Thesis (PhD)--Stellenbosch University, 2002. / ENGLISH ABSTRACT: The major concern in the world of work has been the mismatch between a person's vocational interests and job pursuits. The understanding has been that jobs make differing demands on people and that the abilities of the individual and demands of the job have a bearing on productivity hence the need for matching people to jobs. The most notable attempts to identify and organize vocational interests have been Holland's theory of vocational interests and personality types. Holland asserts that people can be categorized predominantly as one of six personality types, namely, realistic (R), investigative (I), artistic (A), social (S), enterprising (E), and conventional (C). The review of literature has shown that person-environment fit explains the degree of match between occupational pursuits and one's vocational interests and that before a measure can be used across cultures, its construct validity must be established in each culture. The current study was motivated by the Zimbabwe Public Service Commission's quest, in the absence of vocational interest measures tailored for Zimbabwe, for assessment tools that can assist in predicting suitability for recruitment or promotion. The study was therefore designed to (a) develop a measure of vocational interest validated on the Zimbabwe population, (b) build a model for predicting and classifying people into job sectors and (c) assess the adequacy of Holland's RIASEC structure for assessing person-job fit in the Zimbabwe Public Service. To achieve this, a correlational research design was used. The vocational interest measure and the MB-10 were the two instruments used for data collection. A sample of 500 public servants representing six occupational sectors in the Zimbabwe Public Service participated in the study. Statistical approaches to data analysis included reliability and validity analysis, factor analysis and multiple discriminant analysis. The results of the reliability coefficients were within acceptable levels. The subscale reliabilities of the vocational interest measure ranged from ra = 0.85 to ra = 0.89. Overall, the concurrent validity of the vocational interest measure was established. Factor analysis and correlation coefficients statistic assessed the adequacy of the hexagonal ordering of the RIASEC types. Factor analysis was computed resulting in 8 factors being extracted instead of the theoretically conceptualized 6 factors. The 8 extracted factors accounted for 65.88% of the total variance. Holland's theory affirms that occupational types that are more proximate on the hexagon are more similar than types that are more distant. Results of the RIASEC intercorrelations showed strong correlations between Realistic and Investigative (r = 0.69) and moderate correlations with four of the adjacent types. However, low and negative correlations were obtained between Conventional and Realistic (adjacent types) (r = -0.14) and between Investigative and Conventional (alternate types) (r = -0.11). On the other hand, there was strong correlation between Realistic and Social (opposite types) (r = 0.25.) The model for predicting and classifying people into appropriate job sectors was developed. The model's utility was confirmed using the scores from the sample data. The cross validation table obtained an error rate of 0.29, an indication of a relatively good model. Given the results of the present study, it appears the study hypotheses were generally supported. Further work in refining the model is recommended.