Sanford Weisberg

Company S has designed a new mechanical heart valve designed to improve the health of people with heart disease. How can we tell if the new valve works, if it is safe, and if it is better--or at least not worse--than any existing products? To answer questions like these, we need to collect information: how long will the valve last, how strong is the material used to make it, what is the chance of the valve causing complications, what is the increase in life expectancy in patients who get the valve? Statisticians figure out how to collect the information to answer these questions, then figure out how to analyze it, and finally we figure out how to present it so that the conclusions to be reached are clear and truly represent the data. That statisticians can play a role in virtually every area of human affairs makes our subject exciting and challenging. In the last few years, I have worked on projects to equitably apportion the cost of a garbage-burning power plant to local taxpayers; to monitor the quality of air emissions from a large industrial facility; to study sex ratios in offspring of deer as a function of dominance (dominant females tend to have more female offspring); to devise a statistical sampling plan to monitor gas tanks at service stations for leaks; to study nitrogen fixation of various types of plants; to study the strength and other properties of waferboard as a function of the manufacturing conditions; and to study changes in the diets of senior citizens over the last 20 years. In addition to my interests in applying statistics to a wide array of practical and important problems, I maintain a regular program of research into statistical methodology. My main area of research is called statistical diagnostics, which consists of methods designed to help the analyst discover if models or assumptions made in an analysis are consistent with the data actually observed. If so, then an analysis may be reasonable and useful. If not, then either the models, the assumptions, or the data are called into question. Most recently, I have been working on graphical diagnostics that make use of the dynamic capabilities of computer graphics workstations.