Michael L. Stein

Ralph and Mary Otis Isham Professor
Department of Statistics


Michael Stein graduated from MIT in 1980 with a B.S. in mathematics and received the M.S. and Ph.D. in statistics from Stanford in 1982 and 1984. After spending a year at the IBM T.J. Watson Research Center, he joined the faculty at the University of Chicago, where he is now the Ralph and Mary Otis Isham Professor of Statistics and the College. He was chairman of the department 1998-2001.

Most of Stein's research has been in the area of spatial statistics and its applications to environmental sciences and astrophysics. His current research focuses on statistical models and methods for natural processes in space-time and the computational challenges that arise in studying such problems.

Michael Stein is married to Laurie Butler, a professor of chemistry at the University of Chicago. They have one daughter, Ellyn, born in 1995. His outside interests include horseback riding and the guitar.


My research focuses on statistical models and methods for spatial and spatial-temporal processes. In particular, I am interested in the nature of the spatial-temporal interactions implied by these models and on developing statistical methods for assessing these interactions.

My main motivation for studying spatial-temporal processes is to describe variations in the physical environment. In recent years, most of my efforts in this direction relate to problems in climate and weather. Two topics of current interest are methods for combining output from deterministic climate models with observational data to produce realistic spatially and temporally resolved simulations of future temperature and precipitation fields, and ways of estimating temperature extremes that avoid using arbitrary cutoffs for what counts as extreme and that take proper account of seasonality, long-term trends and spatial structure.

Climate datasets (either computer-generated or observational) are generally quite large, leading to two major thrusts of my research. First, when one has lots of spatial-temporal data for the natural environment, it is often apparent that the process is nonstationary in space, time, or both. Thus, the development of nonstationary models plays a central role in my current research. Second, when one has large datasets with complex dependencies, computational issues are critical as they relate to evaluating likelihoods and running simulations. Much of my recent research has considered bringing to bear modern tools from numerical linear algebra to these computational problems.

Other topics of recent research projects include statistical properties of estimates of Gaussian process parameters fitted to deterministic functions and statistical methods for analyzing data from chaotic dynamical systems.