Predictive Validity of the Myers-Briggs Type Indicator (Chapter 5)

Hypothesis 1, that CPS scores would be positively associated with counseling difficulty was supported. However, the association was weak for the total sample. Further analysis revealed that CPS is moderately predictive when used with extraverted clients presenting one of three diagnostic conditions--alcohol/drugs, family violence, or depression.

Predictive Validity of the Myers-Briggs Type Indicator Composite Preference Score to Estimate Counseling Difficulty and Counseling Outcome

Dissertation by
Jerry L. Brinegar, Ph.D.
Child and Family Development
University of Georgia
Athens, Georgia 1992


Hypothesis 1, that CPS scores would be positively associated with counseling difficulty was supported. However, the association was weak for the total sample. Further analysis revealed that CPS is moderately predictive when used with extraverted clients presenting one of three diagnostic conditions--alcohol/drugs, family violence, or depression. Conversely, counseling difficulty with introverted clients presenting with any of these three diagnostic conditions is moderately well predicted by their preference scores alone, but not by their CPS scores.

Hypothesis 2, that CPS scores would be positively associated with successful counseling outcome was not supported for the total sample or in the subsamples. There was very limited evidence to sug-gest that preference scores for introverts with diagnoses alcohol/drugs, family violence, or depression are predictive of counseling outcome.

Follow-up analyses yielded two other findings of interest. Specifically, the relation of extraversion-introversion to counseling difficulty was most evident among males, and among persons with more than a high school education.

Finally, the pilot study was reexamined to determine if the corre-lations between the CPS subscale score of extraversion-introversion and counseling outcome replicated the findings of the present study. As previously discussed, higher CPS scores predicted less successful counseling outcome scores. In addition, strengthening the evidence for the importance of the extraversion-introversion dimension, it proved to be the only single dimension score significantly related with counseling outcome, and the regression analysis more strongly implicated the extraversion-introversion preference score for introverts than for extraverts in predicting counseling outcome.

Implications for the Counseling Utility of the MBTI CPS

The present findings have prompted me to reevaluate my use of the CPS score. Over the course of 15 years of clinical practice that included my development of the concept of a CPS, I determined that 80 was a mid-range cutoff point for the MBTI scores in my sample. Given the data of the present study, correlational findings involving the CPS measure with extraverts led me to examine whether the CPS could be clinically useful in predicting counseling difficulty. Examination re-vealed that a CPS cutoff score at 90 or above predicted 80% of the cases judged to be difficult (4) or very difficult (5), and yielded no cases judged to be easy (2) or very easy (1). A similar analysis for introverts revealed that a preference score of 23 or above yielded the best discrimination of "hits" and "misses." Specifically, 60% of the difficult cases (4 and 5) were correctly identified while 5% of the easy cases (1 and 2) were misidentified as difficult. Thus, from the present data it appears that the CPS for extraverts has good practical counseling utility in correctly identifying difficult cases, particularly when the score is >90, whereas the preference score appears to be somewhat less reliable in discriminating difficult and easy cases.

Discrepancy in the Results Involving Counseling Difficulty and Counseling Outcome

One question raised by the present findings was why did the CPS and extraversion-introversion measures predict the counseling difficulty ratings reasonably well, but not the success of counseling outcome? Analysis revealed that counseling difficulty and counseling outcome were moderately correlated (r = .54, p <.001), yet only the former measure proved to be predicted by either cps or extraversion-intro-version scores. analysis of the relative variability in the two measures suggested a possible reason for the discrepant findings.

Specifically, as seen in table 10, ratings of counseling difficulty varied across the whole range of values (from 1 to 5), with indications that the counseling difficulty ratings were normally distributed. how-ever, in examining the ratings of counseling outcome, each counseling setting reported a high level of success with their clients. that is, out of the 127 cases in the present sample across the five counseling settings, 95 clients (75% of the sample) were rated by the therapists as improved (4) to much improved (5), whereas only two were rated as worse (2) and none as much worse (1).

J.C. Nunnally (75) wrote of this limitation, stating, "in many programs of evaluation (even good ones), the therapists who operate the program are highly committed to their evaluations being successful. They bring to the evaluation their bias that what they're doing produces successful outcomes" (p.104).

The effect of this issue of rater bias (with ceiling effects) is to reduce the range of correlational possibilities between CPS scores with counseling outcome. it is possible that the results of the present study didn't turn out more significantly because there were too few counselors that were willing to judge their cases as not successful. It's also quite possible that the restricted range of counseling outcome ratings may have kept the correlation between counseling difficulty and counseling outcome (of .54) lower than what it would have otherwise been.

It could be argued that client feedback about the success of their counseling might have yielded a more valid measure of counseling outcome. however, clients usually want to please their therapist by indicating they have improved even if they have not (Nunnally, p.105).

This information leads me to conclude that we need other ways of measuring counseling outcome besides (or in addition to) counselor and client evaluations. In future studies it is recommended that third party ratings by trained clinicians not involved in the counseling of the subjects being rated be considered. these data could be gathered in interviews with both counselor and client within one month of counseling termination and a rating be made based primarily on dropouts and premature terminations.

Table 10
Outcome Measure Ratings and Percentages(a) of all Counselors
Counseling Difficulty(b) Counseling Outcome(c)
1 2 3 4 5 1 2 3 4 5
11 38 40 27 11 0 2 30 64 31
(9) (30) (31) (21) (9) (0) (2) (24) (50) (24)

a =Percentages in parentheses
b =Counseling Difficulty:
1=very easy, 2=easy, 3=average, 4=difficult, 5=very difficult
c =Counseling Outcome:
1=much worse, 2=worse, 3=no improvement, 4=improved, 5=much improved

In rating the success or failure of counseling outcome, counselor's have differing standards. These standards, though unpredictable across individual differences of counselors, generally converge in common guidelines that need to be identified in future research regarding the present study and ratings of counseling outcome. Such guidelines include early termination, dropouts, counselor/client agreement and post-coping evaluations by counselor and client. Early termination is generally assumed by most investigators to be a function of resistance or lack of counseling readiness on the part of the client (Mendelsohn & Geller,1967). In addition, it is not common for counseling to be completed successfully in a few number of sessions. Outcome criteria rely heavily on "dropping out of counseling" vs "continuing in counseling" (Russell, Anderson, Atilano & Jurich, 1984, p. 242). If termination is not a collaborative agreement between counselor and client, successful counseling outcome is brought into question. On the other hand, client satisfaction and counselor's agreement usually indicate successful counseling outcome. Furthermore, if the client or counselor rate the presenting problem worse or not resolved, counseling is considered either unfinished or unsuccessful.

Not only was there evidence of a positive bias in the ratings of counseling outcome, the findings from the pilot study, in contrast to the present study, also raise the possibility that the CPS may "work" for some counselors but not for others. Consider the possibility that certain personality types among counselors may rate outcome measures differently. For example, in the pilot study, my ratings of outcome were based on what I consider was pragmatic logical assessment. My MBTI type is introverted-sensing-thinking-judging (ISTJ). Though I did not include the MBTI types of the thirteen therapists involved in the present study, it is my estimation that only one is similar to my personality type--introverted, sensing, feeling, judging (ISFJ). Most, I conjecture, are extraverted-sensing-feeling-judging (ESFJ) and extraverted-intuitive-feeling-perceiving (ENFP) types. Based on the results of my pilot study and the present study, could it be that ISTJ therapists are more willing to assess outcome measures more concretely and objectively than ESFJ's or ENFP's?

This is more than a self serving conjecture since, congruent with this line of reasoning, according to the MBTI manual, different counselor MBTI types do rate outcome measures according to their own preference types. For example, a sensing type counselor would rate outcome measures on the basis of facts and sensory information. An intuitive counselor would rely more on patterns of behavior and possibilities. A judging counselor would rate on the basis of structure fulfillment and organization of the client's response to counseling, whereas a perceiving counselor would rely on the client's spontaneous achievements and ability to adapt to change.

An added complication is in the strength of the counselor's preference scores. Carl Jung thought that individuals should be clear about their preferences. Higher scores, referred to as Clear (25 to 40) and Very Clear (41+) in the manual reflect an individual that is unambiguous and effective in response to life situations, as in evaluating counseling outcome measures. However, Slight (1-9) or Moderate (10-24) preference scores of counselors may reflect an ambivalence in rating outcome measures. A sensing counselor with a score of 25 will have more practiced ability with his or her preference than a sensing counselor with a score of 5. On the other hand, higher scores mean that the counselor is less practiced in using his or her opposite preference in handling assessments. Simply put, high outcome scores reflect minds made up and low scores, minds of ambivalence.

Last of all, though it was not a focus in the present study, counselor-client type similarity and difference needs to be considered as problematic in future studies. An ISTJ therapist may rate the out-come measures of an ISTJ client differently than an ENFP client. Interestingly, Robert Clark Vilas, in Counseling Outcome as Related to MBTI Client Type, Counselor Type and Counselor-client Type Similarity (88), found that greater similarity of counselor and client on the extraversion-introversion dimension was related to greater likelihood of mutual termination (successful counseling outcome). For future studies, it would be helpful to find four or five therapists of different types, i.e. ISTJ, ENFP, ENFJ, ISFJ, and compare their ratings of outcome measures across their sample's client typology.

Discrepancy in Results with Extraverts and Introverts

Given that the major part of Jung's Psychological Types is devoted to the history and description of extraversion and introversion as orientations to life (Briggs Myers, I., McCaulley, M. H., 85), perhaps it is not so surprising that the extraversion-introversion dimension surfaced as the only single dimension score to be significantly correlated with counseling outcome. As reported earlier, analysis revealed the CPS of extraverts to be moderately predictive for counseling difficulty and the preference score of introverts to be weakly (but significantly) predictive of successful or unsuccessful counseling outcome. This raises a difficult question, "why does the CPS work for extraverts and the preference score work for introverts?"

One possible explanation of this discrepancy may be in the difference between extraversion and introversion. Extraverts are more willing to divulge information to the outer world and therefore may have clearer preferences (higher scores) in the other three dimensions. Introverts are more private with personal disclosure and therefore may have less clear preferences (lower scores) in the other three dimensions. Extra-verts, therefore, show more of their personality functions to the outer world--and maybe on self-report tests of these functions--and introverts show their personality functions with more protection and privacy to the outer world. Following this line of reasoning, each and every dimension of extraverts is of utility in reflecting any strong preferences they may hold. Conversely, introverts may minimize somewhat the expression of their preferences in their test scores, and variations in their preferences, thereby, may be less predictive of counseling difficulty or outcome. In any case, only further research with a larger data base will be able to document the validity of this explanation.

Suggestion for Future Research

There are other research avenues to pursue besides the ones already mentioned. For example, counseling difficulty within the present study is understood as the degree of resistance encountered in the natural process of counseling. Resistance in counseling is held as a part of a universal and natural resistance to change. Resistance is viewed in the counseling community as a symptom of health and ego strength as a necessary part of change. As Lynn Hoffman has reported, "resistance might better be called persistence," (Anderson and Stewart, p.4). Resistance, however, can be viewed as dysfunctional persistences to relinguish maladaptive response patterns that lead to elaborate patterns of self destruction. After all, compulsive disorders are among the hardest to treat.

In order to test the validity of the constructs CPS and resistance, future research needs to study the relation of the CPS to other non-counseling measures of resistance to change. A simple approach would be to interview a random sample of noncounseling individuals and ask them to rate themselves as either resistant to change or open to change, rigid or resilient. Each respondent would then be given the MBTI. An analysis of their CPS scores with their ratings could produce supportive evidence for the justification of the CPS.


This study was designed to investigate the relation of a newly developed composite preference score (CPS) with the predictability of the counseling difficulty and counseling outcome. The findings were discussed in light of the results of therapists ratings in one career development center and five counseling group settings. In addition, re-sults of a pilot study were discussed in looking for replication of findings in the present study.

The findings of this study were:
1. Though the correlation was quite modest in size, higher composite preference scores (CPS) were associated, as predicted, with higher ratings of counseling difficulty.
2. The CPS was not predictive of counseling outcome.
3. The CPS was more substantially related when used with extraverted clients presenting one of three diagnostic conditions--alcohol/ drugs, family violence, or depression.
4. Conversely, counseling difficulty with introverted clients presenting with any of these three diagnostic conditions was moderately well predicted by their preference scores alone, but not by their CPS scores.

These findings were sufficiently conclusive in regard to the CPS and extraversion-introversion correlations to counseling difficulty, in-cluding its practical counseling utility, to recommend that a more quantitative view regarding preference scores may yield potentially im-portant knowledge to add to the typology information when working with counseling populations. Ultimately, such information should help counselors estimate counseling difficulty and success of counseling outcome.


Anderson, C.M., & Stewart, S. (1983). Mastering resistance: A practical guide to family therapy, New York, The Guilford Press, p.1-38, 151-206.

Briggs Myers, I., McCaulley, M. H. (1985). Manual: A guide to the development and use of the Myers-Briggs Type Indicator, Palo Alto, California.

Brownsword, A.W. (1987). It takes all types!, San Anselmo, California: Baytree Publication Company.

Goodheart, C.D. (1989). Short-term dynamic psychotherapy with difficult clients, Innovations in clinical practice: A source book, 8, 18-19.

Greenson, R.R., (1967). The technique and practice of psychoanalysis. New York, International Universities Press.

Holland, J.L. (1973). Making vocational choices: A theory of careers, Englewood Cliffs, N.J., Prentice-Hall, Inc.

Instruction Sheet, (1985). Palo Alto, California: Consulting Psychologists Press.

Johnson, S. M., & Greenberg, L.S. (1988). Relating process to outcome in marital therapy. Journal of Marital and Family Counseling, 14, 175-183.

Keirsey, D., & Bates, M. (1984). Please understand me, Del Mar, CA: Prometheus Nemesis.

Kroeger, O., & Thuesen, J.M. (1988). Type talk, New York: Delacorte Press.

McCaulley, M. H. (1990). Contemporary clinical applications of the Myers-Briggs Type Indicator. Innovations in clinical practice: A source book, 9.

McCaulley, M. H. (1990). The power of an idea. Bulletin of Psychological Type, 13.

Meisgeier, C., Meisgeier C., & Meisgeier M. (1989). A parent's guide to type: Individual differences at home and in school, Palo Alto, CA: Consulting Psychologists Press.

Meisgeier, C., Meisgeier, C., & Murphy, E. (1989). A teacher's guide to type: A new perspective on individual differences in the classroom, Palo Alto, CA: Consulting Psychologists Press.

Mendelsohn, G. A. (1966). Effects of client personality and client counseling: A replication and extension. Journal of CounselingPsychology, 13.

Mendelsohn, G. A., & Geller, M. H. (1963). Effects of counselor-client similarity on the outcome of counseling. Journal of Counseling Psychology, 10, 71-77.

Mendelsohn, G. A., & Geller, M. H. (1967). Similarity, missed sessions, and early termination. Journal of Counseling Psychology, 14, 210-215.

Munjack, D.J., & Oziel, R.J. (1978). Resistance in the behavioral treatment of sexual dysfunction. Journal of Sex and Marital Therapy, 4.

Myers, I. B., & McCaulley, M. H. (1985). Manual: A guide to the development and use of the Myers-Briggs Type Indicator, Palo Alto, CA: Consulting Psychologists Press.

Nunnally, J.C. (1975). The study of change in evaluation research: Principles concerning measurement, experimental design, and analysis, Handbook of Evaluation Research, Bever-ly Hills, Ca., Sage Publications, (1).

Russell, C. S., Atilano, R. B., Anderson, S. A., & Jurich, A. P. (1984). Intervention strategies: Predicting family therapy outcome. Journal of Marital and Family Therapy, 10, 241-245.

Stiles, W. B., Shapiro, D. A., & Elliott, R. (1986). Are all psychotherapies equivalent? American Psychologist, 41, 165-180.

Vilas, Robert Clark, (1988). Counseling outcome as related to MBTI client type, counselor type and counselor-client type similarity. Dissertation, The University of Iowa, 202.

Willis, C.G. (1983). Myers-Briggs Type Indicator, Journal of Counseling Psychology, 482-490.

Appendix A

Question Examples for Attitudes and Functions

The following questions are offered as examples of the forced choice format of the MBTI:

Another example asks, "Which word in each pair appeals to you more?"

Appendix B

Definition of Terms

Type refers to one of sixteen personality types:
ISTJ = introverted-sensing-thinking-judging
ISFJ = introverted-sensing-feeling-judging
INFJ = introverted-intuitive-feeling-judging
INTJ = introverted-intuitive-thinking-judging
ISTP = introverted-sensing-thinking-perceiving
ISFP = introverted-sensing-feeling-perceiving
INFP = introverted-intuitive-feeling-perceiving
INTP = introverted-intuitive-thinking-perceiving
ESTP = extraverted-sensing-thinking-perceiving
ESFP = extraverted-sensing-feeling-perceiving
ENFP = extraverted-intuitive-feeling-perceiving
ENTP = extraverted-intuitive-thinking-perceiving
ESTJ = extraverted-sensing-thinking-judging
ESFJ = extraverted-sensing-feeling-judging
ENFJ = extraverted-intuitive-feeling-judging
ENTJ = extraverted-intuitive-thinking-judging.

Omissions are questions in which none of the choices were answered. In research, cases are dropped if omissions exceed 25 in Form G and 55 in Form F.

Counselors refers to licensed clinicians that includes, in this study, a psychiatrist, licensed social workers, a licensed professional counselor, and marriage and family therapists.

Dominant function refers to the function that is assumed to be the strongest preferred function which becomes the controlling force of one's personality.

Auxiliary function refers to the function that is of secondary strength and provides "bench-strength" to the personality type.

Inferior function refers to the fourth function, the one that is op-posite the dominant function. A dominant function of Sensing, for example, would produce an Inferior Function of Intuition. The Inferior function is assumed to be unconscious and to carry the potential of asset and liability for the personality type.

The dominant process is the function which the individual relies on the most. The auxiliary process is used exclusively under circumstances where the individual is in a situation that requires an attitude opposite his or her preference. For example, an introvert is asked to give a speech to a large audience and mingle socially before and afterwards. The introvert who has not developed an adequate auxiliary function may find this task overwhelming and unmanageable. The dominant function is utilized in dealing with the world "out-there." However, the introvert's dominant function is more difficult to ascertain because he or she will be using their favorite (dominant) function internally while using their auxiliary function to deal with the world "out-there." It is converse because the introvert is consistently faced with the external world.

Description of Preferences

The results of the scoring give individuals information on their attitudes and two sets of functions in relating to the outer world (orientation). There are four general attitudes. The extraverted attitude is reflective of an outgoing social personality that is comfortable in crowds and public events. Extraverts live their life out loud. The introverted attitude manifests an individual's inner world of ideas. Introverts are comfortable with solitude, privacy and are selective in social engagements and relationships. Introverts internalize the outside world and personalize it before responding to it.

The two remaining major attitudes to the outer world are Judgment and Perception. In the judging attitude individuals thrive in structure and organization. They are purposefully systematic in making decisions and getting the job done. They are great organizers! The perceptive attitude represents individuals who prefer spontaneity, improvisions and unrehearsed action. They study incoming information with a curious openness and have no 'hurriedness' in their response. Judging people tend to make decisions quickly whereas perceptive people take their time. Both attitudes are valuable in the decision making process.

There are four main functions within these two sets of attitudes. Jung divided each attitude into two categories: judging into thinking and feeling; perception into sensing and intuition. Jung used the terms thinking and feeling to refer to rational functions that are directed toward converting life events into a harmonious relationship with laws of reason. Thinking is the function of logic and analysis. Thinking individuals connect cause and effect in behavior and decision making for action. Thinking individuals are concerned with justice and objectivity. Feeling individuals rely on values, subjectivity and emotions. Feeling types weigh the importance of other's values and seek harmony in relationships. They are more personal in expressing feelings.

Jung associated sensing and intuition types with irrational functions. He taught that these functions are tuned to the direction of life events and carry on most effectively when not restrained by rational criterion. The sensing function refers to perceptions experienced or observed by the senses. Sensing individuals prefer to know all the facts when making decisions and are quick to establish "what's so," what is happening in the present moment. They are realistic in their methodology, practicality and pragmatism. The intuitive function of perception refers to a person's preference for possibilities and relationships by way of insight. Intuitive individuals are future oriented, abstract and creative. An example that I like to use in clarifying the difference between sensing and intuition is in playing Seven-Card-Stud Poker. The sensing poker player will know how many cards have been played, how many are left to be played, how many Aces are showing and have a pretty good idea of the probability of whether or not the next card will be an Ace. The intuitive poker player will play a "hunch," a "third-eye" vision beyond reality that the next card is or is not an Ace.

Appendix C

History of the MBTI

The MBTI, based on the formal profession and theory of Carl G. Jung and his ideas about judgment and perception, was developed by Isabel Briggs Myers and Katharine Cook Briggs, "a brilliant mother-daughter team without the benefit of formal training," who discovered Jung's work in 1923 and began testing his theory by "type-watching" for twenty years. Their aim was to implement Jung's psychological theory of type. They began with testing high school and college students in order to facilitate students finding direction for their lives by understanding their preferences and strengths. In 1962 the MBTI became available to help career counselors "use the growing body of knowledge about type differences in school majors, occupations, and choice of work environments." It also became available to educators to "help apply teaching methods to meet the needs of different types and learning styles and to help teachers, administrators, parents, and students to work more constructively together." Business and industrial settings began to utilize the MBTI to help with staff development in groups and between individuals. The MBTI is used to "improve communications by clarifying the approaches most likely to elicit agreement and cooperation from each type."

Copyright 1992 Jerry Brinegar. All rights reserved.

Click to return to Chapter One.