Rational Choices

[andrewducker]

Interesting article here which talks about decision making in teens and adults.

Teens, apparently, aren't the irrational emotion-driven beings they're frequently made out to be. In fact, they make decisions _more_ rationally than adults, as they weigh up all of the options and think about the pros and cons - whereas adults largely rely on comparing situations to ones in the past and applying fuzzy logic.

The difference is that teenagers have different wants than adults - and are happy to take a short-term benefits and care less about the risks.

What I find particularly interesting is that the article concludes that it's not worth trying to reason with teenagers, as if you're trying to persuade them not to behave in risky behaviour they are likely to not be swayed by the logic, as they don't care about risk. Instead, it indicates we should try to inculcate them against risk entirely.

Sounds about right - if someone holds a different opinion to you, and your logic fails, brainwash them.

Comments

[khbrown.livejournal.com]

"Teens, apparently, aren't the irrational emotion-driven beings they're frequently made out to be. In fact, they make decisions _more_ rationally than adults, as they weigh up all of the options and think about the pros and cons - whereas adults largely rely on comparing situations to ones in the past and applying fuzzy logic."

Not read the article, but...

This would make sense in one way - being relatively lacking in experience and typifications to use, the teenager has to instead make a more thought-through decision initially. Then, over time, if their chosen solution seems to work, it becomes part of the repertoire that they unthinkingly/unreflexively use as an adult. Until, that is, some situation arises where this hitherto tried and tested solution falls down. And then it's whether the failure is fatal or provides a beneficial learning experience.

Risk is always a tricky one, partly because its so difficult to distinguish actual from perceived risks and the system, as it were, has a vested interest in making us more aware of some than others - e.g. the terrorist threat...

[andrewducker]

I recommend reading the article, which is only about 300 words long.

But yes, I largely agree.

[malwen.livejournal.com]

Using old fashioned vocabulary from '80s strategic game programming, what the article says to me is that less experienced humans have
- less effective experience-optimised information filters
- less good experience-optimised evaluation functions
and that therefore, when experience matters, less experienced humans
(a) take longer to make a decision because of a higher noise ratio
(b) make less good decisions.

Well duh.

However, this doesn't mean
- that they are more rational, except inasmuch as their logical processes are more transparent
- that they are less rational, except inasmuch as their evaluation functions are not experience optimised
- that they are pre-disposed to prefer short term benefits
- that they care less about risks
It's more that their evaluation functions are unable to properly process long term benefits. And they really mess up on risk assessment, producing perverse risky behaviour despite being naturally more risk averse.

The article goes on to say that bombarding teens with risk data is a counter-productive strategy, because their evaluation starts off hypersensitive and is quickly broken because it produces results that clearly don't tally with the real world.

I don't think it says teens won't be swayed by logic. Or that teens should be inculcated against risk: they already start off too risk averse, and more risk data results in more cycles of / avoid risk/ get poor results / take risk/ get hurt / heal / learn.

What it suggests, and here I'm not sure I agree, is that a more fruitful approach would be to try to work directly on category based risk management. That is, taking a non-experiential approach to something that we patently mostly 'get' through experience.

[andrewducker]

I don't think it says they have less optimised evaluation functions. It says that they think about things more because they don't have as much experience, that's all.

To train a neural network you keep feeding it situations, and telling it when it's got the right answer, until it associates certain inputs with certain outputs. But it doesn't hone it's logic - it merely associates A with X and B with Y. And that's what the article says adults do - they see A and come to a conclusion, based on what they've done before, without really thinking about how they came to know that Y is the way forward.

I don't think that teen are less able to process long-term benefits, I think they prioritise them lower. Which isn't wrong, per se, just different.

[malwen.livejournal.com]

I don't think that teens are less able to process long-term benefits, I think they prioritise them lower. Which isn't wrong, per se, just different.

The start of the article agrees with you on lower priority, but then contrasts slow teen processing with rapid adult processing. It also suggests that the latter is more effective. The implication is that the lower priority reflects poorer processing.

I don't think it says they [teens] have less optimised evaluation functions. It says that they think about things more because they don't have as much experience...

The article suggests that experienced adults are faster and more successful, by dint of processing which references stored data in a way that differs from teens. I've equated that with better evaluation functions, which may be stretching things.

To train a neural network you keep feeding it situations, and telling it when it's got the right answer, until it associates certain inputs with certain outputs. But it doesn't hone it's logic - it merely associates A with X and B with Y.

I'm not sure that a pure neural network model captures what's going on very effectively. Especially when a person is repeatedly making their own judgments on whether something is the right answer, frequently based on indirect experience, from sources of varying reliability. Often there's a lot of conceptual modelling at the experience evaluation stage, and refinement in post-processing.

And that's what the article says adults do - they see A and come to a conclusion, based on what they've done before, without really thinking about how they came to know that Y is the way forward.

This is what the article says, but I believe there's much more to it than the article implies.

Overall I think the article is confused. It talks as if results were uniform across populations, so doesn't address the fact that both teens and adults have varying degrees of success and rationality. As a result it is schizophrenic about whether gist based approaches are crap, irrational magic or opaque rationality.

[azalemeth.livejournal.com]

*Cue long, angst-filled teenagery post about how irrational adults are usually the ones who are most convinced they are rational*

Thanks for sharing that though :P.