The Government Statistics Section has been publishing twice a year, GovStat News. We have agreed to transform this into a joint Government Statistics Section and Social Statistics Section newsletter. Tentatively, the title is "GovStat/SocStat News" [editor's note: suggestions for a title from readers are welcome]. This issue includes a Social Statistics column by Edie McArthur that starts the collaboration between the two sections [see p. 2]. Over the next year, we expect the joint newsletter to be a more equitable collaboration between the Government Statistics Section and the Social Statistics Section.There are some joint members of the Government Statistics and Social Statistics Sections. These members have already been receiving GovStat News. However, the American Statistical Association members who are members of the Social Statistics Section and not of the Government Statistics Section will receive the newsletter for the first time.
With this issue, we will be cooperating with the Government Statistics Section in producing a newsletter for our members. It only made sense. Many of our members are members of both sections (53% of respondents to a Social Statistics Section survey reported that they were also members of the Government Statistics Section). We share interest in many of the same issues, data sets, and data sources. So why not reach a wider audience? Also the Social Statistics Section knew a good deal when they saw it: Government Statistics had an established newsletter; the Social Statistics Section did not but its membership thought it should (see survey results).Social Statistics Section News In October, Martha Farnsworth Riche, our chair-elect, was confirmed as the new Census Bureau Director. She felt that she could not serve as chair of the section, so that post, by the by-laws of the section, falls to next year's chair-elect, Daniel Weinberg, Bureau of the Census, who will serve in that post for Riche's term as well as his own.
Other officers: 1995 Program Chair, Margo Anderson, University of Wisconsin-Milwaukee; Program Chair-Elect, Susan Miskura, Bureau of the Census; Secretary-Treasurer (1994-95), John Czajka, Mathematica Policy Research, Inc; Publications Officer (1995-97), David Cantor, Westat, Inc.; Council of Sections Representative (1995-97), Daniel Kasprzyk, National Center for Education Statistics.
Analysis of the 1994 Social Statistics Section Member Survey (submitted by Daniel H. Weinberg) In order to get a better reading of what our membership is interested in, the officers of the section developed a questionnaire for the section. Then, in January of 1994, 940 questionnaires were mailed by the American Statistical Association to all members of the Social Statistics Section. As of April 4, 1994, 169 questionnaires (18.0%) had been returned. (The low mail response was probably a reflection of the decision not to include a return-mailer in the survey, but it should be noted that the mailing was to the entire membership, not just a sample. Because of the relatively low response rate, however, findings should be interpreted with caution.)
General findings: When asked about how the section should focus its resources given $1,000 and a list of potential activities, respondents gave highest priority to "Sponsor methodological workshops or topical conferences" ($183) and "Social Statistics Section Newsletter" ($169). Second priority was given to "Co-sponsor continuing education programs with ASA" ($108), "Scholarships for students" ($102), and "Annotated bibliographies of recent publications" ($91). Finally, third priority was given to "Increase savings for future projects" ($62), "Commission monographs" ($55), "Joint Statistical Meetings mixer" ($44), and "Provide technical consulting services" ($41). Ideas for other activities were suggested by respondents as write-ins. These included: specialized training for social science applications; evaluation of Census Bureau Year 2000 plans for methods, content, and continuous measurement; development of Survey Research Methodology for the Census Bureau; regional get-togethers and seminars; oversight of federal government operations and proposed legislation (with reporting of results in Amstat News); more coordination with other sections and chapters; sponsoring a Bulletin Board; and giving awards of books and publications. A full list of recommendations may be obtained from Dan Weinberg, Housing and Household Economic Statistics Division, Bureau of the Census, Washington, DC 20233.
In recent years the number of participants in the luncheon roundtables has declined. Therefore we asked how the membership felt about roundtables. The overwhelming majority (86%) of respondents favor some kind of roundtable, particularly the no-fee kind (84%). However, a significant fraction (47%) also favors luncheon roundtables.The ASA membership is wrestling with the issue of certification of statisticians. Members of the Social Statistics Section who responded overwhelmingly oppose certification, with only 23% favoring it. A substantial number of comments were received, and can be obtained from Dan Weinberg.
Now to some interesting Social Statistics member background information. On average, the respondents have been statisticians for 20 years, members of ASA for 16 years, and members of Social Statistics for 13 years. The primary fields of specialization were statistics (31%), demography (20%), and economics (17%). Among the Social Statistics respondents, 42% reported that their main affiliation was academic, 21% reported the federal government, 13% reported for-profit business.
Question-by-question survey results
1. If the section had $1000 to spend on member services, please indicate below how you would allocate those funds. (N=160) Mean Standard Deviation
A. Increase savings for future projects $62 12.0
B. Co-sponsor continuing education programs with ASA $108 13.7
C. Sponsor luncheon roundtables at the Joint Statistical Meetings $34 6.8
D. Sponsor methodological workshops or topical conferences $183 17.7
E. Social Statistics Section Newsletter $169 18.3
F. Joint Statistical Meetings mixer $44 8.9
G. Awards for distinguished members $ 29 6.4
H. Scholarships for students $102 13.2
I. Membership drive $ 27 6.9
J. Provide technical consulting services $ 41 8.3
K. Annotated bibliographies of recent publications $ 91 14.3
L. Commission monographs $ 55 11.4
M. Develop cooperative arrangements with other professional associations $ 37 6.4
N. Other (please specify) $ 18 8.0
Note that some suggestions provided in the answer to question #2 (Do you have any other suggestions for how the Social Statistics Section can serve you better?) might have received greater support had they been listed explicitly. The answers to question 2 are available from Dan Weinberg, as are the answers to question #5 (If the section were to sponsor continuing education programs, methodological workshops, or topical conferences, what topics or issues should be the basis for such meetings?).
3. Attendance at recent JSM/Social Statistics luncheon roundtables has been low. Do you feel the section should continue to sponsor luncheon roundtables (the section pays for the luncheon of the roundtable leader)? Yes/No
4. The ASA has begun no-fee (non-luncheon) roundtables at the Joint
Should the section sponsor such no-fee roundtables? Yes/ No
Favor luncheon roundtables (Q3): 47% (N=145)
Favor no-fee roundtables (Q4): 84% (N=152)
Favor either kind of roundtable: 86% (N=157)
Favor both kinds of roundtable: 42% (N=144)
6. Attendees at the August business meeting of the section voted to oppose certification of statisticians. How do you feel about the ASA Committee on Certification's proposal (see Amstat News, May, July, October, and December 1993)? Favor/Oppose Respondents overwhelmingly opposed certification, with only 23% favoring it (90% c.i. 17-29%, N=145). A substantial number of comments were received.
7. Number of years you have been a statistician: 20 (mean); 1.0 (std.
dev.); 146 (N)
Been a member of ASA: 16 (mean); 1.0 (std. dev.); 161 (N)
Been a member of Social Statistics: 13 (mean); 1.0 (std. dev.); 140 (N)
8. Main field of specialization: (N=167)
Health Fields, 5%;
"Other" responses include biostatistics, history, management science, market research, mass communications, microsimulation, political science, social policy, survey research.
9. What is your main affiliation? (N=166)
Not-for-profit organization, 7%;
Business (for profit), 13%;
State/Local Gov., 5%;
Federal Gov., 21%
10. Do you also belong to the Government Statistics Section? Yes: 53% (N=158)
11. Other organizations you belong to: (N=169) [multiple responses
American Association for the Advancement of Science, 19%;
American Association for Public Opinion Research, 10%;
American Economic Association, 13%;
American Educational Research Association, 7%;
American Public Health Association, 8%;
American Sociological Association, 17%;
Population Association of America, 27%;
American Psychology Association, 5%;
Institute of Mathematical Statistics, 6%;
International Association of Survey Statisticians (7%)
Eighty-six percent reported that they belonged to at least one other organization; 82 different organizations were mentioned under "other".
12. What annual meetings have you attended in the past year? the past 5
years? (N=169) [multiple responses possible]
Meetings receiving at least 5% attendance in the past year:
Allied Social Science Associations, 5%;
American Educational Research Association, 5%;
American Public Health Association, 5%;
American Sociological Association, 7%;
Joint Statistical Meetings, 38%;
Population Association of America, 15%.
Meetings receiving at least 10% attendance in the past 5 years:
Allied Social Science Associations, 15%;
American Association for the Advancement of Science, 10%;
American Association for Public Opinion Research, 10%;
American Public Health Association, 10%;
American Sociological Association, 16%;
Joint Statistical Meetings, 63%;
Population Association of America, 25%.
Sixty-three different conferences were mentioned under "other".
The Government Statistics Section, the Social Statistics Section, and the Washington Statistical Society have agreed to jointly sponsor an award to honor the late Roger Herriot [see article, GovStat News, June 1994]. The three organizations have each designated a representative to determine how the award shall be given. The three members of the award committee are Dan Kasprzyk (chair), Fritz Scheuren, and Dan Levine. More detailed announcements will be published in Amstat News.The Proceedings of the 1994 program of the Government Statistics Section will be available from the ASA office soon. This publication will include the roundtable discussion on state and local statisticians. The objective of this discussion was to achieve a greater participation of state and local statisticians in the activities of the ASA. All sessions sponsored by GSS were eligible for inclusion in the Proceedings. Bob Lehnen is the new GSS Chair. I wish him well and know he will have a productive year. He can be reached at Indiana University, Public and Environmental Affairs, 801 West Michigan St., Indianapolis, IN 46202-2880; tel. 317-261-3041; e-mail: firstname.lastname@example.org.
The Council of Professional Associations on Federal Statistics looks forward to a busy and productive 1995. In the immediate weeks ahead, the COPAFS executive director, Ed Spar, will unscramble the currently enigmatic reconfiguration of the new Congressional committees whose assignments will touch upon the federal statistical system. Ed will familiarize the new committee members with the needs of the user community and explore the ability of the federal statistical agencies to meet those needs. High among priorities is continued work with the Bureau of the Census on the use of private and state and local government administrative records in the 1995 Test Censuses to improve nonresponse follow-up and to achieve integrated coverage measurement.COPAFS will also continue to work with the Office of Management and Budget on a conference to analyze potential revision of metropolitan area concepts. In addition, as part of the year's activities, COPAFS hopes to co-sponsor a program with the Bureau of Economic Analysis to assess new BEA initiatives on economic indicators, such as national product accounts.
There are several important statistical issues which were discussed in a variety of meetings in September and October (COPAFS, September 9; 2000 Advisory Committee, September 8-9; Census Advisory Committee of Professional Associations, October 20-21; Association of Public Data Users, October 24-26). This is a summary of these presentations and discussions.
The Census Bureau has made a commitment to further research and testing of this data collection vehicle, with a decision to continue or abandon the project scheduled for the end of 1996. The form of Continuous Measurement (CM) currently being discussed involves a sample of 400,000 households per month in the 1999-2001 period, followed by 250,000 per month thereafter. This would provide a 3-year average of data at the census tract/small community level at first, and a 5-year moving average after that, with sampling errors slightly higher than those of the traditional census long form. There are a host of issues surrounding CM.
It has many advantages. Data would be more current. Data collection would be more flexible and responsive to special needs--for example, to assess the effect of natural disasters. A permanent, trained staff would produce higher quality data. Products would come out with only a 6-month time lag. The survey could serve as a sampling frame to find rare populations for follow-on surveys (e.g., regarding specific disabilities), by including appropriate screening questions.
There are also many problems. The current plan expects a 60 percent mail return of a questionnaire with full decennial census content. Telephone interviewing would be used to complete about half of the remaining 40 percent; the rest would be followed up with personal interviewing on a sample basis. Quality issues include variance in small areas, nonresponse coverage, item nonresponse (e.g., respondents skipping the income items), quality of reporting, residency rules, seasonal vacancies, and the reference period for such items as income, employment status, and journey to work. In addition, there are many questions to be answered about the utility and meaning of a data set when the data are collected over a long period of time. Another important point concerns the relationship between CM and other ongoing surveys, including the Current Population Survey (CPS), Survey of Income and Program Participation (SIPP), and American Housing Survey (AHS). With the establishment of a Master Address File, one master sample could be used for CM and the other surveys; the initial CPS, SIPP or AHS interview could provide data to CM.
Census staff have begun discussions with other federal agencies regarding these possibilities. The big question is: will there be a long form in 2000? Does CM, if successful, mean elimination of the long form? Both National Academy of Sciences Panels have stated that the long form does not hurt the census response rate, and the Requirements Panel, in its report Modernizing the U.S. Census, specifically recommends retention of the long form [see p. 9]. Many data users would like to have both, to have the 2000 long form and then to begin CM after the census.
Race and Ethnic Classifications
The Office of Management and Budget has jurisdiction over this issue. Its Statistical Policy Directive 15, establishing the familiar "White, Black, Asian or Pacific Islander, Native American, or Other" racial categories along with the independent identification of "Hispanic," was promulgated in 1977 and has not been revised since. Following a series of Congressional hearings in 1993, OMB's statistical policy group issued a Federal Register notice, soliciting input, and conducted four hearings on the issue during this past summer.
The controversy can be summarized into four distinct sets of issues: Should people be able to identify as "multi-racial" or "multi-cultural"? There is significant support for this choice around the country (including a bill introduced in the Michigan legislature), but it poses clear statistical classification problems. An alternative may be to permit people to mark more than one category, and handle the issue in tabulation and reporting. Should additional racial and/or ethnic categories be established? For example, Arabs and Chaldeans in Detroit often refuse to identify as White in the census. Should native Hawaiians be classified as Native Americans (with their perceived federal government benefits) rather than with Asian and Pacific Islanders? Should there be a classification by race and ethnicity at all? OMB staff are currently preparing a summary of the input received so far, with another Federal Register notice scheduled for early 1995. There will also be a May 1995 CPS supplement that tests several new data collection forms in this area.
Income and Poverty Estimates
Over the past two or three years, several members of Congress have expressed serious concern over the use of decennial census income and poverty statistics for an entire decade, and the major shift in funding that occurs when new data are brought into the allocation of Chapter 1 education funds after each census. Accordingly, Rep. Tom Sawyer introduced legislation that would mandate biennial estimates of income and poverty level at the school district level. The bill passed in the House of Representatives but did not become law in the Congress just concluded; however, $600,000 was appropriated specifically for the Census Bureau to begin work to develop the methods. The statistics to be estimated include the count of persons in poverty by age (under 5, 5-17, 18-64, 65 and older), median household income, and per capita income. The Census Bureau is working now to formulate a strategy to develop these six statistics at the state level using March CPS data; these would then be the controls for any substate estimates. The current plan is to have 1993 state and county estimates in Fall 1996, and to work toward school district estimates in 1998.
Meanwhile, a law entitled "Improving America's Schools Act" did pass in the final days of the last Congress. This Act authorizes the Secretary of Education to use updated estimates rather than just decennial data for allocation of education funds. It also authorizes a National Academy of Sciences panel to study the entire estimation issue.
1995 Census Test
The Census Bureau is stepping up its pace in preparation for the 1995 Test of Census Methods to be conducted in Oakland, CA, Paterson, NJ, and six parishes in rural Louisiana. (New Haven, CT, had also been designated as a test site, but was dropped due to lack of funds.) A wide variety of new methods will be tested, designed to improve response and/or to reduce cost. Each has been specified in an Evaluation Research Design (ERD), which specifies the objectives and how they will be measured.
Non-Federal Content Outreach
The Census Bureau has just about completed its research into federal data needs, and has classified questionnaire items as mandatory, required, or programmatic. Mandatory items are those specifically mentioned in law. Required items are needed to meet statutory requirements for data for which the decennial census is the only or traditional source. Programmatic items meet neither of these qualifications, and are at risk of being eliminated. In addition, proposed items on health insurance coverage and non-cash benefits are in the programmatic category. Finally, a number of housing items that have been traditionally collected on a full-count basis may be moved to the sample questionnaire.The Census Bureau has developed a lengthy questionnaire to solicit information on how the programmatic, and some of the mandated, items are used outside the federal government. It has also contracted with the Association of Public Data Users (APDU) to solicit input. APDU will be using a Census Data Use Description Form along with the census questionnaire. To find out more about this project and/or to assist in the outreach, please contact APDU by voice at 609-258-6025 or by e-mail at APDU@princeton.edu.
The 1995 Joint Statistical Meetings (JSM) will be held in Orlando, Florida, Sunday-Thursday, August 13-17. Mitch Gail, 1995 ASA President, has selected "Statistics in Action" as the theme for these meetings. Authors were encouraged to submit papers on this theme. All sessions on this subject will be designated in the program as theme sessions.
The Government Statistics Section was allotted one invited session for the 1995 JSM and gained an additional invited session in the competition that was held at the mid-January planning meeting of the 1995 Program Committee. Thanks to all of those who contacted me with suggestions for invited sessions. Also, many, many thanks from me and from Carolyn Shettle (1996 GSS Program Chair) to the many GSS members who took the time and effort to organize special contributed sessions and roundtable contributed discussion groups. We look forward to a great program in Orlando!
The first of the two GSS invited sessions will discuss the report of the Committee on National Statistics Panel on Poverty and Family Assistance (scheduled to be released in May). Robert Michael, University of Chicago, the panel chair, will give an overview of the report, Measuring Poverty: A New Approach. Then, three discussants will give their reactions: Nancy Gordon, Congressional Budget Office; Harold Watts, Columbia University (on leave at The Urban Institute); and Daniel Weinberg, Bureau of the Census.
The second invited session will review the issues surrounding OMB's Statistical Policy Directive No. 15, "Race and Ethnic Standards for Federal Statistics and Administrative Reporting" [see Becker article, p. 6]. The session will include papers by: Suzann Evinger, Office of Management and Budget, "Why Review OMB's Statistical Policy Directive on Race and Ethnicity?"; Clyde Tucker, Bureau of Labor Statistics, and Roderick Harrison, Bureau of the Census, "The Research Agenda on Issues Surrounding the Definition of Racial and Ethnic Categories"; and Manuel de la Puente, Bureau of the Census, and Ruth McKay, Bureau of Labor Statistics, "Developing and Testing Race and Ethnicity Questions for the Current Population Survey Supplement." Reynolds Farley, University of Michigan, will discuss the papers.
Lastly, GSS will be organizing another set of roundtable luncheons. Typically, each roundtable has one discussion leader and can accommodate up to nine participants. These are fee events. If you have an idea for a topic for one of the GSS-sponsored roundtable luncheons, please contact Carolyn. Please feel free to contact Carolyn or me for assistance as you prepare for Orlando. We'd be happy to help in any way we can.
FINALLY, DO NOT FORGET!--The 1995 Joint Statistical Meetings will begin on SUNDAY, August 13th. The reason for this is that the 4:00-6:00 p.m. slot on Tuesday is set aside for the Deming Lecture, and no contributed or invited sessions will be scheduled at that time. To allow for the same number of contributed and invited sessions as we have had in prior years, the annual meeting will begin with a 4:00 p.m. session on Sunday, August 13th. The 1995 Program Committee is planning to schedule some of the more interesting invited sessions on Sunday afternoon. So . . . do plan to be in Orlando early so you can catch some of the dynamite sessions that will be scheduled for Sunday. Virginia de WolfBLS, ORE, PSB 49152 Massachusetts Avenue, N.E.Washington, DC 20212-0001, voice: 202-606-7374, fax: 202-606-7426, Internet: email@example.com; Carolyn Shettle, National Science Foundation, Division of Science Resources Studies, 4201 Wilson Boulevard, Suite 965, Arlington, VA 22230, voice: 703-306-1780, fax: 703-306-0510, Internet: firstname.lastname@example.org
This is my last issue of GovStat News, as my 2-year term as GSS Publications Officer has expired. I have enjoyed very much the opportunity to serve the section, and am very pleased that the GSS and SSS will be collaborating to produce future newsletters. The 1995-1996 GSS Publications Officer is Joan Turek, who I am sure will do an outstanding job. She can be reached at the Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services, 200 Independence Ave., SW, Washington, DC 20201; tel. 202-690-5338; Internet: email@example.com.
The Committee on National Statistics Panel on Census Requirements in the Year 2000 and Beyond released its final report, Modernizing the U.S. Census, in November 1994. The panel, chaired by Charles L. Schultze of the Brookings Institution, supported the need for the range of data for small areas and small population groups that the census provides. The panel reviewed the problems of high costs and less-than-complete population coverage that have affected recent censuses. It recommended a design--already being developed by the Census Bureau--that combines traditional counting with increased use of statistical estimation, both in sampling for nonresponse follow-up and in completing the count through an integrated coverage measurement program. The panel also recommended a thorough reengineering of all census operations, in light of the new design, that could identify substantial cost savings without sacrificing quality. Finally, the panel recommended that the year 2000 census include a long form (or perhaps more than one medium-length form) to obtain needed information on the characteristics of the population. For information on obtaining the report or a separately-bound summary, contact CNSTAT, tel. 202-334-2246. 1994 Government Statistics Section Proceedings Order
Please send the 1994 Government Statistics Section Proceedings:
______ copies @ $20 (Member Price) ___________
Add 4.5% Sales Tax (Virginia residents only) ___________
Add shipping charges ($3.50 per copy) ___________
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 ___________
ASA Member Number ______________ (from mailing label)
City/State________________ ZIP _______
Mail form to:
American Statistical Association
732 N. Washington St.
Alexandria, VA 22314-1943
American Statistical Association
732 N. Washington St.
Alexandria, VA 22314-1943