Thomas PPA Methodology
Original Construction Methodology
The first empirical trials were conducted by Hendrickson on a small group of 115 people (67 males / 48 females) in 1958 and the occupational distribution was: 46 college students, 17 teachers, 27 supervisors, 16 other professionals, 13 office workers, 6 miscellaneous.
Frequency distributions of responses were made and words were re-combined in tetrads such that each tetrad contained a word relating to each dimension. Moreover, attempts were made to combine words of relatively equal response strength in order to reduce the effects of social desirability in response patterns. High response words were grouped together with other high response words, low response words with other low response words. 76 of the original 96 words were absorbed in this manner and five extra tetrads were constructed to bring the total once more to twenty four. Of the words retained 39% are the same as in Marston’s original model.
Test, Revise, Retest
The revised Thomas PPA form was administered to a larger and more representative sample group of 500 (388 male/112 female) divided between the following occupations groups: 212 managers, 128 professionals, 62 clerical, 38 salespeople, 34 machine operators, 36 miscellaneous.
A random sample of 100 was drawn from this group to determine split half reliability and inter-correlation among the four factors. The results indicated that the Personal Profile had a satisfactory internal consistency when assessed in this way.
To eliminate non discriminating items from the scoring key, an item analysis was initiated. A random sample of 185 (130 male/55 female) was drawn from a population of 1200 with an occupation distribution of 89 managers, 35 technicians, 26 office workers, 12 engineers, 12 salespeople, 6 staff and 5 miscellaneous. The internal consistency was confirmed and the scoring key adjusted. At this stage, the Marston dimension of SUBMISSION was changed to STEADINESS and the Marston dimension of INDUCEMENT changed to INFLUENCE.
A random sample of 100 (75 male/25 female) was selected to test the new scoring key and the results correlated against the original trials. To develop and add to the reliability, test/retest trials were conducted on a sample of 72 (47 male/25 female) with the new scoring keys.
Submission to the American Psychological Society
This research in the early 1960s and the final version formed part of an extensive paper on the question of behavior in the work place. This paper was submitted to the American Psychological Society both in terms of the results the Thomas PPA achieved and the methodology of the research.