Content uploaded by David Pittenger
Author content
All content in this area was uploaded by David Pittenger
Content may be subject to copyright.
Cautionary Comments Regarding the Myers-Briggs
Type Indicator
David J. Pittenger The University of Tennessee at
Chattanooga
The Myers-Briggs Type Indicator (MBTI; K. C.
Briggs & I. B. Myers, 1998) is a popular mea-
sure of normal personality that its promoters
claim has many applications. M. H. McCaulley
(2000) offered an optimistic and enthusiastic
account of how counselors can use this instru-
ment in corporate settings. The present article
evaluates several of the psychometric limita-
tions and criticisms of the MBTI that warrant
considerable caution when making inferences
from its 4-letter type formula. The author con-
cludes that the MBTI, while offering much intu-
itive appeal, may not yet be able to support the
claims its promoters make.
McCaulley (2000) recently provided an
optimistic and enthusiastic account of how
consulting psychologists can and should in-
tegrate the Myers-Briggs Type Indicator
(MBTI) into consulting work with business
management. More specifically, she sug-
gested that consulting psychologists can
use the MBTI to help employees and man-
agers enhance their interpersonal relations
and thereby improve their ability to work
effectively for the corporation. The funda-
mental assumption presented in McCaul-
ley’s article is that knowledge of cowork-
ers’ personality preferences, or MBTI type,
will facilitate greater respect for individual
differences, aid in assigning work respon-
sibilities, and foster effective collaboration
among employees. Indeed, McCaulley ad-
vocated using the MBTI as a component of
the employee selection process. In addition,
McCaulley described what she believed to
be the striking difference in the worldview
of counselors and managers and offered
recommendations to help counselors con-
vey the utility of the MBTI to corporate
decision makers.
Although the MBTI is an extremely
popular measure of personality, I believe
that the available data warrant extreme cau-
tion in its application as a counseling tool,
especially as consultants use it in various
business settings. McCaulley (2000) of-
fered decisive conclusions that attest to the
validity of the MBTI four-letter type for-
mula and its utility for consulting work.
Unfortunately, she offered little empirical
evidence to support the veracity of those
claims. As such, these conclusions require
additional examination.
Currently, the role of personality assess-
ment in employment settings is receiving
renewed interest. As Hogan, Hogan, and
Roberts (1996) noted, although there re-
mains considerable skepticism regarding
the value of personality assessment in ap-
plied settings, they believe that well-con-
structed measures of normal personality
can serve a useful role in personnel selec-
tion and development. Such an endorse-
ment depends upon measures that evidence
David J. Pittenge is the Associate Provost for
Academic Administration at the University of
Tennessee at Chattanooga. Prior to this position
he was the Department Head of the Department
of Psychology and taught a variety of courses
including Statistics, Research Methods, Learn-
ing Theory, and Introductory Psychology. His
research interests include persistence of human
behavior, coping behaviors, and psychometrics.
Correspondence concerning this article should
be addressed to David J. Pittenger, Department
of Psychology, The University of Tennessee at
Chattanooga, 615 McCallie Avenue, Depart-
ment 2508, Chattanooga, TN 37403. E-mail:
david-pittenger@utc.edu
Copyright 2005 by the Educational Publishing Foundation and the Society of Consulting Psychology, 1065-9293/05/$12.00
DOI: 10.1037/1065-9293.57.3.210
Consulting Psychology Journal: Practice and Research, Vol. 57, No. 3, 210–221
210
construct validity and afford useful predic-
tions. Given the ramifications of applying
personality measures in the work setting, it
is necessary to maintain a critical analysis
of how practitioners use tests in applied
settings. In essence, I do not believe that
the relevant data justify the conclusion that
the MBTI is a direct measure of Jung’s
(1921/1971) theory of personality types.
Furthermore, I do not believe that the cur-
rent scoring procedures for the MBTI allow
one to make important prophecies about
individuals (see also, Pittenger, 1993).
In the remainder of this article, I review
what I believe to be important psychomet-
ric data that raise significant caveats re-
garding the MBTI and its applications. The
primary goal of this review is to provide an
alternative perspective of the MBTI and its
utility as a measure of personality. There
are important consequences when psychol-
ogists apply personality measures in the
work setting. The results of the test affect
the individuals who are tested, the corpo-
rations who use those test results, and our
profession to the extent that we endorse
specific measurement practices and infer-
ences from test scores. Consequently, it is
necessary to maintain a critical analysis of
the issues and ensure a lively dialectic re-
garding the use of normal personality as-
sessments in corporate and other applied
settings.
To organize my arguments, I have di-
vided this article into four parts. First, I
offer an abridged review of the theoretical
structure of the MBTI, its scoring, and its
interpretation. According to Myers, Mc-
Caulley, Quenk, and Hammer (1998),
Briggs and Myers developed the MBTI as a
measure of the personality types first artic-
ulated by Jung (1921/1971). The strong
theoretical background of the MBTI af-
fords many falsifiable hypotheses regarding
the psychometric properties of the MBTI
results. The second section reviews the rel-
evant psychometric data that assess these
hypotheses. The third section reviews the
relevant empirical literature that has exam-
ined the utility of the MBTI for industrial/
organization applications. In the final sec-
tion, I suggest that many of the uses of the
MBTI, as endorsed by McCaulley (2000),
lack empirical support, and that consulting
psychologists should consider these facts
before using the instrument. I also suggest
programs of future research that might ad-
dress several important matters I raise.
Theoretical Foundations of the MBTI
Before proceeding with a review of the
empirical features of the MBTI, it is impor-
tant to examine its theoretical foundation as
an explication of Jung’s (1921/1971) the-
ory of personality. This foundation is es-
sential for interpreting the psychometric
properties of the instrument and the utility
of its application for assessing an individ-
ual’s personality. I will forego a detailed
review of Jung’s theory and the particular
structure of the MBTI as other authors pro-
vide lucid and detailed reviews (Jung,
1921/1971; Macdaid, McCaulley, & Kaniz,
1986; McCaulley, 2000; Myers et al.,
1998).
The important feature that warrants at-
tention is the fact that Jung’s theory and the
MBTI are typologies (McCaulley, 2000).
More specifically, the instrument treats per-
sonality types as distinctive groups. This
perspective suggests that there are quanti-
tatively and qualitatively different popula-
tions of people who express different per-
sonality characteristics. In other words,
these populations will demonstrate relative
homogeneity of variance within groups and
heterogeneity of variance between groups
(Block, 1971; Block & Ozer, 1982).
Most personality tests represent a trait
perspective that characterizes personality
as a construct best measured using a con-
tinuous variable that ranges between two
extremes. For example, common measures
of introversion and extroversion (e.g., NEO
Personality Inventory [NEO-PI], Costa &
McCrae, 1985) treat these constructs as op-
211Consulting Psychology Journal: Practice and Research
Summer 2005
posing poles along a continuous scale.
Scores derived from trait measures of per-
sonality reflect the degree or magnitude of
the personality construct. Although casual
descriptions of the trait may lapse into
“type-as-label” discussions (Block & Ozer,
1982; e.g., Extrovert vs. Introvert), the un-
derlying assumption is that the scale is con-
tinuous and that the “types” refer to the
extremes of the population or scale. More-
over, trait theories presume that the major-
ity of scores cluster near a common point
along the scale, with fewer scores at the
extremes, thus forming a unimodal and rel-
atively symmetrical distribution.
By contrast, the MBTI is a type theory
that emphasizes the 16 unique categories of
personality created by the four type pairs
(Extraversion
1
-Introversion [EI], Sensing-
Intuition [SN], Thinking-Feeling [TF], and
Judging-Perceiving [JP]). The EI, SN, and
TF dimensions derive from Jung’s (1921/
1971) theory of personality. The JP scale is
an embellishment that Myers and Briggs
added to their interpretation of the theory
(see Saunders, 1991, for a history of the
MBTI and a biography of its creators).
The four-letter type formula (e.g., ENFP
or ISTJ) is the most salient characteristic of
the MBIT as it forms the central feature for
interpreting the results of one’s MBTI re-
sults. When scoring the instrument, one
converts the observed score for each of the
four scales to a letter. Because the four
scales each have two dimensions, there
are 16 letter combinations, each represent-
ing a unique and distinct personality pro-
file. In an earlier version of the MBTI,
individuals who scored at the center of one
of the scales received an xrather than a
letter code to illustrate the intermediacy of
the person’s score. The developers of the
MBTI subsequently abandoned that prac-
tice and replaced it with a tie-breaking pro-
cedure to avoid assigning intermediate
scores (Myers et al., 1998).
Because promoters of the MBTI view
the type conditions as mutually exclusive
categories, the practice of test interpretation
focuses on the four-letter type formula. As
Carskadon (1979a) noted,
In general, continuous scores on the MBTI
are probably overused. However much it
may seem to make better intuitive or psycho-
metric sense to use continuous scores in
many instances, it should be remembered
that the type scales are theoretically dichot-
omous in nature, and the indicator was de-
signed with this in mind. In general it is
probably better to use dichotomous classifi-
cation except where there are specific con-
siderations to the contrary. (p. 20)
McCaulley’s (2000) description of the
MBTI clearly endorsed this perspective. In
fact, McCaulley quoted Jung describing the
fundamental nature of type development
and then observed that “the most important
unit of measurement for the MBTI is the
four-letter type formula that indicates the
choices for the four preferences plus their
dynamic interaction” (McCaulley, 2000, p.
121). Both Jung’s theory and supporters of
the MBTI treat personality as an invariant
that is set at birth and tempered by experi-
ence. Thus, normally functioning adults
have well-established, unambiguous, and
stable personality preferences. McCaulley
also affirmed that the MBTI follows type
structure by asserting that the dimensions
measured by the MBTI are dichotomous
and not preferences that exist across a con-
tinuous scale that ranges from one pole to
another. Finally, McCaulley asserted that
the MBTI assesses the fundamental person-
ality features that influence an individual’s
cognitive processes. As I will illustrate sub-
sequently, the type perspective and the di-
chotomous classification procedures are
problematic. In addition, I will show that
the link between the personality dimen-
sions measured by the MBTI, cognition,
and work behaviors is tenuous.
1
The spelling of extraversion, as opposed to
extroversion, is unique to the MBTI and follows
the practice established by Jung (1921/1971).
212 Consulting Psychology Journal: Practice and Research
Summer 2005
Psychometric Properties of the MBTI
Descriptive Statistics
Given the type theory orientation of the
MBTI, one would predict between-group
heterogeneity of variance and within-group
homogeneity of variance for each of the
four type preference dimensions. This pre-
diction implies that the distribution of
scores should be bimodal, with greater rel-
ative difference between the two preference
types compared with the variance of scores
within each type. In addition, the nature of
the type preference predicts that few indi-
viduals should score at the midpoint of the
scale. Although bimodality appears to be an
essential characteristic of the distributions
of scores, it is conspicuously absent. Evi-
dence from several sources (Harvey &
Murry, 1994; Hicks, 1984; McCrae &
Costa, 1989; Stricker & Ross, 1962) indi-
cates a continuous distribution of scores
across each dimension. More recently, Bess
and Harvey (2002) replicated the finding of
unimodality of scores using the item re-
sponse theory scoring scheme promoted in
the current edition of the measure (Myers et
al., 1998). As I have noted elsewhere (Pit-
tenger, 1993), even the data presented in
the MBTI manual provide only tentative
evidence for a discontinuous scale and
greater evidence for a continuous measure
of personality traits.
The lack of bimodality and the high
frequency of scores at the midpoint of the
scales have profound implications for inter-
preting the MBTI’s four-letter type for-
mula. There is no evidence of separate pop-
ulations of personality types using the stan-
dard scoring procedure. Thus, concluding
that an Etype is qualitatively different from
an Itype is indefensible unless there are
corresponding data to suggest that the dif-
ference between the scale scores is suffi-
ciently large to support such a distinction.
Using a conventional 95% confidence in-
terval, the standard error of measurement
(SEM) for each scale is 20 or larger (Har-
vey & Murry, 1994; Pittenger, 1993). From
a statistical perspective, the MBTI four-
letter type formula may imply statistically
significant personality differences where
none exists.
Ipsative and dichotomous scaling proce-
dures, such as those used by the MBTI,
have several disadvantages and limit the
ability to make reasonable prophecies of a
person’s behavior using the four-letter type
formula. As a generality, using dichoto-
mous scores reduces the predictive power
of a continuous scale (Hunter & Schmidt,
1990) and can greatly increase the rate of
Type I errors (Maxwell & Delaney, 1993).
Specifically, converting continuous interval
or ratio scaled scores to an ordinal or nom-
inal scale removes important information,
especially information related to variability.
Cohen (1983) demonstrated that dichot-
omizing a continuous scale reduces the
shared variance with another variable by
.64(r
2
); dichotomizing both scales reduces
the shared variance by .40(r
2
). By implica-
tion, the four-letter type formula limits its
own predictive power, especially if the cri-
terion is a binary condition. Consequently,
counselors who advocate the MBTI as a
component of the job-selection or team-
forming process, as McCaulley (2000) rec-
ommended, may well be promoting a deci-
sion-making tool whose own procedures
reduce its predictive validity, especially in
cases where the final decision is dichoto-
mous (e.g., hire vs. do not hire).
Tenopyr (1988) offered a more em-
phatic caution regarding the use of forced-
choice instruments (of which the MBTI is a
member). Her analysis of the internal con-
sistencies among scales using forced-
choice items revealed the potential for con-
siderable artifactual and inflated reliability
estimates. Tenopyr thus concluded, “It ap-
pears that construct interpretation on the
basis of forced-choice scales should be
made with extreme caution. In the case of
some inventories, it would probably be bet-
ter not to make these interpretations, or at
213Consulting Psychology Journal: Practice and Research
Summer 2005
least not rely on them heavily, with respect
to important decisions about individuals”
(p. 751; italics added). Applying item re-
sponse theory procedures to examine the
MBTI, Harvey and Murry (1994) found
that the dichotomizing procedure produced
between 26% and 32% loss of information
for each of the scales.
These observations are troublesome as
there are notably little data to support the
type structure of the MBTI. Furthermore,
the practice of converting scale scores to
type categories reduces any predictive
value that the test may afford. Conse-
quently, those who use the conventional
MBTI scoring practice to make inferences
about others’ personality risk reaching con-
clusions that cannot be empirically justified.
Test-Retest Reliability
The theoretical framework supporting
the MBTI provides a unique perspective for
interpreting the test–retest reliability of the
instrument. Recall that Jung and the pro-
moters of the MBTI treat personality as an
invariant that is set by adulthood. As such,
one would predict that test-retest reliabili-
ties will be high, especially for the four-
letter preference categories.
There are several reports of the test-
retest reliabilities of the four dimensions of
the MBTI (Carskadon, 1977, 1979b;
Howes & Carskadon, 1979; Stricker &
Ross, 1962). These reports offer a consis-
tent pattern that suggests that the reliability
of the MBTI does not meet expectations
derived from its theory. For example,
Sticker and Ross found that across a 14-
month period the reliabilities ranged from a
low of r(38) ⫽.48 for the TF scale to a
high of r(38) ⫽.73 for the EI scale.
Schuerger, Zarrella, and Holtz (1989) also
demonstrated that the attenuation of the
reliability of the MBTI decreases at a rate
over retest intervals that is comparable to
other measures of personality. More recent
reports (Capraro & Capraro, 2002; Salter,
Evans, & Forney, 1997) confirm previous
reviews of the test-retest reliabilities of the
MBTI scores. Although one might consider
these reliabilities acceptable for conven-
tional trait measures of personality, they are
at odds with a type theory that predicts that
nonpathological personality preferences
should become and remain stable early in
life.
Howes and Carskadon (1979) provided
data that raise additional and equally im-
portant questions regarding the reliability
of the four-letter type score. Their analysis
indicated that a large portion of their par-
ticipants received different type profiles
when retested. Not surprisingly, the great-
est proportion of changes occurred when
the initial preference score was close to the
middle of the scale (1 to 15 points on either
side of the midpoint). When the initial
score was within this intermediate range,
32% of the EI, 25% of the SN, 29% of the
TF, and 30% of the JP labels shifted on the
second testing. McCarley and Carskadon
(1983) replicated these findings and dem-
onstrated that across a 5-week test-retest
interval, 50% of the participants received a
different classification on one or more of
the scales. Indeed, Myers et al. (1998) re-
ported that 35% of individuals had a differ-
ent four-letter type score after a 4-week
interval.
These results are not surprising given
the center-heavy distribution and heteroge-
neity of variance of the scale scores. Nor
are these changes trivial. If we are to pre-
sume that “an extraverted sensing type will
show extraversion differently from an ex-
traverted thinking type” (McCaulley, 2000),
then the alteration of one or more of the
four-letter type formula represents a con-
siderable change in personality. These data
also raise profound questions regarding the
advisability of using the four-letter typing
system, while ignoring the magnitude of
the scale scores, and raise questions regard-
ing the veracity of any type interpretations
for individuals with scale scores close to
the midpoint of the scale.
214 Consulting Psychology Journal: Practice and Research
Summer 2005
Kummerow (1988) and Walck (1992)
provided further evidence for the tenuous
nature of the four-letter type formula. For
example, Walck administered the MBTI to
a large sample of individuals and then
asked the participants to describe, in writ-
ing, their type preferences. Walck found
that as many as a third of the participants
believed that the MBTI four-letter type for-
mula mislabeled them. Specifically, there
were considerable inconsistencies regard-
ing the TF and EI dimensions. Walck also
reported that these disagreements were
most likely to occur when the scale scores
were close to the middle of the scale, thus
raising further questions about the appro-
priateness of using type categories for
individuals who do not evidence a clear
preference for a specific MBTI personality
dimension.
Factor Analytic Analysis
Factor analytic studies of the MBTI re-
veal consistent findings that also question
the construct validity of the instrument.
First, the factor structure is not consistent
with the design or theory of the MBTI. For
example, Sipps, Alexander, and Freidt
(1985) found a factor structure that was
inconsistent with the four factors specified
by the MBTI and found only a marginal
relation between these factors and the
MBTI scales. Other researchers have re-
ported findings that are inconsistent with
MBTI theory (e.g., Johnson & Saunders,
1990; McCrae & Costa, 1989; Saggino,
Cooper, & Kline, 2001; Saggino & Kline,
1996; Sipps & Alexander, 1987; Sipps &
DiCaudo, 1988; Stricker & Ross, 1962;
Thompson & Borrello, 1986a, 1986b).
Similarly, Lorr (1991) found that cluster
analytic techniques failed to produce re-
sults consistent with MBTI theory and, like
others, faulted the instrument’s scoring pro-
cedures used to categorize individuals.
A second problem is the correlation
among the MBTI scales. To reiterate, pro-
moters of the MBTI treat the four scales as
independent dimensions of one’s personal-
ity preferences. If the scales are indepen-
dent measures of different dimensions of
personality, there should be little or no in-
tercorrelation among the scales. The avail-
able data, however, suggest that there is a
sizable correlation among the scales (Berr,
Church, & Waclawski, 2000; Myers et al.,
1998; Sipps & DiCaudo, 1988).
Other inconsistencies arise when one
compares MBTI scores with the scores of
other measures of personality. For exam-
ple, the EI scale of the MBTI correlates
with other measures of extroversion, in-
cluding Eysenck’s (Sipps & Alexander,
1987) or a version of the Big Five measure
of personality (McCrae & Costa, 1989).
This is an interesting finding because
Jung’s (1921/1971) account of extraversion
is substantively different from other theo-
ries of extroversion (Pittenger, 1993).
Perhaps the most interesting finding is
that the Big Five perspective of personality
can readily subsume the variance explained
by MBTI. For example, McCrae and Costa
(1989) concluded that a five-factor model
of personality more efficiently explains the
MBTI factor structure. Furnham (1996)
reached a similar conclusion after examin-
ing the correlation of scores between the
subscale MBTI and the NEO-PI (Costa &
McCrae, 1985).
Although McCrae and Costa (1989) ar-
gued that the MBTI is derivative of their
version of the Big Five, it is possible that
NEO-PI is a special case of the MBTI. By
comparison, the MBTI has emerged from a
theoretically enriched environment, whereas
the Big Five perspective has emerged al-
most exclusively from empirical analysis
(Goldberg, 1990, 1992, 1993). As such, the
MBTI affords greater opportunity for sys-
tematic analysis of its construct validity.
Nevertheless, those who wish to use the
MBTI in corporate settings should recog-
nize the depth of evidence supporting the
predictive validity of other personality in-
struments, specifically those derived from
215Consulting Psychology Journal: Practice and Research
Summer 2005
Big Five measures. As Furnham (1996) and
Hogan et al. (1996) noted, the NEO-PI ap-
pears to have strong support for its utility to
predict and explain work-related behaviors.
Unfortunately, there are no published data
to demonstrate the incremental validity of
either instrument as a method for predicting
job performance or other work-related be-
havior. Therefore, conclusions regarding
the superiority of either the MBTI or other
instruments are, at present, premature.
Summary of the Psychometric
Properties of MBTI
A review of the surface features of the
MBTI reveals several important inconsis-
tencies between predictions derived from
the nomothetic network that defines the
MBTI and the empirical data. In total, these
inconsistencies raise serious doubts about
the inferences and conclusions one can
make from the MBTI. Most notable among
these problems is the tenuous nature of the
dichotomous scoring procedure. I, like Mc-
Caulley (2000), belabor the dichotomous
nature of the instrument because the entire
structure of the MBTI depends upon this
perspective of personality. Consequently,
these inconsistencies are not trivial. The
data related to the distribution of scale
scores, the test–retest reliability, and the
correlation with other measures of person-
ality also raise doubts that one can make
reasonable inferences from a four-letter
type formula, especially when an individu-
al’s scale scores are close to the mean.
Stated from a different perspective, the
four-letter type formula may create the im-
pression that there is meaningful difference
between the personality profiles of two in-
dividuals when no such difference exists.
Analysis of the Utility of the MBTI in
Corporate Settings
A central feature of McCaulley’s (2000)
presentation is the difference in type pref-
erence between managers and counselors.
Although the pattern of differences be-
tween the two populations appears impres-
sive, one must interpret the data with cau-
tion. The information presented by Mc-
Caulley represents the four-letter types and
not the magnitude of the scale scores; con-
sequently, the reader has no information to
determine whether the intensity of these
preferences is statistically significant. There-
fore, a statistically significant
2
in the pat-
tern of type preferences may be artifact of
the forced-choice and dichotomous scoring
procedure used by the MBTI. For example,
Berr et al. (2000) reported the average
scaled scores for their sample of senior
managers. With the exception of the TF
preference pair, all means were within 1
SEM of the center of the scale.
Other research examining the preference
types of managers provides notably little
information about the magnitude of the
type preferences of managers. Gardner and
Martinko (1996) conducted a literature re-
view that examined the use of the MBTI in
management research. They noted that of
the 13 studies that examined the distribu-
tion of MBTI type preferences for manag-
ers, all were limited to reports of simple
frequencies of type preference with no
mention of scale scores. In addition, Gard-
ner and Martinko observed that although
“these studies indicate the proportions of
types within managerial samples, they re-
veal little about managers’ cognitions or
behaviors” (p. 53).
Even if we accept the existence of a
correlation between personality types and
profession, the cause of that relation is, as
yet, unclear. Gardner and Martinko (1996)
offered three reasonable alternative expla-
nations for the distribution of types among
professions, especially those observed in
corporate settings. The first is a conven-
tional hypothesis that links personality to
vocational preferences. This hypothesis
predicts that people with specific personal-
ity types seek out specific professions or
216 Consulting Psychology Journal: Practice and Research
Summer 2005
are selected for a profession based on their
personality. This hypothesis is the central
feature of current research of personnel se-
lection (Hogan et al., 1996). Although the
potential link between personality and
work performance is an interesting pros-
pect, it is one that has yet to be broadly
endorsed by the research community and
should be presented, therefore, with caution.
A second account of the results is that
factors related to social desirability, im-
pression management, or stereotypes about
one’s profession bias self-reports on the
MBTI. This concern is not unique to the
MBTI, however, as it is a common criti-
cism of all personality inventories (Hogan
et al., 1996). Evidence provided by Walck
(1992), however, suggests that many peo-
ple may disagree with their assigned type
category and underscores McCaulley’s
(2000) recommendation that individuals
have the opportunity to determine the va-
lidity of the reported type.
Gardner and Martinko’s (1996) third al-
ternative hypothesis suggests that working
in a specific managerial role may enhance
or alter one’s expression of personality.
Because the majority of the research on
type preference and job classification has
been cross-sectional, it is not possible to
resolve whether it is the personality prefer-
ence that affects the job selection or if the
selected job affects the personality profile.
This problem is not unique to the MBTI,
however, as few researchers have applied
longitudinal methodologies to examine the
relation between personality and job per-
formance. Although this is the most spec-
ulative of the alternative hypotheses, it does
suggest another caution regarding use of
the MBTI.
McCaulley (2000) noted that the use of
the MBTI can quickly become pervasive
throughout an organization as individuals
gladly share their MBTI type formula with
others. Although a free exchange and dis-
cussion of individual differences among
coworkers is admirable, it does raise the
specter of the fundamental attribution error.
In other words, making personality a salient
component of individual interactions
(“Hello, I’m INTJ.”) may exaggerate the
tendency of individuals to make disposi-
tional attributions while ignoring situa-
tional conditions that affect behavior. Such
exaggerations can lead to unwarranted con-
clusions regarding the use of the MBTI.
Elsewhere, I (Pittenger, 1994) have de-
scribed my experiences with a wholesale
application of the MBTI in an academic
setting that resulted in the dissemination of
dubious advice to students regarding their
selection of courses and majors.
Reading McCaulley’s (2000) descrip-
tion of the type preferences suggests that
situational factors have little or no influ-
ence on individuals’ behaviors or cogni-
tions. Indeed, little in the MBTI theory
appears to acknowledge the Person ⫻Sit-
uation interaction that is a common com-
ponent of contemporary social– cognitive
theory (Mischel & Shoda, 1995; Shoda &
Mischel, 2000). For example, Mendoza-
Denton, Ayduk, Mischel, Shoda, and Testa
(2001) recently presented experimental ev-
idence of Person ⫻Situation encoding of
experiences in complex social situations.
This encoding process attenuates accep-
tance of global stereotypes and enhances
the analysis of self in different social con-
texts. Such results echo the findings of
Gardner and Martinko (1996), who re-
viewed evidence that supports the conclu-
sion that context may dominate one’s deci-
sion style. Consequently, those who rely
exclusively upon MBTI personality types
to explain differences among persons may
overlook an important contributor to differ-
ences among persons.
Conclusion
Gardner and Martinko (1996) provided
a comprehensive review of the empirical
literature examining the relation between
the MBTI and various work-related behav-
iors. Their evaluation of the MBTI repre-
217Consulting Psychology Journal: Practice and Research
Summer 2005
sents, at best, a guarded and conditional
endorsement. Although they noted that
there are studies that imply a link between
the MBTI types and some important work
performance measures, they were also care-
ful to highlight the many methodological
limitations of those studies. Furthermore,
their review of the empirical link between
personality types and work performance
measures warrants greater qualification
than McCaulley’s (2000) comments imply.
Specifically, they concluded that, “It is
clear that efforts to detect simplistic link-
ages between type preferences and mana-
gerial effectiveness have been disappoint-
ing. Indeed, given the mixed quality of re-
search and the inconsistent findings, no
definitive conclusion regarding these rela-
tionships can be drawn” (Gardner & Mar-
tinko, 1996, p. 77, italics added). This con-
clusion echoes the conclusion of other au-
thors who question the utility of the MBTI.
Previously, Bjork and Druckman (1991)
and Boyle (1995) reviewed the literature
available to them and found no evidence
supporting the utility of the MBTI. As
Bjork and Druckman noted, “At this time,
there is not sufficient, well-designed re-
search to justify the use of the MBTI in
career counseling programs. Much of the
current evidence is based on inadequate
methodologies” (p. 99).
Furthermore, there is a conspicuous lack
of data demonstrating the incremental va-
lidity of the MBTI over other measures of
personality. Such a finding is not surprising
given the nascent revised interest in the role
of personality measures in work settings.
Nevertheless, given the attention that other
personality perspectives receive from in-
dustrial– organizational psychologists (e.g.,
Big Five, McCrae & Costa, 1989; see also
Hogan et al., 1996), it appears incumbent
on counselors working in corporations to
consider these alternatives as a part of their
consulting work.
Equally important to recognize are the
variables that a psychological test does not
measure. As Berr et al. (2000) noted, the
correlations between conventional mea-
sures of personality and performance indi-
cators are approximately r⫽.20. In other
words, the shared variance between a per-
sonality dimension and work performance
is 4%. Therefore, those who claim a link
between personality and performance must
acknowledge that the majority of work per-
formance reflects many other influences.
Although the MBTI is an extremely
popular test, the available evidence of its
psychometric properties recommends cau-
tion regarding optimistic accounts of its
ability to make empirically verifiable pre-
dictions. On many occasions, Messick
(1965, 1980, 1982, 1989, 1995) called upon
psychologists to recognize that test validity
is an inference one derives from the scores
and not an inherent property of the instru-
ment. Specifically, a personality measure,
such as the MBTI, is not inherently valid or
invalid, only the interpretations made from
the test results are. Moreover, Messick ar-
gued that inferences regarding the validity
of a psychological measure must examine
the potential consequences of using the in-
strument. Given the equivocal data related
to the construction of the MBTI and the
gravity of its application in work settings, I
believe that counselors must consider with
care claims made about the utility and
value of the MBTI as a measure of person-
ality and as a counseling tool.
The available evidence suggests that the
MBTI does measure constructs related to
personality; whether it measures the con-
structs identified by its underlying theory is
not clear. Furthermore, it is not evident that
the instrument can compartmentalize accu-
rately, consistently, and unambiguously in-
dividuals’ personality into the 16 type cat-
egories created by the instrument. Conse-
quently, using the MBTI as a consulting
tool in corporate settings may be, in some
instances, the equivalent of making prom-
ises that one cannot keep. Such a conclu-
218 Consulting Psychology Journal: Practice and Research
Summer 2005
sion leaves us with the question, what is the
utility of the MBTI as a consulting tool?
The popularity of the MBTI as a con-
sulting tool most likely reflects the success
of the publisher’s marketing campaign and
the intuitive and simple sounding nature of
the instrument’s scoring scheme. Com-
pared to a conventional measure of the Big
Five, it is probably comforting to learn that
one tends to be intuitive and feeling, rather
than learning that one has scored high on
the neuroticism and low on the conscious-
ness scales. Consequently, the MBTI can
serve as a nonthreatening vehicle to intro-
duce the concept of individual differences
in personality and the relation between per-
sonally constructs and behavior to a general
audience. The instrument might even serve
as a catalyst for exercises that lead to im-
proved esprit de corps among employees.
However, presenting the data using the
four-letter type formula rather than the
scaled scores is a misrepresentation of the
available evidence. The four-letter type for-
mula may also be an overly simplistic ac-
count of complex personality dynamics and
leave the recipient with a false impression
that there is little left to doubt.
By contrast, other uses of the MBTI,
especially those that affect an employee’s
job status, may have iatrogenic effects. Us-
ing the MBTI to select employees, to assign
employees to work groups or assignments,
or for other forms of employment evalua-
tion are not justified for the simple reason
that there are no available data to recom-
mend such decisions. As such, counselors
promoting applications of the MBTI should
clearly represent the known limitations for
specific uses of the instrument.
It is impossible to endorse applications
of a personality measure without sufficient
empirical evidence. Although McCaulley
(2000) offered many claims regarding the
conformation and utility of the MBTI in
corporate settings, the available evidence
does not support these predictions. There-
fore, consulting psychologists require ded-
icated research that tests the specific pre-
dictions regarding the MBTI and other
measures of personality.
An advantage of the MBTI is that its
strong theoretical structure affords specific
predictions regarding the link between per-
sonality and behavior. At the same time,
there are other measures of personality
whose structure portents useful applica-
tions in corporate settings. Therefore, it
will be useful for future research to exam-
ine the predictive validity of these instru-
ments relative to each other. Furthermore,
it will useful, as Hogan et al. (1996) noted,
that analysis of the predictive power of
personality tests incorporate other measures
of performance, including intelligence,
knowledge and experience, and skill.
The fact that the MBTI evolved within a
strong theoretical structure can also be a
liability so long as the structure of the the-
ory, the scoring procedures, and the canons
of score interpretation do not keep pace
with the available evidence. That the MBTI
shares variance with other measures of per-
sonality suggests that it may not be a
unique measure of personality or that it
measures unique personality dimensions.
Consequently, those interested in using the
MBTI should examine the advantages of
replacing the four-letter type formula with
more traditional magnitude assessments of
personality.
References
Berr, S. A., Church, A. H., & Waclawski, J.
(2000). The right relationship is everything:
Linking personality preferences to manage-
rial behaviors. Human Resource Develop-
ment Quarterly, 11, 133–157.
Bess, T. L., & Harvey, R. J. (2002). Bimodal
score distributions and the Myers-Briggs
Type Indicator: Fact or artifact? Journal of
Personality Assessment, 78, 176 –186.
Bjork, R. A., & Druckman, D. (1991). In the
mind’s eye: Enhancing human performance.
Washington, DC: National Academy Press.
Block, J. (1971). Lives through time. Berkley,
CA: Bancroft Books.
219Consulting Psychology Journal: Practice and Research
Summer 2005
Block, J., & Ozer, D. J. (1982). Two types of
psychologists: Remarks on the Mendelsohn,
Weiss, and Feimer contribution. Journal of
Personality and Social Psychology, 42,
1171–1181.
Boyle, G. J. (1995). Myers-Briggs Type Indica-
tor (MBTI): Some psychometric limitations.
Australian Psychologist, 30, 71–74.
Briggs, K. C., & Myers, I. B. (1998). Myers–
Briggs Type Indicator. Palo Alto, CA: Con-
sulting Psychologists Press.
Capraro, R. M., & Capraro, M. M. (2002).
Myers-Briggs Type Indicator score reliabil-
ity across studies: A meta-analytic reliability
generalization study. Educational & Psycho-
logical Measurement, 62, 590 – 602.
Carskadon, T. G. (1977). Test-retest reliabilities
of continuous scores on the Myers-Briggs
Type Indicator. Psychological Reports, 41,
1011–1012.
Carskadon, T. G. (1979a). Clinical and counsel-
ing aspects of the Myers-Briggs Type Indi-
cator: A research review. Research in Psy-
chological Type, 2, 2–31.
Carskadon, T. G. (1979b). Test-retest reliabili-
ties of continuous scores on Form G of the
Myers-Briggs Type Indicator. Research in
Psychological Type, 2, 83– 84.
Cohen, J. (1983). The cost of dichotomization.
Applied Psychological Measurement, 7,
249 –253.
Costa, P., & McCrae, R. (1985). The NEO Per-
sonality Inventory manual. Odessa, FL: Psy-
chological Assessment Resources.
Furnham, A. (1996). The Big Five versus the big
four: The relationship between the Myers-
Briggs Type Indicator (MBTI) and the
NEO-PI face factor model of personality.
Personality and Individual Differences, 21,
303–307.
Gardner, W. L., & Martinko, M. J. (1996). Us-
ing the Myers-Briggs Type Indicator to study
managers: A literature review and research
agenda. Journal of Management, 22, 45– 83.
Goldberg, L. R. (1990). An alternative “descrip-
tion of personality”: The Big Five factor
structure. Journal of Personality and Social
Psychology, 59, 1216 –1229.
Goldberg, L. R. (1992). The development of
markers for the Big Five factor structure.
Psychological Assessment, 4, 26 – 42.
Goldberg, L. R. (1993). The structure of pheno-
typic personality traits. American Psycholo-
gist, 48, 26 –34.
Harvey, R. J., & Murry, W. D. (1994). Scoring
the Myers-Briggs Type Indicator: Empirical
comparison of preference score versus la-
tent-trait methods. Journal of Personality
Assessment, 62, 116 –129.
Hicks, L. E. (1984). Conceptual and empirical
analysis of some assumptions of an explicit
typological theory. Journal of Personality
and Social Psychology, 46, 1118-1131.
Hogan, R., Hogan, J., & Roberts, B. W. (1996).
Personality measurement and employment
decisions: Questions and answers. American
Psychologist, 51, 469 – 477.
Howes, R. J., & Carskadon, T. G. (1979). Test-
retest reliabilities of the Myers-Briggs Type
Indicator as a function of mood changes.
Research in Psychological Type, 2, 67–72.
Hunter, J. E., & Schmidt, F. L. (1990). Dichot-
omization of continuous variables: The im-
plications for meta-analysis. Journal of Ap-
plied Psychology, 75, 334 –349.
Johnson, D. A., & Saunders, D. R. (1990). Con-
firmatory factor analysis of the Myers-Briggs
Type Indicator—Expanded analysis report.
Educational & Psychological Measure-
ment, 50, 561–571.
Jung, C. G. (1971). Collected works of C. G.
Jung: Vol. 6. Psychological types (H. G.
Baynes, Trans., revised by R. F. C. Hull).
Princeton, NJ: Princeton University Press.
(Original work published 1921)
Kummerow, J. M. (1988). A methodology for
verifying type: Research results. Journal of
Psychological Type, 15, 20 –25.
Lorr, M. (1991). An empirical evaluation of the
MBTI typology. Personality & Individual
Differences, 12, 1141–1145.
Macdaid, G. P., McCaulley, M. H., & Kaniz,
R. I. (1986). Myers-Briggs Type Indicator:
Atlas of type tables. Gainesville, FL: Center
for Applications of Psychological Type.
Maxwell, S. E., & Delaney, H. D. (1993). Bi-
variate median splits and spurious statistical
significance. Psychological Bulletin, 113,
181–190.
McCarley, N. G., & Carskadon, T. G. (1983).
Test-retest reliabilities of scales and sub-
scales of the Myers-Briggs Type Indicator
and of criteria for clinical interpretive hy-
potheses involving them. Research in Psy-
chological Type, 6, 24 –36.
McCaulley, M. H. (2000). Myers-Briggs Type
Indicator: A bridge between counseling and
consulting. Consulting Psychology Journal:
Practice and Research, 52, 117–132.
McCrae, R. R., & Costa, P. T. (1989). Reinter-
preting the Myers-Briggs Type Indicator
from the perspective of the five-factor model
of personality. Journal of Personality, 57,
17– 40.
220 Consulting Psychology Journal: Practice and Research
Summer 2005
Mendoza-Denton, R., Ayduk, O., Mischel, W.,
Shoda, Y., & Testa, A. (2001). Person ⫻
Situation interactionism in self-encoding (I
Am . . .when . . .): Implications for affect
regulation and social information processing.
Journal of Personality and Social Psychol-
ogy, 80, 533–544.
Messick, S. (1965). Personality measurement
and the ethics of assessment. American Psy-
chologist, 20, 136 –142.
Messick, S. (1980). Test validity and the ethics
of assessment. American Psychologist, 35,
1012–1027.
Messick, S. (1982). Test validity and the ethics
of assessment. Diagnostica, 28, 1–25.
Messick, S. (1989). Validity.In R. L. Linn
(Ed.), Educational measurement (3rd ed.; pp.
13–103). New York: Macmillan.
Messick, S. (1995). Validity of psychological
assessment: Validation of inferences from
persons’ responses and performances as sci-
entific inquiry into score meaning. American
Psychologist, 50, 741–749.
Mischel, W., & Shoda, Y. (1995). A cognitive-
affective system theory of personality: Re-
conceptualizing situations, dispositions, dy-
namics, and invariance in personality struc-
ture. Psychological Review, 102, 246 –268.
Myers, I. B., McCaulley, M. H., Quenk, N. L.,
& Hammer, A. L. (1998). Manual: A guide
to the development and use of the Myers-
Briggs Type Indicator. Palo Alto, CA: Con-
sulting Psychologist Press.
Pittenger, D. J. (1993). The utility of the Myers-
Briggs Type Indicator. Review of Educa-
tional Research, 63, 467– 488.
Pittenger, D. J. (1994). The cross-disciplinary
ethical responsibilities of psychology fac-
ulty. Ethics & Behavior, 4, 199 –208.
Saggino, A., Cooper, C., & Kline, P. (2001). A
confirmatory factor analysis of the Myers-
Briggs Type Indicator. Personality & Indi-
vidual Differences, 30, 3–9.
Saggino, A., & Kline, P. (1996). The location of
the Myers-Briggs Type Indicator in person-
ality factor space. Personality & Individual
Differences, 21, 591–597.
Salter, D. W., Evans, N. J., & Forney, D. S.
(1997). Test-retest of the Myers-Briggs Type
Indicator: An examination of dominant func-
tioning. Educational & Psychological Mea-
surement, 57, 590 –597.
Saunders, F. W. (1991). Katherine and Isabel:
Mother’s light, daughter’s journey. Palo
Alto, CA: Consulting Psychologists Press.
Schuerger, J. M., Zarrella, K. L., & Holtz, A. S.
(1989). Factors that influence the temporal
stability of personality by questionnaire.
Journal of Personality and Social Psychol-
ogy, 56, 777–783.
Shoda, Y., & Mischel, W. (2000). Reconciling
contextualism with the core assumptions of
personality psychology. European Journal
of Personality, 14, 407– 428.
Sipps, G. J., & Alexander, R. A. (1987). The
multifactorial nature of extraversion/intro-
version in the Myers-Briggs Type Indicator
and Eysenck Personality Inventory. Educa-
tional & Psychological Measurement, 47,
543–552.
Sipps, G. J., Alexander, R. A., & Friedt, L.
(1985). Item analysis of the Myers-Briggs
Type Indicator. Educational & Psychologi-
cal Measurement, 45, 789 –796.
Sipps, G. J., & DiCaudo, J. (1988). Convergent
and discriminant validity of the Myers-
Briggs Type Indicator as a measure of socia-
bility and impulsivity. Educational & Psy-
chological Measurement, 48, 445– 451.
Stricker, L. J., & Ross, J. (1962). A description
and evaluation of the Myers-Briggs Type In-
dicator (Research Bulletin #RB-62– 6).
Princeton, NJ: Educational Testing Service.
Tenopyr, M. L. (1988). Artifactual reliability of
forced-choice scales. Journal of Applied
Psychology, 73, 749 –751.
Thompson, B., & Borrello, G. M. (1986a). Con-
struct validity of the Myers-Briggs Type In-
dicator. Educational & Psychological Mea-
surement, 46, 745–752.
Thompson, B., & Borrello, G. M. (1986b). Sec-
ondary-order factor structure of the MBTI: A
construct validity assessment. Measurement
and Evaluation in Counseling Develop-
ment, 18, 148 –153.
Walck, C. L. (1992). The relationship between
indicator and type and “true type”: Slight
preferences and the verification process.
Journal of Psychological Type, 23, 17–21.
221Consulting Psychology Journal: Practice and Research
Summer 2005
