William B. Simpson

Lecturer of Business Administration
Principal Statistician

William Simpson is a Lecturer in the Negotiations, Organizations, and Markets Unit and Principal Statistician in Research Computing Services at Harvard Business School. He earned his B.S. in Physics from California Institute of Technology, M.S. in Computer Science from University of Michigan, and Ph.D. in Clinical Psychology from Boston University.

Dr. Simpson teaches the course, Seminar in Applied Statistical Methods, for doctoral students working on dissertation research, and the statistics module of Theory and Practice of Research.

As a statistician with Research Computing Services, he has provided statistical support and data analysis for a wide variety of faculty research projects. He has also been a statistical consultant for a number of organizations including Cambridge Hospital, Children’s Hospital, the Trauma Center at Justice Resource Institute, Abt Associates, Verizon, and the Forum Corporation.

  1. Statistical Methodology

    William Simpson is developing methods of inference to use when assumptions of standard models are not met. He has created a hypothesis test to use for ipsative variables that adjusts for the non-zero correlations among variables expected under the null hypothesis. He has developed a number of extensions of the quadratic assignment procedure -- a method used for dyadic data in social network analysis -- for generalized linear models, multiple networks, and network panel data.

  2. Clinical Research

    Dr. Simpson has been a consulting statistician for clinical research, especially in studies of post-traumatic stress disorder. Most recently, he has been involved in a clinical trial comparing a pharmacological treatment and an exposure-based psychological treatment for post-traumatic stress disorder.