MBTI for Skeptics

By Ryan Smith, Eva Gregersen, and Sigurd Arild

A number of common critiques of Jungian typology and the MBTI are often voiced in no uncertain terms by people who have comparatively little knowledge about these fields. In this article we purport to answer the most common of these critiques.


  • MBTI is pseudoscience and no better than astrology.
  • There are personality tests that are more scientific than the MBTI.
  • In terms of empirical evidence, types don’t exist. Human traits are distributed on a bell curve (normally), not as two camel humps (bimodally).
  • People who take the MBTI multiple times often come out as different types (Low Test-Retest Reliability).
  • MBTI descriptions contain vague and flattering statements that anyone would agree to (Forer Effect).
  • Psychologists and academic researchers don’t take the MBTI seriously.
  • Jung’s cognitive functions were good. It was Isabel Myers and the MBTI that simplified the theory so that it came out wrong.
  • The MBTI makes use of pseudoscientific practices like handwriting analysis (Graphology) and face reading (Physiognomy).
  • Four dichotomies don’t say everything about human personality. Even if the MBTI isn’t pseudoscience, it is often used by amateurs and often used to explain things that it can’t explain.

MBTI is pseudoscience and no better than astrology.
False. By the common definition of pseudoscience (i.e. that there is no empirical evidence for the claims that are set forth by the theory), the MBTI is not pseudoscience. No scientific study has ever found that there is no validity to the MBTI. On the contrary, every study ever carried out has found that, while there are problems with the MBTI, there is nonetheless an acceptable level of empirical evidence supporting its claims. By contrast, astrology has repeatedly been found to have no empirical validity. Thus no skeptic who says he is “just following science” can claim that the MBTI is no better than astrology or that the MBTI is pseudoscience without relinquishing his status as a skeptic.

There are personality tests that are more scientific than the MBTI.
True. The MBTI is an old instrument, its earliest roots dating back almost a century, namely to C.G. Jung’s typology. The MBTI was developed prior to most modern means of validating personality tests and so it is not surprising that the MBTI does not pass such validation with flying colors. In principle this should mean that the stage is wide open for a better personality test to supersede the MBTI as the world’s preferred personality test, but so far that hasn’t happened.

One alternative personality test, which has a very high degree of empirical validation and which is often touted as the replacement for the MBTI, is the Big Five test. However, the Big Five instrument has been around for some 35 years now and has consistently failed to stir the same fascination as the MBTI. Common critiques of the Big Five test include that it is inaccessible, boring, and hard to relate to oneself and to one’s everyday life.

Part of the reason that the MBTI is more popular than the Big Five is that people find it easier to relate to 16 distinct ‘types’ than to the series of free-flowing numeric scores that are the result of a Big Five test. Another reason is that the Big Five constricts itself to making claims that can be verified by direct observation. In other words, the Big Five only tells you what it has measured. It does not tell you anything about what these measurements mean. Thus people tend to find the insights yielded by the Big Five to be either boring or commonsensical. To put it plainly, the Big Five only describes personality traits – it does not explain them the way the MBTI does.

In terms of empirical evidence, types don’t exist. Human traits are distributed on a bell curve, not as two camel humps.
True. Empirical evidence suggests that the measurable dimensions of personality are not distributed bimodally (as two camel humps) but fall along a continuum (on a bell curve). This is not what we would expect if types did in some sense exist as blueprints for the human psyche. This discrepancy with the empirical data is the single biggest weakness with regards to the scientific validity of the MBTI. The flaw has been known for quite some time and no empirical solution has yet been found. However, one should note that this way of postulating ‘types’ that do not exist as distinct, empirical entities is no different from the way that the Diagnostic Manual of psychiatrists and psychologists postulate ‘types’ such as narcissistic, compulsive, and paranoid personality ‘types.’ Additionally, one can get around the impasse by regarding the types as a heuristic device (i.e. as a mental shortcut to describe a large body of material) rather than regarding the types as distinct empirical entities.

However, when criticizing the MBTI, we shouldn’t just ask ourselves if we can find fault with it. We can, because no personality test is perfect. So rather than asking ourselves whether the MBTI has any faults at all, we should ask ourselves how the MBTI does compared to other personality tests.

As we have said, the MBTI fails to account for the fact that human traits fall on a continuum (on a bell curve rather than as two camel humps). But the Big Five also has a major weakness and that is the so-called lexical hypothesis. Unlike the MBTI, which is based on the cognitive theory of C.G. Jung, the Big Five was literally derived by looking up adjectives in a dictionary. Again, the Big Five describes personality traits but does not explain them. The strength of the Big Five approach is that you get a high level of empirical validity. But the downside is that, without a cognitive theory to fall back on, the Big Five is essentially inductive and relies on circular logic (e.g. “he is conscientious because he delivers his work on time, and he delivers his work on time because he is conscientious”). By contrast, the MBTI, for all its flaws, was developed deductively.

So while the MBTI has the weakness of ‘types’ not being supported by empirical data, the Big Five has the weakness of not having a cognitive theory to underpin it. This means that the MBTI and the Big Five are not exclusive to each other, but rather complementary to each other. At CelebrityTypes we use both the Big Five and the work of C.G. Jung in our own work. If we want to get an accurate description of somebody’s behavior, we use the Big Five model. And if we want to get an idea of why somebody does what he does, we use the Jungian theory, which is the theory that the MBTI was based on.

People who take the MBTI multiple times often come out as different types (Low Test-Retest Reliability).
True. As we have already noted, human traits are distributed on a bell curve, and not as two camel humps. Because the MBTI takes a continuous scale that goes from 0-100 and divides it into two at the 50% mark, people are bound to come out as different types more often that they would if the MBTI did not use types.

For example, suppose we have a test subject called Amy. Amy takes the MBTI two times and both times she scores somewhere close to the middle on extroversion.  However, the first time she took the test, Amy scored below 50 and therefore came out as an ISFJ. But the second time she took the test, Amy scored just over 50 and was therefore deemed an ESFJ.

amyUsing the MBTI in this scenario thus results in a failure to test and retest reliably as the same type. However, if Amy had taken the Big Five instead of the MBTI, the same variance in score would not have caused her to ‘change type': Even if she scored both above and below the 50% threshold, she would still be the same old, middle-of-the-road, ambivert Amy.

The difference can be visualized like this:


As you can see, a substantial part of the critique that the MBTI has poor Test-Retest Reliability is really the same critique that we examined above (bell curve vs. camel humps), just phrased differently. To put it technically, it is the nature of dichotomous scoring and not the MBTI’s measurements themselves that produces the failure to test and retest reliably. As we have already said, the failure to account for the fact that ‘types’ do not exist as empirically distinct, measurable entities is the MBTI’s biggest weakness. That weakness rears its head again here.

However, as we will discuss below, the MBTI is not an end in itself. It is an attempt to quantify C.G. Jung’s cognitive theory empirically. And as most people who have grappled with Jung’s theory will readily attest, his concepts are by no means as straightforward as the Big Five’s ideas that someone who delivers his work on time is ‘conscientious’ and someone who talks a lot is an extrovert. Jung’s concepts of the cognitive functions are not only staggeringly complex, they also stand in a dialectical relationship to each other where they alternatively imply, negate, and complement each other.

In other words, the complexity that goes into the definition of Jung’s cognitive functions is well beyond what we are currently able to prove or disprove empirically. Since Jung’s cognitive theories were not laid out in a format suited to empirical analysis, the advocates of Jung’s ideas can always claim that the fault lies not with the concepts but with the process of translating and interpreting the concepts into a format that will suit empirical testing.

For example, according to the MBTI, Amy was either an extrovert (ESFJ) or an introvert (ISFJ). But her empirical scores said that she was actually somewhere in the middle. Now if you are critical of Jung, you will say that Jung’s theory thus fails to account for the fact that there are people out there who are ambiverts. But if you are sympathetic to Jung, you will say that according to his theory, people extrovert some parts of their personality and introvert others. So under a pro-Jungian view, even if Amy’s scores are “just in the middle” that doesn’t disprove Jung’s theory, but only the particular interpretation or adaptation of it.

Interestingly, one research paper published in 2014 has found evidence that people who score “just in the middle” on personality traits like those used by the Big Five are beset by so-called internal “dynamic processes.” So the reason that Amy was “just in the middle” could actually be because some parts of her personality were extroverted while others were introverted, just as Jung said.

On the basis of their study, the researchers find that the Big Five has “problems of interpretation for mid-range trait scores” as well as “a compromised theoretical foundation.” That the Big Five has trouble interpreting mid-range trait scores is of course just what every MBTI enthusiast wants to hear, and no doubt some MBTI enthusiasts will construe these results as a victory for the MBTI. But so far, the study is the first of its kind and it’s still too early to say anything definite about what’s going on.

MBTI descriptions contain vague and flattering statements that anyone would agree to (Forer Effect).
True. This is why we don’t feature type descriptions prominently on our site. The world is full of people who have been seduced into thinking they are a certain type simply because they were exposed to a vague and flattering portrait at the time they were introduced to the MBTI.

One reason that the MBTI makes use of such statements is that it was developed to be used in organizations where it was (and is) necessary to get the respondents to accept the system and to self-identify as a given type. The Big Five, which was from the onset developed for use by researchers, did not need to win the acceptance of the test takers in a similar manner.

However, since psychologists are now seeking to employ the Big Five in organizations, they have since found it necessary not only to develop descriptions, but also to postulate types(!). Tellingly, perhaps, these Big Five descriptions also contain vague and flattering statements that induce the Forer Effect. For example, one Big Five type description, developed by a professor of psychology, runs as follows:

“Individualistic types consider themselves to be unique and more intelligent than most people around them. In extreme cases they might be regarded as eccentric, but in most cases they are perceived by others as complex, well-read, imaginative, and industrious.”

Previous studies have shown that almost everyone considers themselves to be unique, complex, and smarter than the average person. At the same time, most people harbor a fear of being ostracized from the group, which is why they are bound to feel that someone has read them correctly when this fear is aired in a non-threatening manner. Finally, almost everyone regards themselves as “imaginative,” which is one reason that people over-identify as N types in the MBTI system.

Two more points: While the MBTI descriptions are indeed peppered with vague and flattering statements, they don’t just induce the Forer Effect. There is usually enough that is concrete in the MBTI portraits for people to be able to recognize the overall ambiance of the type that is being described.

Likewise, it should be noted that just like how type descriptions are not the meat of the Big Five, type descriptions are not the meat of the MBTI either. That is the cognitive theory of C.G. Jung. However, Jung’s theory is very poorly understood and it is not profitable to work on it, as opposed to re-hashing the same old truisms that you find on every MBTI website. Jung had the Rockefeller family sponsoring his work, but today we are not so lucky.

Psychologists and academic researchers don’t take the MBTI seriously.
True and false. Some professors (such as Bryan Caplan) take it seriously and others (such as Adam Grant) don’t. Furthermore, reputable academic publishers (such as Wiley and various psychological journals) often publish scientific studies that make use of the MBTI. It true that the scientific mainstream has converged on the use of Big Five as the standard for academic research, but that doesn’t mean that no scientists are using the MBTI or that the MBTI is unable to capture some things that the Big Five can’t.

The MBTI has often been criticized by the academic community and as a rule, the criticisms have been balanced and fair. However, some of the researchers who criticize the MBTI are completely ignorant of its functioning as well as alarmingly dishonest in the way that they quote from the scientific source material. Likewise, it is demonstrable that some (but not all) of the MBTI’s academic critics have used sensationalism, double standards, and straw men where they should have been offering an academic critique.

When a critique of the MBTI is featured in newspapers and other popular media, it is rarely the unemotional “the MBTI does not give comprehensive information on all domains, but it does measure major personality dimensions and we see need for further research”-type of critique that researchers write for each other in peer-reviewed journals (and which is the type of assessment that is supported by the data). No; what most people have seen when they’ve seen an academic attack the MBTI is almost always a sensationalized version of the critique that overplays or inflates the weaknesses of the MBTI.

Interestingly, one of the most hard-headed academic researchers that ever lived was H.J. Eysenck. He was famous for following the evidence wherever it led. He was also famous for saying that “if it cannot be measured, it does not exist.” And he approved of both Jungian typology and the MBTI (albeit with the qualifications that critical minds always put forth).

Jung’s cognitive functions were good. It was Isabel Myers and the MBTI that simplified the theory so that it came out wrong.
False. Jung’s cognitive functions are a set of postulated psychic processes that are very hard to falsify or quantify. By themselves, they are just words in a book. Jung’s ideas need some form of operationalization in order to be brought into the realm of empirical science. The MBTI is one such possible operationalization of the Jungian theory, the Gray-Wheelwright type indicator is another, and our own Jung Type Test is a third. There are tons of such proposed operationalizations of the Jungian theory out there and the three mentioned here don’t even begin to cover the tip of the iceberg.

However, the most well-validated and extensively tested of these instruments is the MBTI. Almost all instruments that are in use today make use of one or more specific innovations fostered by Isabel Myers, even when they claim to be independent of (or even antagonistic to) the MBTI.

The MBTI makes use of pseudoscientific practices like handwriting analysis (Graphology) and “face reading” (Physiognomy).
False. Both Jung’s typology and the MBTI were developed exclusively as instruments for ordering psychological material and no publication by Jung or any official publication regarding the MBTI has ever relied on handwriting analysis or “face reading” in relation to people’s personality types. They are not even close.

In fact, where Freud was open to using the physical characteristics of the person as a clue to his personality, Jung would repeatedly say that his typology was based exclusively on psychological premises. In the same way, Jung’s authorized biographer, E.A. Bennet, said that we should be careful to note that Jung’s typology “omits any reference to bodily characteristics.” We have examined the question of “face reading” in more detail here.

As for handwriting analysis, that is indeed another pseudoscientific practice. And again, there is no instance of either Jung or the official MBTI publications positing that handwriting has any significance in relation to personality types.

What “face reading” and handwriting analysis have in common with the MBTI is that some enthusiasts – who are not the official proprietors of the MBTI – enjoy mixing these pseudoscientific practices (which have no scientific validity) together with the MBTI (which has acceptable levels of scientific validity).

Conflating the two is like blaming the sweet kid on the block for the actions of his troublesome friends. Now you may say that the sweet kid has some responsibility for not hanging out with the shady kids, but in the real world the publishers of the MBTI can’t control who uses their instrument and for what purposes, just as physicists can’t control how quantum mechanics is espoused by the New Age crowd.

Four dichotomies don’t say everything about human personality. Even if the MBTI isn’t pseudoscience, it is often used by amateurs and often used to explain things that it can’t explain.
True. For example, unlike the Big Five, the MBTI does not say anything about how nervous a person is. Likewise, compared to the MCMI personality inventory, neither the MBTI nor the Big Five says anything about how paranoid a person is. Different personality inventories capture different elements of the personality and the MBTI should not be taken to say everything about the personality. That’s why we always encourage our visitors not to constrict themselves to the MBTI, but to study personality broadly.

However, it is true that there is a tendency among MBTI enthusiasts not to study personality broadly, but to think that the MBTI does indeed explain everything. In our experience, even so-called “MBTI master practitioners” are rarely aware of any complementary perspectives besides the MBTI. This means that the majority of type practitioners, both offline and online, tend to think that the MBTI says everything about the human personality when in fact it doesn’t.

Again it is tempting to blame the MBTI for the dilettantish nature of its fanbase. But when a fool misapplies a tool we don’t usually blame the tool. Nor should we here.

At CelebrityTypes, our contribution has by and large been to take a “minimalist” and purist approach to type where we try to constrict ourselves only to use the system as it was intended: It says something about the arrangement of the four functions and their orientations. All sorts of other factors that pertain to the personality are in effect irrelevant to the system.


MBTI for Skeptics © Ryan Smith, Eva Gregersen, Sigurd Arild, and CelebrityTypes International 2014.

Myers-Briggs Type Indicator and MBTI are trademarks of the MBTI Trust, Inc.

CelebrityTypes.com is an independent research venture, which has no affiliation with the MBTI Trust, Inc.


  1. Michael Harper says:

    Wasn’t Jung’s Theory of Typology developed in a clinical setting and used to identify the polar extremes of psychosis? Such that these extremes, when fully manifested, would be considered undesirable. Wasn’t the goal of his practice to help his patients move from these ‘toxic’ polarities to a more healthy middle ground or balance? Further, if true, wouldn’t we expect to see most scores outside clinical practice falling into a bell curve and not a camel hump?

    For both psychologists and executive coaches assessments of type or trait are not an ends in themselves but rather tools for intervention. Roger Pearman has done great work in operationalizing type development (developing more of one’s opposite or mitigating overuse of one’s type).

  2. admin says:

    Thanks for a good comment.

    It was developed from clinical experience but the people who came to see Jung were not always pathological the way people who come to see psychiatrists today are. Therapy with Jung was also a pastime of the upper classes, including at least one Rockefeller. So it wasn’t that everyone who was a type was psychotic, or even neurotic.

    It is true that Jung saw types as a ‘problem’ to be negated, as we have also dealt with here and here. However, almost everyone after Jung, including some of his closest associates, have abandoned this normative aspect of his typology. So yes, to Jung, the goal was to become whole through the negation of extreme type preferences, but today that is a minority view. (Which is not to say that it is wrong.)

    Jung would agree that most people are indeed somewhere in the middle on most preferences. But like so much else, he did not work out exactly what that would mean in relation to his typology. Thus he left the door open for the popular “you are either X or Y” interpretation of his typology.

  3. Brian Mingus says:

    It’s worth pointing out that whether there are types or continuous distributions of traits is an open research questions. The main reason is that the observed distributions (such as in the Big 5) are underdetermined, meaning that there are a number of ways that discrete cognitive functions could produce distributions that look normal. The most obvious of these reasons is social desirability – there may simply be normal ranges of acceptable behaviors in society and nearly all types can manage to more-or-less behave in what is considered a socially appropriate manner.

    There are also other reasons to suspect that there might be types. The first reason is an evolutionary argument – one of the best ways to occupy a niche is to occupy it fully. Thus, we have introverted and extroverted versions of each of the cognitive functions competing with each other. And the combinations of these types result in a four dimensional binary interaction space, with each type occupying a corner (or niche) and competing with all the others by fully occupying that niche. The way evolution would have selected for the distribution of these types in the population is via group selection, which is also thought to determine the ratio of sexes. Group selection results in what is essentially a meta-parameter that determines the distribution of traits in an offspring’s offspring. The utility of this becomes apparent when we start to discuss more abstract concepts that have been associated with aspects of types. For instance, group selection might have found an optimal ratio of strategists (such as INTJ) vs. tacticians (such as ISTP) that tends to work in groups of nearly any size.

    There is also a more brain-based reason to suspect that their might be types. In cognitive modeling terminology each cognitive function might be considered to be a neural attractor, and it may not be possible or even make sense for two opposing attractors to be simultaneously active. For instance, while one person might be able to use both introverted thinking and extroverted thinking, they might not be able to use both at the same time because these attractors are in competition. Furthermore a given person has, according to the theory, an inborn preference for using one or the other, and while an introverted thinker might be able to emulate the behavior of an extroverted thinking, or do something with their brain that resembles the description of extroverted thinking, neuroimaging has not yet tested whether an introverted thinker can even activate this attractor. Nor has it tested whether the cognitive functions do or do not exist as attractors, leaving Jung’s completely untested by modern cognitive neuroscience.

    This attractor-based reasoning also comes to bear on the notion of an “ambivert” (which I take to mean falls in the middle of one of the four letter code distributions) and whether such a concept even makes sense. Consider someone who is confused as to whether they are an ENTP or an ENTJ as based on the description of P vs. J. They think they are rather orderly with a strong spontaneous streak as well. Now consider what this ambiversion means in terms of Jungian theory: The ENTP has the function stack Ni Ti Fe Si whereas the ENTJ has Te Ni Se Fe. While both of these types use Ni (introverted intuition) in the top half of the stack, the stack is completely flipped when it comes to the orientation of thinking (Ti vs Te), as is the preference for sensing (Si vs Se). So this ambiversion is then theoretically manifested as flip-flopping between the orientation of thinking and sensing, in addition to flip-flopping between the ordering of the functions overall. While it’s not impossible that an ENTP who just had a child and has been forced to live a more structured and planned life might be in such a chaotic state, personality has been found to be largely stable and it seems unlikely that such chaotic flipping between attractor states would be found in a normal, stable person. And just because an ENTP has a child and has J traits doesn’t mean they are suddenly expressing Te Ti Ni Se Si Fe, but rather might mean that they are an ENTP who needs to make sure that stuff gets done on time, which means that the environment has more influence on their personality at the present time than their genes do.

Leave a Reply