[Blogathon] Bimodal Personalities and the Myers-Briggs

[Part 1 of a topic suggested by @mmace134]

The Myers-Briggs Personality Test is very, very popular. It’s used as a predictor of career paths, of relationship styles, of leadership ability. Basically, you take the test and answer a bunch of questions like “Do you enjoy having a wide circle of acquaintances?” (To which you can only respond “Yes” or “No”. Based upon tens of answers, you’re given a four letter code, like INTJ or ENFP. Each letter codes for a specific trait, with two possibilities for each.

I/E: Extraverted/Introverted

S/I: Sensing/Intuitive

T/F: Thinking/Feeling

J/P: Judging/Perceiving

This all seems fairly reasonable–we can all agree that some people are more extraverted and some are more introverted. Some people make decisions based on feelings and some people don’t.  The problem is, the test is based on the idea that people are one or the other–that is, that most people are overwhelmingly extraverted or overwhelmingly introverted. If that was the case, we would expect a graph of scores of introversion and introversion to look like this:



That’s a bimodal graph–one with two peaks. Those should represent the extraverts and the introverts (or the Thinking people and the Feeling people, or the Sensing and the Intuitive people).

The problem is, what we actually get is this:


…a unimodal graph. A bell curve. A normal distribution.

Most people fall near the center of  the Introversion/Extraversion, Thinking/Feeling, Sensing/Intuitive, and Judging/Perceiving spectrums.  Of course, there are people scoring very highly for one or the other–but they’re not the norm. This wouldn’t be terribly problematic if the Myers-Briggs didn’t insist on divvying people up into one or the other, which they do by splitting the responses down the middle.

That means if I’m just slightly to the left of center, because I’m pretty extraverted, but I don’t enjoy having a huge group of friends, I’m in the Introverted category, along with everyone who thinks gatherings of more than three people are hell. But aren’t I closer in type to people who are just to the left of center? Yep.

And nearly half of the research on the Myers-Briggs is done by institutions that benefit or publish the test in the first place. [*makes skeptical face*]

Want a validated personality test? Try the Big Five Personality Inventory

Have a post topic? Tweet me! Even more importantly, donate to Secular Students!

[Blogathon] Bimodal Personalities and the Myers-Briggs

12 thoughts on “[Blogathon] Bimodal Personalities and the Myers-Briggs

  1. 1

    I have struggled with depression for most of my adult life (I’m 56 now), and my experience with Meyers Briggs is that my results vary depending upon how well I have my illness controlled at the time. The big one around here now is based on colours. I’m no longer working so I have no direct experience with the test, but from all accounts it is just as useful (less) as the others.

  2. 2

    Anyone with a basic understanding of psychology can manipulate that test to give whatever answers they want, making it completely useless for most of its intended purposes.

  3. 3

    Several of my friends love Myers-Briggs because they happen to be outliers on two or more dimensions, making it a useful tool for describing their [somewhat rare] personalities to others. I, on the other hand, do not score consistently on any dimension of that test besides the “N”… I always come out as intuitive. I’m an easily overwhelmed person and have anxiety issues, which make me lean towards introversion, but I’m also rather enthusiastically social in a lot of ways… I don’t like being the center of attention, but I like being around people, especially playing hostess.

    On the other hand, my big five results are pretty consistent. Yay for more valid tests!

    I’m a O80-C25-E53-A38-N90 Big Five!!

  4. 4

    I’ve had a lot of experience in analyzing political-quiz scores, though it’s amateur experience. It helps to get a number for each dimension of quiz result, otherwise one doesn’t have much to work with.

    From the looks of it, the MBTI is not very useful without numerical values of its personality dimensions.

    There’s also the question of how well-supported it is. The Big Five are the result of asking lots of questions to lots of subjects and then looking for correlations in their answers. Has anyone tried to use that approach to test the MBTI?

  5. 6

    Have you read Barbara Ehrenreich’s Bait and Switch? In the course of her searching for a corporate job, she had to take (at the behest of her “coaches”) the MTBI test along with something called the Enneagram:

    Even the more superficially rational of these tests, the Myers-Briggs Personality Type Indicator, possesses not a shred of scientific respectability according to Annie Murphy Paul’s 2004 book, The Cult of Personality. It was devised, in the early forties, by a layperson — a homemaker in fact — who had become fascinated by her son-in-law’s practical, detail-oriented personality, which was so different from her own, more intuitive, approach. Inspired by the psychoanalyst Carl Jung’s notion of “types” — which were by no means meant to be innate or immutable — Katharine Briggs devised a test to sort humanity into sixteen distinct types, all of them fortunately benign. (There were no psychopaths, of the kind who might show up at work one day with an automatic weapon, in Briggs’s universe.) To her eternal frustration, the test never won respect from the academic psychology profession, and not only because of her outsider status. Serious psychologists have never been convinced that people can be so readily sorted into “types.”
    Leaving aside the validity of “types,” the Myers-Briggs Type Indicator has zero predictive value even in its own terms. In one study, undertaken by proponents of Myers-Briggs, only 47 percent of people tested fell into the same category on a second administration of the test. Another study found 39 to 76 percent of those tested assigned to a different “type” upon retesting weeks or years later. Some people’s “types” have been found to vary according to the time of day. Paul concludes that “there is no evidence that [Briggs’s] sixteen distinct types have any more validity than the twelve signs of the zodiac.”

    1. 6.1

      @cactuswren I know the Enneagram, too – it’s weirdly esoteric. However, I do think that these personality types make more sense than the zodiac signs, considering that one sorts you by self-observation and the other by your date of birth.
      I think that all personality tests (and IQ tests and things like that) have to be taken with a huge grain of salt, but like most people I love getting to know more about myself, which is why I can never keep myself from taking them. ūüėČ I do think there might be some merit in the more complex ones, at least in providing food for thought.

  6. 7

    I’ve had many an issue with the MB test. I routinely score at close to or exactly 50/50 for E/I and J/P. I tried the big five test and it seemed like many of the same questions framed slightly differently. Some of the results I agreed with, others less so.

    One area I think personality tests fail is the difference between a person’s private and public face. I can only speak anecdotally, but while left to my own devices, I tend to be very introverted, I also have a lot of wide-ranging interests, and the only way I can pursue them is to get out there and stick my public face in places my private face freaks out about slightly when considering. So it’s not that I’m naturally organized or outgoing, it’s that I work really freaking hard at being so.

  7. Wes

    I have a friend who is a professional in the field of personality studies, and is somewhat of an expert on the MBTI. I sent him this post, and he gave me permission to post his response (with the caveat that it was intended as a personal response, so please forgive anything that comes off as harsh or aggressive, as it was not intended that way):


    Short version, there are a number of incorrect assumptions going on. ¬†Not that I really blame Kate, a number of her misconceptions are perpetuated by a number of “so called’ mbti experts.

    For example, this statement is wrong:
    The problem is, the test is based on the idea that people are one¬†or¬†the other‚Äďthat is, that¬†most people¬†are overwhelmingly extraverted or overwhelmingly introverted

    Yes, the “test’ places someone on one side or the other, that is NOT the same thing as saying “most people are overwhelmingly extroverted or introverted”.

    Hence all the fancy graph stuff in her post is not relevant.  Further, the test is using dichotomies but it is to get to a WHOLE type and the individual elements are not truly intended to be interpreted without the other traits.  MBTI experts are supposed to spend time covering this, but it is far easier to discuss the individual facets.  Most presenters barely even cover what the theory is actually supposed to be talking about.  My fancy boxes that you saw in my presentation are an attempt help explain the true theories more easily.  For example
    ENTJ – Extroverts their Thinking as their dominant function and Introverts their Intuition as a secondary.
    INTJ РIntroverts their Intuition as their dominant function and Extroverts their thinking as a secondary.  

    Flat out, the default assumption is that all people both extrovert and introvert. ¬†That doesn’t sound like a bimodal graph at all to me.¬†

    The MBTI has gone through stupidly large quantities of validation. ¬†Even if we assume Kate’s statement “nearly half of the research on the Myers-Briggs is done by institutions that benefit or publish the test in the first place.” ¬†That still leaves half of the research¬†done¬†by people that don’t benefit.

    Yes, big five or five factor model uses traits as opposed to the myers briggs using “type”. ¬†There are differences in how these two type of systems work. ¬†Being frustrated that a type system doesn’t work like a trait system is like getting frustrated with potatoes because they don’t grow on trees. ¬†Like the bimodal graph above does not properly apply to a non-trait system.

    Let’s step a moment into support for the systems. ¬†Many Psychologists and Psychiatrists (most of who get paid for helping to fix people) find it much easier to use a trait system which will give you a score, that they can then point out that a person is lacking “extroversion” or has too much “neuroticism” and then work on it to improve your score. ¬†In general, the traits are seen from a good score to a bad score rather than both sides having equal value. ¬†A type system is not as useful in that kind of environment. ¬†All the systems are¬†incentivised¬†by money at some point in our capitalist system.¬†

  8. 9

    As the previous post pointed out, just because someone is one or the other does NOT make the person overwhelmingly one or the other.

    However, if you start with a ridiculously flawed premise it does follow that you will end up with a annoyingly pointless post.

  9. 10

    IN short bullets:

    a) The MBTI does not say you “are one thing or the other” that MBTI results show a _preference_.

    b) It’s a type preference, not a behaviour preference. (not a trait)

    c) The MBTI has been validated

    d) Wes’s friend (above) explained everything quite nicely.

Leave a Reply

Your email address will not be published. Required fields are marked *