This. Cerebral people tend to denounce personality types as pseudoscience, astrology, Barnum effect etc. which they likely are.
But the usefulness of personality types is that they are shorthand for people who are into them to describe themselves without going through every story in their lives. It gives people a kind of crude vocabulary to convey their rough self image without using a lot of words and without the need for repeated exposure. It is kind of like an aggregate statistic that is wrong but still useful.
Oddly enough, that is extremely valuable guide to reading their behaviors. They may self identify as a ENFP or whatever while being something else, but even knowing what they perceive themselves to be is extremely useful especially in relationships. In romantic relationships it helps you navigate unknowns by anticipating them (trust but verify still applies but this gives you some Bayesian priors).
Just because a tool is not scientific doesn’t mean it’s not useful and to be summarily dismissed. (Note the subtle point: I’m not saying all non scientific tools are useful — just that some are) You just have to know how to extract the signal.
P.s. the big 5 personality test is more scientific and reproducible and reports along dimensions in percentiles. This is useful for research but no one ever mentions their big 5 scores in casual conversation. (I don’t even remember what my scores are)
> the big 5 personality test is more scientific and reproducible and reports along dimensions in percentiles.
It's also pretty useless for genuinely understanding people because it's claims are so shallow. You like talking to people, therefore you are extraverted, which means you like talking people. Brilliant.
The equivalent claim in Myers-Briggs on the other hand is something more akin to "you like talking to people", therefore you are extraverted which means that your mental processes are tuned to process a high quantity of external input, which you can then use to make predictions well outside of the domain of talking to people.
> personality types as pseudoscience, astrology, Barnum effect etc
Personality types at least have a plausible mechanism - different processes in the brain. If you think about it, the idea that people have different mental processes is actually a more sensible starting point than the idea that everybody thinks in the same way.
1. There's utility in information compression. Extraversion encompasses a lot more than "talking to people"; it encompasses things like positivity, social proactivity, energy level etc. If you say "someone is more extraverted than 99% of the population" you can make a lot of inferences other than "likes to talk to people". The Big Five aren't everything, but they convey a lot.
2. There are theories underlying the Big Five; the Big Five and its variants are so well-replicated and appear in so many settings that they have become something to be explained, which is generative. E.g., there's a theory that extraversion basically refers to basal level of active, positive emotion, neuroticism refers to basal level of negative emotion; agreeableness to empathic ability processes etc. Maybe these are or aren't right, but there are non trivial theories about them that are very elaborated. You just don't hear about them as much necessarily because you don't hear about the Big Five as much, and because there are multiple theories for the Big Five -- as opposed to the Myers Brigs, which is tied to one account.
Except the meyers briggs claim is probably not true while the big 5 claim probably is.
> Personality types at least have a plausible mechanism - different processes in the brain. If you think about it, the idea that people have different mental processes is actually a more sensible starting point than the idea that everybody thinks in the same way.
Except the MBPI results show very poor repeatability and don't seem to be particularly predictive of anything.
As much as people hate on IQ, a large fraction of what an IQ test measures seems to be intrinsic. The same is true for e.g. the Big 5.
The percentile thing is not to be dismissed either. If we were to do the MBPI thing with IQ, then everyone under 100 would be "stupid" and everyone over 100 would be "smart" which is silly because obviously someone with a 99 IQ is much more similar to someone with a 101 IQ than either is to someone with an 85 or 115 IQ.
> Except the MBPI results show very poor repeatability and don't seem to be particularly predictive of anything.
Which is either an issue with the theory or the experiments, right?
In general I agree that one should be skeptical of theories that aren't validated by experimentation, but in this specific case there is plenty of reason to believe that it is the experiments themselves that are deficient:
- MBTI theory is complex. An order of magnitude more so than something like the big 5. This also makes it more useful / predictive.
- MBTI theory posits internal thinking processes. These cannot be directly observed. Most experiments don't even discuss this, they just rely on the unreliable type indicators (questionnaires) blindly. Of course they don't find significant results
- Most experiments that do discuss theory, discuss the simpler (and IMO plain wrong) trait model of MBTI that is more similar to the big 5 model. Under this model, MBTI terms (e.g. Introversion, Feeling) apply to people. The more sophisticated and plausible model (that almost all MBTI practitioners and theorists work with) applies these terms to Cognitive Processes (i.e. thinking mechanisms), and only then describes people in terms of which thinking processes they prefer (use more often/easily).
- The theory itself is still often described in terms of Jargon that came from Jung. This makes it hard to understand what is being described, even though it very similar to other theories in mainstream academic psychology/philosophy.
> If we were to do the MBPI thing with IQ, then everyone under 100 would be "stupid" and everyone over 100 would be "smart" which is silly because obviously someone with a 99 IQ is much more similar to someone with a 101 IQ than either is to someone with an 85 or 115 IQ.
This is actually a misconception. An MBTI type really describes "A baseline trajectory of preferences towards different mental processes across their lifetime". So:
- Individuals will differ (due to life circumstances, experience, etc). The MBTI claim, is that modulo life experience, this individual would develop in this way.
- A given individual will vary how they think over their lifetime (in a predictable way!).
- A given individual will vary how they think over short time scales (minutes) in a random (unpredictable) way. But that over longer time scales patterns will emerge.
Notably, dual-process theory has the same binary structure, and a LOT of experimental evidence that the bimodal distribution holds up in practice.
> - MBTI theory is complex. An order of magnitude more so than something like the big 5. This also makes it more useful / predictive.
If it's more predictive, where are the studies showing it? Big 5 has some fairly good studies (and of course lots of terrible no good very bad studies[1]) that show e.g. conscientiousness predicts future academic performance.
> - MBTI theory posits internal thinking processes. These cannot be directly observed. Most experiments don't even discuss this, they just rely on the unreliable type indicators (questionnaires) blindly. Of course they don't find significant results
Can they be indirectly observed? Are there studies doing so? Otherwise this is just another non-falsifiable prediction.
> - Most experiments that do discuss theory, discuss the simpler (and IMO plain wrong) trait model of MBTI that is more similar to the big 5 model. Under this model, MBTI terms (e.g. Introversion, Feeling) apply to people. The more sophisticated and plausible model (that almost all MBTI practitioners and theorists work with) applies these terms to Cognitive Processes (i.e. thinking mechanisms), and only then describes people in terms of which thinking processes they prefer (use more often/easily).
I am (possibly unfairly) regarding this as a no-true-scottsman. You may be right, but I've seen enough of "Oh, sure that MBTI wasn't good becaus X/Y/Z", but my special MBTI A/B/C is better!" type comments that I lack the motivation to continue investigating them. In addition, this is somewhat irrelevant to TFA because the MBTI types are exactly what the internet is obsessed with.
> - The theory itself is still often described in terms of Jargon that came from Jung. This makes it hard to understand what is being described, even though it very similar to other theories in mainstream academic psychology/philosophy.
Agreed. It doesn't help that a lot of Jung's theories are just absurd.
>> If we were to do the MBPI thing with IQ, then everyone under 100 would be "stupid" and everyone over 100 would be "smart" which is silly because obviously someone with a 99 IQ is much more similar to someone with a 101 IQ than either is to someone with an 85 or 115 IQ.
> This is actually a misconception. An MBTI type really describes "A baseline trajectory of preferences towards different mental processes across their lifetime". So:
> - Individuals will differ (due to life circumstances, experience, etc). The MBTI claim, is that modulo life experience, this individual would develop in this way.
This is probably not testable, and I'm not sure how it's useful to say "If you didn't have any life experience you Alice would be like X and Bob would be like Y"; if MBTI measures something that will have a net effect on life outcomes, than it should be revealed in the data.
> - A given individual will vary how they think over their lifetime (in a predictable way!).
> - A given individual will vary how they think over short time scales (minutes) in a random (unpredictable) way. But that over longer time scales patterns will emerge.
It seems likely that these statements are true, but I disagree that there is any reason to believe that MBTI will predict these patterns very well, and particularly not better than Big 5 (as your GP comment states).
> Notably, dual-process theory has the same binary structure, and a LOT of experimental evidence that the bimodal distribution holds up in practice.
Which makes me ask why MBTI doesn't have such experimental evidence about bimodal distributions showing up.
> Cerebral people tend to denounce personality types as pseudoscience, astrology, Barnum effect etc. which they likely are.
At one point back in college I made up my own astrological sign and insisted people (frankly, women; the major preoccupation of a 20-something dude) accept it. When I explained carefully how my birth sign didn't fit me at all, they tended to accept it.
But the usefulness of personality types is that they are shorthand for people who are into them to describe themselves without going through every story in their lives. It gives people a kind of crude vocabulary to convey their rough self image without using a lot of words and without the need for repeated exposure. It is kind of like an aggregate statistic that is wrong but still useful.
Oddly enough, that is extremely valuable guide to reading their behaviors. They may self identify as a ENFP or whatever while being something else, but even knowing what they perceive themselves to be is extremely useful especially in relationships. In romantic relationships it helps you navigate unknowns by anticipating them (trust but verify still applies but this gives you some Bayesian priors).
Just because a tool is not scientific doesn’t mean it’s not useful and to be summarily dismissed. (Note the subtle point: I’m not saying all non scientific tools are useful — just that some are) You just have to know how to extract the signal.
P.s. the big 5 personality test is more scientific and reproducible and reports along dimensions in percentiles. This is useful for research but no one ever mentions their big 5 scores in casual conversation. (I don’t even remember what my scores are)