This project examined the use of quantitative statistical models to bridge the gap between commonly available university data and knowledge for decision making. A method using a multivariate statistical model is discussed in the context of academic quality in higher education. A "problem-oriented" data base designed to study quality and using quality determiners, quality indicators and proxy quality indicators was built using data from the 170 doctoral institutions in the United States. Using this tool, the study examined five institutional profiles: University of California-Berkeley, 1982 and 1986; University of California-Berkeley and Michigan State University, 1986; Average State of California Doctoral Institution, 1982 and 1986; Average State of Michigan Doctoral Institution, 1982 and 1986; and University of California-Berkeley 1986 Actual and Simulated Profiles. The paper suggests two ways of using data to improve decision making: (1) by showing how administrators could view the impact of their past decisions on the academic quality of their institutions; and (2) by simulating, for administrators, the impact of alternative decisions on the academic quality of their institutions. Concerning the profiles, the study noted the following characteristics; that they: are stable over time; reflect national trends; can uncover major institutional, state, and regional policy differences; show a high association between faculty compensation and faculty reputational quality; reflect the strength of the relationship between the Quality Determiners and the Quality Indicators, and reinforce the strategies currently used by high-quality universities. The paper includes 18 references and 5 tables. (JB)