5 ms·

How to avoid cognitive biases when you get paid to think

48 points by paraschopra 9y ago

thisisit 9y ago
> “80% fat-free” and the other one says “20% fat added”, which one will you pick? The 2nd framing is awkward. I am no marketing expert but it seems the non-awkward one will be only 20% Fat > Hack #4: Watch out if you are happy about a decision The advice here is - if you are giddy about the idea, google a contrarian view. So, how does someone actually become happy or excited about his/her idea?
- paraschopra 9y ago
  >The 2nd framing is awkward. I am no marketing expert but it seems the non-awkward one will be only 20% Fat Great point. >The advice here is - if you are giddy about the idea, google a contrarian view. So, how does someone actually become happy or excited about his/her idea? In my personal experience, I haven't found any correlation between giddiness about an idea and it being right or effective. (Especially with startup ideas). So I watch out for excitement about an idea, write out why it would fail, wait for a couple of days and revisit it. If I'm still excited, I pursue it.
  - analog31 9y ago
    This is an interesting point. Everything that we read about reason, treats it as a static problem: The analysis either weighs for or against the decision, or is inconclusive, right now. But we actually reason in time. Allowing ourselves to spend more time with a decision shouldn't change the facts or the logic, but it could improve our reasoning anyway. I describe myself as a "slowpoke" when it comes to reasoning my way through certain subject matter, notably politics. I am certainly willing to change my opinions, but I might not change them right away based on a single argument. You might have to wait a year, or a decade, for me to decide.
    - paraschopra 9y ago
      >But we actually reason in time. I love the way you phrase it. It's a wild hypothesis but I suspect that good (Bayesian style of) thinking requires metabolic energy in our brain that's a magnitude of order or more as compared to intuitive A -> B style of thinking. This is why perhaps our brains avoids the former.
      - analog31 9y ago
        What is Bayesian thinking? When I took statistics in college, Bayes' theorem, and conditional probability, were presented in fairly straightforward mathematical terms. There was no special form of thinking associated with it, but maybe Bayes had other things up his sleeve.
- MarkMc 9y ago
  > I am no marketing expert but it seems the non-awkward one will be only 20% Fat I agree, but that doesn't really alter the author's point: Saying '80% fat free' is likely to sell better than 'only 20% fat'. Daniel Kahneman makes a similar point: When a medical operation was described as having a 90% survival rate two thirds of doctors recommended the procedure. When it was described as having a 10% mortality rate only a third of doctors recommended the procedure.
- vorotato 9y ago
  Just as an aside, milk in the US is a living example of n% fat.
thethirdone 9y ago
> You take a mammogram and it comes out as positive. What is the probability that you have cancer? Lower than 1%. I'm a guy. Though this would have a similar issue for anyone who isn't a 40-50 year old woman. On a more serious note, I feel that trap should set off alarm bells for anyone who has taken a statistics course. The more dangerous fallacy along the same lines is that probabilities often brings up intransitive. There can be three medicines: A,B, and C that in a sense A > B and B > C, but C > A. This type of issue can really sneak up on you.
- DrScump 9y ago
  Lower than 1%. I'm a guy. Any male deemed a candidate for mammography probably faces a much higher probability than that. Men get breast cancer, and those who do are more likely to die from it than women who do.
  - true_religion 9y ago
    Because they are caught at a later stage than women since men are not routinely screened.
- paraschopra 9y ago
  >On a more serious note, I feel that trap should set off alarm bells for anyone who has taken a statistics course. Yes, but the problem is that most physicians get it wrong. That's what worries me. I agree that the mammogram problem is a bit artificial, but like you say similar problems come up in different forms. It's often not that people have less cognitive abilities (doctors _are_ intelligent), it's just that their intuition convinces them the answer they have in mind is the right one. There's a related research that showed that a similar problem given in _less_ legible font has less cognitive errors: https://www.theatlantic.com/technology/archive/2012/02/the-value-of-making-reading-hard/252743/ https://www.theatlantic.com/technology/archive/2012/02/the-v...
colorincorrect 9y ago
friendly reminder that cognitive biases are heuristics that are right most of the time
- siddharthdeswal 9y ago
  Interesting point. Could you add more detail to your comment please?
  - colorincorrect 9y ago
    the basic idea is that heuristics help us make quick and dirty inferences such that we don't need to overload our "brain processing power" every time we are tasked with a decision. ex: the gambler's fallacy is mostly a fallacy if events are i.i.d, which often isn't the case in real life situations. many are labeled as a "bias" once we make the observation that it failed us. but of course we should try to optimize regardless. this is tangentially related: https://www.thecut.com/2016/06/how-only-using-logic-destroyed-a-man.html https://www.thecut.com/2016/06/how-only-using-logic-destroye...
marklgr 9y ago
System I/System II [1] has been all the rage lately, making it to a wide audience with Kahneman's bestseller [2], but it is not without criticism. For instance, the usual two-column list of System I/System II traits makes things much more clear-cut than they really are. Furthermore, both Systems can be inter-related and influence each others. The System I/System II narrative is an old idea that can be quite convincing, and it's not without merits, but one might wonder if all its appeal doesn't call for a suggestion of the original post: watch if you are happy with a belief/theory. [1]: https://en.wikipedia.org/wiki/Dual_process_theory https://en.wikipedia.org/wiki/Dual_process_theory [2]:https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow
hotdox 9y ago
It is very funny to read how he avoids pitfalls of bad decision-making in a role of a CEO. He has many priveleges which his employees doesn't have. Like not visit too many meetings, didn't answer some mails or choose time to make decisions. Yes, it works for him, but almost noone else can use these tricks. EDIT:fixed my double negative

48 points by paraschopra 9y ago

> “80% fat-free” and the other one says “20% fat added”, which one will you pick? The 2nd framing is awkward. I am no marketing expert but it seems the non-awkward one will be only 20% Fat > Hack #4: Watch out if you are happy about a decision The advice here is - if you are giddy about the idea, google a contrarian view. So, how does someone actually become happy or excited about his/her idea?

paraschopra 9y ago
>The 2nd framing is awkward. I am no marketing expert but it seems the non-awkward one will be only 20% Fat Great point. >The advice here is - if you are giddy about the idea, google a contrarian view. So, how does someone actually become happy or excited about his/her idea? In my personal experience, I haven't found any correlation between giddiness about an idea and it being right or effective. (Especially with startup ideas). So I watch out for excitement about an idea, write out why it would fail, wait for a couple of days and revisit it. If I'm still excited, I pursue it.
- analog31 9y ago
  This is an interesting point. Everything that we read about reason, treats it as a static problem: The analysis either weighs for or against the decision, or is inconclusive, right now. But we actually reason in time. Allowing ourselves to spend more time with a decision shouldn't change the facts or the logic, but it could improve our reasoning anyway. I describe myself as a "slowpoke" when it comes to reasoning my way through certain subject matter, notably politics. I am certainly willing to change my opinions, but I might not change them right away based on a single argument. You might have to wait a year, or a decade, for me to decide.
  - paraschopra 9y ago
    >But we actually reason in time. I love the way you phrase it. It's a wild hypothesis but I suspect that good (Bayesian style of) thinking requires metabolic energy in our brain that's a magnitude of order or more as compared to intuitive A -> B style of thinking. This is why perhaps our brains avoids the former.
    - analog31 9y ago
      What is Bayesian thinking? When I took statistics in college, Bayes' theorem, and conditional probability, were presented in fairly straightforward mathematical terms. There was no special form of thinking associated with it, but maybe Bayes had other things up his sleeve.
MarkMc 9y ago
> I am no marketing expert but it seems the non-awkward one will be only 20% Fat I agree, but that doesn't really alter the author's point: Saying '80% fat free' is likely to sell better than 'only 20% fat'. Daniel Kahneman makes a similar point: When a medical operation was described as having a 90% survival rate two thirds of doctors recommended the procedure. When it was described as having a 10% mortality rate only a third of doctors recommended the procedure.
vorotato 9y ago
Just as an aside, milk in the US is a living example of n% fat.

thethirdone 9y ago

> You take a mammogram and it comes out as positive. What is the probability that you have cancer? Lower than 1%. I'm a guy. Though this would have a similar issue for anyone who isn't a 40-50 year old woman. On a more serious note, I feel that trap should set off alarm bells for anyone who has taken a statistics course. The more dangerous fallacy along the same lines is that probabilities often brings up intransitive. There can be three medicines: A,B, and C that in a sense A > B and B > C, but C > A. This type of issue can really sneak up on you.

DrScump 9y ago
Lower than 1%. I'm a guy. Any male deemed a candidate for mammography probably faces a much higher probability than that. Men get breast cancer, and those who do are more likely to die from it than women who do.
- true_religion 9y ago
  Because they are caught at a later stage than women since men are not routinely screened.
paraschopra 9y ago
>On a more serious note, I feel that trap should set off alarm bells for anyone who has taken a statistics course. Yes, but the problem is that most physicians get it wrong. That's what worries me. I agree that the mammogram problem is a bit artificial, but like you say similar problems come up in different forms. It's often not that people have less cognitive abilities (doctors _are_ intelligent), it's just that their intuition convinces them the answer they have in mind is the right one. There's a related research that showed that a similar problem given in _less_ legible font has less cognitive errors: https://www.theatlantic.com/technology/archive/2012/02/the-value-of-making-reading-hard/252743/ https://www.theatlantic.com/technology/archive/2012/02/the-v...

colorincorrect 9y ago

friendly reminder that cognitive biases are heuristics that are right most of the time

siddharthdeswal 9y ago
Interesting point. Could you add more detail to your comment please?
- colorincorrect 9y ago
  the basic idea is that heuristics help us make quick and dirty inferences such that we don't need to overload our "brain processing power" every time we are tasked with a decision. ex: the gambler's fallacy is mostly a fallacy if events are i.i.d, which often isn't the case in real life situations. many are labeled as a "bias" once we make the observation that it failed us. but of course we should try to optimize regardless. this is tangentially related: https://www.thecut.com/2016/06/how-only-using-logic-destroyed-a-man.html https://www.thecut.com/2016/06/how-only-using-logic-destroye...

marklgr 9y ago

System I/System II [1] has been all the rage lately, making it to a wide audience with Kahneman's bestseller [2], but it is not without criticism. For instance, the usual two-column list of System I/System II traits makes things much more clear-cut than they really are. Furthermore, both Systems can be inter-related and influence each others. The System I/System II narrative is an old idea that can be quite convincing, and it's not without merits, but one might wonder if all its appeal doesn't call for a suggestion of the original post: watch if you are happy with a belief/theory. [1]: https://en.wikipedia.org/wiki/Dual_process_theory https://en.wikipedia.org/wiki/Dual_process_theory [2]:https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow

hotdox 9y ago

It is very funny to read how he avoids pitfalls of bad decision-making in a role of a CEO. He has many priveleges which his employees doesn't have. Like not visit too many meetings, didn't answer some mails or choose time to make decisions. Yes, it works for him, but almost noone else can use these tricks. EDIT:fixed my double negative