Creation of a Department of Statistics: 1968–2011
There was little organized research activity in statistics prior to 1968 at the University of Pittsburgh, outside of the Department of Biostatistics. There were isolated individuals in various departments either doing statistical research or with an interest in statistical research, but no center for such activity existed. Non-calculus introductory statistics courses were taught at the undergraduate and graduate levels in several departments as a preparation for courses in research methodology in their fields, and the Department of Mathematics taught two-term calculus-based probability and mathematical statistics courses for undergraduates and for graduates.
In other universities, mathematical statistical research usually started in departments of mathematics. An Italian probabilist, Giuseppi Pompilj, spent a few years in the Department of Mathematics in the 1960s, but until Henry Block was hired as an assistant Professor in Mathematics in 1968, there was no one in mathematics interested in statistical research. In 1971, Asit Basu came as an Associate Professor from Northwestern, followed in the same year by Tom Savits. Basu was successful in inviting senior statisticians with research interests in reliability to visit for short stays, notably Milton Sobel, Frank Proschan and John Gurland. J. K. Ghosh was also a visitor for several years. The new chair of the department, Gene Deskins, encouraged development of this statistical research group and also supported the change of the name of the department to the Department of Mathematics and Statistics. In 1976, P. R. Krishnaiah was hired to coordinate statistical research in the university and to provide consulting. Krishnaiah in turn brought in C. R. Rao as a University Professor. Several other statistics faculty members were hired. Soon after Allan Sampson joined the group as associate professor in 1978, he and Block set up a comprehensive statistics program with a Bachelors of Statistics and Masters in Applied Statistics. By 1982, the group had grown and prospered to the point that the Conference Board of Associated Research Councils ranked the group as the most improved statistics group in the country, also ranking it in the top third of statistics groups. In that year, David Stoffer and Satish Iyengar were hired by the statistics group and the Center for Multivariate Analysis was established. The Journal of Multivariate Analysis was published out of the Center for Multivariate Analysis.
The Center of Multivariate Analysis was intended to be a focal point for statistical research in the University of Pittsburgh. Its permanent members were Rao and Krishnaiah (who served as director). A list of 30 team members prepared at the time includes seven members of the Department of Mathematics and Statistics, seven members of the Department of Biostatistics, three from Psychology, two members each from the Departments of Economics and Sociology, and one member from each of the departments of Electrical Engineering, Industrial Engineering, Psychiatry, Medicine, Anthropology, Neurological Surgery, and Radiology. Completing the list were T. L. Saaty (Graduate School of Business) and N. H. Timm (Planning and Budget, and author of a respected textbook on multivariate statistical analysis).
Some of the faculty members in the statistics group wanted to create a separate Department of Statistics, but Rao and Krishnaiah were concerned about additional administrative responsibilities (including an undergraduate major). However, Krishnaiah’s premature death in the summer of 1987 and Rao’s departure for Penn State (taking the Center for Multivariate Analysis with him), changed matters greatly. In 1989, Leon Gleser came as a full professor with the understanding that the statistics group would work to separate from the mathematicians and create a department of their own. Although the university administration agreed in principle, it took 8 years of hard work to complete the job. A key step was the consolidation of statistics service courses in the School of Arts and Sciences under the control of an Applied Statistics Education Committee, for which the new Department of Statistics took a leadership role. In January 1997, the Department of Statistics was officially created, with Allan Sampson as its first chair. In 1999, Satish Iyengar became the chair of the department, and with the exception of two 1-year leaves of absence, when Henry Block acted as chair, has continued to serve as chair to the present day.
Although the Department of Statistics was planned as a small department, there was expectation of some modest growth as enrollments grew and new research grants were obtained. Unfortunately, soon after the establishment of the department, a freeze on tenure-stream appointments in the School of Arts and Sciences was imposed. The department did add two non-tenure-stream faculty members, justified by an explosive growth in their service course enrollments. The two new faculty members thus added, Senior Lecturers Carl Bodenschatz and Nancy Pfenning, have since won prestigious Arts and Sciences and university-wide teaching awards. Dr. Bodenschatz also was responsible for reviving and growing the undergraduate major, and helped devise a new 5-year BS/MA Applied Statistics Degree that has attracted some outstanding students. In 2008, Dr. Gleser (who has acted as Graduate Director throughout the department’s existence) received the Provost’s Award for Mentoring of graduate students.
Even while they were in the Department of Mathematics, the senior faculty in the department had close collaborative and consulting ties with the Department of Psychiatry. Their collaborations grew, enabling the department to justify and create two joint 50–50 tenure-stream positions with Psychiatry. These positions have attracted a series of bright and creative young faculty interested in working in a cross-disciplinary environment: Xin Ming Tu, Hernando Ombao, Wesley Thompson, and currently Yu Cheng. Some of the senior faculty members also maintain close research collaborations with faculty and research groups in the Department of Psychiatry.
The Department of Biostatistics and the Department of Statistics also have maintained close ties. Three faculty members from the Department of Statistics have secondary (joint) appointments with Biostatistics. Students from the two departments can take courses from either department, and a course on life testing is cross-listed between the two departments. Faculty members from one department often serve on the dissertation committees of the other department. Also, some faculty members in the Department of Biostatistics were previously faculty members (Bryant, Wieand) or students (Weissfeld) in the statistics group, before it became a department. Greg Yothers, a faculty member in the Department of Biostatistics, and Kaleab Abebe, who has a secondary appointment in Biostatistics, both received their PhDs from the Department of Statistics. The two departments were rated together in the recent ratings of departments of statistics done by the National Research Council.
In its 14.5 years of existence (1997–2011), the Department of Statistics has graduated 50 PhDs, 99 Master’s degrees, and 63 BS degrees in Statistics, and has made significant research contributions in time series, function analysis, reliability and life testing, statistical dependence and statistical inequalities, adaptive clinical trials, measurement error models and inference, statistical meta-analysis, and modeling and inference in neuroscience (see www.stat.pitt.edu for further details). One of its faculty members (Gleser) was Executive Editor of Statistical Science from 1998 to 2000, and another (Stoffer) recently completed a 2-year term as Program Director of the Statistics Division, Mathematical Sciences, of the National Science Foundation. The department is also well represented on editorial boards and FDA, NIH, and NSF review panels. Thus, although the department remains small (eight tenure-stream faculty, two teaching faculty), its impact on research and training in the field of statistics is substantial.