A Construct Validity Study of the Myers-Briggs Type Indicator

Abstract

This study assessed the construct validity of the Myers-Briggs Type Indicator, MBTI (Myers, 1976). The rationale was that friends or relatives can make judgments about an individual which will be associated with his/her predominant personality type. Forty-eight subjects rated themselves on two seven-point Likert scales designed to assess behavioral styles. These inventories (designated, "Behavioral Style Inventories") were designed for this study based on the operational definitions of type outlined by Myers in the manual to the MBTI. One inventory assessed perceptions held by subjects of themselves (Form S); the other was a measure of perceptions of their ideal selves (Form I). Forty-five of the subjects were also independently rated by their spouses who used a similar form of the inventory (Form M). All subjects then took the MBTI. Scores were converted to type categories and compared by using the coefficient of agreement for nominal data, Kappa. Self-typing on Form S of the Behavioral Styles Inventory (BSI) showed no agreement with the MBTI. Self-typing based on use of Form I (measure of ideal self) of the BSI showed no agreement with that from the MBTI; whereas, Kappa coefficients comparing the type categories of extrovert-introvert (E-I), sensation-intuition (S-N), and thinking-feeling (T-F) between the MBTI and the form of the BSI filled out by the spouses all showed significant positive values: E-I, Kappa = .70, P < .001; T-F, Kappa = .78, p < .001; S-N, Kappa - .43, p < .01. The Kappa for the J-P scale was not significant (Kappa = -.08. The construct validity of the MBTI scales of Extroversion-Introversion, Sensation-Intuition, and Thinking-Feeling was supported, whereas that of the Judging-Perceiving scale was not.