3 ms·

It’s astrology because it takes unimodal distributions and splits them into two groups above and below the mean. This is a cardinal sin in statistics because it

It’s astrology because it takes unimodal distributions and splits them into two groups above and below the mean. This is a cardinal sin in statistics because it implies the existence of distinct populations that don’t really exist and so any predictions you try to make are no better than a coinflip.

rbanffy 2y ago
They just said MBTI doesn’t predict anything. It’s a measurement tool, with limitations, and you should use it understanding those. Much like an IQ test, or a having or not a PhD, it doesn’t hold much predictive power.
- Joker_vD 2y ago
  If you can't make any useful predictions with it then whatever MBTI measures (or it believes it measures) is... also useless?
  - _yb2s 2y ago
    I attempted to explain what it is useful for above, but will elaborate. The valuable result is the extreme opposite of your result- it is a therapy or personal development tool that directs you to explore and understand parts of yourself you reject or disapprove of (your “shadow”). It simply quantifies something obvious that you already know: how you consciously see yourself- in a way that guides you to explore aspects of yourself that you otherwise would choose to avoid because it’s difficult and possibly even painful to do so. What it measures is actually pretty banal and obvious if you are familiar with Jungian archetypes- one can easily guess the result for yourself and other people without using the test.
    - disgruntledphd2 2y ago
      > The valuable result is the extreme opposite of your result- it is a therapy or personal development tool that directs you to explore and understand parts of yourself you reject or disapprove of (your “shadow”). Look, I too have read Jung and agree with what you're saying here, but this is emphatically NOT how MBTI tends to be used in corporate settings, which is what everyone else in this thread is talking about.
      - _yb2s 2y ago
        I agree, it is being misunderstood and misused. Instead of using it as a tool for deep introspection, it seems to be used for the condescending purpose of “fabricating a mystical 3rd party authority get our idiot employees to notice that people have different personalities and everyone else isn’t just a carbon copy of them.” Which sounds like an irredeemably awful workplace situation, and I don’t envy people in it.
_yb2s 2y ago
It actually does not do that- each dimension which represents two opposite archetypes (e.g. introvert vs extrovert) is a continuum, and the output of the test isn’t actually a single category but a point on that line for each dimension. The full output is actually a 4 dimensional vector. Importantly- everyone has both traits, it is only telling you which you more consciously identify with. Many people will be near the middle and not strongly on one end of a dimension. Again, astrology claims to predict the future based on nonsense. MTBI makes no such claims, it predicts nothing, and it’s chosen dimensions are very transparently arbitrary value judgements, among limitless possible archetypes that one could choose.
- chongli 2y ago
  I think you may have missed my point so I’ll give an example: Suppose you take a random sample of adult humans and sort them by height. What you’ll see is a bimodal distribution which reflects the fact that there are actually two distinct populations represented in the sample: men and women. The sexes can be considered types and you will get reasonably accurate results for predicting height from sex and vice versa. However, if you instead take a sample of only men and sort them by height you will see a unimodal distribution. If you then try to split this sample into two types above and below the mean, you will find that it doesn’t tell you anything! This is what I mean when I say MBTI astrology. The archetypes are meaningless! Someone who is at 51% on the extroversion scale will get labeled an Extrovert and someone at 49% will get labeled an Introvert, but in practice these two are virtually indistinguishable and fall squarely within the margin of error. Since the distribution is unimodal the majority of the population clusters around the means, exactly what you would expect to find with the hypothesis that personality archetypes do not exist.
  - _yb2s 2y ago
    I don’t think I missed your point but you missed mine. If someone is 51% or 49% it is misleading to classify- which is why the test actually outputs a vector. Archetypes are fairly arbitrary concepts that isolate individual aspects among thousands that could potentially explain personality traits- they don’t claim to exist in real people. It is critical to realize that in the dimensions on MBTI everyone is actually both, but some people strongly identify with one and reject or despise the part of themselves that represents the other.

rbanffy 2y ago

They just said MBTI doesn’t predict anything. It’s a measurement tool, with limitations, and you should use it understanding those. Much like an IQ test, or a having or not a PhD, it doesn’t hold much predictive power.

Joker_vD 2y ago
If you can't make any useful predictions with it then whatever MBTI measures (or it believes it measures) is... also useless?
- _yb2s 2y ago
  I attempted to explain what it is useful for above, but will elaborate. The valuable result is the extreme opposite of your result- it is a therapy or personal development tool that directs you to explore and understand parts of yourself you reject or disapprove of (your “shadow”). It simply quantifies something obvious that you already know: how you consciously see yourself- in a way that guides you to explore aspects of yourself that you otherwise would choose to avoid because it’s difficult and possibly even painful to do so. What it measures is actually pretty banal and obvious if you are familiar with Jungian archetypes- one can easily guess the result for yourself and other people without using the test.
  - disgruntledphd2 2y ago
    > The valuable result is the extreme opposite of your result- it is a therapy or personal development tool that directs you to explore and understand parts of yourself you reject or disapprove of (your “shadow”). Look, I too have read Jung and agree with what you're saying here, but this is emphatically NOT how MBTI tends to be used in corporate settings, which is what everyone else in this thread is talking about.
    - _yb2s 2y ago
      I agree, it is being misunderstood and misused. Instead of using it as a tool for deep introspection, it seems to be used for the condescending purpose of “fabricating a mystical 3rd party authority get our idiot employees to notice that people have different personalities and everyone else isn’t just a carbon copy of them.” Which sounds like an irredeemably awful workplace situation, and I don’t envy people in it.

_yb2s 2y ago

It actually does not do that- each dimension which represents two opposite archetypes (e.g. introvert vs extrovert) is a continuum, and the output of the test isn’t actually a single category but a point on that line for each dimension. The full output is actually a 4 dimensional vector. Importantly- everyone has both traits, it is only telling you which you more consciously identify with. Many people will be near the middle and not strongly on one end of a dimension. Again, astrology claims to predict the future based on nonsense. MTBI makes no such claims, it predicts nothing, and it’s chosen dimensions are very transparently arbitrary value judgements, among limitless possible archetypes that one could choose.

chongli 2y ago
I think you may have missed my point so I’ll give an example: Suppose you take a random sample of adult humans and sort them by height. What you’ll see is a bimodal distribution which reflects the fact that there are actually two distinct populations represented in the sample: men and women. The sexes can be considered types and you will get reasonably accurate results for predicting height from sex and vice versa. However, if you instead take a sample of only men and sort them by height you will see a unimodal distribution. If you then try to split this sample into two types above and below the mean, you will find that it doesn’t tell you anything! This is what I mean when I say MBTI astrology. The archetypes are meaningless! Someone who is at 51% on the extroversion scale will get labeled an Extrovert and someone at 49% will get labeled an Introvert, but in practice these two are virtually indistinguishable and fall squarely within the margin of error. Since the distribution is unimodal the majority of the population clusters around the means, exactly what you would expect to find with the hypothesis that personality archetypes do not exist.
- _yb2s 2y ago
  I don’t think I missed your point but you missed mine. If someone is 51% or 49% it is misleading to classify- which is why the test actually outputs a vector. Archetypes are fairly arbitrary concepts that isolate individual aspects among thousands that could potentially explain personality traits- they don’t claim to exist in real people. It is critical to realize that in the dimensions on MBTI everyone is actually both, but some people strongly identify with one and reject or despise the part of themselves that represents the other.