'd stopped at before. (rephrasing dobbs, reasonably I believe) It's known that some teenage passengers and drivers encourage or engage in risky driving for the thrill, that includes the social aspect of a shared thrill - in this situation the risk has also multiplied for now the death of passengers should be taken into account, whilst the reward has also increased. Might be zero sum, don't know, more research needed, but it’s clear that both reward and risk have increased due to the presence of others.

Now I hope that all you guys can see, this isn't a case of me being some tired old man who doesn't understand the joys of risk taking and thrill seeking. One should not forget that whilst I may be a somewhat respectable sober adult, I have already been through the teenage years under discussion and hence am able to recall those feelings that have been presented as if I am incapable of relating to. On the other hand a thrill seeking teenager is likely to be less aware of what it is to be a respectable sober adult, having never been one. In short I've been both a thrill seeking teenager, a thrill seeking adult and also a respectable sober adult. I do see both sides of this discussion.

Comment author: thomblake 10 January 2012 03:22:40PM 2 points [-]

On the other hand a thrill seeking teenager is likely to be less aware of what it is to be a respectable sober adult, having never been one. In short I've been both a thrill seeking teenager, a thrill seeking adult and also a respectable sober adult. I do see both sides of this discussion.

That's exactly what I'd expect a respectable sober adult to say.

Comment author: Peacewise 10 January 2012 04:14:47PM -1 points [-]

That's exactly what I'd expect a respectable sober adult to say.

Then you have the fortunate ability to accurately predict accurate statements.

Comment author: gwern 10 January 2012 07:46:30PM * 2 points [-]

instead I mean that when the risk is death of oneself or others then the risk is so high as to outweigh most, if not all rewards.

And how does one judge when the reward is outweighed? Hm, that wouldn't be subjective would it...?

If one gets better at balancing those things during development, ie. growing up, that reveals one has a lack in balancing those things, which are to do with judgement - and poor judgement is one form of overconfidence.

Never reason from a price change. Are teenagers overconfident about how much daytime sleep they need, and their circadian rhythm shifts due to them getting "better at balancing impulse, desire, goals, self-interest, rules, ethics, and even altruism, generating behavior that is more complex and, sometimes at least, more sensible"?

Dobbs continues "If offered an extra reward, however, teens showed they could push those executive regions to work harder, improving their scores."

This draws out a reason why they underestimate risk - because if the reward isn't "extra" or perceived as higher to the teen, then they likely won't push those executive regions to work harder, as is revealed in the articles previous paragraph, which I haven't quoted.

This practically demonstrates the opposite: that the reward is the important part! What rewards do experimenters usually offer? Pretty lousy ones. Why is it surprising that teens might not work as hard as conscientious saps - I mean, mature adults. (I am reminded of how money rewards can improve IQ test performance by half a standard deviation or so.) This is like the usual criticism of the PISA test scores: of course Americans will underperform, since not a single one of the test-takers cares about what score they get. Incentives matter, which is what I've been saying all along.

This can actually be interpreted either way, Steinberg chooses a higher reward, whilst one can just as reasonably choose an underestimation of risk – a rationale for which is alluded to in article is that social feelings/thoughts are more sensitive for the teenagers, hence the brain is more focussed on social cognition than risk estimation (as the risk estimation processes require more effort), hence less brain cycles on risk estimation – hence risk underestimation.

Wow. So your explanation for a clear-cut reward link is... they get distracted and can't estimate risk as accurately.

Comment author: Peacewise comment score below threshold [+] (27 children)

Comment author: Vaniver 11 January 2012 02:18:24AM 3 points [-]

The article reveals that reward is why the teenagers underestimate risk. The article reveals that teens perception of reward motivates their impulsiveness.

No. The value of a decision is gain minus cost; if the cost remains the same but the gain increases, then that can swing the value of a decision from negative to positive. Thus, they can be more impulsive while maintaining the same beliefs about risk.

Comment author: Peacewise 11 January 2012 02:44:15AM -2 points [-]

Thus, they can be more impulsive while maintaining the same beliefs about risk.

I'll unpack that... Thus, they can be overconfident while maintaining the same beliefs about risk. Being impulsive is being overconfident, impulsive is a lack of estimating risk, which is underestimating risk.

Comment author: Vaniver 11 January 2012 03:06:31AM 4 points [-]

Being impulsive is being overconfident, impulsive is a lack of estimating risk

I think we're looking at different dictionaries, so I'll abandon the word impulsive and try with a more object-level phrase. They can drive less carefully while maintaining the same beliefs about risk.

Comment author: Peacewise comment score below threshold [+] (23 children)

Comment author: Peacewise 11 January 2012 03:20:03PM * -5 points [-]

I think we're looking at different dictionaries, so I'll abandon the word impulsive and try with a more object-level phrase.

Hilarious, the point you have abandoned has +2, whilst my point that forced the abandoning still has -1. anyways...

They can drive less carefully while maintaining the same beliefs about risk.

and if those same beliefs are already an underestimation of risk? strike 1, just clipped the outside of the plate.

Let's unpack that last quote in context of driving... a-yawn-gain. they can drive less carefully. Less carefully, is about less care - what is "care", that's about

Care = Feel concern or interest; attach importance to something: "they don't care about human life". (dictionary.com)

So they feel less concern, they attach less importance to driving. What's the key word there, hmmm? "Less" well that's a term, in context that goes with "under"-estimate. Do you think? I do. Strike 2 - straight up the middle of the plate. Batter says, I didn't see that. Too bad says the ref.

Let's examine the opposite side, to include a process for minimising disconfirmation bias. They drive less carefully. Ok, I'm flipping my brain. The less carefully has nothing to do with underestimating risk, actually in this flip it's about overestimating risk... why do I say overestimate - well apparently that's part of the argument opposing my viewpoint, check above.

Well what's the dictionary say about what "Over estimate" means

o·ver·es·ti·mate/ˌōvərˈestəˌmāt/ Verb:
Estimate (something) to be better, larger, or more important than it really is. (dictionary.com)

hang on, hang on - overestimate = estimate something to be more important than it really is. Does overestimate sound at all like "less care"? No it doesn't, contradiction found, conclusion is Driving less carefully is about underestimating risk. Strike 3. Yer outta here!

Now, here's a thing. When the teenagers judge the reward highly, sufficiently highly to outweigh the risk of death - they have underestimated the risk. Perception of reward and risk are not in opposition, they go hand in hand.

Now let's look at the rest of the sentence.

They can drive less carefully while maintaining the same beliefs about risk.

The implication in context, is that it's reward driving the behaviour, supposedly being the entire reason for the behaviour, one significant context of the reward perception was peer involvement (see article). Let's try that one.

They can drive less carefully with more people in the car, while maintaining the same beliefs about risk, because they perceive the rewards are higher.

That fits the counter argument to my viewpoint... but hang on, now with more people in the car the risk of death is multiplied. So factually the risk has increased - yet the behaviour is supposedly all due to the reward, now if the behaviour is truly all to do with the reward, then yep the teen has discounted the risk - for the risk increased and it's not changing the behaviour.

So in that situation we've got another example where a teen has underestimated the risk due to a perception of a higher reward.

Am I being too anecdotal for you guys? Of course, discount outgroup behaviour whilst permitting the same ingroup. The article is itself filled with anecdotes... maybe we should just dismiss the entire article... stop press no no, don't do that there's no counter to my op then, lets just pick and choose the parts of it that support the counter, dismiss those that don't - both in the research and the anecdotes.

Please by all means, chuck up the -1, I'm considering them badges of honour now.

Comment author: Grognor 09 January 2012 09:55:26PM 12 points [-]

Since gwern is, well beyond what I thought was typical of him, refusing to call a horse a horse, I'm going to say it: man, you're so lame.

First understand that I noticed a disagreement going on between someone I've never seen before, a Mr. Peacewise, and a Mr. Gwern, whom I despise ever so lightly. He's a jerk on IRC, you see. It would have made me feel better for him to be wrong and you to be right, you see. I wanted that, in my gut (though not by Tarski).

But man. Gwern posts an article with a perfectly reasonable conclusion attached, and you take a slice of anecdotal evidence to say just the opposite, which just happens to precisely match your preconceptions, and then in the ensuing discussion, instead of recognizing this, you accuse gwern first of being insulting and then of being labored by cognitive biases, meanwhile with no evidence, literally none, that you are right and he is wrong.

LAME.

Comment author: Peacewise comment score below threshold [+] (13 children)

Comment author: shokwave 10 January 2012 08:38:37AM 1 point [-]

One can lead a horse to water but one can't make it drink.

This is patently false.

Comment author: nshepperd 10 January 2012 01:38:50PM 0 points [-]

I feel like "technically false" would be more accurate. If it's just you, the horse, and a puddle, it's surely going to be at least difficult to convince it to start slurping it up if it doesn't want to.

Comment author: wedrifid 10 January 2012 01:41:31PM 0 points [-]

I feel like "technically false" would be more accurate. If it's just you, the horse, and a puddle, it's surely going to be at least difficult to convince it to start slurping it up if it doesn't want to.

"You can lead a horse to water but you can't make him drink if you aren't very imaginative and your resources are artificially limited".

Comment author: TheOtherDave 10 January 2012 02:56:04PM 5 points [-]

If they're sufficiently limited, you can't even lead a horse to water.

Comment author: nshepperd 13 January 2012 11:33:14PM 2 points [-]

There is a sense of "drink" which encompasses raising a glass to your mouth and ingesting the liquid in it, or in the case of horses, lowering their head to the water, taking some into the mouth and swallowing it.

Sticking a tube down a horse's throat certainly achieves something, but not precisely this.

Comment author: dlthomas 13 January 2012 11:54:52PM 0 points [-]

Presumably your goal is a hydrated horse, however.

Comment author: wedrifid 14 January 2012 12:20:07AM 0 points [-]

You could alo be trying to drug the horse or provide him with nutritional supplementation in liquid form.

Comment author: CuSithBell 14 January 2012 12:22:02AM 1 point [-]

Another example of the danger of explicit goal maximizers.

Comment author: Peacewise comment score below threshold [+] (0 children)

Comment author: Peacewise 23 January 2012 01:38:29PM * 1 point [-]

Do you recall this line in the Matrix?

MORPHEUS: I told you that I can only show you the door. You have to step through it.

Thats what i hoped would be understood by the previous, one can lead a horse to water but one cant make it drink.

Comment author: shokwave 10 January 2012 08:49:50AM 5 points [-]

Just politely, I don't think this style is going to have good results for you. A more robust approach would be: when something does not deserve your respect, ignore it.

Comment author: Peacewise 10 January 2012 09:29:18AM -1 points [-]

You are indeed correct shokwave, thanks.

Comment author: wedrifid 10 January 2012 09:42:55AM 4 points [-]

Hey thanks Grognor, I'll take your ad hominems about both Gwern and myself with the lack of respect they deserve.

Those are insults, not ad hominem fallacies. It is a social violation and not a logical one.

Comment author: Grognor 11 January 2012 08:42:42AM 2 points [-]

I'm just going to point out that I'm surprised that this comment of mine has not only been voted up, but voted up quite strongly, considering it rests in a nest of 'hidden' comments. I fully expected this comment to rest somewhere in the neighborhood of -2 karma.

I'm not sure whether to call this a pleasant surprise. I'm really just confused.

Comment author: wedrifid 11 January 2012 09:18:46AM 1 point [-]

I'm just going to point out that I'm surprised that this comment of mine has not only been voted up, but voted up quite strongly, considering it rests in a nest of 'hidden' comments. I fully expected this comment to rest somewhere in the neighborhood of -2 karma.

I recall being surprised too, back when I saw your comment (then at +6). Usually I expect that sort of comment to be negative.

Mind you I upvoted it myself because you nailed it. You were a little bit of a jerk in the comment but I thought it was entirely appropriate to the circumstance - and balanced by at least agreeing with Gwern in the current battle.

Peacewise comments on We Change Our Minds Less Often Than We Think - Less Wrong

Article Navigation

Comments (115)

Nearest Meetups

Virtual Study Room

Recent Comments

Recent Posts

Recent Rationality Quotes

Recent Wiki Edits

Top Contributors, 30 Days

Recent Karma Awards

Comments (115)

Top Contributors, 30 Days

You'll need to login or register to do that

(Don't worry, it only takes a few seconds)

Create

Login