ASA Leaders Reminisce: Barbara Bailar
In the 21st installment of the Amstat News series of interviews with ASA presidents and executive directors, we feature a discussion with 1987 ASA President Barbara Bailar.
Barbara Bailar enjoyed working at the U.S. Census Bureau. Starting as a GS-7, she worked her way up to Associate Director for Statistical Standards and Methodology. When she left the bureau, she became the executive director of the American Statistical Association. She then served as ASA president in 1987, becoming the only person to serve as both president and executive director of the association.
Upon leaving the ASA, Bailar became the Senior Vice President for Survey Research at the National Opinion Research Center in Chicago. When she retired in 2001, she and her husband, John, returned to Washington, DC.
Q: You worked at the Census Bureau for almost 30 years. During your time there, what major statistical innovations did the bureau make?
A: The Census Bureau has been a leader in the development of sample surveys and censuses. A very distinguished staff of statisticians had developed sampling for use in both. Instead of trying to get all the information from everyone in the census—on income, education, etc.—the census statisticians used a sample and collected only a few items from everyone. During the late 1950s, Morris Hansen, William Hurwitz, and William Madow wrote a two-volume work [Sample Survey Methods and Theory] on sampling and illustrated different sampling methods for different populations.
Following the continued development of sampling, the bureau undertook work for other federal agencies. The staff knew how to select samples, the work force to collect the data was there, and the statisticians knew how to process the data and present it clearly to folks in other agencies.
A major innovation in the 1960 census was the move from personal enumeration, in which census field persons knocked on each door and interviewed the residents, to a mail out-mail back census. Several experiments were run showing the effect of enumerators on census data was decreased when people answered for themselves.
Of course, the introduction of computers changed everything. The huge ENIAC was first used in processing some of the 1950 census data. It was used on the economic censuses and all sample surveys. Most statisticians had to code their own inquiries. By the time I left the bureau, personal computers were on almost every desk.
Statisticians at the bureau began branching out when studying the census undercount, and hypotheses were introduced and tested. A sampling scheme was used to measure the undercount, as well as a method of demographic analysis. Statisticians also looked at question wording and derived tests to detect which wording was most effective.
It was an exciting time at the bureau.
Q: How do you believe recent and emerging technologies will change the way the U.S. decennial census is executed?
A:Following the examples of the past, the bureau will continue to develop new methodologies. Basically, in the United States, it is the belief that the best information comes from the persons themselves. However, information is available from other records that people have filled out. Perhaps the bureau could use some records to fill in missing items, instead of spending the time and money to follow up with nonrespondents several times. I also believe the bureau will find ways to use computers to do even more. Having seen censuses go from punch cards to sensing devices, I think there will continue to be improvements in data capture.
The bureau has made changes to every aspect of census-taking. The design of the questionnaire changes with every census. Its delivery has changed from having enumerators contact everyone to the U.S. Postal Service delivering most questionnaires. People fill out the questionnaire—by pencil, pen, and computer. Undoubtedly, computers will be used more in future censuses. The processing of the census will be subject to more advances. The use of record systems will help. Allowing people to respond by way of electronic means would also be helpful and is being tested.
Q: Could you please explain the controversy that arose in the late 1980s over the use of post-enumeration surveys to adjust census results?
A: The bureau has known for years that there is an undercount in the census and that it is concentrated in young men of color. The bureau published its results and asked advisory groups to discuss the problem. Cities thought they were losing money because there were more of the uncounted within their boundaries. Many asked why the census was not corrected and made fairer to the cities if the bureau knew there was an undercount and how big it was. One problem was that the undercount was measured at the national level, and not in fine detail. An adjustment demanded data on a very local level so people could be added to the right blocks, counties, and states. The Post-Enumeration Survey was designed to get estimates of the undercount at lower levels.
The bureau decided in the late 1980s that it would try to adjust the census if the Post-Enumeration Survey data were valid. This caused a big uproar not because of statistical considerations, but because of political considerations. The undercounted young men of color resided primarily in cities, which would benefit from this adjustment. City dwellers tend to vote for the Democratic party, and the Republicans—who were in power—said there would be no adjustment.
Q: What was the most challenging issue you faced during your term as ASA executive director? How did the ASA overcome this challenge?
A:There had been very little change in the ASA bureaucracy for many years. Any new executive director was going to want to make changes and face the words, “We’ve always done it this way.” Sure enough. So, there were some minor issues that arose because of this.
The issue that was most challenging was the ASA’s relationship with the IRS. Within a few months of my arrival, I was called by an IRS agent, who said we were not reporting income and expenses correctly for our publications and owed a big sum of money. We were visited by an IRS team, who came daily to review our operations. It turned out that many other associations were also being investigated for the same problem. Every day, a bus load of IRS employees arrived in Alexandria to visit the many associations that reside there. After a few months, with some changes on the ASA’s part and some relaxing demands by the IRS, we came to an agreement and did not owe any money. What a relief!
Q: You have the rare distinction of having served the ASA both as president and as executive director. What are the differences in the challenges faced by the holders of these offices? What are the differences in the opportunities enjoyed by the holders of these two offices?
A:I enjoyed both positions very much. The president is chosen by the members and chairs the board of directors. Every president has some ideas he/she thinks would make the ASA stronger and attract new members. One of the things I did was to establish the position of an office of public affairs. This was not very controversial and just had to work its way through the board. Other challenges surrounded the issues of hiring new staff and setting up new publications. All of these were successfully managed.
The executive director is selected by the board and reports to the president and board of directors. This means that every year, as there is a new president, the executive director has a new boss. Almost all the interactions are harmonious and there is good agreement. However, when the two do not agree, difficulties can arise.
The executive director represents the ASA in many associations around Washington and meets a lot of interesting people and sees the way in which statistics can be used to help others. The executive director also works with government employees and learns about some of their basic problems. One of the things I most enjoyed about being executive director was getting to know committee and section members much better and work with them at solving problems. Also, the ASA staff was always a pleasure to work with.