Statistics Department

Modern Statistics has emerged from the probabilistic reasoning of the nineteenth century, from the design and analysis of agricultural experiments in the early twentieth century that gave acceptance to this NEW field of study and from the coherent organization of surveys to replace censuses for many governmental purposes. Now the thrust of the discipline is toward the problems of making efficient use of very large and very small experiments and observational studies, toward drawing information from high-dimensional observations interrelated in complex ways, and toward predicting events for complex systems and quantifying the uncertainties inherent in these projections. Modern statistical methodology, too, has emerged from its roots in calculator computation and approximation to focus on computationally intensive and/or high-dimensional problems in visualization, analysis, and approximation. The statistical topics of design and analysis continue to exist, but the problems are much harder, more likely to depend directly on one or more substantive disciplines involved in a research investigation and to demand greater precision in the final objective, whether estimation or prediction.

Thus over the past half century, as with other disciplines, the mathematics used in Statistics has become more sophisticated and more diverse; and computing has become an essenital part in almost every aspect of statistical theory and practice. Also, the linking of a mathematical/statistical formulation to the substantive field is now preeminent, for an inaccurate understanding leading to a gross approximation in the statistical formulation is no longer good enough to be satisfactory.

But the need for inference in the presence of uncertainty and some understanding of the magnitude of the uncertainty in that inference remains at the core of the discipline. As research has become more interdisciplinary, and as technology continues to become more powerful but also more intricate, the ONE-VARIABLE-AT-A-TIME approach used in experimentation has become patently untenable and the BRUTE FORCE method of trying every combination of factors is clearly infeasible. So the need for efficient inference and for assessment of its quality is expanding.

The Statistics programs at Case Western Reserve University reflect these changes in the discipline with strong cores to underpin the fundamental concepts of statistics, with courses in a variety of methodologies, with serious connection to a substantive area for application of statistics and practice in scientific collaboration, and with forums for the development of articulate oral (and written) presentation.