Government Statistics Section

Newsletter

June 1994

Contents

Toronto in August

Health Care Reform and Health Information

Notes from the GSS Section Chair Session on Continuous Measurement at ASPA Executive Committee

Notes Census Advisory Committee Meeting

Untimely Death of Roger A. Herriot

Legislating Data Needs - Report of Conference Session

COPAFS Looks to the Future

Call for 1995 Papers 1993-1994 Proceedings

Order Form

Toronto Program Highlights

U.S. and Canadian Government Statistics

Gerald Gates, Bureau of the Census 1994 GSS Program Chair Virginia de Wolf, Bureau of Labor Statistics 1995 GSS Program ChairThe theme of this year's Joint Statistical meetings is "Increasing Statistical Awareness through Promotion, Preparation and Perception." Many of the GSS-sponsored sessions are tied to this theme either directly or indirectly. Our 1994 program is the result of a collaboration between the Government Statistics, Social Statistics, and Survey Research Methods Sections. Together, we have arranged a program of invited and special contributed sessions with well-known speakers on topics of immediate interest to our combined membership. For example, each section has allotted one of its invited sessions to address methodological and policy aspects of planning for the year 2000 census. In addition, the three sections are jointly cosponsoring an invited session with COPAFS, August 17 at 8:30 a.m., entitled "Statistical Statesmanship." This panel discussion, organized by Monroe Sirken, will bring together statistical statesmen Ingram Olkin, Vincent Barabba, Ivan Fellegi, Edward Spar, and Miron Straf to discuss the meaning of statistical statesmanship, how well it is being done, how it can be done better, and how people can be trained in statistical statesmanship.GSS has one invited session this year, on August 15 at 2 p.m., which we have devoted to a topic of great concern in the U.S. and Canada. The session involves an invited paper coauthored by Joe Knott of the U.S. Census Bureau and John Leyes of Statistics Canada, entitled "New Research in Canada and the United States to Integrate Administrative Records into Census Taking." The paper will be discussed by Bruce Phillips, the Privacy Commissioner of Canada, Lawrence Thompson, the Principal (Continued on Page 2, Col. 1)Deputy Commissioner of the Social Security Administration, and Paula Bruening, a privacy expert at the Office of Technology Assessment of the U.S. Congress. This session promises to be a lively debate on the difficult issues involved in balancing privacy concerns, meeting agency mission requirements, and making efficient use of technologies and resources.The GSS also has several special contributed sessions highlighting a variety of interests for government statisticians. One session organized by Robert Klein of the Department of Veterans Affairs will focus on research issues involving studies of the nation's veterans, including alternative data sources, survey design, data collection and data analysis. Bruce Klein of the Food and Nutrition Service has organized a session on microsimulation modeling as it is currently applied to guide federal health and welfare program changes. Carrol Kindel of the National Center for Education Statistics has a session on library statistics, which will look at issues of structure, coverage and comparability. Bob Lehnen of Indiana University has organized a panel session entitled "Informing Education Policy: The Role of International Policy Indicators," which will address the major statistical policy issues associated with creating a system of international policy indicators and report on the usefulness of these indicators to inform education policy. Finally, Benchieh Liu will participate in a session he organized to explore various methodological and analytical aspects of selected government surveys and programs.This year, the GSS will sponsor two roundtable discussion groups, which are similar to roundtable luncheons but without the lunch. These discussion groups do not require preregistration, and you can come and go as you please. Both GSS roundtable discussions are scheduled for Wednesday morning at 8:30 a.m. One of the discussions will be led by Charles Alexander and Ed Spar and will deal with plans for developing a continuous measurement program designed to make the small-area census data more current. The other discussion will feature Maria Gonzalez and Leonard Gaines, who will explore the role for state, local, and provincial government statisticians in ASA.The GSS has five luncheon roundtables scheduled for Wednesday, August 17. Two of the discussion leaders, Connie Citro and Barry Edmonston, are from the National Academy of Sciences' Committee on National Statistics (CNSTAT). Connie will discuss the pros and cons of alternative poverty measures. Barry will describe ongoing efforts to assess the needs for data currently collected in the U.S. decennial census. George Duncan, Carnegie Mellon University, will talk about the ramifications of the 1993 CNSTAT report on confidentiality and data access for legislation currently being considered in the U.S. Martin Podehl will outline the experiences of Statistics Canada in marketing official statistics. Maryanne Webber, Project Manager of Statistics Canada's Survey of Labour and Income Dynamics, will lead the fifth luncheon roundtable on the role of longitudinal data in policy research. Other roundtable luncheons of interest to GSS members include microsimulation, data sharing, race and ethnicity, and information needs in an era of U.S. health care reform. This year the Abstracts for the meetings will have a new feature describing each roundtable luncheon. The GSS program committee is very excited about the roundtables it has organized and urges all members to make luncheon reservations early to ensure a seat.The remainder of the GSS program consists of regular contributed paper and poster sessions on topics of general interest to our members. Also, do not forget to review the Social Statistics and Survey Research Methods Sections offerings in the May issue of AMSTAT News for those that are cosponsored by GSS. Finally, be sure to mark your calendars to attend the GSS business meeting on August 16 from 6:00 7:00 PM. This will be a chance to meet your elected officers and discuss your views on section activities and business matters.

Health Care Reform and Health Information

Ed Hunter, Chief Planning, Budget, and Legislation Staff National Center for Health Statistics 6525 Belcrest Rd. - Room 1100 Hyattsville, MD 20782The national debate over health care reform is approaching the boiling point in Washington. More than a dozen Congressional subcommittees are weighing decisions that might--or might not--lead to wholesale changes in the U.S. health care system and, as a result, in our nation's health information systems. The work of these subcommittees is focused on piecing together legislation that can pass both Houses before members leave the Capitol to face voters in November.Among the questions yet to be resolved are: (1) Will health care reform result in universal coverage for health insurance? (2) How will coverage for the uninsured be financed? (3) If standard health insurance packages are mandated, how generous will the benefits be? and (4) To what extent will the government (federal and state) be involved in regulating activity in the health care system? Major proposals still under consideration include modifications to the President's proposed Health Security Act and more modest or market-oriented alternatives proposed by moderate Democrats and Republicans.Depending on the nature of the reform that is enacted, there is a wide range of implications for health information and statistics: National Framework for Health Information The President's plan devotes considerable attention to development of a national health information system. This system would provide information necessary for the operation of the health care system, but would also become a "public utility" serving multiple purposes. It would be based on three components: (1) enrollment information obtained on each individual eligible for health insurance coverage; (2) a standard, minimum data set collected by health providers for each patient encounter with the health care system; and (3) administrative data on the characteristics and operations of health care plans and providers. This information would be processed through regional data centers, which would have strong privacy protections. These centers would, in the Administration's planning, create anonymous, person-level public use files for researchers; allow limited access to individual data by researchers with institutional review board (IRB) approval; and facilitate the use of network data to create sampling frames for federally-sponsored surveys.In addition to the President's plan, many of the other leading health care reform proposals include provisions for health information, and particularly for the standardization and uniformity of health records. Questions over the federal role in sponsoring the aggregation of such data, and on the levels of privacy protection afforded to individual data, are yet to be fully addressed in the Congress.

Monitoring Implementation and Impact through Surveys The information network described above would not be available for analysis until well into the implementation period for health care reform. Even when operational, it will not provide critical information to assess the health status of the population and their access to health care. As a result, planning is underway for new or modified survey systems. The National Center for Health Statistics, the Agency for Health Care Policy and Research, and other federal agencies are developing survey plans that will constitute a comprehensive monitoring system to assess the impact of health care reform on the population and on the health care system. Among the first steps:(Continued on Page 4, Col. 1) The National Employer Health Insurance Survey (NEHIS) was started in April 1994 and is designed to provide state-level information on health benefits provided by employers. The National Health Interview Survey was modified in July 1993 to include new supplements on health insurance and access to care, which will be used to monitor change as reforms are implemented. A process has been established that will lead to the development of a set of key indicators for monitoring health care reform.Other steps will be determined as the nature of health care reform becomes more clear, and as data priorities are addressed through the budget process.

### Notes from the GSS Section Chair

Maria Gonzalez Statistical Policy Office U.S. Office of Management and Budget Washington, D.C. 20503I want to call your attention to the roundtable discussion at the ASA Annual Meeting in Toronto that we have organized on the topic of, "Is There a Place in ASA for Statisticians Who Work for State/Provincial/Local Governments?" This session will be held on Wednesday, August 17, 8:30 to 10:30 a.m. I strongly encourage GSS members who are interested in this topic to attend this session.More generally, GSS would like to encourage the participation of state/local/provincial statisticians in the section and ASA. We wonder whether a broader outreach is needed in order to integrate these statisticians into the activities of ASA. Ideas from section members on this topic are most welcome.On a sad note, one of our section members, Roger A. Herriot, died recently as the result of an accident. GSS will miss him.

Session on Continuous Measurement at American Society for Public Administration Meeting

Peter Bounpane, Assistant Director U.S. Bureau of the Census Washington, D.C. 20233While the periodic population census in the United States has many advantages, it suffers the problem of data that can become outdated very quickly. One way to solve that difficulty is to take the census on a more frequent basis, and in fact, some countries do take a census every five years. In the United States, however, the discussion has been on the use of intercensal surveys rather than more frequent censuses. The issue was brought to the foreground when Leslie Kish suggested rolling samples.Currently, the U.S. Census Bureau is considering a modification of rolling samples, called continuous measurement. Continuous measurement would mean a "small scale" census in years ending in zero supplemented by intercensal surveys to gather current information that was traditionally gathered on the sample questionnaire (long form) in the census.Obviously, such a change is of importance to many data users and to the Government Statistics Section. Our section has arranged for a panel discussion of this important issue at the upcoming meeting of the American Society for Public Administration (ASPA). The panel is scheduled for 8 a.m. on July 25 in Kansas City at the ASPA national training conference. The goal of the panel is to learn more about continuous measurement and to have key stakeholders discuss its advantages and disadvantages. Clearly, this topic fits well with the ASPA meeting theme of "New Ideas, New Challenges, New Conditions."Congressman Thomas Sawyer from Ohio has been invited. Harry Scarr, the acting Director of the Census Bureau, will discuss the Census Bureau position on continuous measurement. Michael Knight of the Wisconsin State Data Center will discuss the view from the states, and John Blodgett of Missouri will discuss the local view. It promises to be an interesting and useful session.

### Executive Committee Notes Alan R. Tupek Secretary-Treasurer, GSS National Science Foundation 4201 Wilson Boulevard Arlington, VA 22230

Report of GSS Executive Committee Meeting, December 8, 1993 General Business

Al Tupek reported that GSS membership continues to remain around the 750 level. Approximately half of our members are joining through the joint Gov't Stat/Social Stat membership option. Connie Citro reported that the December newsletter will be mailed 3rd class after Christmas; the June issue will be mailed 1st class to be sure to reach members before the annual meetings.

COPAFS Report

Ed Spar reported on planning for the next decennial census. Pete Bounpane said that most of the discussion has been methodological, when what we need is to acquire more information on customer data needs. The key issue for GSS is to make sure people who should be participating in the decennial census planning process do so. It was agreed that we need to find out more about the Census Bureau's plans for a "continuous measurement" data collection system and to find ways to inform census data users about the pros and cons [see article, p. 4].Dan Levine said that the Census Bureau has been given permission to reactivate their advisory committees [see article, p. 6]. Wray Smith reported that the Statistics of Income (SOI) Division at IRS will remain intact, headed by an SES position.

Certification

Pete Bounpane said that GSS will not take a position or write ASA on the issue of certifying statisticians. Individual members are encouraged to participate in the debate.

New Business

Bob Lehnen has been working on international education indicators. He would like GSS to invite members of the OECD and other international organizations to join the GSS.

Report of GSS Executive Committee Meeting, March 10, 1994 General Business

Al Tupek reported that GSS membership increased to 787 as of March 2, 1994 (804 as of April 4). The section had a net surplus of $488.83 for the year ending 12/31/93, and $2,087.91 cash on hand.

1994 ASA Meetings

Jerry Gates reported that the GSS invited session, "Reinventing the Population Census," will be held on Monday at 2:00 p.m. Jerry reported that 26 papers were submitted for contributed sessions and that 5 special contributed sessions have been arranged. Ginny de Wolf is organizing 5 luncheon roundtables. There will also be 2 morning roundtables. One is a session, led by Maria Gonzalez, on involving state and local statisticians in ASA, a topic that is also of interest to the ASA Membership Committee according to Ginny. Maria will contact Ivan Fellegi for names of Canadian local government statisticians who might participate in the Toronto roundtable. She will also work with Jerry Gates and Data User Services Division at Census to obtain a list of state and local statisticians to contact for participation in the roundtable. Ginny said that the Social Statistics Section, the Section on Survey Research Methods, and the GSS will get together to plan for the 1995 annual meetings.

Proceedings

We discussed the value of producing GSS proceedings from the annual meetings and methods we could use to promote and sell more copies of the(Continued on Page 6, col. 1) proceedings. Penny Young said it was not cost effective to send promotional material to small mailing lists. Ginny suggested advertising our proceedings at meetings such as the upcoming COPAFS conference in May. Connie Citro and Ed Spar agreed to draft a flyer for advertising GSS proceedings. Maria agreed to contact Richard Rockwell, Chair of the Social Statistics Section, to again discuss the idea of publishing a joint proceedings for the 1994 annual meetings.

Increasing Membership

Bob Lehnen discussed ideas for developing new markets for new GSS/ASA members. A suggestion was made to ask the session chairs at the annual meetings to encourage attendees to join the GSS. Bob Lehnen will provide a letter and labels to Barbara Milton for advertising GSS and its Toronto program among international organizations.

### First Meeting of Reconstituted Census Advisory Committee

Patricia C. Becker Southeast Michigan Census Council 17321 Telegraph, Suite 204 Detroit, MI 49219The Bureau of the Census recently received a charter for the reconstituted

Census Advisory Committee of Professional Associations.

This committee is a consolidation of the four previously chartered advisory committees of the American Economic Association, American Marketing Association, American Statistical Association, and Population Association of America. The four individual professional association committees have a long history, ranging in age from 75 years (ASA) to 29 years (PAA). There are now 36 members, nine from each of the four associations.I was recently appointed to the Committee through the Population Association and attended the first meeting on April 14-15. It was an enjoyable day and a half, and we were told that our suggestions will be helpful to the Bureau. Here are some highlights of the meeting:The Census Bureau has several initiatives under way connected to reinventing government. These include: Statistics 2000, a program in which Census, the Bureau of Labor Statistics, the Bureau of Economic Analysis, and the National Agricultural Statistics Service will share data with each other, which will reduce business reporting burden by eliminating duplicate data requirements and by facilitating and simplifying reporting. This, of course, will also reduce costs. To accomplish this laudable goal, new data sharing legislation is required. The Office of Management and Budget is to draft the bill. CASIC, which stands for Computer-Assisted Survey Information Collection. Its vision is described as "a future in which Census Bureau surveys are designed, conducted, and managed using a compatible set of automated technologies to improve timeliness, data quality and cost effectiveness." CASIC is a collection of technologies, including CAPI (computer-assisted personal interviewing), CATI (computer-assisted telephone interviewing), and others, that use modern computer, telephone, and fax systems to collect and retrieve data. The Census Bureau has received a HAMMER AWARD for CASIC. This honor recognizes efforts to reinvent government, increasing efficiency and reducing cost. DIAMOND is Data Information Access for Modified On-line Network Delivery. This project is in the Communications office and deals with setting up a Bureau-wide customer information system, getting census data onto the Internet and serving Congress in its role as data user.There were several agenda items dealing with the 2000 census design and the 1995 test. Much of this material was also presented at the March COPAFS meeting [see article, p. 12] and 2000 Census Advisory Committee meeting. In summary, the(Continued on Page 9, col. 1) Bureau's "principles for change" in 2000 include greater accuracy to reduce the differential undercount, reduced cost, avoiding any requirement for constitutional change, and involving stakeholders in the decision-making process. There are several suggested legislative changes, including taking the Master Address File (MAF) outside of Title 13 so that it may be shared with appropriate local officials to increase coverage accuracy, moving census day to the first Saturday in March, and changes in the hiring procedures. Content determination will proceed along a parallel path to the 1995 test and other work on methods. 1990 content items have been divided into three groups, based on their legislative mandate: Mandatory items are those specified in law that cite the decennial census as the source. Required items are used in statutory-based programs, such as formula funds distribution, but the legislation does not specifically mention the decennial census as the source. Programmatic items are not required by legislation. Place of residence five years ago, the questionnaire item which permits migration analysis, is in this category.

Continuous Measurement (CM) was also discussed. Committee members raised major concerns about methodology, stability of the funding base, and the general concepts involved in shifting from "single point in time" data collection to merging data collected year round and aggregated over several years. The Census Bureau appears committed to an investigatory research program on CM, leading to a decision by the end of fiscal 1997 on whether it can replace the year 2000 long form.An update was provided on the redesign of the Survey of Income and Program Participation (SIPP) and the Current Population Survey (CPS), two of the Bureau's most important current programs in the demographic area.The PAA "subcommittee" in which I participated adopted five recommendations:(1) that the Census Bureau build on the successful experience of the 1992 Economic Census in regard to explaining the mandatory nature of the decennial census;(2) that certain improvements be made in the plans for testing Integrated Coverage Measurement (for evaluating the quality of the census count);(3) that the Census Bureau immediately prepare and implement a detailed plan for soliciting non-federal input into the year 2000 content determination process;(4) that some of the components of the year 2000 plan be evaluated with regard to cooperative ventures between the Census Bureau and state/local governments; and(5) that the PAA subcommittee supports legislation to open the MAF to non-census government agencies and supports continuous decade-round MAF updating.The next meeting will be in October.

### Untimely Death of Roger A. Herriot

Contributed by Daniel Kasprzyk National Center for Education Statistics 555 New Jersey Ave., NW Washington, D.C. 20208-5651Roger Herriot, Associate Commissioner for Standards and Methodology at the National Center for Education Statistics and a member of GSS, died suddenly last month of a pulmonary embolism resulting from an accidental fall. His untimely death at the age of 52 robbed the federal statistical community of a creative intellect that was always trying to "reinvent" the ways and methods of the statistical system.Roger made significant contributions during his career in a number of areas. Chief among these was his work in the area of income statistics. Here Roger played a leading role in improving the measurement and analysis of the economic well-being of the U.S. population through measuring theseparate components of cash and noncash income, poverty status, and income before and after taxes. He expanded the collection of income data in the March Current Population Survey and was instrumental in the development and success of the Survey of Income and Program Participation.Recently at the National Center for Education Statistics, Roger made significant contributions in improving public access to education data and in developing a database on the nation's school districts for use by policy makers and educational researchers.Underlying Roger's federal career was a strong commitment to providing quality data to the public using the best available data systems--whether census data, survey data, administrative data, or some combination of the three. In particular, Roger's understanding of the potential benefits of the statistical uses of administrative records was in evidence throughout his career:(1) in the decennial census for a variety of evaluation programs;(2) in the population estimates program, for which he had responsibility when he was Population Division chief at the Census Bureau; and(3) in the augmentation of survey data with administrative data to expand analysis capabilities--evident in the landmark and unique Census-IRS-Social Security Administration exact-match study of the early 1970s, the Income Survey Development Program (the predecessor to the SIPP) of the late 1970s, and the Survey of Income and Program Participation of the 1980s.In short, Roger Herriot made many significant contributions throughout his career as a government statistician. The federal statistical system will miss his vision, creativity, and enthusiasm.

### Legislating Data Needs--Can Statisticians Produce What Politicians Demand?

Remarks from a Panel Discussion American University Conference on Statistics for Public Policy Contributed by Gerald Gates, Census Bureau Program and Policy Development Office Washington, D.C. 20233On May 13, 1994, five panelists representing Congress, a federal program agency, a federal statistical agency, and two private organizations that analyze and prepare reports needed for public policy presented their views on meeting the nation's data needs. The setting was the American University Campus--host for a Conference on Statistics for Public Policy co-sponsored by AU and the American Statistical Association. This stimulating panel discussion was sponsored by the Government Statistics Section and proved to be an important event for understanding the tradeoffs that are often made when statisticians are asked to provide "acceptable" data to answer urgent policy questions. The panel and the conference participants brought to light several issues that we thought would be of interest to our members.The panel was chaired by Gerald Gates (Census Bureau) who summarized the current data needs facing lawmakers. He characterized five situations when laws require data: (1) allocating resources based on statistics; (2) using statistical series as signals to start or stop programs; (3) defining populations to be served based on statistics; (4) establishing standards that are based on statistics; and (5) requiring evaluations of past performance or prediction of future impact using specified statistics. In meeting these legislative requirements, he cited several sources of data and their limitations: (1) surveys and censuses, which are burdensome and costly; (2) administrative records, which may be difficult to access and offer conceptual problems; (3) exact and statistical matches of surveys and administrative records, which present privacy concerns and may be subject to methodological constraints; and (4) modeling techniques, which are inexpensive but less precise. Finally, he raised several issues faced by statisticians when meeting the data requirements imposed by legislation: What are the risks in using experimental, untested methodologies? How do we produce acceptable data with limited funding and short lead times while keeping respondent burden to a minimum? Can we gain the cooperation of other agencies to make use of existing data? Are we able to produce statistically defensible results that meet conceptual and reliability standards? What are the implications of providing data that may be used improperly? Can we defend our data decisions in court? And, how well are data concerns being communicated between Congressional staff and the providers of statistics?The first panelist, David McMillen, staff assistant to the U.S. Senate Committee on Governmental Affairs, Subcommittee on Government Information, stated that the major problem is communication. What politicians ask for may not be what they need. Consequently, when they say "can you do this?" the answer may be "no," but had they asked "how can you do this?" the answer may be "we can provide these data at this cost." Often Congress has no idea how much data cost. A dialogue is needed so that both sides understand the costs and the possibilities. Further, a trust must be developed that recognizes that Congress is capable of using data properly and that statisticians operate under cost and burden constraints that limit what data they can provide. William Butz, Associate Director for Demographic Programs, Bureau of the Census, remarked that politicians want data that are good, cheap and quick. Federal statistical agencies can probably deliver on only two of the three at one time. Mr. Butz cited several examples (with good and bad results) in which legislation required data from the Census Bureau: development of intercensal poverty estimates to allocate money for school-age kids in poverty; legislation to encourage first-time home buyers; legislation to create enterprise zones; and legislative uses of the 1990 census. He concluded that (1) the system is not broken but it is also not neat or coordinated; (2) statistical agencies canand should be pushed by Congress to meet data needs in innovative ways; (3) statistical concerns need to be considered when Congressional decisions are being made; (4) statisticians need to be proactive to ensure that (3) happens; (5) regular forums are needed (for example, conferences) with Congressional people who support legislation requiring data; and (6) OMB needs to be proactive in ensuring that all this happens.Wendell Primus, Deputy Assistant Secretary for Planning and Evaluation, Department of Health and Human Services, stated that statisticians cannot produce what politicians demand, under politicians' constraints. Timeliness is the key constraint because information often drives legislation. Some questions statisticians are asked to answer are complicated and sensitive--such as the change in poverty rates. Providing the answers can place agencies in the middle of controversy and possibly affect their funding. Questions can't always be answered. We provide our best estimate even if we cannot assess its reliability. For instance, under health care reform, will Medicare costs go down? Dr. Primus stated a need for administrative data to supplement survey data. We also need to think about randomized designs. Finally, confidentiality is a big constraint to linking the data needed to answer policy questions. Finding solutions that accommodate important legal and practical concerns will not be easy. Harold Beebout, Research Director, Mathematica Policy Research, Inc., focused his remarks from the perspective of a third party that does special evaluative studies. He noted that Congress supports much statistical research in agencies and consulting firms like Mathematica. He indicated a need for a dialogue in order to get the questions right. Statisticiansneed to help Congress pose the right question so it can be answered by the data available, within the needed time frame, and with the needed reliability. To understand what data are available and what is possible, both parties must be represented by people with both analytical skills and statistical skills. Dr. Beebout indicated that data must be reliable enough to be helpful but their limitations need to be fully spelled out. What is needed is a good system of review. Finally, statisticians need to build an infrastructure capable of anticipating policy questions that may arise.The final panelist, Judy Gueron, President, Manpower Demonstration Research Corporation, found that statisticians can meet politicians' demand but still have a long way to go. She raised several questions that arise in legislating data needs: (1) Can you get high quality evaluation data? Yes. Use a random assignment-control group approach when appropriate. However, a definitive answer is very important for legislators. Competing results from weak evaluation programs can be harmful to the program.(2) How important is assuring access to administrative data? Very, as such data can provide large samples for subgroups. Dr. Gueron suggests that despite concerns about privacy, administrative records can be used safely with limits on retention of identifiers.(3) Can public officials distinguish "good" data? Politicians like to emphasize the good news from data by highlighting general findings that may misstate the real findings. Statisticians need to work with politicians to understand the difference.(4) Can we afford the information we need? Yes, if we develop short-term indicators that lead to long-term solutions.(5) Does it take too long to get good data? To be fast, we need to give the system time to work. Let the start-up bugs work themselves out and look to the long-term gains.Questions raised by the audience hit on several important topics including what do we do when Congress stifles research because of a political agenda. Experimental designs need to be part of how statistics are provided. And, we need to be sure that the media also understand how to interpret statistics.*The author would like to thank Wendy Alvey of the Internal Revenue Service whose notes were invaluable in preparing this report.

### COPAFS Looks to the Future on Many Statistical Fronts

Evelyn S. Mann ASA Representative to COPAFS 77-11 35th Avenue Jackson Heights, NY 11372

COPAFS continues to present an opportunity for members to react to the Census Bureau's rapidly advancing plans to radically restructure the year 2000 census. Two scenarios are being considered by Bureau experts who were invited to present progress reports at the March meeting. Considerably advanced are the specifications for the 1995 Test Census to be conducted in three urban areas and one rural area. Innovations are designed to reduce costs and improve coverage. They include special coverage questions to ensure a complete listing of household members. Special treatment of historically difficult-to-enumerate areas will be tested. The counting of the homeless will be restricted to locations where they are serviced, eliminating a count of those on the street or in casual locations. A short-form questionnaire will be widely available in a "be counted" campaign as will mailed-out Spanish-language questionnaires. A sample of households will be reenumerated during the census, providing the basis for a built-in adjustment for undercount. Some of these features to be tested in 1995 are controversial and will be politically challenged. Of concern to the user community is the possible abbreviation of census content, restricting the questionnaire to only those items that are mandated by law, eliminating some of the valuable information that has enriched the census as well as illuminated important socioeconomic trends.Paralleling the above plans, but not suitable for testing in 1995, is a "continuous measurement" system, which would eliminate all long-form questions in the decennial census year, presumably in exchange for a more complete head count. Data would be collected on such variables as employment, income, educational achievement, and journey-to-work patterns in a series of sample surveys taken over the decade. Users are concerned with the problems of data comparability by area afforded by this technique. This plan is being presented as a serious alternative to the more traditional form of the census, but it may be scuttled by lack of assurance that it will be financially supported over a decade by Congress.The COPAFS members were also exposed to the results of the changes in the Current Population Survey's questionnaire design and data collection, which has led to a more accurate count of unemployment, especially among females. Members were also informed about the international effort to revise the Standard Industrial Classification (SIC) system, so that a common industry classification is established for the United States, Canada, and Mexico, which will eventually provide the information base to improve trade among North American nations.

### Call for Papers for 1995 Meetings

Virginia de Wolf 1995 GSS Program ChairYes, it is time to start planning for next year's Joint Statistical Meetings in Orlando! I can hear you saying to yourself, I have not even gone to the 1994 meetings in Toronto and they want me to think about the 1995 meetings. GSS needs your ideas to help make our 1995 program both topical and stimulating. What issues do you think are of interest to statisticians who work for federal statistical agencies, programmatic agencies, and state and local agencies, as well as to users of government statistics?The ASA program has five formats to support the exchange of ideas and interaction among participants: invited sessions; special topic contributed sessions; contributed paper sessions; poster sessions; and roundtable discussions. Invited sessions can be panel discussions as well as paper sessions. Special topic contributed sessions share a common theme with 35 presenters discussing a single topic. Individual papers are organized into contributed paper sessions of 56 papers. ASA also has two roundtable formats: luncheons (for which participants are charged a fee), and informal roundtables, without lunch, for topics that lend themselves to individualized discussion among groups of 810. Got another suggestion for a session format? If so, do not hesitate to contact me since new ideas are always welcome.The 1995 ASA President, Mitchell Gail, has selected "Statistics in Action" as the theme for the 1995 annual meetings. Try to capitalize on this theme as you plan to write a paper and/or organize sessions. If you have an idea for a paper or discussion topic that you would like to see on the 1995 ASA program, please send you ideas to the 1995 GSS Program Chair at the following address:Virginia A. de WolfBureau of Labor StatisticsORE, Room 49152 Massachusetts Ave., N.E.Washington, DC 202120001email: dewolfv@ore.psb.bls.gov2026067374 (voice) 2026067426 (FAX) 1993-1994

Government Statistics Section Proceedings Order Form

Please send the Government Statistics Section Proceedings (1993 and/or 1994):______ copies 1993 @ $20: Member Price; ______ copies 1993 @ $38: Non-member price______ copies 1994 @ $20: Prepublication Price (price good through 8/18/94) Subtotal ___________ Add 4.5% Sales Tax, (Virginia residents only) ___________ Add shipping charges ($3.50 per copy) ___________ Total ___________ Check or Money Order (Payable to American Statistical Association, Drawn in U.S. funds on a U.S. Bank) Credit Card: Visa MasterCard (no other card acceptable) Card No: ____________________ Exp. Date _________ Name Printed on Card: Signature: _______________________________________ ________________________________Date ___________ Telephone ______________________ ASA Member Number ______________ (from mailing label)Name _______________________________________________________Address _______________________________________________________ City/State ___________________________ ZIP _______ Country ______________________(Province)

Mail form to:

Nicole Brin, American Statistical Association

732 N. Washington St., Alexandria, VA 22314-1943

GovStat News

American Statistical Association

732 N. Washington St.

Alexandria VA 22314-1943

GSS Invited and Contributed Paper Sessions - 1994 Joint Statistical Meetings - Toronto Monday, 8/15 Tuesday, 8/16 8:30 a.m. Selected Research Issues on the New Interest in Public Library Data Nation's Veterans Type: Special Contributed Type: Special Contributed Chair: Paul Planchon Chair/Organizer: Robert Klein Organizer: Carrol Kindel 10:30 a.m. Microsimulating Federal Program Choices: Options for Change Type: Special Contributed Chair: Constance Citro Organizer: Bruce Klein 12:00 noon GSS Poster Sessions 2:00 p.m. Reinventing the Population Census: Data Collection Issues Balancing Privacy and Efficiency for Government Surveys and Censuses Type: Invited Type: Contributed Chair/Organizer: Lois Alexander Chair: Thomas Jabine 4:00 p.m. Sample Design and Estimation for Key Government Statistical Indicators Type: Contributed Chair: Virginia de Wolf Methodological and Quality Issues in Government Statistics Type: Contributed Chair: Wendy Alvey 6:00 p.m. GSS Business Meeting

GSS Invited and Contributed Paper Sessions - 1994 Joint Statistical Meetings - Toronto

Wednesday, 8/17 Thursday, 8/18 8:30 a.m. Federal and Non-Federal Data Needs The Use of Administrative Records from Continuous Measurement in Government Statistics Type: Roundtable Discussion Type: Contributed Discussants: Charles Alexander Chair: Gerald Gates and Ed Spar Is There a Place in ASA for Statisticians Who Work for State/Provincial/Local Governments? Type: Roundtable Discussion Discussants: Maria Gonzalez and Leonard Gaines Statistical Statesmanship* Type: Invited Panel Chair/Organizer: Monroe Sirken 10:30 a.m. Informing Education Policy: Special Government Statistics: The Role of International Issues, Methodology, and Policy Indicators Applications Type: Special Contributed Type: Special Contributed Chair/Organizer: Robert Lehnen Chair: Alex Michalos Organizer: Ben-chieh Liu 12:30 p.m. GSS Luncheon Roundtables 2:00 p.m. Modeling, Analysis, and Use of Government Statistics Type: Contributed Chair: Maria Gonzalez *Co-sponsored with the Social Statistics Section, the Survey Research Methods Section, and the Council of Professional Associations on Federal Statistics (COPAFS).NOTE: For comments or queries on this chart or any part of this newsletter, contact Constance F. Citro, Committee on National Statistics, 2101 Constitution Ave., NW, Washington, D.C. 204 18. Phone: 202-334-3093; FAX: 202-334-3751; Internet: CCITRO@NAS.EDU.