Shiskin Award Recipients Honored for Work on Major Surveys

William R. Bell and Robert M. Groves, recipients of the 2008 Julius Shiskin Memorial Award for Economic Statistics, will be recognized at NABE’s 50th Annual Meeting in October for their contributions to survey methodologies and technical initiatives that have become part of key economic statistical measures.

In brief interviews conducted in early August via e-mail, the two honorees talked about what drew them to their work in the statistical system’s critical infrastructure and the major challenges maintaining sound economic data in times of budget constraints.

Groves, director of the Survey Research Center of the University of Michigan’s Institute for Social Research, said that he has serious concerns about the ill effects of shrinking data agency budgets.   Given what he views as a growing problem of recruiting graduates to serve in federal statistical agency jobs, he also “fears for our country’s future in economic statistics.”

Bell, a senior mathematical statistician at the U.S. Census Bureau, described his research on seasonal adjustment, small area estimation, and time series modeling.  In recent years, his work has focused on the Small Area Income and Poverty Estimates (SAIPE) program at Census. SAIPE data are used by states, counties, and school districts.

The 2008 Shiskin Award Committee, chaired by NABE member Robert Parker, recognized Bell for his “important contributions to the theory and practice of seasonal adjustment, small area estimation and time series modeling.”  Groves is recognized for his “important contributions to the theory and practice of survey methods for the conduct of sample surveys of both households and establishments.”

Groves Leads UM Survey Research Center

Groves said he was drawn to survey methodology because of his interest in combining the study of computer science and sociology.  “ I took one class in survey sampling [at the University of Michigan], met Leslie Kish, and fell in love with both the statistical and nonstatistical sides of survey research.  This was a logical conclusion of being exposed to survey data analysis in an undergraduate course taught by James Davis at Dartmouth; Davis was the father of the general social survey and former director of the National Opinion Research Center at the University of Chicago,” he said.

When asked what changes in survey methods have been the most important in recent years, Groves underscored the benefits of  “the explosions of modes of data collection over the past decade.”  He added:  “Economic surveys are ahead of surveys of other types to integrate the Internet and electronic data transfer and other modes.  This has helped, in my opinion, to stem the tide of overall lower cooperation rates in the United States with survey requests.”

Director of the UM’s Survey Research Center since 2001, Groves also is professor of sociology, research professor at its Institute for Social Research, and research professor at the Joint Program in Survey Methodology at the University of Maryland.  “The results of his research have been incorporated into many surveys, including the current employment survey and consumer expenditure surveys at the Bureau of Labor Statistics, household income surveys at the Census Bureau, and the surveys conducted by national statistical offices in the United Kingdom, Sweden, and The Netherlands,” according to the Shiskin committee.

Groves has co-authored five books and more than 50 articles over the past two decades.  His 1989 book Survey Errors and Survey Costs was named one of the 50 most influential books in survey research by the American Association of Public Opinion Research (AAPOR).  His book Nonresponse in Household Interview Surveys (with M.Couper) received the 2008 AAPOR Book Award. He is also a long-time contributor to improving federal statistics through his work for the Committee on National Statistics of the National Academy of Sciences.

Groves Laments Effects of Tight Budgets

Budget constraints facing federal data agencies have affected quality and curtailed initiatives to improve data series, Groves said, and he has observed that many recent graduates are not interested in pursuing jobs in the federal statistical system.  “The recent past has not been good to federal statistics, in my opinion.  Several agencies have been administratively layered, to remove them further from the heads of the departments they serve.  They are often now overseen by program areas, whose interests are not well-allied with those of the statistical agency,” he said.

“The amount of interaction between the statistical agency head and the secretary of the department is small.  The agency becomes invisible to the appropriators in the legislature and budget suffer,” Groves said.  Even though agencies that produce the principal economic indicators have “probably fared a little better, times are not good.”

Having spent much of his time in recent years working with the next generation of researchers and statisticians, Groves said “the most challenging part of the educational side of my life is recruiting graduates to consider public service in the federal statistical system.  The last two decades have been totally successful in communicating to younger persons that nothing exciting is happening in federal statistical agencies, that such a career is a second-best option.  I fear for our country’s future in economic statistics.”

NABE members and other interested data users have an important role to play as data agencies work to evaluate and improve statistics, Groves said.  “NABE members are respected and listened to, within the statistical agencies.  Further, they form one of the most visible interest groups that can communicate to the government that independent, high quality statistics are key to the nation,” he said.

Bell Joined Census To Work on Time Series

While working on his Ph.D. in time series analysis at the University of Wisconsin, Bell said that he made a connection with the Census Bureau through his advisor, George C. Tiao (a 2001 Shiskin Award recipient), who was a member a committee set up by the American Statistical Association to help Census establish a research fellowship program. “One of the topics of interest for research under this program was seasonal adjustment. So when I finished at Wisconsin I went to the Census Bureau to work (as a research associate) with Steven Hillmer, another former student of George Tiao's, who had already done important research on seasonal adjustment and had been awarded an ASA/NSF/Census Fellowship,” he said.

Soon after arriving at Census, Bell became a member of what was then a newly formed time series staff in the bureau’s Statistical Research Division, then headed by David Findley, a 1996 Shiskin Award winner. There Bell worked on seasonal adjustment and time series problems for many years, until the mid-1990s when he became involved in the Small Area Income and Poverty Estimates (SAIPE) project.

Bell pointed out “there are connections between these two topics (time series analysis and small area estimation). In fact, a significant part of my work in time series has been related to modeling time series that are produced by repeated sample surveys and that have substantial amounts of sampling error. This can be thought of as a small area estimation problem, though with time series data.”

His work on seasonal adjustment “has provided methods and software for estimating time series models accounting for trading/working day effects, the effects of moving holidays, and the automatic identification and estimation of certain kinds of outlier effects,” according to the Shiskin committee.

Bell Played Key Role With Statistical Community

As an original member of the Census Bureau team charged with developing the program that became SAIPE, Bell “played a key role as a liaison to the statistical community, most particularly the National Academy of Sciences (NAS) panel established by Congress to determine that the basis for the estimates was scientifically sound,” according to the Shiskin committee. After a favorable review by the NAS panel, the Department of Education has used the SAIPE data to allocate billions of dollars to school districts under the agency’s Title I program. For more information on the SAIPE program go to: http://www.census.gov/hhes/www/saipe/index.html

Bell continues to oversee statistical methods used in the SAIPE, including changing methods, such as use of data from the American Community Survey, to further improve the estimates.  From 1988-1995 Bell served as an associate editor of the Journal of Business and Economic Statistics (JBES) and he was elected an ASA fellow in 1993.

The Shiskin committee also noted that Bell is well known for an article he co-authored with Steven Hillmer on “Issues Involved with the Seasonal Adjustment of Economic Time Series,” published in the 1984 Journal of Business and Economic Statistics.  It was reprinted in the JBES 20th anniversary issue as one of the ten “most influential papers” published in the journal.