UBC Theses and Dissertations
A statistical analysis of some aspects of the Johnson temperament analysis Hammett, Joseph Francis
This study examines a number of aspects of the Johnson Temperament Analysis based on the responses of 48 male graduates in the Department of Education at the University of British Columbia. Item difficulty is considered from the reports of 39 students. The mean number of difficult items per student is 21.6 or 12% of the total. It is found that the difficulties can be classified under five headings: (a) Lack of experience with the topic. (b) Difficulties of comprehension. (c) Difficulties of evaluation. (d) Response dependent upon circumstance. (e) Two or more notions in one question. It would appear that a number of these difficulties could be avoided if a revision of the Analysis were undertaken. The adjustment of the group in terms of the Analysis is considered. The means are in the acceptable zone or better in all traits except Cordial, where the mean lies at the borderline between acceptable and improvement desirable. Age differences are not found to be statistically significant. There is some slight advantage of better adjustment in terms of the Analysis for the older group. Reliabilities of the trait tests, calculated by the split half technique are found to range from .32 to .82. In general these support the published figures of the test author. A phenomenon of difference between the first and second halves of the Analysis is noted. No definite conclusions are presented to account for this, but it is suggested that further research might determine the advisability of having unscored practice items at the b beginning of such inventories. The inter-correlations of the trait tests are calculated by the product-moment formula and are found to be comparable to the tetrachoric r's reported in the Manual of Directions, The former are factored by the Centroid Method and five factors extracted. The axes are rotated by the single-plane method and a simple structure obtained. This is discussed in terms of Thurstone's criteria and is considered to be reasonably satisfactory. An attempt is made to identify the factors obtained, one being named with a high degree of assurance. It is Factor II, Social Introversion-Extraversion. Factors I and III are identified with some assurance as Emotionality and Rathymia respectively. The remaining two factors are most tentatively, and with considerable reluctance, named as Objectivity and Cooperativeness. It is suggested that, if a shorter form of the Analysis be required, traits D and H would give a fair index of the general adjustment of the individual in terms of the five factors named. It is felt that, in the case of the 48 students used as the basis of this study, the nine traits defined by the test author are not distinctly measured. Since the group is small and homogeneous, no generalizations to the general population are made.
Item Citations and Data