### Abstract:

The purpose of this study was to test the predictive ability of the Personality Assessment System (PAS) with special emphasis placed on predicting the personality patterns of two groups of mathematics students. Two ancillary purposes were also identified. These included the possible contributions to academic/vocational counseling and the addition of the findings to a PAS validation data bank that has been established.The sample included 26 selected volunteers from Purdue University and 26 selected volunteers from Ball State University. These students were identified into two distinct groups. The Purdue group was composed of senior students who were majoring in pure mathematics, while the Ball State group was composed of students who were majoring in mathematics education. Each group contained an equal number of males and females.Each student was administered the Wechsler Adult Intelligence Scale (WAIS) from which a personality profile was derived. This profile was based on the theoretical constructs of the Personality Assessment System.The hypotheses in the study were in the form of predictions. The predictions attempted to identify the personality patterns of the two groups of mathematics students.The resultant data were analyzed using percentages, a t test for independent groups, and a chi square analysis. Trends in the data, rather than statistically significant outcomes, were the main focus of the analysis.The predictions generated by the researcher correctly identified eight of 26 or 30% of the specific expected PAS patterns of the Purdue group. When the original predictions for the Purdue group were expanded to include all possible mathematics patterns, 23 of 26 or 88% of the Purdue group was correctly identified. The predictions correctly identified eight of 26 or 30% of the specific expected PAS patterns of the Ball State group. When the original predictions for the Ball State group were expanded to include all possible mathematics patterns, 15 of 26 or 57% of the Ball State group was correctly identified.Trends were suggested by the data for other types of measurement. The most consistent trend indicated by the data was the preponderance of IRU Primitive personality patterns in the Purdue group as compared to the general population, and the preponderance of IRA Primitive personality patterns in the Ball State group as compared to the general population.Another trend indicated by the data was the effect of a low Digit Symbol score on academic achievement in the Purdue group. Members of the group with a low Digit Symbol score consistently obtained a Grade Point Average below the mean Grade Point Average of the group.A final trend in the data was found between the combined groups of mathematicians and the general population on the R-F dimension of the PAS. The general population tends to be F or flexible, while the mathematicians tend to be R or regulated.The findings indicated that the PAS could adequately identify the personality patterns of a group of mathematicians. Rather than one "mathematical personality" being identified, the trends in the data point to many mathematical personalities. Recommendations for further research were made.