A new class of matching problems that models centralized college admissions via standardized tests is presented. The allocation mechanism that is used in real-life applications of this problem in Turkey is analyzed. It is shown that this mechanism, multi-category serial dictatorship, has a number of serious deficiencies, most notably inefficiency, vulnerability to manipulation, and the potential of penalizing students for improved test scores. Exploiting the relation between this class of problems and the celebrated college admissions model (Gale & Shapley [4]), an alternative mechanism is proposed that overcomes these deficiencies. This mechanism-the Gale-Shapley student optimal mechanism-is characterized as "best" in this context. Journal of Economic Literature Classification Numbers: C71, C78, D71, D78. (C) 1999 Academic Press.