MEASURE FOR MEASURE: Looking at an Australian Critique of the MBTI

Although the MBTI continues to be popular over a wide range of applications and disciplines, its reception in the field of psychometrics has been cool, to say the least, and this is quite well documented in the literature, as well as being presented that way in MBTI Qualifying/Accreditation Programmes.

Somewhat surprisingly, this view doesn’t seem to be held by some well-known figures in the area: Anne Anastasi and Harrison Gough for example and from listening to them in Boston in 1997, Theodore Millon and John Mayer.

Mayer, incidentally, is one of the authors of the concept of Emotional Intelligence who, from what I heard, wouldn’t agree with some of the key claims made about his idea in the best-seller by Daniel Goleman, but that’s another story.

Historically, it seems that when a psychometric critique of the MBTI is published, the MBTI community as a whole ducks, and hopes the critique will go away. A vigorous counter-challenge seems out of the question.

Dan Foster’s series in the Australian Journal of Psychological Type (Vols 6 and 7) is partly about that. Whether this happens because “the cost of disagreeing is too high” as Bruno Latour (1985) would have it, is another matter.

People who want to respond also may not think they know enough.

This view sometimes presumes that the MBTI critiquer does know, which from my research and experience is more unlikely than likely.

It often comes down to whether the MBTI is liked or disliked (Geyer 1995).

This might seem a bit cavalier, but there can often be negative consequences for the MBTI regardless of the quality of the critique, and there might be no response from people using or researching the MBTI who might be expected to know.

In Australia, the most current example of this is the article by Boyle, set for review below, and its response.

Shortly after its publication, the Army Reserve banned the use of the MBTI, citing his paper.

Quite recently, I have had it referred to me from two sources as a significant article in the field.

This is notwithstanding serious errors of fact which to my mind, are compounded by a confusing writing style.

If you want to check this out, read the relevant issue of Australian Psychologist, or ask the author, at Bond University, for a copy.

Boyle’s stated purpose is to suggest that psychometric limitations of the MBTI, together with a lack of local norms, lead him to suggest that “routine use of the MBTI is not recommended” notwithstanding its popularity in applied settings.

He begins the article by describing the MBTI as designed to “quantify” Jung’s theory of types, which is not strictly true, given Myers and McCaulley’s description of the purpose of the MBTI as not to measure, but to sort (p.1985).

Correctly defining the purpose of the four-letter MBTI codes, Boyle then says that “for a simple.... description of ... personality”, the MBTI seems appropriate, indicating that he hasn’t read the theory parts of the MBTI Manual (a book he refers to in the article) particularly closely, as the MBTI provides insight into Jung’s quite complex view of personality.

Referring to the scoring method, Boyle describes MBTI items as “forced choice (true/false)”, a statement again contradicted in the MBTI Manual, where no such claim is made.

In describing the dimensions/dichotomies of the MBTI, Boyle seems to prefer descriptions other than Myers and McCaulley, sketchily and with predictable results.

He neglects to add the energy source component of Extraversion/Introversion, and doesn’t acknowledge Jung’s construct as the original formulation.

His description of Sensing /Intuition is similarly loose, in defining Intuition.

In an incongruous (read it and you’ll see what I mean) sentence that looks like it’s been cut and pasted in the wrong spot, he loosely aligns T-F with field dependence-independence, and then seems to dismiss the notion.

J-P for him (following a review in Test Critiques) determines which function is dominant and which is auxiliary, which is not true at all, and he indicates that the Auxiliary is used only in exceptional circumstances, which is news to this introvert.

Pearman and Fleenor (1996) have recently done some work in this area, which, in fairness, Boyle would not have had access to at the time of writing his paper.

Boyle then launches into a brief description of the 5-Factor model of Personality, currently the big thing in personality research, although it is certainly not new, dating back to the 1940s, and places MBTI scales within this construct.

It’s quite possible to do the reverse, of course and some current research indicates that the MBTI is a better instrument than the Big 5 favourite, the NEO-PI (see e.g. Newman (ed) 1995 from a 1993 presentation).

Incidentally, there are no local norms postulated for the NEO-PI, which, from Boyle’s opening perspective, puts it into the same boat as the MBTI.

One of the interesting things about Boyle’s article is the lack of reference to recent work and lack of research into the sources he uses.

He discusses MBTI Forms without mentioning Forms J and K (both out since 1989) and the two self-scoring Form Gs (1989 and 1993 respectively), but mentioning Form AV, withdrawn from sale in 1987.

It’s hardly a critique if you’re not up to date in your knowledge.

Most of the critique, in fact relates to data from the 1985 MBTI Manual, and there’s lots of research done since then.

This tendency to ignore later data means that Boyle must understate available MBTI research MBTI significantly, given the 4000+ articles in the CAPT databank.

The next parts of this article deal with scoring issues. As we all know, the bipolar scoring method used in the MBTI follows from the theory; the notion of traits, continuous scores, normal distributions and amounts is a different theory of personality altogether. So it’s predictable when the scoring method is critiqued from the trait perspective.

Boyle does a reasonable job in the standard way, but he is unaware of Rytting, Ware and Prince (1994) on Bimodal Distribution, Tischler on Factor Analysis (1994) let alone the large number of studies across cultures.

I just wonder why psychometricians and others haven’t looked at why the MBTI does work, rather than stating it can’t.

In the validity section of his paper, Boyle also misleads the reader in quoting Isabel Myers (p77, 1962) as referring to the MBTI as “affording hypotheses for further testing and and verification rather than infallible expectations of all behaviors”.

Apart from adding the plural to "behavior", Boyle fails to note that Myers is referring to all self-report personality inventories, not just the MBTI, and is talking about the uses of the Indicator, not its predictive validity.

There’s also an implication in the text that the rest of the paragraph in which this statement is contained relates to Myers’ statement, particularly with other words supplied in parentheses: “causal” “surface traits”, but they do not.

Boyle doesn’t state where they come from at all and his understanding of type features as surface traits, as I read this part of the article, is a gross misunderstanding of the data he has read.

Reference is also made to the work of Bjork and Druckman (1991) which states as Boyle says “the instrument’s popularity is not consistent with research evidence” (p73).

I have this report and there is no evidence presented that would justify this assertion, there is simply a statement made to that effect, and the authors of the report didn’t back up their claim

In fact data presented in this report tended to support the usefulness of the MBTI for participants in the sample referred to. (See also Hammer (ed) 1997p34).

Boyle concludes by saying that “overall, the MBTI provides a psychometrically simple description of Jungian personality types”.

It’s hard to state that when he shows little interest in investigating the theory, or noting the explanations in his own provided references for unscored questions, or research items, for instance (McCaulley 1981 p312).

Boyle also concludes that “the current enthusiasm for the MBTI is certainly not warranted on psychometric grounds” (p73).

From this reviewer’s point of view, it’s hard to see how he would know, as the limited literature he’s chosen to have at his disposal is selectively assessed and he remains ignorant, in my view, of what Jung’s theory of types actually is, consequently, therefore of the claims and operation of the MBTI.

I found that none of his own 5 references in the paper relate to the MBTI. It doesn’t fit his framework, which is fine, as we all have our preferences.

Nonetheless, I think this is poor scholarship, and largely unsubstantaited opinion, which is fine except that this is supposed to be a serious critique and has been taken as such by many.

Kenneth L. Feder, in examining fraudulence in archaeology, points out that if someone writes inaccurately in a field of knowledge with which you are familiar, then the likelihood is high that they’ll be inaccurate in fields unfamiliar to you (pp7-8, 1990).

I am not a professional psychometrician, but the data here on type that is in my field of expertise is so misreported that I feel justified in querying the rest of the assertions, particularly as I have found published research on the MBTI to be so uneven in its quality.

Perhaps we should leave the last word to Isabel Myers in referring to data presented in her 1962 Manual where “certain basic preferences as to the use of perception and judgment have been portrayed from three widely different angles, -- theory, statistical research, and, lastly, personal observation, without which both theory and statistics remain purely intellectual exercises (p76, 1962)”.

Whether people like it or not, the MBTI presents ideas that can and have been validated by personal observation.

Whether it’s a factor of either or both the MBTI and human beings is incidental: the point of good science is to find out why this happens, not say it doesn’t because they don’t like the instrument.

Let’s critique the MBTI, by all means, but let’s do our homework beforehand.

SELECTED REFERENCES:

R.A. Bjork and D. Druckman: In the Mind’s Eye (National Academy Press 1991)

Kenneth L. Feder: Frauds, Myths and Mysteries (Mayfield 1990)

Peter Geyer: Quantifying Jung (MSc Thesis, Melbourne University, 1995)

Allen L. Hammer (ed) MBTI Applications: A decade of research on the MBTI (CPP 1996)

Bruno Latour Science In Action (Open University 1986)

Mary H. McCaulley Jung’s Theory of

Psychological Types and the MBTI (CAPT 1981)

Mary H. McCaulley& Isabel Briggs Myers, Manual: The Myers-Briggs Type Indicator (CPP 1985)

Isabel Briggs Myers, Manual (1962) The Myers-Briggs Type Indicator (CPP 1975)

James Newman (ed) Measures of the Five Factor Model and Psychological Type (CAPT 1995)

Roger Pearman and John Fleenor Differences in observed and self-reported qualities of psychological types (Journal of Psychological Type Vol 39 1996 pp3-17)

Marvin Rytting, Roger Ware and Russ A. Prince: Bimodal Distributions in a sample of CEOs: Validating Evidence for the MBTI (Journal of Psychological Type Vol 31 1994 pp16-23)

Len Tischler The MBTI Factor Structure (Journal of Psychological Type Vol 31 1994 pp24-31)

© Peter Geyer 1997, 1999

This article was originally written for the Australian Journal of Psychological Type Typetexts and Materials column Vol 6. No. 2. 1997, and has been reformatted and edited for this workbook.

Photograph courtesy of Jamie Johnston, CAPT Library. PETER GEYER (INTP) is a consultant, researcher and writer in the field of C G Jung's theory of psychological types. He conducts MBTI Accreditation programs and presents internationally on a regular basis. Peter is a life member of AusAPT and a professional affiliate of the Australian Psychological Society.