On the Myers-Briggs Type Indicator ®

Many people presume, when you tell them you're a psychologist, that you spend your time doing Freudian psychoanalysis. That presumption can be avoided somewhat by specifying "personality psychology," but this title often leads to another misunderstanding -- that you spend most of your time working with the Myers-Briggs Type Indicator®. In truth, I've never worked with the Myers-Briggs and I don't know any psychologists who use it. I've only seen it mentioned in peer-reviewed journals a handful of times and none of those references were in influential articles. Yet my work with the SAPA Project has made it clear that a large percentage (most?) of the "personality" resources available on the web have something to do with either the MBTI® or one of the unofficial MBTI® knock-offs. Several of these sites are very popular (though some are much better done than others - this one is particularly good). For each of the 16 personality types, there are a surprising number of active user groups and forums, some with 1000's of members. Not bad for a measure that is largely ignored by the research community. So, why the disconnect between the science of personality and the ways that non-scientists evaluate it?

Well, we should start by acknowledging that there are a number of viable theories for personality and it would be inaccurate to say that only one is "right." It is sometimes possible to say that one theory is more right than another, but it's often difficult to make a compelling case for or against a specific theory before you can figure out how to measure it. Carl Gustav Jung proposed a theory of psychological types in 1921 [1] and it seemed as likely to be correct as any going theory at the time (or even now, if I'm being impartial). Unfortunately, he did not propose a methodology for testing his theory, much of which revolves around the role of unconscious factors affecting motivation. The Myers-Briggs Type Indicator® was developed decades later on the basis of Jung's ideas (more or less -- there are some discrepancies [2]). Evaluating the reliability and validity of measurement tools developed on the basis of theory is a straight-forward process thanks to the development of several well-defined metrics. It's on the basis of these metrics that most active psychologists are skeptical of the MBTI®.

An excellent review of these criticisms can be found on the Wikipedia entry for the Myers-Briggs Type Indicator®. In fairness to the Myers & Briggs Foundation, they claim that the reliability is "quite good" and the validity is "proven," but none of the three sources they provide in support of these claims were published by independent groups[3]. In other words, these are not arms-length, peer-reviewed citations. Independent researchers have been unable to replicate these values (or, if they have, they don't appear to have been published). This lack of objective support hasn't made much difference however -- the popularity of the MBTI® has remained very high among the general public.

It has also been proposed that the MBTI® is popular because the Myers-Briggs is easier to understand than other personality models. I think this claim is completely backwards - it seems to me that the Jungian model is considerably harder to understand than the Big Five, which was empirically developed around the words that people already use when describing personality. Frankly, I don't even fully understand Jung's model and I've given it more thought than I care to admit. And in any case it seems to have been developed out of a motivation to reconcile theoretical discrepancies between the theories of Adler and Freud. While that's not an unreasonable motivation, it is problematic now that both of their theories have fallen out of favor as well (due to lack of empirical support and measurement difficulties). I recognize that many people are drawn to Jung's work across a wide range of topics, but he is still only one voice among the many who have studied personality over the years. More than a few of the other voices have belonged to researchers of remarkably high intellect and knowledge who have made the study of personality their primary focus rather than a short-lived theoretical pursuit. Several have also gone to great lengths to develop theories that encompass the full range of personality. It's not clear that this is true of Jung's theory. The MBTI®, for example, does not generally account for the dimension of Emotional Stability (aka Neuroticism), despite the evidence that this dimension is predictive of several important life outcomes.

It seems to me that the Myers-Briggs is the most recognizable brand in personality assessment because of their typological approach and their marketing. As the saying goes, "there are two types of people in this world: those who like typologies and the rest of us." It turns out that people find it fun/pleasing/satisfying/relieving/simplifying to be told, "you are this type of person" even though this is not generally how nature works. People vary in degrees across a massive number of individual differences. And most people are about average on most of those differences. This tendency towards averageness is exacerbated when all those little differences are aggregated into a few groups (four dimensions in the case of the Myers-Briggs). Typologies like the Myers-Briggs don't account for these normal(-ish) distributions and, while infinite variety can be overwhelming at times, I find this willful oversight to be equally distressing. It's no more reasonable to describe personality in typological terms than it is to say that everyone is either tall or short, fat or skinny, smart or stupid, strong or weak.

But there are plenty of people who vote against me with their pocketbooks. It costs about $50 a person to take the "official" copyright version of the MBTI® online (see here) and this is a substantial discount to the recommended one-on-one assessment. Those seeking qualifications to administer the MBTI® are required to undergo training on its ethical use at considerable personal expense. These costs suggest to me that those who invest in learning their MBTI® type are less likely to evaluate the framework critically and more likely to become word-of-mouth marketers among their friends and co-workers. It is also possible to find unofficial measures online that claim to mimic the MBTI® but, in addition to having questionable ethical and legal status, these measures are not likely to be any more reliable or valid than the copyrighted version.

This all leads to an important question: should personality scientists take steps to make their measures more appealing to the general public? The answer is apparently "no" as little action has been taken to date. That's probably as it should be. The most simple step would be to "typologize" more valid and reliable personality scales. This already happens to some extent, as a cognitive shortcut, but it shouldn't be encouraged. That said, it would be useful to determine the extent to which some types or profiles of personality traits occur more frequently than would be expected at random. These occurences might suggest some overlap between the traits or maybe even a common, underlying mechanism. There are several promising lines of research working in this direction already.

In any case, it seems unlikely that personality scientists will embrace the Myers-Briggs Type Indicator® as part of mainstream personality research without revision. And this probably means that there will continue to be a gap between popular opinion about personality and the most valid theoretical models for the foreseeable future.

[1] Jung, C.G. (1923). Psychological types: or the psychology of individuation. Oxford, England: Harcourt, Brace.
[2] Myers-Briggs Type Indicator. (n.d.). In Wikipedia. Retrieved August 28, 2013, from http://en.wikipedia.org/wiki/Myers-Briggs_Type_Indicator
[3] Myers & Briggs Foundation. (2003). Reliability and Validity of the Myers-Briggs Type Indicator® Instrument. Retrieved August 28, 2013, from http://www.myersbriggs.org/my-mbti-personality-type/mbti-basics/reliability-and-validity.asp.

This page last modified on August 28th, 2013.