Is MBTI a pseudo-science in your point of view?

[GavinElster]

No, it's not as long as you're reasonable in how you interpret the test. If you interpret the test as saying there are 4 *dimensions* N/S, T/F, I/E, and J/P, namely continuous with room for grey (like at minimum probably N n X s S), which is how the Big 5 is interpreted, then actually MBTI 'works' for the same ho-hum reasons the Big 5 works, really. It's based on the idea of grouping intercorrelated personality-oriented terms into large scales (that is, for instance, taking lots of related but not identical facets of personality like for instance tough-tender and logic-feelings).

McCrae and Costa's main criticisms seem to be if you regard the MBTI as predicting discrete dichotomies and if you regard the dichotomies as actually related closely to Jung's types to the point of predicting they exist the way described by Jung in nature (rather than just loosely motivated by them).



There are advantages to the Big 5, like its claim to 'completeness,' since supposedly you could claim the Big 5 really summarizes all of personality in the most parsimonious and useful way....however, at the end of the day, there are lots of more specialized psychometric tools, designed to describe facets of personality the Big 5 is too general about, and one could say the MBTI is probably a good balance between Big 5 and Jung. Although, methodologically it is way more similar to the former, it's pretty clear the kinds of personality terms it considers are motivated from Jung (vs the Big 5's from a dictionary.

 

[rav3n]

The reason why the Big Five is more popular with the psych community is because it's trait based where they can diagnose premised on any combination of traits.

 

[Sineva]

MBTI is a statistics-based "science". Just as is astrology. (I'm referring to the personality aspect of astrology, not to lucky-days-predictions)
I have absolutely no problems with statistics-based sciences, because pretty much every social science concerning politics/humanities/psychology/philosophy/marketing even economics to a degree is mostly statistics based. They focus on observing very broad and loosely defined phenomenons in human behavior, and try to draw conclusions from it.
i.e. people born on 10th May tend to do [this] and [this]
people having and adhering to a certain set of values tend to do [this] and [this]
people living in poor economic conditions tend to have [these] and [these] political views
people living in good economic conditions tend to avoid visiting [these] and [these] cafes or shops
white people tend to do [this] and [this]
black people tend to do [this] and [this]
certain socio-economic conditions tend to lead to wars
certain socio-economic conditions tend to lead to revolutions
people who focus on the consistency of a logical process (Ti) tend to outwardly behave in a more relaxed manner
people who focus achieving goals (Te) tend to outwardly behave in a strict or overbearing manner
if [these] conditions are met, a military battle can usually be won, and if [these] conditions are met, it is usually better to withdraw
Concerning astrology, whenever I meet Virgos or Pisces we instantly become good friends, but then gradually drift apart with the Pisces due to their lack of rigidity and attachment. Whenever I meet Aries/Leos we constantly end up fighting and playing domination games. The Leo eventually learns to tolerate me, the Aries usually bitches out and leaves. Whenever I meet a Cancer, our relationship begins with me hurting their feelings, but eventually becoming best buds. This is a statistic I observed myself way too often to be brushed off as "coincidence", and it does fall into the general prediction of zodiac sign inter-communication and relationships.

... you get the drift.

It's not accurate science, because it's based on broad and loosely defined observation. But so-so many others things are too, and are useful nonetheless.

 

[BlueScreen]

Um, sorry, but not to put too fine a point on it, you're just wrong. As the leading Big Five psychologists long ago acknowledged, the MBTI dichotomies are tapping into four of the Big Five dimensions — and those are substantially-hardwired clusters of personality characteristics with respect to which identical twins raised in separate households are much more likely to match than less-genetically-similar pairs. In order for the items on an MBTI or Big Five test to do a reasonably effective typing job, there's no need for them to come anywhere close to covering the waterfront in terms of all the ways those hardwired preferences — not to mention combinations of preferences — can affect someone's personality, and they certainly don't. To just take introversion as one example: psychologists have been studying introversion for decades. Is it really your understanding that when the MBTI types someone as an introvert, we don't know anything more about them than the specific stuff in the E/I items on the MBTI test? If you're interested, you might want to check out this TC Wiki article (already linked to earlier in this thread) for more about the reliability/validity/etc. of the MBTI, and how it relates to the Big Five.

I looked at the link and it seems like a lot of text and name dropping rather than getting to the point. Can you summarise the scientific part?

In what way do you believe that the letter "I" or introvert is deterministic rather than being a statistical classification? What can you predict about a person definitely by knowing it?

I believe people sit at various points on a scale in terms of E/I rather than it being a clear dichotomy. Their reasons for being at a certain point on the scale are not always the same and could be any of the contributing factors that are used to classify a person as E/I. Therefore, the questions used to determine the classification may say something definite about a specific person, but wouldn't whether they fall on the E/I side of the centreline just determine what statistical category they are classed in, not provide any specific insight about the individual?

It might also be worth having a read of the first paragraph here: Big Five personality traits - Wikipedia

 

[Peter Deadpan]

I'm too lazy to read the thread so I'm sure this has been said many times, but it is a pseudoscience per the definition of the term.

Also, MBTI is the actual test, not the body of research, and it's horribly flawed for several reasons, such as user bias, user error, misinterpretation, lack of consistent definitions of functions, lack of objective self-awareness, etc etc etc.

 

[GavinElster]

It's not accurate science, because it's based on broad and loosely defined observation. But so-so many others things are too, and are useful nonetheless.

Well even science is approximate -- you test a theory up to certain amounts of error.... which tells you how likely it is the theory is wrong. People in science also tend to say this is likely, not that this is certain, but by likely they mean our best physics is extraordinarily likely


I don't think being statistical is any better or worse accuracy-wise than being physical, because at the end of the day, accuracy is a matter of a mathematical calculation saying something...whethr that something is about soft things like personality terminology or about tangible concrete stuff doesn't make a difference.

After all, ultimately mathematics is more certain than the hard sciences.

Thus, when we look at statistically based things, there are broad levels of variation from extraordinarily accurate to very arbitrary.... and this has nothing to do with the hard-ness. If anything it probably has more to do with the number of variables/our ability to keep track of them. Physics is very accurate because the equations fit very exactly, and it's also simpler than biology, which is pretty messy.

Big 5 replicates the same structure across a ton of populations -- it's a LOT more accurate than general rule of thumb. However, it's a lot less precise than say, fundamental physics. ... because again, the latter is studying the most fundamental building blocks of the universe, not fuzzier higher-order properties of macroscopic bodies (us!)

In what way do you believe that the letter "I" or introvert is deterministic rather than being a statistical classification? What can you predict about a person definitely by knowing it?

McCrae-Costa themselves note that it's not deterministic. They rather criticize the idea that the MBTI should be read as predicting discrete types rather than a statistical distribution.

As to the utility of providing this kind of statistical insight, that's most obvious from things like measures of general intelligence. it's an amazing fact that high general intelligence correlates with doing well in many different tasks. If we wanted to be uber-definite, ironically we'd MISS this insight, because each person will have a different result on different tasks most likely....but the point is we can quantify overall how well they do.

And in such situations, it's actually a fact that what we really care about IS that they do fuzzily speaking 'well' on many apparently unrelated things. This is key in developing intelligent AI and such. You'll never understand how to build intelligent machines if you care only about definite events, and that should be relegated to physics, which predicts the equivalent of where the zeroes and ones will be in the AI at a specific time. But that's just number crunching -- you don't get insight into the more intelligent facets without the fuzziness.
It's very similar with stuff like N/S -- no one item is a perfect measure of it, but your overall score is a good measure of a general higher-order thing.

(In fact, I think it was Turing who observed an intelligent machine should be allowed to make mistakes.....else you can't learn anything complex and have to stick to very simple tasks.)

 

[BlueScreen]

Well even science is approximate -- you test a theory up to certain amounts of error.... which tells you how likely it is the theory is wrong. People in science also tend to say this is likely, not that this is certain, but by likely they mean our best physics is extraordinarily likely I don't think being statistical is any better or worse accuracy-wise than being physical, because at the end of the day, accuracy is a matter of a mathematical calculation saying something...whethr that something is about soft things like personality terminology or about tangible concrete stuff doesn't make a difference. After all, ultimately mathematics is more certain than the hard sciences. Thus, when we look at statistically based things, there are broad levels of variation from extraordinarily accurate to very arbitrary.... and this has nothing to do with the hard-ness. If anything it probably has more to do with the number of variables/our ability to keep track of them. Physics is very accurate because the equations fit very exactly, and it's also simpler than biology, which is pretty messy. Big 5 replicates the same structure across a ton of populations -- it's a LOT more accurate than general rule of thumb. However, it's a lot less precise than say, fundamental physics. ... because again, the latter is studying the most fundamental building blocks of the universe, not fuzzier higher-order properties of macroscopic bodies (us!) McCrae-Costa themselves note that it's not deterministic. They rather criticize the idea that the MBTI should be read as predicting discrete types rather than a statistical distribution. As to the utility of providing this kind of statistical insight, that's most obvious from things like measures of general intelligence. it's an amazing fact that high general intelligence correlates with doing well in many different tasks. If we wanted to be uber-definite, ironically we'd MISS this insight, because each person will have a different result on different tasks most likely....but the point is we can quantify overall how well they do. And in such situations, it's actually a fact that what we really care about IS that they do fuzzily speaking 'well' on many apparently unrelated things. This is key in developing intelligent AI and such. You'll never understand how to build intelligent machines if you care only about definite events, and that should be relegated to physics, which predicts the equivalent of where the zeroes and ones will be in the AI at a specific time. But that's just number crunching -- you don't get insight into the more intelligent facets without the fuzziness. It's very similar with stuff like N/S -- no one item is a perfect measure of it, but your overall score is a good measure of a general higher-order thing. (In fact, I think it was Turing who observed an intelligent machine should be allowed to make mistakes.....else you can't learn anything complex and have to stick to very simple tasks.)

So MBTi predicts nothing and is a mistake?

For some reason I'm not really that amazed that people can create two statistical categories and find correlations between them. I'm not really sure that finding correlations provides any real insight either.

If instead you said that those who spend more time on reflective things are often more well read or that those who make intuitive jumps in thinking are often seen as more intelligent, it could almost be called insight, if you weren't pointing out the obvious. Obscuring those mechanisms with letters doesn't make it any more amazing.

 

[GavinElster]

For some reason I'm not really that amazed that people can create two statistical categories and find correlations between them. I'm not really sure that finding correlations provides any real insight either.

Well literally every empirical claim is gonna be correlations lol... anything exact is more in the realm of pure logic! The real question is how ambitious a claim about correlations you're making... stuff like general intelligence is making a very extraordinary claim, namely that your ability to succeed in extraordinarily distinct tasks ALL correlate.

If instead you said that those who spend more time on reflective things are often more well read or that those who make intuitive jumps in thinking are often seen as more intelligent, it could almost be called insight, if you weren't pointing out the obvious.

These are exactly the sorts of things that come out of the MBTI and/or the Big 5 -- that different tendencies are all part of a broader macro-tendency, and a high score in the broader tendency lets you predict the likelihood of success in the micro ones.

This is exactly like saying someone with high general intelligence is likely to kick someone's ass at most intellectual tasks, regardless of their specific content.


MBTI / Big 5 not only tell you there are these macro-variables but also that they're mutually independent macro-variables despite intercorrelations between specific facets or items.

So MBTi predicts nothing and is a mistake?

There's definitely no way that could be what my post was saying, given my analogies with general intelligence

 

[BlueScreen]

These are exactly the sorts of things that come out of the MBTI and/or the Big 5 -- that different tendencies are all part of a broader macro-tendency, and a high score in the broader tendency lets you predict the likelihood of success in the micro ones.

So it's like any meaningless statistical category I could create? Why on tests don't people answer all or even most of the questions in any category the same way even though they are grouped as supposedly closely related aspects? Wouldn't you say it is because people aren't all or none of these things and are instead just some of them in their own personal way? What can happen will happen, etc. Wouldn't you also say that by being high in any category it is self fulfilling that you will have most of the traits that the questions consider.

The traits, i.e. the actual information, is still the thing that matters. The categories are just a useful way of reducing information to make things easier to work with, discuss, etc. Categorisation and statistics are useful tools, but nothing magical. Similarly MBTi can be a useful tool in some ways, but it hardly defines personality or humanity. It's just one way of categorising personality preferences.

 

[GavinElster]

So it's like any meaningless statistical category I could create

Well you can't create a statistical category: either there ARE certain correlations occurring in nature, or there aren't. But more importantly, the Big 5 is far from a random set of correlations. See below.

The traits, i.e. the actual information, is still the thing that matters.

How the traits are related is the complex information. Anyone can simply list a set of traits, but finding complex relationships among them is the hard part -- your statement that ultimately there just are traits is a little like saying ultimately there are just particles and physics, so why study economics. Well, we're interested in fuzzier higher order variables because they capture complex patterns formed by those atoms. Knowing the atoms themselves tells you very little about the patterns formed by them. The Big 5 involves the ambitious claim that there are 5 macro-variables which summarize the covariance of various personality descriptors in populations internationally. And the key is this 5-variable structure has to replicate again and again in many populations --- this starts suggesting biological bases and evolutionary strategies that each variable may relate to.....with the actual dictionary-based traits potentially more incidental than the macro-variables themselves.
At this point, we're not talking a simple correlation -- we're talking a more ambitious claim in linear modeling about how the various variables of personality are interrelated.


Things like the general intelligence factor are similar -- they're ambitious because they make the claim that many facets of intelligence are surprisingly interrelated. There are debates among serious cognitive scientists as to how much there even is something called general intelligence --- that is, to what extent are facets of intelligence related vs independent (i.e. more like separate modules).

That's far from meaningless and has profound implications for the future of artificial intelligence. All very heavily related to statistical psychometric testing.


In fact, something you should take into account is the question: what IS a trait? It's just a way of describing... but the question is which traits are most powerful in their descriptive capacities. If the general intelligence people are right, if you want to build a robot which can prove Fermat's Last Theorem, it's plausibly a lot better to try to first understand how to create an intelligent robot than to compartmentalize completely specific to mathematics from the start.

The macro-variables discovered by some reasonably serious statistical mathematics may well have descriptive and predictive power that the micro-traits fed into constructing them did not.




It's funny to hear you speak dismissively of the summarizing, because when you go to the doctor for a messed up heart, he's just 'summarizing' in a 'useful way' what's wrong with you when he says your heart is dead. He could just as well have described every particle's state in your body in terms of quantum physics. But the point is that wouldn't be very insightful. The higher order claim that your heart is dead is what matters here.

Similarly, the point with the Big 5 is to discover not just different ways we can colloquially describe personality, but large, internationally replicating trends that all of personality falls into. The point is the macro-variables and their interrelations could potentially be what we really care about if we want the most fundamental facets of personality.

Categorisation and statistics are useful tools, but nothing magical. Similarly MBTi can be a useful tool in some ways, but it hardly defines personality or humanity. It's just one way of categorising personality preferences.

The Big 5 does make a claim as to discovering some of the most fundamental variables personality should be studying. This of course relates to the MBTI because incidentally, 4/5 of the Big 5 dimensions correspond pretty closely to the MBTI.

Obviously there are other ways of categorizing personality, and some are probably as good as or better than the Big 5 and so on. Heck, the HEXACO may win out. But the point is there's a reason you don't see every random tool created by a guy in his basement compete with the Big 5, as flawed as it may be, it's a lot less laughable attempt than most people throw out.

Nobody is saying that it's to be worshipped, of course.

 

[BlueScreen]

Well you can't create a statistical category: either there ARE certain correlations occurring in nature, or there aren't. But more importantly, the Big 5 is far from a random set of correlations. See below. How the traits are related is the complex information. Anyone can simply list a set of traits, but finding complex relationships among them is the hard part -- your statement that ultimately there just are traits is a little like saying ultimately there are just particles and physics, so why study economics. Well, we're interested in fuzzier higher order variables because they capture complex patterns formed by those atoms. Knowing the atoms themselves tells you very little about the patterns formed by them. The Big 5 involves the ambitious claim that there are 5 macro-variables which summarize the covariance of various personality descriptors in populations internationally. And the key is this 5-variable structure has to replicate again and again in many populations --- this starts suggesting biological bases and evolutionary strategies that each variable may relate to.....with the actual dictionary-based traits potentially more incidental than the macro-variables themselves. At this point, we're not talking a simple correlation -- we're talking a more ambitious claim in linear modeling about how the various variables of personality are interrelated. Things like the general intelligence factor are similar -- they're ambitious because they make the claim that many facets of intelligence are surprisingly interrelated. There are debates among serious cognitive scientists as to how much there even is something called general intelligence --- that is, to what extent are facets of intelligence related vs independent (i.e. more like separate modules). That's far from meaningless and has profound implications for the future of artificial intelligence. All very heavily related to statistical psychometric testing. In fact, something you should take into account is the question: what IS a trait? It's just a way of describing... but the question is which traits are most powerful in their descriptive capacities. If the general intelligence people are right, if you want to build a robot which can prove Fermat's Last Theorem, it's plausibly a lot better to try to first understand how to create an intelligent robot than to compartmentalize completely specific to mathematics from the start. The macro-variables discovered by some reasonably serious statistical mathematics may well have descriptive and predictive power that the micro-traits fed into constructing them did not. It's funny to hear you speak dismissively of the summarizing, because when you go to the doctor for a messed up heart, he's just 'summarizing' in a 'useful way' what's wrong with you when he says your heart is dead. He could just as well have described every particle's state in your body in terms of quantum physics. But the point is that wouldn't be very insightful. The higher order claim that your heart is dead is what matters here. Similarly, the point with the Big 5 is to discover not just different ways we can colloquially describe personality, but large, internationally replicating trends that all of personality falls into. The point is the macro-variables and their interrelations could potentially be what we really care about if we want the most fundamental facets of personality. The Big 5 does make a claim as to discovering some of the most fundamental variables personality should be studying. This of course relates to the MBTI because incidentally, 4/5 of the Big 5 dimensions correspond pretty closely to the MBTI. Obviously there are other ways of categorizing personality, and some are probably as good as or better than the Big 5 and so on. Heck, the HEXACO may win out. But the point is there's a reason you don't see every random tool created by a guy in his basement compete with the Big 5, as flawed as it may be, it's a lot less laughable attempt than most people throw out. Nobody is saying that it's to be worshipped, of course.

How familiar are you with statistics? Because this is all far simpler than what you are making it. I wasn't dismissing that the categories can be used for useful things, just saying that in MBTi they provide no more insight than the questions which assign a person to the category.

This wasn't a general rule for all systems, but a statement that MBTi is about categorisation rather than being a window to some greater truth or the soul. The categories are chosen to be good descriptors as you have said. If we agree, then we agree .

On a slightly different, but far more interesting topic, I'm not sure we would ever actually program personality into artificial intelligence. Personality seems like a chaotic system with too many contributing factors. The better approach may be to identify the "cells" or simple rules that generate the right system for it to develop.

On that topic, the somewhat mechanical study of personality that we often see around here doesn't really do it justice. Your post gave me a bit of a laugh because I'm a scientist and you may have assumed certain things based on that, but I'm actually arguing here because of an appreciation of people and that you can't accurately capture over 7 billion different personalities with just 4 letters. As I noted in my earlier post, what can happen will happen. If you pretend anything different in relation to personality then you are likely to be disappointed.

 

[Lark]

It might be awesome to have a range of action figures which were based on the MBTI types and have accessories and vehicles which are stereotypical to each type.

 

[iwakar]

No, but MBTI fandom does have lots of pseudo-experts, which may amount to the same damaging outcome.

 

[Jaguar]

1. focus on the consistency of a logical process
2. focus on achieving goals

They're not mutually exclusive.

 

[GavinElster]

just saying that in MBTi they provide no more insight than the questions which assign a person to the category.

OK, so if your point is to say that MBTI isn't to be worshipped and ditto for Big 5, I certainly don't dispute that. It's not magic. At the same time, what you directly say here isn't true as stated as far as I can tell (unless your point is specific to the MBTI but doesn't apply to the Big 5-- then I need some clarification what you mean/sorry if I'm missing that--- in that case you should explain why what you say applies to the MBTI, despite it not applying to the Big 5 maybe? I obviously get you're singling out the MBTI, but it's not clear to me if the argument you use for the MBTI also would apply to the Big 5 --- in which case I'd suspect the argument fails) for the same reason that it would be incorrect to say a test for general intelligence that tested based on various specific tasks is telling you no more than that you could complete those specific tasks with the success rate you achieved. Rather, the whole point is it's saying MORE, namely that it's likely you will succeed at many quite unrelated tasks beyond the ones you were specifically tested for.

Similarly, the questions that measure a Big 5 dimension can be changed quite a bit to measure that dimension, meaning the insight received is exactly NOT the same as the sum of the question content.
If you are high on extraversion, just as if you're high on general intelligence, there are things that can be predicted about you that weren't covered on the exact question content of the specific inventory used.

This is really just getting at the same idea as saying there's such a thing called just being smart, and that's not reducible to any specific skill at any specific task. And there are situations where 'being smart' is the right predictor to use, rather than a more direct measurement of a specific task.



(As a note, if you're interested in singling out the MBTI over the Big 5 here, we can go into that -- though MBTI is produced by the same principle of extracting macro variables from the test items, so I don't see how the same argument as does apply to the Big 5 doesn't apply here .)

Your post gave me a bit of a laugh because I'm a scientist and you may have assumed certain things based on that, but I'm actually arguing here because of an appreciation of people and that you can't accurately capture over 7 billion different personalities with just 4 letters.

Well I don't know your background, so that wasn't it; I certainly agree you can't capture 7 billion people with 4 letters. However, it seems a bit of a leap to go from there to saying OK, the MBTI or the Big 5 is just some random statistical set of categories... certainly some models are more powerful than others, and certainly a linear model which replicates repeatedly across many populations internationally is saying something nontrivial, even if it isn't telling you the answer to the Meaning of Life

 

[Norexan]

Type is how you think NOT -> what you do with your life. It is Ego or REAL YOU.
Type should be use for discovering your weakness in you and other people and start to think:
"Why someone think in that way? How can I improve my life? How can I understand other people?"

Also events can damage types and deformed them. Ter function can damage person because it is "fantasing" function i.g when the person read his type which is FALSE one or ISFP thinks about himself that he is INTJ because of ter Ni.
Now some type are more sensitive for change, some are not, some are more sensitive on hurting, some are not. Some are doers, define self by doing stuff, some define themselves with magic of their quick minds which is why Introverted Instuitives (dom Ni) are so obsessed with MBTI. Some types have a lot of autonomy , some are not i.g Si and Fe are complete depend on the world itself (now imagine xSFJ dependence) to the point they cannot live it without others, Te as well but not in that extreme form but still Te are not autonomous type because they need tasks and structure of the tribe to work. And the reason why Ne user can see themselves as introvert is because Ne is the most autonomous type.

 

[Turi]

Does the MBTI actively seek to challenge its claims and search for evidence that might prove these claims false, yes or no?

 

[reckful]

Does the MBTI actively seek to challenge its claims and search for evidence that might prove these claims false, yes or no?

That's not really a "yes or no" question in this context, because the real answer is that you're asking a somewhat inappropriate question.

Knowledgeable people disagree on exactly what the "hard science" and "soft science" terms mean, and where the boundaries are, but here's a quick sum-up from Wikipedia's "Hard and soft science" article:

Hard science and soft science are colloquial terms used to compare scientific fields on the basis of perceived methodological rigor, exactitude, and objectivity. Roughly speaking natural sciences (e.g. biology, chemistry, physics) are considered "hard", whereas the social sciences (e.g. economics, psychology, sociology) are usually described as "soft".

Precise definitions vary, but features often cited as characteristic of hard science include producing testable predictions, performing controlled experiments, relying on quantifiable data and mathematical models, a high degree of accuracy and objectivity, higher levels of consensus, faster progression of the field, greater explanatory success, cumulativeness, replicability, and generally applying a purer form of the scientific method.​

For a more longform discussion of the issues involved, you may want to check out the Stanford Encyclopedia of Philosophy's Science and Pseudo-Science article.

On top of all the errors currently involved in measuring, e.g., whether somebody's an introvert and how introverted they are, and how multifaceted and/or otherwise untidy the "introversion" cluster may be, there's the all-important fact that whether someone is (to some extent) hardwired to be introverted is just one among many factors that can potentially contribute to their personality, and to what they'll choose to do on any particular occasion.

So, generally speaking, personality psychologists aren't in a position to conduct "controlled experiments" where they type someone and test a falsifiable prediction about exactly what that particular person will do in a particular experimental scenario. And those kinds of limitations are true of economics and many of the other "social sciences" as well.

To somewhat oversimplify, personality typologies like the MBTI and Big Five establish their "validity" by way of studies where the types of subjects in a suitably large sample are found to correlate significantly with various other things. As one rather dramatic example that I expect you've seen before, here are the self-selection ratios that Myers reported for a study involving 705 Cal Tech science majors:

INTJ 3.88
INFJ 2.95
INTP 2.92
INFP 1.97
ENTJ 1.56
ENTP 1.42
ENFP 1.09
ENFJ 1.08
ISTJ 0.68
ISTP 0.50
ISFP 0.49
ISFJ 0.43
ESTP 0.22
ESTJ 0.12
ESFJ 0.18
ESFP 0.02

Stat spectrums that tidy are what you call a personality psychologist's dream. What they indicate (and the sample size was pretty large, at 705) is that the MBTI factor that has the greatest influence on somebody's tendency to become a Cal Tech science major is an N preference, and the MBTI factor that has the second greatest influence is introversion, with the result that the spectrum tidily lines up (from top to bottom) IN-EN-IS-ES.

Keeping in mind that twin studies indicate that the MBTI is tapping into four substantially-genetic dimensions of personality, the results of that sample suggest that there are relatively hardwired dimensions of personality that can make a person of one type (e.g., an INTJ) something like 30 times more likely than another type (an ESTJ) to end up as a science major at Cal Tech.

And I assume you'd agree that if someone had ascertained the zodiac signs of those same 705 Cal Tech science majors, it's very unlikely that the distribution of zodiac signs for those students would have proven to be substantially different than the distribution in the general population. And that's because the zodiac belongs in the "pseudoscience" bin, and the MBTI belongs in the "soft science" bin.

 

[Turi]

[MENTION=18736]reckful[/MENTION] - there is nothing inappropriate about that question - to me, there's either science (falsfiable, able to be challenged, actively seeks to prove it's claims wrong) and then there's everything else that's just not science (non-falsifiable, can not be challenged, doesn't seek to prove itself wrong).

I'm unaware of any occasions that CPP Inc. or whoever, has openly sought to prove itself wrong.

Who says Sensing must oppose Intuition? Who says Thinking must oppose Feeling? Who says Judging must oppose Perceiving?
Introversion and Extraversion actually has some support behind the dichotomy, the others - not so much, the only support comes from the use of a false dichotomy, correct me if I'm wrong.

Can I challenge my "type" in any way? If I were to be given the official MBTI in a work or school environment or whatever - am I able to challenge the result?
Some might suggest this doesn't make sense as I'm the one who selected the answers - to that, see: problems with forced dichotomies and black-and-white either/or questions that don't allow for people to select answers that truly fit who they are.

Can I challenge your "type" in any way? Can I approach CPP Inc and be like, here is my argument against reckful being an INTJ within your model - where do we go from here? Can I challenge results? Can I prove the MBTI incorrect or wrong?

If I can challenge the MBTI, what is the process? Do they just change my type? Do they just print out another result for me and voila that's it? Or will they rally against it and seek to disprove me?
Can I challenge someone as being an "intuitive"? Can I rally against the notion that somebody indeed prefer N over S? Where can I take my argument? What happens?

Is the MBTI falsifiable, yes or no?

I have real issues here, because there's nothing in the MBTI that you can challenge, it makes no certain predictions, it has no hypothesis to challenge - which to me, means it's just not science. It's based on "preferences" - I can score INFJ like I did on the official test, and still enjoy chilling with friends, still enjoy playing sports, still enjoy making logical decisions and using my head over my heart, and still prefer to take in as much information as possible before making a decision, and there's nothing anyone can do to prove I'm not an INFJ.

I can be the worlds biggest ESTP ever, and nobody can challenge my INFJ status due to "preferences".
This means it's not science, to me.

I'm sure there's more scientific-minded people out there that are able to express my concerns in a more apt fashion than I am - but the question of "Does the MBTI actively seek to challenge its claims and search for evidence that might prove these claims false, yes or no?" is not inappropriate - there's either science - or, there's pseudoscience - that's what this thread was about. Arguably, you branching off into hard-science and soft-science land is what's inappropriate as it's somewhat of an attempt to provide the MBTI with more scientific credibility than it deserves.

I mean the question was "is the MBTI a pseudo-science in your point of view?"
The answer to this, is a resounding "yes" and all you need to do to figure this out is ask yourself, does the MBTI follow the scientific method, yes or no?

 

[reckful]

I mean the question was "is the MBTI a pseudo-science in your point of view?"
The answer to this, is a resounding "yes" and all you need to do to figure this out is ask yourself, does the MBTI follow the scientific method, yes or no?

Did you read my post? Did you look at the self-selection ratios in that Cal Tech sample?

No, Turi, the respectable (dichotomy-centric) districts of the MBTI are not pseudoscience. They're soft science — like the majority of psychology, economics, etc.

And they deal with probabilities (not "certain predictions," as you put it), and they use the scientific methods appropriate to the field of personality psychology, which involve reliability and validity, among other standards.

And in those departments, as I think you know, the official MBTI can claim to be more or less on a par with the leading Big Five tests.

And there's more on that in this TC Wiki article that I already linked to.