The relationship between students' and instructors' personality types and the students' ratings of their instructors was investigated using data obtained from the Vocational Preference Inventory (VPI) for personality types and the Student Perception of Instruction (SPI) scale for students' ratings. Thirteen instructors and 464 students from two community colleges participated. The data were analyzed by the Johnson Max cluster analysis, multivariate analysis, and one-way analysis of variance procedures. The following questions were investigated: (1) When students are clustered according to the homogeneity of their ratings across the first six items on the SPI, how many groups will emerge in a class? (2) Will these cluster groups represent a predominance of student raters with matching personality types? (3) When ratings in the 13 classes are examined separately, will students with matched personality types have similar ratings on the six items on the SPI? (4) When students and instructors are grouped according to their personality types, to what degree will the interaction of personality types of students and instructors influence the ratings on the seven items of the SPI? and (5) To what extent are ratings on the "Overall Rating of the Instructor" item related to the level of congruence of student's/instructor's personality types? The results were the following: Sixty-five percent of the students clustered into groups in the 13 classes; however, there was no evidence that personality type was the basis on which clusters were formed. In only one of the 13 classes did students with matched personality types have similar ratings on any of the rating items. It was concluded that this class was not a representative sample of the groups. When data were grouped for question four, statistically significant interactions occurred between Realistic students and Investigative instructors on five of the seven rating items; however, these interactions did not support Holland's assumptions that students would respond more positively to instructors who had personality types similar to the students'. Finally, although the fifth analyses yielded statistically significant interactions, again these were not in the direction expected and did not support Holland's assumption of personality/environment congruence and satisfaction. / Ed. D.