The link between academic performance in secondary education and the subsequent performance of students studying architecture at university level is commonly questioned by educators and admissions tutors. This paper investigates the potential for using measures of cognitive style and spatial ability as predictors of future potential in architectural design education. The research investigates the relationship between the academic performance of three cohorts of architectural students and their cognitive style (as measured by the Cognitive Style Analysis), spatial ability (measured by the redrawn Vandenberg mental rotation test), and performance in secondary education (measured by the make-up of their portfolio of secondary qualifications). The results from the research provide little evidence to suggest that any of the measurements were good predictors of eventual performance. Neither were there any significant interactions between the test scores, performance, and gender. Nevertheless, there was evidence suggesting that students with certain cognitive styles were less likely to complete the course, particularly those with a verbaliser cognitive style. Furthermore, a significant proportion of female, wholist students also failed to complete the course. (Contains 1 figure and 7 tables.)