Perceptual and Motor Skillr, 1983, 56, 127-132. @ Perceptual and Motor Skills 1983 THE GHISELLI SELF-DESCRIPTION INVENTORY: CONCURRENT VALIDITY VERSUS SUPERVISORS PERCEPTIONS ROBERT B. BOWIN California State College, Sun Bernardino' Summary.-The Self-description Inventory was designed to measure the traits of middle managers. 27 students who were' managers completed the inventory and were evaluated by their immediate supervisors as more or less successful. Managers' scores and supervisors' rankings of their success generally did not correlate, suggesting a problem of concurrent validity. However, supervisors' rank of importance of the Self-description Inventory traits in determining more or less successful managers also did not correlate, suggesting a problem in perception. An earlier study (Bowin & Leonard, 1981) reported that data for a sample of 24 middle managers who completed the Ghiselli Self-description Inventory did not distinguish between those middle managers who were ranked by their supervisors as more versus less successful. The present study replicates the original with new analysis as to a possible explanation for the difference in findings. The Ghiselli Self-description Inventory is an empirically based instrument designed to identify the traits and talents of middle managers. Previous studies present varying degrees of support and non-support (Frederiksen, Jensen, & Beaton, 1972; Brief, Aldag, & Chacko, 1977; Miner, 1978). The inventory is a questionnaire of 64 word-pairs from which the respondent chooses one each of 32 pairs that least describe and one each of 32 pairs chat most describe the individual. GhiselLi (1971) found, of the 13 traits he ultimately selected to be assessed, six formed a cluster described as playing a major role in managerial talent with another three forming a group classified as minor. The remaining four traits were considered not pertinent. Ghiselli classified the following traits as major: Supervisory ability, need for occupational achievement, intelligence, the need for self-actualization, selfassurance, and decisiveness. Three traits classified as minor include the lack of the need for security, the lack of working-class affinity, and initiative. The remaining four traits were considered by Ghiselli to have no pertinence to managerial talent and include need for high financial reward, the need for power, maturity, and masculinity/femininity. Of the 13 traits, Ghiselli found that supervisory ability ". . . has a unique position with respect to managerial talent. It is the trait which plays the most important role, and stands out clearly and is apart from all of the other traits" (Ghiselli, 1971, p. 99). '5500 State College Parkway, San Bernardino, CA 92407. The methodology by which Ghiselli determined the relative importance of the traits required that each satisfy three conditions: First, on the average, managers should obtain the highest score, line workers lowest, and line supervisors (foremen) should fall in between. Second, there should be a substantial relationship for managers between the trait and their success. Third, the relationship between the trait should be highest for managers with line supervisors in the middle, and the lowest for workers. For each of the three conditions above, Ghiselli assigned zero to five points indicating the degree to which it was satisfied. These points were added together resulting in an overall score for the trait (Ghiselli, 1971, p. 98). METHOD The present study was conducted in an evening undergraduate/graduate class in business policy that was part of a degree-oriented program conducted at a high technology company located in the Pacific Northwest. There were 31 students in the class, of whom 27 were employed by the high technology firm. Of the 27 participants, 21 were graduate students and 6 were undergraduates. Their tuition was paid by the company and there was a backlog of several hundred prospective students waiting to enter the program. These students were selected on the basis of their potential for managerial success and were a l l in middle management. During the first few weeks of classes, the 31 students were administered the Ghiselli Self-description Inventory. Students were told beforehand that the inventory was part of a continuing study of management by the author, would be presented only as group data, and that all information would be coded to protect the individuals' identity. Upon conclusion of the course, the names and addresses of the students' immediate supervisors were obtained through the cooperation of the company's personnel department. These supervisors were then mailed a covering letter with a return envelope requesting their cooperation in the study by completing two forms. One form requested the supervisor to indicate an evaluation of their subordinate (the scudent) as to whether rhat individual was considered to be more successful or less successful. Ghiselli recognized that the supervisor's decision would be subjective and the present study utilizes the same procedure of two categories, more or less successful, as in the original research (Ghiselli, 1971, p. 25). The other form was a Likert-type scale listing the 13 traits requesting the supervisor to rank the importance of the 13 traits in considering whether their subordinate was a more or less successful manager. The scale ranged from 1 (not important) to 5 (very important). All the supervisors responded by completing both forms for the sample population of 27 student/managers. The non-parametric Kruskal-Wallis one-way analysis of variance by ranks (Siegel, 1956) was used in this study as in the original one to analyze con- 129 GHISELLI SELF-DESCRIPTION INVENTORY current validity. The Pearson correlation coefficient from the Statistical Package for the Social Sciences (SPSS) was used to interpret the supervisors' responses to the Self-description Inventory and their ranking of trait importance. RESULTS The data presented in Table 1 suggest that the scores do not differentiate becween those middle managers who were ranked by their supervisors as more or less successful. This finding replicates the earlier study. Further, this finding is supported whether only the major traits are considered or whether the major and minor traits are considered or whether all 13 traits are considered. This is important since Ghiselli (1971, p. 99) originally ranked the importance of the traits in determining managerial talent as to major, minor, TABLE 1 Major Traits Less More Major and Minor Traits More Less All 13 Traits More Less and not pertinent. Table 1 demonstrates the lack of a significant relationship between supervisor's rankings and subordinate's scores regardless whether the major traits are considered, or both the major and the minor traits, or all 13 traits. T o reject the null hypothesis of no difference between the two groups considering only the major traits, one would have to be willing to accept a level of significance greater than .50 and less than .70. Considering both the major and minor traits, the level of significance range decreases to greater than .20 and less than .30. With all 13 traits (including the four regarded as not pertinent), the range decreases further to greater than .10 and less than .20. The scores in Table 2 generally demonstrate lower correlations with the supervisors' ranking of more or less successful managers than in Ghiselli's original study. An interesting divergence is the two major traits with negaand need for self-actualization tion coefficients, supervisory ability (-.lo) (-.12). While neither of these correlations is statistically significant, Ghiselli obtained a positive correlation (.46) for supervisory ability, deeming it the dominant element of managerial talent. The only significant correlation was that of the minor trait, Initiative (.35). In total, the results of Table 1 and Table 2 are not supportive of the concurrent validity of the inventory. TABLE 2 CORRELA~ONS BETWEENSELF-DES(XIPTION INVENTORY TRAIT SCORES AND SUPERVISORS' RANKINGOF MORE VERSUS LESS SUCCESSFUL MIDDLEMANAGERS Major Traits r Supervisory Abiliq -.lo Need for Occupational Achievement .24 Intelligence .04 Need for Selfactualization -.I2 Self-assurance .23 Decisiveness .22 Major Total .24 Minor Total .23 Total 13 Traits .25 * P < .05. tGhiselli's sample correlations. rat .46 .34 .27 .26 .19 .22 Minor Traits Need for Job Security Working Class Affinity Initiative Not Pertinent Traits Need for High Financial Reward Need for Power Over Others Maturity Masculinity/Femininity - f ro -.05 -.30 .12 -.I7 .35* .15 . - r fa -.I6 -.I8 .13 .03 .12 .13 -.03 -.05 To determine why the present study does not demonstrate concurrent validity with the Ghiselli research, Table 3 analyzes the supervisors' rank importance of the 13 traits in determining a middle manager as more or less successful. With a rating of 5 being very important, an arbitrary scale would be 4 and above a major trait, 3 and above a minor trait, with the remainder not pertinent traits. Using this schema, the supervisors' average ratings would realign the traits to consider the most important major trait to be Initiative (4.70) followed by Supervisory ability (4.44), Decisiveness (4.33), Need for ~ c & ~ a t i o n aachievement l (4.33), Maturity (4.22), and Need for self-assurance (4.07). The minor traits would then be Intelligence (3.77) and Need for self-actualitation (3.51). However, only two traits from this revised listing are statistically significant, Supervisory ability (.013) and Need for job security (.029) with another, Maturity (.061), close. Yet, the correlations for the Supervisory ability and Maturity traits ate negative. Also, Ghiselli found that it was a lack of the need for job security which is a part of managerial talent not the need for job security. The supervisors in the present study at- GHISELLI SELF-DESCRIPTION INVENTORY 131 TABLE 3 BETWEENSUPERVISORS' RANKEDIMPORTANCE OF GHISBLLI TRAITS DETERMINATION OF MOREOR LESS SUCCESSFUL MIDDLEMANAGERS CORRELATIONS AND Traits P M Rank Importance Ghiselli Traits Major Traits Supervisory Ability Need for Occupational Achievement Intelligence Need for Self-actualization Self-assurance Decisiveness Minor Traits Need for Job Security Working Class Affiniry Initiative Not Pertinent Traits Need for High Financial Reward Need for Power Over Others Maturity Masculinity/Femininitg tach differing degrees of importance to the traits than the supervisors in Ghiselli's research. DISCUSSION The replicated part of this study supports the earlier finding that the Ghiselli Self-description Inventory does not distinguish between supervisors' ranking of middle managers as to being more or less successful. The correlation coefficients of the present study are generally lower than those of the Ghiselli study. Some correlations are negative when Ghiselli's are positive and vice versa. This could suggest a concurrent validity problem with the small sample, since the instrument appears not to measure what it purports to do. The supervisors' ranking of the importance of the 13 traits is quite different from the rank in Ghiselli's research. So it is not surprising that the correlations differ. It is interesting to note that the supervisors valued the supervisory ability trait as important yet downgraded that importance when determining whether a middle manager was more or less successful. Further, the supervisors found two traits highly important that Ghiselli classified one as minor and the other not pertinent. Perhaps the supervisors are behaving in accord with that old saying, "do what I say, not what I do." This suggests that the supervisors perceive other traits or combinations not present in the small sample, which take precedence in their evaluation. However, it is pos- 132 R. B. BOWIN sible that the small sample does not have concurrent validity. Yet, two small samples remain two small samples. Regardless, it appears prudent to approach the application of the Selfdescription Inventory with caution. There is a need for further research with larger samples before definitive conclusions may be drawn. REFERENCES BOWIN,R. B., & LEONARD,M. R. The Ghiselli Self-description Inventorg as a predictor of managerial success. Psychological Reports, 1981, 49, 291-294. BRIEF, A. P., ALDAG,R . J., & CHACKO,T. I. The Miner Sentence Completion Scale: an appraisal. Academy o f Management lournal, 1977, 20, 635-643. FREDERICKSEN, M., JENSEN,O., & BEATON,A. E. Prediction of organizational behavior. New York: Pergamon, 1972. GHISELLI,E. E. Expjor&ions i n mnagerial talent. Pacific Palisades, CA: Goodyear, 1971. MINER,J. B. The Miner Sentence Completion Scale: a reappraisal. Academy o f Managemertt Journal, 1978, 21, 283-294. NIB, N. H., HULL,C. H., JENKINS,J. G., STEINBRBNNER, K., & BENT, D. Stlistical package for the social sciences. New York: McGraw-Hill, 1975. SIEGEL,S. Non-parametric statistics for the behavioral sciences. New York: McGrawHill, 1956. Accepted December 14, 1982.