Robustness and Reliability in the Analysis of Count Data

  • No photo
    Deepak Somaya

    Associate Professor and Stephen and Christy King Faculty Fellow
    Gies College of Business, University of Illinois at Urbana-Champaign
    Director of Graduate Studies (Business Administration)
    Academic Director, Illinois Consulting Academy
    Associate Professor, College of Law
    Phone: 217-333-6873
    Email Deepak
    Deepak’s Website

    Deepak Somaya is an Associate Professor of Strategy and Entrepreneurship and the Stephen and Christy King Faculty Fellow in the College of Business, University of Illinois at Urbana-Champaign. He also holds a courtesy appointment in the University of Illinois College of Law. Deepak received his Ph.D. in Business Administration from the Walter A. Haas School of Business at the University of California at Berkeley, his MBA from the Indian Institute of Management (Calcutta), and his B.Tech. in mechanical engineering from the Indian Institute of Technology (Bombay). In his research, he studies how companies strategize about and derive competitive advantage from their knowledge assets, including the use of inter-organizational relationships to achieve these objectives. Deepak’s research has been published in several journal articles, book chapters and conference proceedings, and received several awards including a best dissertation award (Technology and Innovation Division, Academy of Management), conference best paper awards, and the 2012 California Management Review Best Article Award. He teaches courses on Strategic Management, Technology Strategy, and Strategic Human Capital. Deepak currently serves on the editorial boards of the Academy of Management Review, Journal of Management, Strategic Entrepreneurship Journal and Strategic Management Journal.


Abstract:
Management research often seeks to explain and predict phenomena that are represented by count data (integer non-negative outcomes), which has led to a proliferation of count models in empirical work. Examples of such phenomena include alliances, board directorships, initial public offerings, lawsuits, licenses, start-ups, patents, and product introductions. The appropriate use of count data models entail several critical issues, which we catalog in two broad subsets – those related to the functional specification of the mean and those related to distributional assumptions. We examine alternative empirical approaches towards these critical issues and recommend best practices. By so doing, we hope to facilitate robust and reliable modeling and interpretation of count phenomena in future research.

Digital Reader: Resources Recommended by the Speaker:

  • Colin Cameron and Pravin Trivedi. 1998. Regression Analysis of Count Data. Cambridge, U.K.: Cambridge University Press.
  • Rainer Winkelmann. 2013. Econometric Analysis of Count Data. Fourth Edition. New York: Springer.