The Myers-Briggs Type Indicator ™ has had a distinguished history as a personality assessment instrument. Consulting Psychologists' Press1, publisher of the MBTl®, report that over 2 million copies are sold every year, making the Myers-Briggs the most popular personality test in print today. But MBTI users well trained in psychometrics have long recognized that the instrument has some unfortunate shortcomings which compromise its usefulness.
- Only 94 items of the 126 on Form G of the MBTl actually are scored in determining type just about 75% of the questions. In practical application, completing Form G takes about half an hour. If the 32 superfluous items were omitted, the questionnaire could be completed in substantially less time, much to everyone's benefit.
- Paired items on the MBTl (e.g. T, F) often have drastically different levels of validity for the dimensions they are supposed to be measuring. In practical terms, that means that, for example, the T selection may be scored and the F answer not, or one may be double scored and the other not, or one may be double scored and the other single scored. From one question to another, this situation may lead to "false positive errors" (e.g. the T selection being chosen by many "true" Fs) or "false negative errors" (e.g. the T selection being shunned by many "true" Ts).
- As a consequence of this item weighting problem, a subject may actually answer more questions reflecting a given preference and still be classed as preferring the opposite.
- On the T-F scales, men and women are not scored for responses to the same questions, or their answers may receive different weightings-not just because more men tend to favor the thinking function and more women tend to favor feeling--but rather because the items on these scales may contain some inadvertent sexual bias in wording or emphasis.
- In calculating an examinee's type, any MBTI preference score of 15 or less is considered to be statistically dubious (i.e., the person may well prefer the opposite )--and most examinees score at least one out of four preferences in this questionable range. For example, in one study conducted by The Leadership Dimension and Steven Coats of AT&T (RoBards & Coats, 1990,) 1090 Form G answer sheets were analyzed to reveal that nearly 65% showed at least one preference score of 7 or less and over 35% showed two or more scores in this same insignificant preference range! In our experience, using thousands of MBTls with hundreds of groups of all kinds, we have learned to develop our programs around the unfortunate and unsettling expectation that 75 to 100% of the four-letter type codes tentatively assigned to a given group of people on the basis of Form G results will not reflect the "true types" eventually selected by the clients. As trainers and psychologists, we have found the concept of type to be highly valid and incredibly powerful in many counseling, training, teaching and consulting applications, but we have had to confront the fact that the MBTI is often painfully weak at determining "true type" in any efficient or expeditious way.
- As clearly stated by Isabel Myers, the numerical preference score reflects only the degree of certainty that the individual actually prefers that choice. Nonetheless, the use of numerical scores misleads both administrators and examinees into thinking that the numbers reflect the frequency, competency or intensity of an individual's behavior. More than once, we've been asked by an eager trainee: "How high was my thinking score this time?"
- The construction of the MBTI randomizes items addressing the four preference pairs, ostensibly to prevent examinees from recognizing the pattern of their choices and to minimize the possibility that subjects might "pick" their type to satisfy some unknown motive or respond to perceived gender- or organization-related expectations. At the same time, the supportive materials for administering and interpreting the MBTI emphasize that the instrument is only an "indicator," and that the individual is the best judge of his or her type. These two elements of the MBTI's construction' and interpretation are mutually contradictory. First, the examinee is intentionally confused by the test's construction, and then he or she is asked to unscramble weak preference scores by picking the type description which best fits. From a practical standpoint, this conundrum may render a training session catastrophic.
- Although the MBTI is an indicator of preferences, not a test per se, it is administered like a test; it has to be scored like a test; the instructions ask the examinee to use a #2 pencil and stay inside the little bubbles on the answer sheet and not to mark both choices on the items; and the computerized report form returned by the publisher looks, to the average person, like highly objective, official and final test results.
For these and other reasons, we invented The Insight Game—an instrument which may be helpful to type users in a variety of counseling, training or teaching settings.
Going step by step through our difficulties with the MBTI and the resolution to these problems afforded by The Insight Game:
- All 36 choices (72 items) of The Insight Game are evaluated in determining the person's 4-letter type code. No items may be skipped. Sorting the cards, counting the cards in various stacks, re-evaluating the preferences by examining the content of all the cards in a pile, and picking a type description which fits takes most people ten minutes or less after instructions have been given. Even when people experience difficulty in choosing between two preference pairs, the entire process rarely takes more than fifteen minutes.
- Paired items (e.g. T, F) both are valid for the preferences being measured, even though some acceptable variance in the level of validity for each element may be present. (See detailed analysis of the eight "scales" in the appendix of this paper.) For the sake of scoring simplicity, all items are equally weighted.
- The individual's type is estimated by a simple majority of items selected for four preference pairs. For example, in no case can a person pick more T choices and then be tentatively classed as an F.
- On the T-F scale, as might be predicted from the results of the MBTI, more men prefer T and more women prefer F, but the T -F items, taken together, comprise valid sub-indicators of thinking and feeling preferences, both for men and women. (See detailed analysis of men's and women's responses in the appendix of this paper.)
- The validity of an individual's preferences is determined directly by the trainee, not by a numerical score. In the instructions to examinees, the professional may encourage each participant to review the content of cards in the "majority" stacks (E or I, S or N, T or F, and P or J) to see if those statements sound right for him or her. Next, the examinees are encouraged to locate the appropriate card in the Type- Temperament Sorting Deck which matches those four letters. If that description doesn't suit, the professional guides the examinee to find the "best fit" card. The Insight Game is a process, as much as it is an instrument for determining type.
- The result of The Insight Game is merely a four-letter code and the examinee's own selection of his or her "best fit" type. Because no computer-coded answer sheet, report form, graph or numerical scale is used, the client is not distracted by numbers associated with his or her various preferences.
- The game board and the sorting deck make it clear to the client that only four choices among two-preference pairs are being made by distributing the cards. No effort is made to hide the significance of the items or the purpose of the "game." At any time in the sorting, the "player" is free to review the cards stacked in any pile and move them, if he or she chooses. In the end, the client may read the cards in his or her preference stacks and find that the statements "tell a story" about the self.
- The Insight Game is clearly and unquestionably a game. It is non-threatening. It requires no #2 pencil, no test booklet, no answer sheet, no scoring templates, no computerized analysis. It treats students, clients, counselees and trainees as adults participating actively in their own type determination.
Three hundred and twenty adults who were familiar with the concepts of type and had been previously examined with the MBTI, were sent a four-page preference sorter. The four pages were labeled "Extraversion-Introversion," "Sensing-Intuition," 'Thinking-Feeling," and "Perception-Judging." Thirteen pairs of forced-choice statements were presented in columnar form on each page. In each case, statements related to a given preference were grouped together under the appropriate heading.
The 104 statements included in this preliminary instrument were composed by a licensed clinical psychologist with ten years' experience in using type in a variety of counseling, consulting and training settings.
A covering letter was included with the sorter, explaining the purpose of the study and offering participants an incentive for completing the questionnaire promptly. In addition, subjects were requested to provide the experimenters with their four-letter type code as determined by the MBTI and with their "true type," if different. Enclosed with the package were sixteen thumbnail type sketches which the subjects were asked to examine and critique.
One hundred fifty two responses to this inquiry were received. Ten submissions were eliminated from consideration because the subjects bad omitted more than four items on any section of the sorter or had substantially altered the content of more than four items on any section, or a combination of those two actions.
Tables 1-4 present standard type tables showing (1) ALL: True Type of Participants in Study, (2) MEN: True Type of Participants in Study, (3) WOMEN: True Type of Participants in Study, and (4) True Type of Participants Removed from Study.
After responses to the items were analyzed, 9 were selected from each preference to represent the final sorter. The original data were then re-evaluated with the reduced number of items. The obtained type code was then compared with the, individual's "true type" (or MBTI type, if no other type was listed).
The results are shown in Tables 5-9: (5) ALL: Participants Accurately Typed by The Insight Game, (6) MEN: Participants Accurately Typed by The Insight Game, and (7) WOMEN: Participants Accurately Typed by The Insight Game. Tables 8 and 9 give the type break-down for (8) Participants Accurately Typed by The Insight Game, but NOT by the MBTI, and (9) Participants Accurately Typed by the MBTI, but NOT by The Insight Game.
As these results show, 133 of the 142 participants were accurately typed by items used to develop The Insight Game, while 130 were "true-typed" by the MBTI. Of the 568 possible preference selections, questions used for The Insight Game scored 559 hits, while the MBTI was correct on 556 choices, a statistically non-significant difference.
Click here to open Research Study Tables
Myers Briggs Type Indicator and MBTI are registered Trademarks of Consulting Psychologists’ Press, Palo Alto, CA.