Survey Research Methods Section Newsletter: Winter 1995-1996



Highlights
It s time to start planning your summer vacation ... in Chicago! SRM is already gearing up for an exciting program at the 1996 Joint Statistical Meetings -- see below for a taste of what is to come; also, see page 3 for information on the latest SRM short course. Here s hoping we ll see you there!

InterCASIC 96 isn't far off ! A call for contributed papers is out -- for updated information on latest SRM-sponsored International Conference on survey methodology, see below.

SRM is going electronic! The SRMSNET is well underway -- see page 6 -- and we ll soon be on the World Wide Web, as well -- see below. Help us as we move towards the 21st century!



1996 JSM in the Windy City
As you surely know by now, the 1996 Joint Statistical Meetings will be held in Chicago, IL, August 4-8, 1996. The SRM Program is already shaping up for the occasion. The Section has been awarded 6 invited paper sessions:

Organization of special contributed, regular contributed, and poster sessions and round table luncheons is now underway. More details will be provided in the next newsletter.


SRM Joins the World Wide Web
SRM is pleased to announce that this Spring it will become one of the first Sections to join ASA at their new Web site. The Council of Sections, with practically unanimous support from the Sections, has established an Internet node at ASA on the World Wide Web. As part of that effort, SRM is developing its own home page; right now you can reach the ASA Web site by accessing http://www.amstat.org/.

The SRM home page is now being finalized and both suggestions and assistance are very welcome. As it stands now, the SRM Web page will include the following items:

Other potential items to be added later might be: We are also looking for a permanent "webmaster" to coordinate and update the SRM presence on the Web. All volunteering, information, and comments should currently be addressed to Bill Kalsbeek, who agreed to help us get this effort off the ground (thanks, Bill!) -- e-mail: bill_kalsbeek@unc.edu.



What Is a Survey?
By now all Section members should have received the first three brochures in the newly updated series on survey methodology -- What Is a Survey? How to Plan a Survey, and How to Collect Survey Data. (Additional copies are available from the ASA Office for $.50 each.) The series is aimed at improving survey literacy for individuals who participate in surveys or use survey results. By promoting an understanding of what is involved in carrying out sample surveys of human populations, hopefully, the series will lead to more and better uses of survey data.

Several more pamphlets in the series are now in the works. They include: Judging the Quality of a Survey; What Are Focus Groups? More About Mail Surveys; How To Conduct Pretesting; What Is a Margin of Error? and What About Surveys in the Media? The Section would like to continue to produce relevant additions to the series on other topics, in the years ahead. Suggestions for potential future topics -- as well as offers to help in this effort -- are welcome!

These pamphlets are intended to be a work in progress. Your feedback has been invaluable and continues to be needed. Please pass along any suggestions for improving the brochures to Fritz Scheuren at George Washington University -- phone: (202) 994-6358; fax: (202) 994-6917; or e-mail: scheuren@aol.com.



InterCASIC 96
Dreaming of a warm weather getaway....? InterCASIC 96 is coming to San Antonio, Texas December 11-14, 1996! The International Conference on Computer-Assisted Survey Information Collection (CASIC) is the sixth in a series of meetings on survey methodology organized and co-sponsored by the Survey Research Methods Section. The focus is being broadly defined to include general applications of computer technology to the entire process of collection, capture, and preparation of survey data.

The Organizing Committee is now soliciting abstracts for contributed papers and proposals for live demonstrations. (See Call for Papers on page 7 for specific details.) Suggested topics for papers include:

The InterCASIC 96 focus on technology also makes this year s conference a natural as an interactive forum with appropriate equipment for live demonstrations of automated systems. Proposals are, therefore, being sought for those who would like to come and demonstrate hardware or software for computer-assisted survey information collection.

For more information on the conference, itself, contact Mick Couper -- phone: (301) 405- 9523; fax: (301) 314-7912; or e-mail: mcouper@survey.umd.edu -- or Lee Decker -- phone: (703) 684-1221 x 145; fax: (703) 684-8069; or e-mail: lee@amstat.org. See, also, the World Wide Web page at http://www.wam.umd.edu/~jpsm/casic.html .



Treasurer s Report
The SRM treasury remains quite healthy. As of June 30, 1995, we had cash on hand of $34,394. Our largest source of revenue is from the Proceedings, as is the largest portion of our expenses. We have also set aside monies to cover publication of the pamphlets in the What Is a Survey? series and to serve as seed money for the InterCASIC 96 conference in San Antonio. Meanwhile, we continue to receive revenue from past conferences sponsored by the Section. Our dues -- at $2.00 a year -- remain among the lowest available. Membership figures for the Section were 1,424, as of September 30, 1995.



List-Assisted Telephone Surveys Short Course Planned
The ASA Committee on Continuing Education has approved a short course on List-Assisted Telephone Surveys: Design, Estimation and Administration, to be offered at the 1996 Joint Statistical Meetings. The course, to be taught by Robert Casady, Bureau of Labor Statistics; James Lepkowski, University of Michigan; and Clyde Tucker, Bureau of Labor Statistics, is the first SRM-sponsored short course to be offered in a few years.

The course will examine sample design and estimation issues for an expanding class of surveys by telephone -- list-assisted telephone sample surveys. Many telephone sampling methods have been employed, with samples being selected either from lists of available telephone numbers or by randomly generating numbers from the conceptual set of all possible telephone numbers. Most list samples have severe coverage problems, while random digit dialing methods have substantial inefficiencies, due to the large number of sample elements that are not working residential telephone numbers. List-assisted telephone sample designs use a list of telephone numbers as a basis for stratifying the conceptual set of all possible telephone numbers; thus, design efficiency is substantially improved, while desirable coverage properties are maintained.

The course will present basic features of list samples, random digit dialing methods, and list- assisted techniques. Several specific design options for list-assisted samples will be considered in detail, including two-stage sampling from directory listings, stratified sample selection from partitioned sets of telephone numbers, optimum allocation of samples across strata, and truncated sample designs. Such estimation topics as nonresponse adjustment and post-stratification will also be addressed. Finally, administrative aspects of sample selection will be covered, as well.

The course is aimed at survey researchers, survey practitioners, and graduate students with prior training or experience in the general area of sample surveys. For more information, look for details in the JSM Preregistration Packet, to be sent out in late Spring by ASA.



Waksberg Wins Herriot Award!
Congratulations go out to Joseph Waksberg, Westat, Inc., who became the first recipient of the Roger Herriot Award for Innovation in Federal Statistics! The Award, sponsored by the Social Statistics Section, Government Statistics Section, and the Washington Statistical Society, recognized Waksberg for his recent contributions to improving procedures for rare populations, methods for random digit dialing, and understanding of recall error.

Nominations are now being sought for the 1996 Herriot Award. The award honors individuals who develop unique approaches to the solution of statistical problems in Federal data collection programs. It is intended to reflect the special characteristics of Roger Herriot s career --

As such, individuals at all levels -- from entry to senior -- from the Federal government, private sector, or academic community may be nominated on the basis of the significance of a specific contribution. For nomination forms or information contact Daniel Kasprzyk -- phone: (202) 219-1588; fax: (202) 219-1728; or e-mail: daniel_kasprzyk@ed.gov. Completed nominations are due May 17, 1996.



Nine New Fellows!
Nine Section members were inducted as Fellows of the American Statistical Association at the 1995 Joint Statistical Meetings in Orlando. Hearty congratulations go out to Susan W. Ahmed, Kenneth R. Brewer, Brenda G. Cox, Don A. Dillman, Ralph E. Folsom, Richard John Kryscio, Richard F. Potthoff, Keith F. Rust, and Tommy Wright!

Many thanks to Cathryn Dippo (Chair) and the SRM Committee on ASA Fellows for their efforts on behalf of our outstanding Section members.



New Report on Electronic Data
The Office of Management and Budget s Federal Committee on Statistical Methodology has just released the latest in its Statistical Policy Working Paper series of reports on issues of data quality affecting Federal statistics. The report -- Statistical Policy Working Paper 24: Electronic Dissemination of Statistical Data -- summarizes the current state-of-the-art in electronic dissemination of Federal statistical data available for public use. While the report is intended to serve as an aid to statistical agency managers in decision-making about electronic dissemination of publicly-available statistical data, it contains an agency-by-agency summary of current practice, in the Appendix, which might be especially of interest to the survey research community, as a whole.

The 87-page report is available in both hardcopy and electronic versions. For more information on obtaining this or other reports in the series, see the next article.



Statistical Policy Working Papers Now On-Line
The Office of Management and Budget s Federal Committee on Statistical Methodology is pleased to announce that its Statistical Policy Working Paper series of reports on issues of data quality affecting Federal statistics is now available in electronic format. The first 22 reports (see the box) can be accessed via the Internet at http://www.bts.gov/statistical policy. Reports Numbers 23: Seminar on New Directions in Statistical Methodology (1995) and 24: Electronic Dissemination of Statistical Data (1995) will be added to this collection shortly.

At this time, most of the tabular and graphical information from the original works only appear in hardcopy form. However, the National Science Foundation has produced an Adobe "Acrobat" version of Report No. 24, which will contain the original graphics. This report can be found at http://www.nsf.gov/sbe/srs/stats.htm. (For information on obtaining hardcopy versions of these reports, contact M. Denice McCormick Myers -- phone: (202) 690-0922; fax: (202) 720-8738; or e-mail: dmmyers@nass.usda.gov.)




Federal Committee on Statistical Methodology Statistical Policy Working Papers

No.Title & Year
1. Report on Statistics for Allocation of Funds, 1978
2. Report on Statistical Disclosure and Disclosure-Avoidance Techniques, 1978
3. An Error Profile: Employment as Measured by the Current Population Survey, 1978
4. Glossary of Nonsampling Error Terms: An Illustration of a Semantic Problem in Statistics, 1978
5. Report on Exact and Statistical Matching Techniques, 1980
6. Report on Statistical Uses of Administrative Records, 1980
7. An Interagency Review of Time-Series Revision Policies, 1982
8. Statistical Interagency Agreements, 1982
9. Contracting for Surveys, 1983
10. Approaches to Developing Questionnaires, 1983
11. A Review of Industry Coding Systems, 1984
12. The Role of Telephone Data Collection in Federal Statistics, 1984
13. Federal Longitudinal Surveys, 1986
14. Workshop on Statistical Uses of Microcomputers in Federal Agencies, 1987
15. Quality in Establishment Surveys, 1988
16. A Comparative Study of Reporting Units in Selected Employer Data Systems, 1990
17. Survey Coverage, 1990
18. Data Editing in Federal Statistical Agencies, 1990<
19. Computer Assisted Survey Information Collection, 1990
20. Seminar on Quality of Federal Data, 1991
21. Indirect Estimators in Federal Programs, 1993
22. Report on Statistical Disclosure Limitation Methodology, 1994
23. Seminar on New Directions in Statistical Methodology, 1995
24. Electronic Dissemination of Statistical Data, 1995



Quality Journal Wants Input
Data Quality journal is interested in producing half-hour educational videos on data quality in Statistics, Quality, Engineering, Social Science, Medicine, etc. Ideas and suggestions are being solicited. Provide input by e-mail to dataqualty@aol.com.



FTC Hosts Privacy Discussion Group On-Line
The Federal Trade Commission's Bureau of Consumer Protection has established a Privacy Principles Mailing List as a forum for discussing the privacy principles that should be applicable to consumers' use of the Global Information Infrastructure. The Bureau seeks commentary and resource materials from consumers, industry, privacy advocates, and others who are concerned about questions such as the following:

Additional related topics for discussion are welcome. The dialogue that takes place in response to these questions will be invaluable to the Bureau. It will facilitate the creation of a set of voluntary principles governing the use of consumer information in transactions in cyberspace. To post messages to this list, simply address your mail to: privacy@ftc.gov. To subscribe to this list, send a SUBSCRIBE command (in body the of the message) to: privacy-request@ftc.gov.



Looking Ahead -- Your Input is Needed Now!
It is not too early to start planning for the 1997 Joint Statistical Meetings, which will be held in Anaheim, CA, August 10-14, 1997. The first decisions on invited paper sessions for the 1997 JSM will be made at the meeting of the Program Committee in Chicago this August. If you have any ideas for topics or would like to organize a session, please contact Mary Mulry, the Section s 1997 Program Chair. Remember, the more complete the material she can present, the better.

Steve Miller, Chair of the SRM Tutorial Committee, is also looking for your ideas. He is actively seeking suggestions for short courses for the Anaheim meetings. We are interested in both new techniques and more general presentations of current methods. Some of you may have been discouraged by past problems with short courses which were canceled at the last minute. A recent change in policy by ASA now permits the Sections to subsidize under- subscribed short courses which they wish to sponsor. So, do give us a try again -- we would like to see the Section become a bigger part of the ASA Continuing Education program...



Statistical Methodology at Work
When Canadians in Quebec Province went to the polls last Fall to vote whether to secede from Canada and establish a separate French Canada, many of us watched the results with much interest. The results were an extremely close vote, with 50.6% voting not to secede from Canada and 49.4% voting in favor of a separate French Canada. Derek Leebosh, AAPORNET s Canadian correspondent, commented on the methodology which led polling companies to almost unanimously predict a dead heat in such a hotly contested election -- and that is exactly what happened!

Leebosh noted that all of the polls gave the yes side a lead of between four and seven points among decided voters, with yes hovering around 47-48% and no around 40-41%. However, all companies forecasted a dead heat, because they all used formulas to allocate the undecideds on a disproportionate basis. Among their assumptions were that undecided voters would be inclined to err on the side of caution on an issue of such great consequence -- caution, in this case, being the status quo; that the undecided voters tended to be among groups that were more likely to vote no -- such as women and older people; and that French speakers may have responded to pollsters with politically correct yes votes, when, in fact, they were intending to vote no, leading to slightly inflated yes counts during the final days of the campaign. Leebosh concluded that, in this election, it was the pollsters who clearly won!



Reminder...
Remember, all presenters must submit a draft by June 1, 1996, in order to remain on the program. A lot of effort goes into developing interesting and homogeneous sessions and trying to minimize the subject matter and person conflicts which inevitably arise. Please bear in mind that withdrawing a paper after it has been scheduled often upsets the delicate balance which has been achieved. This is a particularly serious problem when papers are withdrawn after the program has gone to press. So, if you have submitted an abstract for the Chicago meetings, try to keep your research on track and the Program Chair -- Cathy Dippo -- informed.



SRMSNET is Up and Running!
The SRMSNET went on-line this past summer, thanks to the efforts of Bob Groves and the Maryland/Michigan/Westat Joint Program in Survey Methodology. Since July 1995, 126 subscribers have signed up and 21 messages have been sent out, including the text of the last SRM Newsletter.

Needless to say, we are a little disappointed with the limited use of the bulletin board that we've seen so far. The intent was for the SRMSNET to be an active vehicle for communication of such topics as:

We invite you to try out the SRMSNET to post useful information and seek input from your colleagues. Recently a request went out for a dialogue on methodological issues pertaining to analysis of longitudinal survey data -- what kinds of questions can be better answered by longitudinal surveys and what methodological difficulties have been encountered? If you are not a current user of the SRMSNET, sign on and join us in this and other exchanges. To subscribe to the SRMSNET, send an e-mail message to cs178@umail.umd.edu with a message to subscribe SRMS-LIST in the body of the message. Once confirmation is received, mail can be sent to the bulletin board by addressing e-mail messages to SRMSNET-LIST@umail.umd.edu.

Your participation can help make the SRMSNET a real success!



Call for Papers
InterCASIC 96 -- Contributed papers and demonstration proposals are now being accepted for the International Conference on Computer-Assisted Survey Information Collection (CASIC), to be held in San Antonio, TX, December 11-14, 1996. Presentations on current and future CASIC methods are being solicited. Send abstracts (300-500 words) with $100 submission fee* to Lee Decker, Conference Planner, American Statistical Association, 1429 Duke Street, Alexandria, VA 22314-3402 by May 1, 1996. Demonstration proposals should also include an abstract (with submission fee*), describing the proposed demonstration content and any hardware and software requirements; clearly indicate it is a demonstration. All abstracts should include name, address, telephone number, fax number, e-mail address, work affiliation of author(s), and primary contact person. Be sure to indicate that the submission is for InterCASIC 96. For more information on the conference, see the feature article on page 2.

gad@olive.mscc.huji.ac.il.



Orlando JSM
The SRM Program for the Orlando Joint Statistical Meetings (JSM) was a great success! We sponsored 5 invited sessions, 12 special contributed sessions, 20 regular contributed sessions, 5 poster presentations, and 6 round table sessions. Attendance was also strong at Section sessions throughout the meetings. Thanks, especially, to Nancy Mathiowetz, for her outstanding efforts as the Section Program Chair!



Newsletter Deadline....
Items for publication in the Spring/Summer issue of the SRM Newsletter should be submitted no later than June 1, 1996. Items should be sent to Wendy Alvey at Statistics of Income, PO Box 2608, Washington, DC 20013-2608; phone: (202) 874-0455; fax: (202) 874-0964; or e-mail: wendy.alvey@wpgate.irs.gov.



On the Ballot in 1996...
The following nominees will appear on the ballot for SRM officers:
Section Chair-Elect: Eugene Ericksen, Temple University
Don Rubin, Harvard University
Program Chair-Elect: Mike Hidiroglou, Statistics Canada
Elizabeth Martin, U.S. Bureau of the Census
Secretary-Treasurer: Lars Lyberg, Statistics Sweden
Rajendra Singh, U.S. Bureau of the Census
Publications Officer: F. Jay Breidt, Iowa State University
Charles Proctor, North Carolina State University
Council of Sections Rep: Ralph Folsom, Research Triangle Institute
Colm O Muircheartaigh, London School of Economics

Look for your ballot in the mail this March and don t forget to vote!



Upcoming Conferences
ARC 1996 -- The Bureau of the Census will hold its 1996 Annual Research Conference and CASIC Technologies Interchange at the Key Bridge Marriott Hotel in Arlington, VA, March 17-21, 1996. The theme for this year s conference is Survey of Methodologies for the 21st Century and it will feature invited and contributed presentations (with discussion) on methodologies and technologies for improving surveys and censuses. Among the topics on the program are disclosure avoidance methodologies, longitudinal data, reduction of respondent burden, small area estimation for labor force programs, use of administrative list data, automated coding, generalized tools and systems, record linkage issues, electronic data dissemination, and metadata. For further information, contact Maxine Anderson-Brown, Conference Coordinator, Bureau of the Census, Washington, DC 20233; phone: (301) 457- 2308; or fax: (301) 457-3682.

Data Editing Workshop and Exposition -- The Federal Committee on Statistical Methodology, the Bureau of Labor Statistics, the Joint Program in Survey Methodology, and the Washington Statistical Society will jointly sponsor a one-day workshop and exposition on editing survey data at the Bureau of Labor Statistics Conference and Training Center in Washington, DC, March 22, 1996. The event will feature both formal presentations and informal demonstrations of data editing -- those procedures designed and used for detecting erroneous or questionable survey data, with the goal of correcting the data or identifying potential estimation problems which can be handled by such means as reweighting. For more information, contact Mark Pierzchala -- phone: (703) 235-5218; or e-mail: mpierzchala@ag.gov -- or David Pierce -- phone: (202) 452-3895; or e-mail: m1dap00@frb.gov.

Symposium on Recent Developments in Sampling Methods -- The Department of Statistics and Operations Research, Leonard N. Stern School of Business, New York University will sponsor a one-day symposium on sample designs and the analysis of resulting data, at New York University, Friday, May 17, 1996. Among the topics featured are data confidentiality, Bayesian approaches to survey sampling, adaptive sampling strategies, resampling methods for complex surveys, and estimation for longitudinal surveys. For more information, contact Mark Handcock -- phone: (212) 998-0475; fax: (212) 995-4003; or e-mail: symposium96@stern.nyu.edu.

Kullback Memorial Research Conference -- The Department of Statistics, George Washington University, will sponsor a two-day research conference in memory of Solomon Kullback, at the Washington Marriott Hotel, May 24-25, 1996. The Conference will focus on applications of the Kullback-Leibler (K-L) information measures in statistical research today. For more information, contact the Dept. of Statistics -- phone: (202) 994-6356; fax: (202) 994-6917; or e-mail: statp1@gwuvm.gwu.edu.



Section Election Results
Congratulations and best wishes to the following newly-elected officers of the Section:



RIC Statement of PROFESSIONAL & ETHICAL STANDARDS for
Marketing and Opinion Research


Properly conducted, marketing and opinion research has great value to its sponsors -- and to the public, as citizens and consumers. The value rests, in the end, on the intelligence, creativity and technical skills of research practitioners and, most critically, on their integrity and ethical standards.

No code can cover all of the ethical issues for all research activities. But, regardless of the technique or the nature of the problem, certain rules describe good research practice. The rules relate to:

What follows are recognized standards for good practice in each of the four areas. Underlying each set of standards -- for how research should be done, how it should be reported, how the public ought to be treated and how the business of research should be conducted -- is a basic principle: These principles cover the most basic of the responsibilities of those who do marketing, opinion, and related research. They are the ethical foundation for research practice.



Research Industry Coalition Establishes Professional and Ethical Standards
At the 1995 Joint Statistical Meetings, the Survey Research Methods Section, on behalf of the American Statistical Association, endorsed a Statement of PROFESSIONAL & ETHICAL STANDARDS for Marketing and Opinion Research put forward by the Research Industry Coalition (RIC). (See page 9.) The RIC has membership from:

One of the goals of the RIC is to raise public consciousness of poor survey practice. The Coalition meets several times a year. Our current representatives are Chuck Cowan and Warren Mitofsky. Innis Sande had been one of our representatives up until this year.

Although some of us might find that the code is not worded strongly enough, it was felt that endorsing this code would be better than not giving it any endorsement at all, as it was unlikely that a more strongly worded code would have been acceptable by all the Coalition members.

Maybe this code will generate some discussion on the SRMSNET!! (See Letters to the Editor for further discussion of the RIC Guidelines.)




Letters to the Editor
The following exchange pertains to the SRM decision to support the new Professional and Ethical Standards for Marketing and Opinion Research, recently drafted by the Research Industry Coalition (RIC).

Further discussion of these issues appears on the SRMSNET -- join us there with your comments...


Return to :
American Statistical Association (ASA)
ASA Survey Research Methods Section


Send comments and suggestions.