Interplay Between Sample Survey Theory and Practice: An Appraisal

Friday November 18, 2005

The speakers at JPSM's November 18 lectures, Jon Rao (left) and Alastair Scott (right), with the Director of JPSM, Roger Tourangeau. 

The Joint Program in Survey Methodology and the Statistics Consortium at the University of Maryland sponsored two lectures on Friday, November 18. The first was by Alastair Scott on the “The Analysis of Population-based Case Control Studies.” The second, titled the “Interplay Between Sample Survey Theory and Practice: An Appraisal,” was by J. N. K. Rao.  Rao's talk was the third in JPSM's Distinguished Lecture Series.  Barry Graubard from the National Cancer Institute and Graham Kalton from Westat and JPSM served as discussants on Scott's talk, and Phil Kott from the National Agricultural Statistical Service and Mike Brick from Westat and JPSM discussed Rao's talk.  Both talks were held in LeFrak Hall on the University of Maryland, College Park campus. 

Scott's lecture concerned the use of complex sampling designs in population-based case control studies, particularly for obtaining the controls.  Although most statistical packages now have special survey modules that can be used to carry out weighted analyses of the data from such studies, weighting tends to be inefficient in these situations, since large differences among the weights are typical.  Fully efficient likelihood methods can be developed for some special designs but, apart from the case of simple stratified sampling, these methods require special software and are difficult to implement.  Moreover, there are questions about their robustness to model breakdown.  The lecture examined the question of robustness in more detail and suggested alternative procedures that are both reasonably efficient and can be implemented using standard survey software. 

Alastair Scott has an MSc in Mathematics from Auckland and a PhD in Statistics from the University of Chicago.  He lectured at the London School of Economics for a number of years before joining the University of Auckland in 1972. He has been here ever since, apart from a number of visiting appointments at universities in North America and the U.K. 

Rao's lecture discussed the impact of the practical problems encountered in the design and analysis of sample surveys on the theory of survey sampling.  In addition, he discussed how sample survey theory has influenced practice, often leading to significant improvements.  The lecture examined this interplay between theory and practice over the past 60 years.  He also presented examples where he argued that new theory was needed or that existing theory needed to be applied in practice.  The talk was based on Rao's Waksberg invited paper, which will be published in Survey Methodology. 

Rao has been at the forefront of research in sampling theory and methods since 1960.  He has made fundamental contributions to the so-called classical theory of sampling, to the foundations of sampling during the debates of the 1960s and 70s, to a variety of aspects of variance estimation, to the analysis of complex survey data, and to small area estimation.  Over this forty-year period his work has been, and continues to be, at the cutting edge of research.

The Joint Program in Survey Methodology offers graduate training in survey methodology and survey statistics, leading to Master's and doctoral degrees.  In addition, it offers certificate and citation programs for non-degree students.  JPSM is funded by a consortium of federal agencies through a contract with the Census Bureau and receives additional funding through cost-sharing agreements with the University of Maryland, the University of Michigan, and Westat.  Graham Kalton heads the Distinguished Lecture Series committee at JPSM and Partha Lahiri is the Director of the Statistics Consortium.