International Journal of Education 2012, 4 (4): 247-266This article is about the Predictive Validity of the University Communication Skills (CS) examination on the Students’ Overall Academic Performance. It is based on the study which was conducted at the Sokoine University of Agriculture (SUA). The objectives of the study were: firstly, to determine the extent to which the CS examination predicts the overall performance expressed in terms of the University GPAs; secondly, to identify whether the predictive validity varies by sex, type of school, and type of degree programme; and thirdly, to examine whether the CS examination is a better predictor for the university GPAs than students’ entry qualifications. The respondents were 358 finalists from six programmes selected randomly out of the 20 degree programmes at SUA. The findings indicated that the University CS examination positively predicted the overall students’ performance expressed in terms of GPAs (r=0.416, p< 0.01). Furthermore, the predictive validity of the CS examination scores on the University GPAs varied by sex and was higher for males (r=0.483, p<0.01) than females (r=0.272, p<0.01). Moreover, the CS examination scores predicted the University GPAs better than the students’ university entry points. However, the type of school attended at both CSEE and ACSEE levels and type of degree programme enrolled at SUA had insignificant influence on the predictive validity of the CS examination.