A Defense of MBTI (Part 2/2)

Part 1 covered the issues of MBTI with respect to reliability, validity, and missing key variables like neuroticism. Reliability was shown to be strongly present, and validity using objective measures like job performance was seen as ephemeral, while missing variables were seen as being complementary rather than as competing. We made clear the distinction between personality and character. In this section we turn to a few other criticisms:

  1. MBTI doesn’t predict job preferences.
  2. Even if it did predict preferences, it doesn’t predict happiness.
  3. Why should we use MBTI when psychologists don’t?
  4. and finally, should we really be “boxing people in” to rigid type definitions?

You are What You Do

You  may have heard that certain types are drawn to certain careers, f.ex. business students are predominantly Thinker-Judgers (TJ’s)  who tend to choose careers of ‘power and authority’, while Feeler-Perceivers (FP’s) tend to avoid them (Bayne 1997). Garner and Martinko’s reviews found that

“Significant correlations of the MBTI scales with various interest, personality, academic and observational measures, further document the MBTI’s criterion-related validity”.

However, Pittinger critiques this by stating that type preferences are always just equal to “normative” type preferences, or population distributions. He claims, for example, that

“the proportion of ESTJs in the teaching profession is the same as the proportion of ESTJs in the general population, or 12 percent. This similarity suggest that there is nothing special about the type of person who becomes an elementary school teacher.”

That is definitely something to consider, yet his claim is simply not true because

  1. According to the “MBTI’s Manual”, roughly 12 percent is the proportion of males that are ESTJs, but most teachers are female where this type represents 6 percent.
  2. Most elementary school teachers are not ESTJ’s but ISFJ’s and ESFJ’s (see table below).
  3. Moreover, teacher distributions depend on the level of education. For example, SFJs prefer pre-school education and are much less common as college professors. By focusing on ESTJs at elementary schools, Pittinger makes the type preferences look more normal than they actually are. If he were to examine the same thing in High School teachers, he might reach a different conclusion.

Still, let’s adopt Pittinger’s idea. I took on the challenge of scouring the literature for “Type Tables” and of comparing them to gender adjusted populations. Detailed results of the discussion can be found by downloading this Excel file: TypeTables.xlsx.

Adjusting for Gender

The first table shows some occupations, the ratio of males and females, type distributions, and the size of the study. The second table divides percentages of each type by the expected percentages with the same male-to-female characteristics, thus taking into account large differences in types that occur by gender. For example, nurses are predominantly female, but given that 19.4% of women are ISFJ’s, you would need an even higher percentage to show anything interesting… And interesting it is.

Using a combined sample 379 individuals, a high proportion of nurses tended to be thinking-sensor types. Green squares indicate that a type was heavily over-represented in a particular occupation; similarly, red squares means that they were under-represented. I also computed Chi-squared tests for significance, because, well… that’s what a geek would do.

TypeTables

Click to Enlarge. Download Source Here: (TypeTables.xlsx)

Type Table Take Aways

Nearly 50% of retail managers were ESTJ’s, yet Intuitive Feelers like INFP’s and ENFP’s were heavily under-represented. This makes sense because sensing types tend to be patient with routine activities and like established ways of doing things, while intuitive types tend to become impatient with routine activities and may get their facts wrong.

Moreover, there were significantly many more Elementary School teachers who were ISFJ’s, ESFJ’s, and ESTJ’s which is exactly what MBTI theory predicts. ISFJ’s, for instance, lead with introverted sensing (a memory-based process) and enjoy teaching children because doing so reminds them of their own childhood and provides them with a sense of comfort and security. Researchers have said that Sensing-Judging types like things to be under control, appreciating predictability to spontaneity, or that they have an “abiding respect and sense of personal responsibility for doing what needs to be done in the here and now” (Rushton 2007).

It is also interesting that ESFP’s (sometimes called “The Entertainer”) clearly avoided becoming Lawyers and Software Engineers. And ISTJ’s who become experts on any learned skill, seem to be able to fit into pretty much anything that I investigated. Overarchingly, there’s a consistent story that MBTI reflects our interests and our learning styles, and that is a powerful beginning to our journey of self-discovery and our ability to unlock our potential.

 

The Pursuit of Success and Happiness

Pittinger claims that there is

“no data to suggest that specific types are more satisfied within specific occupations than are other types”.

My own thorough investigation of these claims confirmed that personality traits by themselves have little to do with job satisfaction. Research can show that being Extraverted and a Judger makes you happier across the board, but I wasn’t able to find any evidence, for instance, that ISFPs are unhappy programmers.

A study of anesthetists and two nursing studies concluded that satisfaction is a result of being

  1. Extraverted
  2. Easy going and low in neuroticism (which makes you negative and unhappy in general)
  3. Conscientious

This pattern appears to be independent of occupation, thus unexpectedly running contrary to the so-called “person-job fit hypothesis”, and it has been replicated many times (f.ex. Rahim 1981, Furnham and Zacherl 1986). So while personality plays some role, it only marginally accounts for satisfaction, and not in the way you might intuitively expect.

Selection and Survivorship Bias

However, studies that try to measure satisfaction and performance have major issues that cannot be emphasized enough.

SelectionBiasSelf-selection bias occurs when individuals select themselves into a group, thereby causing unusual characteristics. An INTJ saleswoman who wants to go into sales, does so because of her own personality idiosyncrasies or because of peculiarities about her specific job. Yet, in all of the world’s typology studies, all of those idiosyncrasies would be washed away and mistakenly attributed to type. It’s not surprising that an exceptional ISFP is able to code up Java like an average INTJ.

Another way self-selection bias enters the picture is if the INTJ is mistyped as an ISFP, which would make the ISFP look more competent than he or she actually is. Here self-selection bias attracts a disproportionately large number of mistyped individuals into the study sample.

Survivorship bias is the logical error of looking only at the people that “survived” a selection process, like getting that dream job. Performance and satisfaction studies are not randomized controlled experiments so they leave out potentially hundreds of INTJ’s who tried a sales job and were fired after only a few miserable weeks.

Apple and Google started their business in the founders’ home garage, yet buying every tech stock that ever came out of a garage is an abysmal investment strategy (Baker 2011). Also check out all the highly-rated conspiracy books on amazon. Are they really that great? Or do they get 5 stars because only a conspiracy theorists would buy them? You may recognize these issues as the “silent evidence” from Nassim Taleb’s Fooled By Randomness.

Ultimately, self-selection and survivorship bias means that these studies are all underpowered to detect the effect that personality has on performance and satisfaction. If this effect is strong, which I suspect it is, then these studies are close to useless (thinly discussed in the literature, but see f.ex. Bealing 2006 and Borges 2002 for further discussion).

Direction and Magnitude

In spite of these biases, let’s discuss the evidence as is, granting them the credibility they don’t actually deserve. Recent evidence suggests that both personality factors (like Judging) and ability (like intelligence) need to be taken into account and can explain as much as 20-30% of the variance in “real life” outcome measures.

“Thus there is now increasing evidence that both ability (high general intelligence) as well as personality factors, particularly stability (vs. Neuroticism) and Conscientiousness predict exam success, but also employee absence, job satisfaction, career success across the life span as well as actual earnings. Together personality and ability factors may account for as much as 20-30% of the variance in “real life” outcome measures. In this sense they have extremely salient predictive validity.” (Furnham 2007)

You may have objected that this quote refers to the Big Five factors, and not to MBTI, but as we’ll see below the two models are actually quite similar. The Conscientiousness factor of the Big Five is a lot like the Judging dimension in MBTI.

pablopicassoSo this takes us back to my new favorite notion that leading a happy and successful life has two parts: finding your gift and then making something out of it, exactly like the Picasso quote to the left. Although, he may actually have it the other way around, because “meaning” implies an establishment of worth or value and I don’t believe that our value is established by our ability to find our gift. Rather, I think that our “purpose” is to find our gift and it is only then that meaning ensues. The etymology of the purpose is suggestive of one’s intentions and aim, like an archer finding her target. A physics geek (or Minions fan) might think of life as a vector, which has both a direction and a magnitude (“Ohh Yeaah, Baby! Oh Yeah!“).

The Paradox of Pursuits

Happiness has a funny zen-like quality in that pursuing it actually makes you unhappy. We desire money, yet earning it is unsatisfying. Similarly, our personalities influence our desires towards certain professions, yet achieving those desires may have little effect. MBTI can tell us what direction we might want to take, but without an emphasis on virtues and meaning, it can’t tell us whether we will be successful.

Good Counselling Bad Selection

MBTI has nearly always been regarded more as a counselling than a selection instrument (Furnham 2007). It is more useful in guiding someone towards an industry and it is less useful in evaluating somebody’s candidacy for a particular job. Even if MBTI doesn’t help to establish meaning in our lives, I’d say that being able to establish our purpose is a pretty darn useful thing, and that is the point that Pittinger is missing.

Psychologists Eat Oranges (and MBTI is a tangerine)

It has been pointed out that psychologists don’t use MBTI, so why are we? Well… because a logician would call that an appeal to authority and the problems should be obvious in stating that because “Psychologists don’t use MBTI” then it must be false.

MBTI is a special case of the more general FFM (Five Factor Model, aka Big Five or OCEAN) and can be understood within this system. As Bayne notes in his book

“The importance of the research on five factor theory for the validity of the MBTI rests on the fact that four of the Big Five factors are closely related to the preferences.”

So even if the Big Five dominates in academia, Bayne adds:

“Research on them is therefore in effect research on the four preferences and supports – in a ‘piggyback’ way – this aspect of the MBTI’s validity”.

Thus, the fact that “Psychologists don’t use it” (but in fact use something very similar) is a disingenuous argument when it makes it appear that MBTI is something completely different than what is actually being used. The chart below shows how very correlated the two methodologies are. As a point of reference, anything less than 0.09 should be considered random and not statistically significant.

Part of Table 3, McCrae and Costa (1989, p. 30), N=468, cited in Bayne (1994). Correlations all given as positive and all other correlations omitted for simplicity of presentation.

Part of Table 3, McCrae and Costa (1989, p. 30), N=468, cited in Bayne (1994). Correlations all given as positive and all other correlations omitted for simplicity of presentation.

False Dichotomies

MBTI is often heavily criticized because it forces people into artificial and predefined buckets. And it is true. They are completely arbitrary and there are no bimodal distributions in the continuous scores to justify different classifications. “Bimodal” is a fancy way of saying that you are typically either one or the other, like these ants:

Histogram of body lengths of 300 weaver ant workers

Histogram of body lengths of 300 weaver ant workers

One could quite rightly argue that MBTI scores don’t support any kind of statistical structure that could be used to justify the 16 types. Rather, the scores suggest a continuous range of traits and that the types are a false dichotomy.

Tall People Do Exist

However, the same could be said for tall and short people, where there aren’t any bimodal distributions that allow us to naturally classify a person as one or the other. Commensurately with MBTI, there’s a continuous range, yet, that doesn’t mean that classifying people as “tall”, “average” or “short” is not a meaningful and useful exercise. Consider how we generally buy clothes based on arbitrarily chosen sizes (S, M, L, etc..). Below is an example of K-Means clustering from a Stanford Machine Learning class I took on Coursera.org. As you can see the arbitrary circles or “clusters” don’t actually exist, but we still perform these kinds of statistical exercises because they lower the cost of manufacturing while simultaneously providing most people with a reasonable fit.

KMeansShirt1 KMeansShirt2

 

 

 

 

 

 

As with personality types, the cutoff points for shirt sizes are completely arbitrary, but can be chosen so as to be useful. Carl Jung was acutely aware of this:

“Every individual is an exception to the rule… the classification of types according to introversion and extraversion and the four basic functions [is not] the only possible one… The four functions are somewhat like the four points of the compass; they are just as arbitrary and just as indispensable. Nothing prevents our shifting the cardinal points as many degrees as we like in one direction or the other, or giving them different names. It is merely a question of convention and intelligibility. But one thing I must confess: I would not for anything dispense with this compass on my psychological voyages of discovery” – Carl Jung

Myers (1962) herself recommended that the MBTI is best viewed

“as affording hypotheses for further testing and verification rather than infallible expectations of all behaviors”

Ultimately, classifying people into types is no stranger and no less useful than describing somebody as tall or short, and moreover, our test results may not even be the final story.

Boxing Hurts

BoxingGloveOne problem with the above (as great as it sounds to me) is that it only deals with the benefits. The availability of small shirt sizes is great for my small stature, but if someone says to me “Hey, shorty! Can you ever see over the steering wheel?” that kind of verbal pugilism might cause a little rebellion.

Labels are powerful and obviously serve as building blocks to language. Some societies don’t have a word for blue (<- amazing video) and are thus unable to easily distinguish blue from green. They can open up our powers of perception, but labels can also degenerate into stereotypes, racism, and war. In the corporate world, MBTI might be used to discriminate or to impose limiting beliefs on employees. And unfortunately that dumps us unto a moral battlefield about whether or not we should get rid of labels and be more politically correct.

My take on the potential evils of “boxing people in”, and yours may be different, is that

  • There is nothing in the theory that says that everyone in their type must behave according to a certain rule. In fact, Carl Jung says the opposite, that everyone is an exception to the rule.
  • Myers says that MBTI gives you little more than a hypothesis (not a theory with infallible expectations)
  • MBTI doesn’t define you, rather the opposite is true. First and foremost, your identity is your own and all that MBTI can do is to describe you.
  • MBTI gives us a system of labels, or a vocabulary, to discuss personalities and this is a powerful tool.
  • With great power comes great responsibility and it is up to the user to abide according to the highest moral standards. Boxing people in is a user-error not a software bug, so don’t blame Microsoft for this one.

Whether you blame The Myers-Briggs Foundation for their “software bug” or not will probably depend on your moral instincts. Advocates of the Kantian categorical imperative would blame the user every time. A consequentialist like Bentham or a moral relativist might do more to weigh the pros and cons. When it comes to the morality of it, the debate will never end.

But when it comes to the statistical theory and if it is data you want to talk about, I can tell you that a bimodal distribution isn’t a prerequisite for utility.

Just the Happy Positives

CarlJungTreeMBTI is often said to be so popular because it only tells you the positives and that is what people want to hear. Yet, this couldn’t be farther from the truth. Deeply embedded in the theory, it is clear that MBTI is about trade-offs, which is why no type is better than another. It explains your talents in light of your weaknesses and vice versa. MBTI resembles Newton’s 3rd Law of Motion (every action has an equal and opposite reaction) in that every talent is gained by suppressing another. In the Newtonian spirit, Carl Jung said that “No tree can grow to heaven, unless its roots reach down to hell”.

An INTP, for example, achieves her creative genius only because of her doubtful, insensitive, and forgetful mind. She sacrifices being aware of her environment in order to cultivate her passions and MBTI is so powerful because it connects these two seemingly disparate modes of being and offers an explanation, but not an absolution, for her weaknesses. The weaknesses are unequivocally present in any MBTI personality report and it would be unbalanced to shine the light on either side of the scale.

The weaknesses are also key to resolving conflict between two different types in a team effort, so that a person’s seemingly persistent, annoying hesitancy to reach a conclusion and to move on to other pressing needs can be understood in light of creative strengths that allow that person to contribute in other meaningful ways.

Conclusion

Critics who claim MBTI is astrological pseudo-science take their position too far, yet it is inappropriate to use the tool as a selection criteria, especially as a means to establishing the worth of somebody’s candidacy. There are too many variables to assess someone’s vocational ability and relying too heavily on MBTI as a selection-criteria can cause one to mistake the forest for the trees.

do feel that MBTI’s reputation at HR departments exceeds its abilities and in this respect Boyle and Pittinger raise very legitimate concerns. However, I also found them to be guilty of occasionally cherry-picking and ignoring the statistical significance of some results. For example, most researchers have found MBTI to be very consistent and reliable with test re-test rates of about 80%. This is 13 times higher than the probability you might expect from a field like astrology, which creates serious problems for those who think that people only love their MBTI description because they’ve been duped by the Forer effect.

Its validity, however, is more ephemerally defined, leaves us wanting additional research, and has narratives that could arguably be spun both ways. It is unsatisfying that a self-reported questionnaire like MBTI could not be objectively and exogenously validated.

In spite of this, however, types tend to choose occupations that are aligned with their personalities and interests. Whether a person is ultimately happy and is successful at work appears on the surface to have less to do with their personality and more to do with their virtues: their work ethic, intelligence, emotional stability, etc. However, when thinking jointly of personality and ability, they seem to explain 20-30% of variance of things like performance and job satisfaction. And this is in spite of there being a considerable amount of self-selection and survivorship bias, which makes many of these studies seriously underpowered to measure what they purport to measure.

By focusing on preferences, MBTI misses things like Neuroticism, but it does so on purpose. Neuroticism, we argued was a personality pathology that doesn’t fit into a framework based on Newtonian trade-offs and sacrifices. There are many traits, and especially virtues, which complement the theory rather than compete with it.

No one should box anyone in or stereotype their potential. I know it happens prevalently in the corporate world, but that wasn’t the intention of MBTI’s creators. Like so many technologies, we each have an individual as well as a collective responsibility to use them for good.

“Not anyone can become a great artist, but a great artist can come from anywhere.” – Jan Pinkava (Ratatouille)

References

  1. Baker M., Bradley B., Wurgler J.; Benchmarks as Limits to Arbitrage: Understanding the Low-Volatility Anomaly; Financial Analysts Journal, 2011, Vol. 67, No. 1
  2. Bayne, Rowan; The Myers-Briggs Type Indicator: A Critical Review and Practical Guide; 1997
  3. Bealing W., Baker R., Russo C.; Personality: What It Takes To Be An Accountant; The Accounting Educators’ Journal, Volume XVI (2006), pp. 119-128.
  4. Boyle Gregory; Myers-Briggs Type Indicator (MBTI): Some psychometric limitations, 1995
  5. Borges N., Savickas M.; Personality and Medical Specialty Choice: A Literature Review and Integration; Journal of Career Assessment 2002; 10; 362
  6. Bratko, D. T. Chamorro-Premuzic, Z. Saks; Personality and school performance: Incremental validity of self- and peer-ratings over intelligence, Personality and Individual Differences 41 (2006) 131–142
  7. Carlson, John; Recent Assessments of the Myers-Briggs Type Indicator; Journal of Personality assessment, 1985, 49, 4.
  8. Carparo, Robert; Carparo, Mary; Myers-Briggs Type Indicator Score Reliability Across: Studies a Meta-Analytic Reliability Generalization Study; Educational and Psychological Measurement 2002; 62; 590
  9. Furnham A., K.V Petrides, Chris J Jackson, Tim Cotter; Do personality factors predict job satisfaction?; Personality and Individual Differences, Volume 33, Issue 8, December 2002, Pages 1325–1342
  10. Furnham A., Georgia Dissou, Peter Sloan and Tomas Chamorro-Premuzic; Personality and Intelligence in Business People: A Study of Two Personality and Two Intelligence Measures; Journal of Business and Psychology, Vol. 22, No. 1 (Sep., 2007), pp. 99-109
  11. Furnham A., Zacherl M.; Personality and job satisfaction; Personality and Individual Differences, Volume 7, Issue 4, 1986, Pages 453-459
  12. Grant, Adam; “Goodbye to MBTI, the Fad that Won’t Die”; Huffington Post 2013; Accessed Mar 22, 2016.
  13. Pittinger, David; Measuring the MBTI… And Coming Up Short
  14. Pittinger, David; Cautionary Comments Regarding the Myers-Briggs Type Indicator; Consulting Psychology Journal: Practice and Research, Vol. 57, No. 3, 210-221
  15. Mark F. Schilling, Ann E. Watkins, William Watkins; Is Human Height Bimodal?; The American Statistician, Vol. 56, No. 3, (Aug., 2002), pp. 223-229
  16. Moutafi J., Furnham A., and Crump J.; Is Managerial Level Related to Personality?; British Journal of Management, Vol. 18, 272–280 (2007)
  17. Podsakoff, P.M. & Organ, D.W.. Self-reports in organizational research: Problems and prospects. Journal of Management, 1986, 12: 531-544
  18. Rahim, A.; Job Satisfaction A Function Of Personality-Job Congruence. A study with Jungian Psychological Types. Psychological Reports: Volume 49, Issue , pp. 496-49
  19. Roelofs, J., Huibers, M.; Effects of neuroticism on depression and anxiety: Rumination as a possible mediator; Personality and Individual Differences ,Volume 44, Issue 3, February 2008, Pages 576–586
  20. Rothmann, S, Coetzer E P; The Big Five Personality Dimensions and Job Performance; Journal of Industrial Pscyhology (2003).
  21. Rushton S., Morgan J., Richard M; Teacher’s Myers-Briggs personality profiles: Identifying effective teacher personality traits; Teaching and Teacher Education 23 (2007) 432-441
  22. Thomas, A, Buboltz, W., Winkelspecht, C;  “JOB CHARACTERISTICS AND PERSONALITY AS PREDICTORS OF JOB SATISFACTION“, Organizational Analysis, (2004) Vol. 12 Iss: 2, pp.205 – 219

About Esotariq

Quantitative Finance Professional with a passion for happy living, self-improvement, nutrition, and minimalist running over maximalist distances.