Washington Statistical Society
        Washington Statistical Society on Meetup

January 2005


Electronic Voting Is Coming

In November, members of the WSS Board of Directors participated in a test of the electronic voting system that will be used in the 2005 WSS annual election this spring. Developed by our webmaster, Dan Jacobs, with direction from a committee of Board members, the new system will allow WSS members to cast their ballots from a website in much the same way that they do for ASA elections. As a cost-saving measure, electronic voting will replace paper ballots except in special circumstances. We hope that this easy-to-use system will increase election participation beyond the 10 to 15 percent turnout that we have experienced in recent years.

Return to top

Reminder About Electronic Newsletter

The WSS currently mails paper copies of its monthly newsletter to about half of the Society members while transmitting an electronic version (both as text and an attached pdf) to the remainder. Because of the high and growing cost of mailing a monthly newsletter (which exceeds a member's annual dues), we plan to eliminate the paper newsletter except as a special option once we resolve some remaining issues. We have asked that those who currently receive a paper newsletter and wish to continue receiving the paper version should notify the WSS secretary while those who want to switch to the electronic newsletter immediately should contact Vince Massimini and provide an email address and ASA member number. To date, only 12 current paper recipients have indicated that they wish to continue to receive the paper newsletter, but most of the rest have not yet asked to be switched to the electronic version. When the electronic version becomes the default option, we will mail paper copies only to those who have indicated that they wish to continue receiving the paper version. Electronic newsletters will be sent to the email addresses in our records. Members for whom we lack valid email addresses will not receive either the print or electronic version. If you do not currently receive the electronic version, you may need to contact Vince (svm@mitre.org) to ensure that your receipt of the newsletter is not interrupted.

Return to top

CDC/ATSDR Symposium Comes To Washington Area

Statistics and Public Health Policy
March 1-2, 2005
Bethesda, Maryland

Tenth Biennial CDC and ATSDR Symposium on Statistical Methods

Registration Fees

Before February 1, 2005:
Free for CSC employees
$200 for WSS affiliates
$150 for full-time students
$250 for others
   After February 1, 2005:
Free for CSC employees
$250 for WSS affiliates
$200 for full-time students
$300 for others

This marks the first time that the Atlanta-based Centers for Disease Control (CDC and the Agency for Toxic Substances and Disease Registry (ATSDR) will hold their Symposium in the Washington area, providing a great opportunity for Washington area statisticians and health policy analysts to get acquainted with recent research in an area of growing importance nationally. A one-day short course, "Dealing with Frame Error and Nonresponse in Surveys," taught by William Kalsbeek of the University of North Carolina, will precede the Symposium. As in 2003, the WSS is cosponsoring the Symposium, and WSS members will receive discounted registration fees for both the Symposium and the short course. There is an additional discount for students. Registration for both events is free to CDC employees.

The program for the Symposium will feature four invited paper sessions, an invited panel session, nine contributed paper sessions (three concurrently), and more than a dozen contributed posters. The first invited paper session, on "Sampling, Surveys and the Role of Observational Data in the Development of Public Health Policy," will include presentations by Robert Groves, Graham Kalton, and Nathaniel Schenker. The second invited paper session, "Determining the Effects of Public Health Interventions," will feature presentations by Monroe Sirken and William Kalsbeek. The third invited paper session, on "Bioterrorism, Biometrics and Security," will have presentations by Joe Scanlon, Martin Kulldorff, and Henry Rolka. The fourth invited paper session, "Evaluating Causation," will include presentations by Susan Murphy and Alan Zaslavsky. The invited panel session, organized by the Center for Discrete Mathematics and Computer Science (DIMACS) at Rutgers University, will feature a discussion of "Statistical Issues in Public Health Surveillance for Bioterrorism Using Multiple Data Streams." Participants will include Howard Burkom, Gregory F.Cooper, Martin Kulldorff, David Madigan, and Henry Rolka.

Please download the registration form (pdf) for both the Symposium and the short course. You are encouraged to register as soon as possible, as space is limited. Registration for the short course will be capped at 200 while registration for the Symposium will be limited to 400. Openings will be filled on a fist come, first served basis. Note, too, that registration fees will increase on February 1.

One Way Ground Transportation
to the Bethesda Marriott
Airport Super Shuttle Taxi
Reagan National (DCA) $25 $35
Dulles (IAD) $25 $45
Baltimore (BWI) $29 $65

The Bethesda Marriott site, located on Pooks Hill Road off of Rockville Pike, has extensive parking available. The hotel courtesy shuttle is no longer available. The hotel is accessible by Metro rail and you should visit www.wmata.com for fares from Reagan National Airport or other stations. Taxi cabs are available at Medical Center and the adjacent Grosvenor and Bethesda stations. The table to the right are estimates of one-way fares to the hotel from each of the three major airports for the Super Shuttle and taxis. Super Shuttle can be reserved by calling 1-800-BLUEVAN or go to www.supershuttle.com.

Return to top

First Interdisciplinary Symposium on Statistical Challenges and Opportunities in Electronic Commerce Research

Location and Date: May 22-23, 2005, Robert H. Smith School of Business, University of Maryland, College Park

Chairs: Wolfgang Jank and Galit Shmueli

Steering Committee: Stephen Fienberg (CMU), Donald Rubin (Harvard), Chrysanthos Dellarocas (University of Maryland), Erik Brynjolfsson (MIT), Joni Jones (University of South Florida)

Abstract Deadline: February 28, 2005

Website: www.smith.umd.edu/dit/statschallenges/

Contact: Wolfgang Jank wjank@rhsmith.umd.edu

Description: Electronic commerce produces an increasing amount of data-related questions and problems. Modern web-crawling technologies, which allow for a convenient collection of data from the Internet, result in huge databases. The openness of online marketplaces allows competitors to observe each others moves.

Traditional statistical models are not designed for the amount of data found on the web. They are also not suited to take into account the dynamics of online transactions as competitors react to each others moves in real time. This workshop focuses on identifying problems and research questions related to empirical research in electronic commerce by bringing together researchers from Information Systems, Statistics and related fields to help better understand how these various lines of work connect to one another and how, together, they can contribute to the modernization and enhancement of empirical research methods for electronic commerce and our digital society at large.

Return to top

From the ASA

Missing paternal demographics: A novel indicator for identifying high risk population of adverse pregnancy outcomes


One of every 6 United Status birth certificates contains no information on fathers. There might be important differences in the pregnancy outcomes between mothers with, versus those without partner information. The object of this study was to assess whether and to what extent outcomes in pregnant women who did not have partner information differ from those who had. We carried out a population-based retrospective cohort study based on the registry data in the United States for the period of 1995-1997, which was a matched multiple birth file (only twins were included in the current analysis). We divided the study subjects into three groups according to the availability of partner information: available, partly missing, and totally missing. We compared the distribution of maternal characteristics, maternal morbidity, labor and delivery complications, obstetric interventions, preterm birth, fetal growth restriction, low birth weight, congenital anomalies, fetal death, neonatal death, post-neonatal death, and neonatal morbidity among three study groups. There were 304466 twins included in our study. Mothers whose partner's information was partly missing and (especially) totally missing tended to be younger, of black race, unmarried, with less education, smoking cigarette during pregnancy, and with inadequate prenatal care. The rates of preterm birth, fetal growth restriction, low birth weight, Apgar score <7, fetal mortality, neonatal mortality, and post-neonatal mortality were significantly increased in mothers whose partner's information was partly or (especially) totally missing. Mothers whose partner's information was partly and (especially) totally missing are at higher risk of adverse pregnant outcomes, and clinicians and public health workers should be alerted to this important social factor.

Visualization and exploratory analysis of epidemiologic data using a novel space time information system


Recent years have seen an expansion in the use of Geographic Information Systems (GIS) in environmental health research. In this field GIS can be used to detect disease clustering, to analyze access to hospital emergency care, to predict environmental outbreaks, and to estimate exposure to toxic compounds. Despite these advances the inability of GIS to properly handle temporal information is increasingly recognised as a significant constraint. The effective representation and visualization of both spatial and temporal dimensions therefore is expected to significantly enhance our ability to undertake environmental health research using time-referenced geospatial data. Especially for diseases with long latency periods (such as cancer) the ability to represent, quantify and model individual exposure through time is a critical component of risk estimation. In response to this need a STIS a Space Time Information System has been developed to visualize and analyze objects simultaneously through space and time. In this paper we present a "first use" of a STIS in a case-control study of the relationship between arsenic exposure and bladder cancer in south eastern Michigan. Individual arsenic exposure is reconstructed by incorporating spatiotemporal data including residential mobility and drinking water habits. The unique contribution of the STIS is its ability to visualize and analyze residential histories over different temporal scales. Participant information is viewed and statistically analyzed using dynamic views in which values of an attribute change through time. These views include tables, graphs (such as histograms and scatterplots), and maps. In addition, these views can be linked and synchronized for complex data exploration using cartographic brushing, statistical brushing, and animation. The STIS provides new and powerful ways to visualize and analyze how individual exposure and associated environmental variables change through time. We expect to see innovative space-time methods being utilized in future environmental health research now that the successful "first use" of a STIS in exposure reconstruction has been accomplished.

Women - the Invisible Victims


Latin America and the Caribbean have up-to-date statistics on inflation, trade, GDP growth and other economic indicators. But there are few to no hard figures on violence against women, a problem that reportedly affects as many as four or five women out of 10 in the region. With a few exceptions like Mexico and Chile, there is a "blackout" surrounding the issue, Sonia Monta o, the head of the Economic Commission for Latin America and the Caribbean's (ECLAC) Women and Development Unit, told IPS. That is because "violence against women is hidden, and is not a priority for the public and political agendas, except on days like today," she said, referring to the fact that Nov. 25 is the International Day for the Elimination of Violence Against Women. According to the few official statistics available, based on different methodologies, around 70 women are killed as a result of domestic violence in Chile every year, while there are roughly 300 fatalities in Colombia. In the cities of Mexico City and Sao Paulo, Brazil, the annual average is 100 and 80, respectively. In Colombia, United Nations officials reported that the number of complaints of domestic violence against women rose from 50,000 a year to 60,000 between 2000 and 2003, although the number of court cases involving spousal abuse actually declined, from 8,000 to 4,000.

Number of Women Living with HIV/AIDS Increases in Each Region of the World


Close to half of 37.2 million adults living with HIV are women, according to new UNAIDS/WHO Report

A new report released today shows that the number of women living with HIV has risen in each region of the world over the past two years, with the steepest increases in East Asia, followed by Eastern Europe and Central Asia. In East Asia there was a 56% increase over the past two years, followed by Eastern Europe and Central Asia with 48%. Women are increasingly affected, now making up nearly half of the 37.2 million adults (aged 15-49) living with HIV worldwide. In sub-Saharan Africa, the worst-affected region, close to 60% of adults living with HIV are women - or 13.3 million. These latest findingd were published in AIDS Epidemic Update 2004, the annual report by the Joint United Nations Programme on HIV/AIDS (UNAIDS) and the World Health Organization (WHO). The joint report was released today in advance of World AIDS Day, commemorated worldwide on the first of December. Here is a description of how UNAIDS and WHO come by their figures here: http://www.unaids.org/Unaids/EN/Resources/Epidemiology/How+do+UN AIDS_WHO+arrive+at+estimates/

Return to top

SIGSTAT Topics for 2005

January 12, 2005 - Resampling Methods, Speaker: Peter Bruce

Peter Bruce, developer of Resampling Stats software, will give an introduction to resampling methods and their history. He will discuss William Gossett's original simulations that led to the development of the t-distribution and work by Fisher and Pittman (all in the early part of the last century), as well as the development of the bootstrap. The talk will include illustrations of the main resampling methods: the bootstrap and permutation procedures. This will be an overview, suited to those who are not experienced resamplers, not an in depth technical talk.

February 9, 2005 - Graphics with R, Speaker: Mike Fleming

The ease of making graphs in R is one of its strongest attractions. Some examples of making graphs were discussed in Part I. In this second part, the topics of making graphs which are annotated with equations, of putting multiple graphs on a single page, and of using polygons will be discussed.

March 9, 2005 - PROC UCM - Unobserved Component Models, Speaker: Charlie Hallahan

Unobserved Component Models are very general time series models that incorporate trends, seasonality, cycles, regression effects, and autoregressive effects. Trends and seasonality can be allowed to change randomly. UCMs can be considered as generalizations of ARIMA and smoothing models. The basics of UCMs will be covered as well as some examples. PROC UCM is new in SAS/ETS version 9.

April 20, 2005 - PROC ROBUSTREG - Robust Regression Method, Speaker: Charlie Hallahan

The experimental ROBUSTREG procedure in SAS/STAT version 9 provides resistant (stable) results in the presence of outliers by limiting the influence of outliers. ROBUSTREG provides four robust methods: M estimation by Huber (1973), Least Trimmed Squares (LTS) estimation by Rousseeuw (1984), S estimation by Rousseeuw and Yohai (1984), and MM estimation by Yohai (1987).

May 11, 2005 - PROC QLIM - Qualitative and Limited Dependent Variable Models, Speaker: Charlie Hallahan

The QLIM (Qualitative ans Limited dependent variable Model) procedure analyzes univariate and multivariate limited dependent variable models where dependent variables take discrete values or dependent variables are observed only in a limited range of values. This procedure includes logit, probit, tobit, and general simultaneous equations models. The simultaneous equations model can contain discrete choice and limited endogenous variables as well as continuous endogenous variables.

SIGSTAT is the Special Interest Group in Statistics for the CPCUG, the Capital PC User Group, and WINFORMS, the Washington Institute for Operations Research Service and Management Science.

All meetings are in Room S3031 (Food Safety and Nutrition Room), 1800 M St, NW from 12:30 to 1:30. Enter the South Tower and take the elevator to the 3rd floor to check in at the guard's desk.

First-time attendees should contact Charlie Hallahan, 202-694-5051, hallahan@ers.usda.gov and leave their name. Directions to the building and many links of statistical interest can be found at the SIGSTAT website, www.cpcug.org/user/sigstat/

Return to top

Note From The WSS NEWS Editor

Items for publication in the March WSS NEWS should be submitted no later than January 26, 2005. E-mail items to Michael Feil at michael.feil@usda.gov.

Return to top

Click here to see the WSS Board Listing (pdf)
Return to top