Are these personality types really scientifically grounded?
I once was sort of compelled to take a Myers-Briggs test and I found it utterly flawed. So much of one’s personality depends on the contact and environment, yet Myers-Briggs wants to force you into one type.
Making things way too simple. But maybe simple is what was sought here rather than really resolving personality.
That's a good question, and your reaction to the Myers-Briggs is justified! It is not evidence-based and is roundly rejected by modern personality psychology. The work the author of this piece was describing on the other-hand is one of the most well-researched phenomena in the history of psychology.
It is based on a massive amount of research looking at clusters of personality adjectives found in dozens of languages that tend to cluster along five dimensions. And the same five clusters are found in virtually all languages (sometimes to include a 6th cluster often called "honesty/humility").
And to your point about types, this system doesn't assign types to people, but instead places them on a continuous dimension that is considered to be much more useful.
And to your point about complexity, the 5 dimensions are higher level structures. Within each dimension is a branching tree of complexity that is where the current work lies.
His article defends a modified myers-briggs that is much closer to the big 5. Myers-briggs with personality traits on a distribution is extremely similar to big 5 minus neuroticism.
Thank you for this deep answer, and for confirming my impression of Myers-Briggs!
I can see how a continuous scale can already fix a few issues. However, isn’t it still too simplistic to say “personality feature X correlates with real-life success measure Y?”
Take for example “Openness”. It’s not hard to imagine how the very same person can be very open with one group of people, but act totally differently in another group.
Is Openness then really a feature of personality?
Or maybe rather a description of the social behaviour of one person in a specific context?
>Is Openness then really a feature of personality? Or maybe rather a description of the social behaviour of one person in a specific context?
This is a great question! And it touches on what's call the "person-situation" debate, first popularized by Walter Michel[1]. Essentially there was a major debate about whether traits were a thing or we should just be focused on context-specific behavior. Mischel believed that traits weren't useful, and situations ruled.
In many ways this debate has been resolved by the "person as a density distribution" [2]. People do vary a lot within, and that variation is vitally important. However, their average or "set point" upon which they vary is also important and predictive.
When we talk about the correlation between a trait and some external marker, we are only looking at correlations with averages or set-points, which as you mention is only a limited picture of that person, but it is still useful and valuable to look at. For example, I'm a fairly open person. This means that on average I'm more open in more situations. And this is predictive. It's also important to note that this only loosely predicts what I'll do in any particular situation or at any particular timepoint.
Excellent—that sounds almost like a Bayesian approach, where personality corresponds to the prior, the context in which the person behaves is the likelihood, and the posterior is a distribution of behaviours.
I guess a personal goal should then be to seek the context in which one’s strengths match best the demands.
I would suggest an MCMC approach but I guess the evaluation of individual outcomes is too costly to make this practical ;)
> Take for example “Openness”. It’s not hard to imagine how the very same person can be very open with one group of people
Lest this be misunderstood, note that "Openness" in this context is not about being honest. It is short for "Openness to Experience". Roughly, it's about creativity and curiosity.
Good point. Still, a person that radiates creativity and curiosity in an environment that they find inspiring could be the exact opposite in another environment that they find less suitable. Hence my point that this personality trait stuff is at least as much dependent on context as on personal traits.
That's widely accepted as part of personality theory. The idea is that you have some "base setting," not that you will literally try anything (or refuse to try anything). It should also be noted that personality is one aspect of your psychology. It doesn't absolutely determine how you will behave under a given set of circumstances. Example: you might be very open to experience, but as a child you were bit by a dog. Next time you see a dog, your experience-based fear could easily overwhelm your natural curiosity.
When I first learned about Myers-Briggs, about 15 years ago, I took the test and I got I wanna say INTJ. About a week later, I took it again and got INTF. A week after that, I took it again and got ENTP. That's when I stopped trusting it, and I can't take it seriously whenever anybody brings it up. Let alone the fact that it was designed by two housewives in the 1910s, based on books by Carl Jung he disavowed himself.
As I've grown older, I've developed a theory about why people like the zodiac sign, Myers-Briggs, blood type personalities and all that other crap:
* Sense of belonging - wanting to belong to an in-group
* Understanding oneself - these categorizations let you explain your personality really quickly, if not really in depth
* Avoiding responsibility for themselves - instead of recognizing your flaws, it's easier to say "well I'm a Libra, what did you expect?"
I can understand the need for the first two, but that last point always makes me angry.
That is a problem of interpretation by amateurs more than a problem with the test itself. Myers-Briggs gives you PERCENTAGES of every trait, then it uses 4 letter labels to make things simpler for people, but you are supposed to understand that a person that scored 49% introvert in the introvert-extrovert spectrum is going to be far similar to a person that scored extrovert with 51% than to a person that scored introvert 10% because its a spectrum not an absolute value.
That is the reason that modern tests like Big Five did away with the labels, because people were incapable to reason beyond them and kept making dumb assumptions like "He scored introvert so he must hate parties or I scored introvert the first time but extrovert the second (a potential 1% difference between the labels without the percentages associated to them) so the test is crap".
> you are supposed to understand that a person that scored 49% introvert in the introvert-extrovert spectrum is going to be far similar to a person that scored extrovert with 51% than to a person that scored introvert 10% because its a spectrum not an absolute value.
This doesn't fit well with my mental model of how Myers Briggs works (regardless of the veracity of Myers Briggs).
As I understand it: Myers Briggs maps to the Jungian cognitive functions[0] and the side that you fall on the J/P dichotomy will invert these.
An INTJ would be: Introverted Intuition, Extraverted Thinking, Introverted Feeling, Extraverted Sensing
An INTP would be: Introverted Thinking, Extraverted Intuition, Introverted Sensing, Extraverted Feeling
Getting near the 50% mark on the J/P aspect appears to show low confidence in any part of the model
Myers and Briggs were inspired by Jung. And some enthusiasts insist on interpreting it how you do. But basically every test treats the axes as continuous.
The Big 5 Personality Inventory is pretty much the gold standard in the field of psychology. The traits it measures are reasonably (though not completely) stable over a person's lifetime, have decent predictive power, and have been tested and studied rigorously for decades now. They have nearly nothing in common with Meyers-Briggs.
In particular, each of the traits is measured on a separate spectrum—none of them are treated as binary the way M-B does, and they don't seek to pigeonhole you based on your levels of the different traits into some oversimplified bucket.
See [1]. It's a lot, I don't fully understand it myself. I find it convincing enough to say it's scientifically grounded, in the linguistic sense at least. I wouldn't say this research is grounded in the biological sense (not the five factor model anyway, perhaps the concept of "personality" is though).
Myers-Briggs is pseudoscientific bs. The only reason it’s so famous is that it’s an easy snake oil to peddle by consultants.
It was literally concocted by two persons without scientific method or context. It’s fiction. Some fiction mirrors life quite well, but that does not turn it into a scientific tool.
The only time Myers-Briggs has some utility is as a facilitation tool when a consultant needs to convince a roomfull of people that diverse teams are good, and that cognitive diversity is an asset and what it might look like. I.e. it’s a usefull proxy for a scientific measure when we’ve been brainwashed to trust only high-modernist fantasies and have no space for actual humanist rumination.
it's claimed that myers-briggs is reliably repeatable, but my personal experience contradicts this, with a flip-flopping of 3 of the 4 attributes over time. it doesn't seem too reliable to me.
I once was sort of compelled to take a Myers-Briggs test and I found it utterly flawed. So much of one’s personality depends on the contact and environment, yet Myers-Briggs wants to force you into one type.
Making things way too simple. But maybe simple is what was sought here rather than really resolving personality.