AMSTATNEWS the Membership Magazine of the American Statistical Association •

August 2012 • Issue #422 AMSTATNEWS The Membership Magazine of the American Statistical Association • http://magazine.amstat.org

ASA Announces Statistics Poster and Project Competition2012 Winners

ALSO: The ASA Fellow Award— Revisited

ASA Group Reviews Mathematical Education of Teachers II

features AMSTATNEWS 3 President’s Corner AUGUST 2012 • ISSUE #415 5 M ASA-SIA How Do I Write a Book Proposal? Executive Director Ron Wasserstein: [email protected] 6 SAMSI Announces 2013–14 Programs Associate Executive Director and Director of Operations 7 Statistics Without Borders Participates in Science Diplomacy Stephen Porzio: [email protected] Program Director of Science Policy Steve Pierson: [email protected] 7 Staff Spotlight

Director of Education 8 The ASA Fellow Award—Revisited [email protected] Rebecca Nichols 11 JASA Highlights Managing Editor New Approach to Analysis of Cancer Clinical Trials in June JASA Megan Murphy: [email protected] 13 Technometrics Highlights Production Coordinators/Graphic Designers Computer Experiments Motivate New Approach to Bayesian Melissa Muko: [email protected] Computation Kathryn Wright: [email protected] 14 Maintaining Quality in the Face of Rapid Program Expansion Publications Coordinator Val Nirala: [email protected] 16 Statistician's View Advertising Manager Misuse of Statistics a National Problem Claudine Donovan: [email protected]

Contributing Staff Members Naomi Friedman • Eric Sampson columns Amstat News welcomes news items and letters from readers on matters of interest to the association and the profession. Address correspondence to 17 175 Amstat News, Managing Editor, American Statistical Association, 732 North Washington Street, Alexandria VA 22314-1943 USA, or email amstat@ Broadening and Deepening Statistical Thinking: amstat.org. Items must be received by the first day of the preceding month Educating for 2014 and Beyond to ensure appearance in the next issue (for example, June 1 for the July issue). Material can be sent as a Microsoft Word document, PDF, or within an email. The ASA will celebrate its 175th anniversary in 2014. In preparation, column “175”— Articles will be edited for space. Accompanying artwork will be accepted written by members of the ASA’s 175th Anniversary Steering Committee and other ASA in graphics file formats only (.jpg, etc.), minimum 300 dpi. No material in members—will chronicle the theme chosen for the celebration, status of preparations, WordPerfect will be accepted. activities to take place, and, best yet, how you can get involved in propelling the ASA Amstat News (ISSN 0163-9617) is published monthly by the American toward its bicentennial. Statistical Association, 732 North Washington Street, Alexandria VA 22314- 1943 USA. Periodicals postage paid at Alexandria, Virginia, and additional mailing offices. POSTMASTER: Send address changes to Amstat News, 732 Contributing Editor North Washington Street, Alexandria VA 22314-1943 USA. Send Canadian Richard L. Scheaffer, professor emeritus at the University of Florida, has address changes to APC, PO Box 503, RPO West Beaver Creek, Rich Hill, ON L4B 4R6. Annual subscriptions are $50 per year for nonmembers. Amstat dedicated much of his academic work to the improvement of statistics News is the member publication of the ASA. For annual membership rates, education throughout the school and college curriculum. He directed see www.amstat.org/join or contact ASA Member Services at (888) 231-3473. the first ASA Quantitative Literacy program and the task force that developed the Advanced Placement Statistics Program. He also has served on American Statistical Association numerous education committees and advisory boards. 732 North Washington Street Scheaffer Alexandria, VA 22314–1943 USA (703) 684–1221 • FAX: (703) 684-2037

ASA GENERAL: [email protected] 19 MASTER'S NOTEBOOK ADDRESS CHANGES: [email protected] Experiences as a Biostatistician at an Academic Health Center AMSTAT EDITORIAL: [email protected] ADVERTISING: [email protected] This column is written for statisticians with master's degrees and highlights areas of employment that will benefit statisticians at the master's level. Comments and suggestions should WEBSITE: http://magazine.amstat.org be sent to Megan Murphy, Amstat News managing editor, at [email protected]. Printed in USA © 2012 American Statistical Association Contributing Editors Cindy Weng is a biostatistician II, who provides statistical consulting for pediatric clinical and translational research scholars and general pediatricians regarding statistical methods and the conduct of statisti-

Promoting the Practice and Profession of Statistics cal analyses in support of the Study Design and Biostatistics Center at the University of Utah. She earned an MPH degree with an emphasis in The American Statistical Association is the world’s largest biostatistics and epidemiology. community of statisticians. The ASA supports excellence in Weng the development, application, and dissemination of statistical James Grady has many years of research experience as the lead science through meetings, publications, membership services, biostatistician for numerous NIH-funded collaborative studies involv- education, accreditation, and advocacy. Our members serve in ing clinical and translational science in large-scale, population-based industry, government, and academia in more than 90 countries, studies. He is a regular reviewer for the National Institute of Dental advancing research and promoting sound statistical practice to and Craniofacial Research and currently serves as the director of inform public policy and improve human welfare. the biostatistics center for the Connecticut Institute of Clinical and Translational Science. Grady columns 21 STATtr@k Preparing for a Career as a Sports Statistician: Two Interviews with People in the Field STATtr@k is a column in Amstat News and a website geared toward people who are in a statistics program, recently graduated from a statistics program, or recently entered the job world. To read more articles like this one, visit the website at http://stattrak.amstat.org. If you have suggestions for future articles, or would like to submit an article, please email Megan Murphy, Amstat News managing editor, at [email protected].

Contributing Editors Jim Albert is professor of statistics in the department of mathematics and statistics at Bowling Green State University. His interests include Bayesian modeling, the statistical analysis of sports data, and statistical education. He is the editor of The American Statistician and writing a text on data analysis, probability, and statistics for prospective teachers. Albert Online Articles The following articles in this issue can be found online at http://magazine.amstat.org.

The Council of Professional Associations on Federal Statistics (COPAFS) acts as the advocate for departments the development and dissemination of high-quality 25 education federal statistics. Member organizations include First eCOTS Claims Success professional associations, businesses, research institutes, and others interested in federal statistics. ASA Announces 2012 Statistics Poster and Project Through COPAFS, members have an opportunity Competition Winners to review and have an effect on issues including timeliness, quality, confidentiality, and the relevance ASA Group Reviews Mathematical Education of Teachers II of data. To view the highlights of their June 1, 2012, meeting, visit http://magazine.amstat.org/ blog/2012/07/10/copafs-812. Minutes and copies of the overheads used by the presenters can be found at www.copafs.org.

The following ASA members passed away recently: Kesar Singh, Earl S. Pollack, and Genichi member news Taguchi. Visit http://magazine.amstat.org/blog/ 35 Awards and Deadlines category/membernews/amstatpeople/obits to read about their lives and careers. For a complete 37 People News listing of the most recent members’ obituaries, visit 41 Section • Chapter • Committee News Statisticians in the News at www.amstat.org/about/ statisticiansinthenews.cfm. 43 Professional Opportunities

Join the ASA Community http://community.amstat.org/Home

Like us on Facebook

2 amstat news august 2012 president's corner A Major Trend: The Rise of Undergraduate Programs in Statistics

Statistics majors in a class on statistical programming taught by Rebecca Ottesen in the department of statistics at Cal Poly San Luis Obispo

Photo courtesy of Rebecca Ottesen

uring August, college Minton, who pointed out that numbers of majors. Although students will flock to the lack of such programs dimin- their programs vary in emphasis, campus for the start of ished the recognition of statistics their stories paint an exciting pic- Dthe academic year. You might be as a discipline. However, it was ture for our profession. I’ll start surprised to know that record not until 15 years ago, when with two departments that have numbers are registering as statis- more high-school students began historically focused on training tics majors. This column taking Advanced Placement (AP) undergraduate majors. explains why—and what it Statistics, that the numbers of At Cal Poly San Luis Obispo, means for these students and college majors and minors in sta- Bob Smidt has seen a dramatic our profession. tistics began to climb. Today, that increase in the caliber and num- number is at an all-time high. ber of students applying as statis- What Has Changed? In 2001, our association tics majors—from five in the late Most statisticians of my vintage endorsed a set of curriculum 1990s to 60 in the last two years. learned about statistics as gradu- guidelines for BS degrees in sta- Smidt credits this to AP Statistics, ate students, having earned an tistics (www.amstat.org/education/ noting that “high-school students Robert Rodriguez undergraduate degree in anoth- pdfs/BS-curriculum.pdf). These who have had a good experience in er field, such as mathematics. guidelines make up one of many this class are more likely to apply.” Although bachelor’s degree pro- ways the ASA statistical educa- In response to this increase, grams in statistics existed as early tion community has strength- the Cal Poly program has evolved. as the 1940s, the number of ened the teaching of statistics at It now emphasizes writing and degrees granted remained small the undergraduate level. communication skills, statistical Join the ASA Community for decades. programming, senior projects, http://community.amstat.org/Home A landmark in the evolution What Trends Are and a capstone class that exposes of undergraduate statistics pro- Departments Seeing? students to working with clients. grams was a 1983 article in The To learn more, I contacted sta- Graduates are highly employable, American Statistician by Paul tistics departments with large often receiving multiple offers

august 2012 amstat news 3 from companies that value the I found yet another example success. Hon Ming Quek, who is statistical contributions they make of rapid expansion at Carnegie entering graduate school this fall, as analysts and programmers. Of Mellon University (CMU), cited practical applications, profi- these graduates, 30% to 50% go where the department of statistics ciency in statistical software, and to graduate school and 20% to has 150 majors, up from 40 in practice in writing reports. For 30% earn a PhD. 2006. Rebecca Nugent explains, Hannah Pileggi, now in graduate At Brigham Young University, “Exposure to statistics in AP cours- school, “the opportunity to partic- Del Scott reports an increase from es is one reason why 17-year-olds ipate in research was invaluable.” 139 majors six years ago to 240 now decide to major in statistics, this year. To accommodate fresh- along with a growing recognition Major and Minor men who have taken AP Statistics, that statistical skills are crucial in Conclusions his department developed a new today’s market.” In addition to a There are three points to glean introductory course that focuses major in statistics, CMU offers a from these stories. First, the on experimental design and is major that combines economics number of students majoring or taken concurrently with calculus. and statistics. minoring in statistics is soaring A strong background in math- The CMU program emphasiz- because of positive experiences in ematics is essential for majors con- es working on research problems AP Statistics courses. The word is sidering graduate school in statis- by using both classical methodol- out that statistics is a “must.” tics or biostatistics. ogy and newer, computationally Second, successful under- The department now offers intensive methods. Students par- graduate programs anticipate and three undergraduate degrees with ticipate in design competitions, deliver the training their students distinct goals: statistical science work in groups, and are required will require when they move into as preparation for graduate work, to write reports and give presen- employment or graduate studies, actuarial science as preparation tations. Thirty to forty percent of whether in statistics or another for an actuarial career, and applied students go to graduate school, field. For business-minded stu- statistics/analytics as preparation and many find employment in dents, graduate programs in ana- for corporate environments in banking, financial services, and lytics are an increasingly attrac- which strong programming skills insurance. tive option. are in demand. Third, the growing on-the- What Are Graduates job contributions of majors and What About Majors Saying? minors are making statistics in PhD-Granting For a student perspective on more visible and more valued by Departments? these trends, I asked CMU grad- employers, colleagues, and society. Growth is also the story in uates why they decided to major In view of the strong demand the department of statistics at in statistics. Nora Albert, cur- for skills needed to analyze Big Purdue, where the number of rently at the U.S. Census Bureau, Data—in business, government, majors has doubled from 200 answered with one word: versa- and scientific research—we should to 400 in the past five years. tility. She elaborated, “I knew update the ASA curriculum guide- Rebecca Doerge notes that the that by majoring in statistics I lines for BS degrees to reflect these undergraduate program in sta- would have a lens through which needs. Defining appropriate statis- tistics emphasizes a strong theo- I could view any industry.” tical training is one way our asso- retical foundation and compu- Daniel Frank, now a quanti- ciation can take a leading role in tational skills. Most graduates tative finance analyst at Bank of the arena of Big Data. either head to graduate school America, replied, “I decided that Wherever our majors and or work in pharmaceutical com- the possibilities with a solid statis- minors land after graduation, panies and other businesses with tics background are virtually lim- we should make sure they see strong data-related needs. itless. Every field, from in-depth themselves as members of our At Purdue, another option for science to industry-focused jobs, profession. The best place for undergraduates is a program in requires statisticians in one way or this to begin is the classroom. actuarial science offered jointly another. The beauty of this degree “As a teacher,” says Nugent, “the by the departments of statistics is that you can move anywhere most fun is watching students and mathematics. An emerging and do anything because your get excited about what they are career opportunity for these stu- skills will always be in demand.” doing and turning into members dents is health care engineering, I also asked students what of the statistics community.” the specialization of Purdue’s aspects of their undergraduate Regenstrief Center. studies best prepared them for

4 amstat news august 2012 ASA–SIAM Titles from the How Do I Write a Book ASA-SIAM SERIES on statistics and applied probability Proposal?

ou’ve decided to write a book, exercises, or new material not yet or you already have a draft, and published elsewhere? Will you be now you need to find a pub- offering supplementary online Ylisher. Between you and a publishing material such as data sets, code for contract lies the book proposal form. particular software, or slides? Every publisher has one, and most of the information included is standard. Purpose and Audience Here’s what you can expect to be What will the main purpose of asked when proposing a book to the your book be? Is it a monograph ASA-SIAM Series on Statistics and for researchers and practitioners? ASA & SIAM MEMBERS Applied Probability. Is it an introductory book for GET 30% OFF LIST PRICES! people seeking to learn about a The Basics new area? Is it a textbook that To start, you’ll need to supply the full provides exercises and problems Data Clustering: for students to solve as home- Theory, Algorithms, names and addresses of all authors, editors, or contributors, along with work or for self-study purposes? and Applications the title you’re planning to use. Don’t If it’s a textbook, for what level Guojun Gan, Chaoqun Ma, worry if you haven’t entirely settled of course is it intended, and and Jianhong Wu on a title yet. Provide one that’s what are the names of courses 2007 · xviii + 466 pages · Soft · ISBN 978-0-898716-23-8 List $114.00 · ASA/SIAM Member $79.80 · Code SA20 workable, and you can always revise for which it would be reasonable it with help from reviewer and/or to expect instructors to use it? The Structural Representation of publisher suggestions. Proximity Matrices with MATLAB Competition Lawrence Hubert, Phipps Arabie, Production Information To help us get an idea of the place and Jacqueline Meulman Next, you’ll need to estimate how your book would have in the mar- 2006 · xvi + 214 pages · Soft · ISBN 978-0-898716-07-8 many pages the finished book will be ket, you’ll be asked to mention List $85.50 · ASA/SIAM Member $59.85 · Code SA19 and note whether you’ll need color competing books from other pub- Experimental Design art so we can get an idea of wheth- lishers and describe how your book for Formulation* er we can produce your book at an differs from each. We’ll also ask affordable price for customers. We’ll you to list titles from our catalog Wendell F. Smith 2005 · xx + 367 pages · Soft · ISBN 978-0-898715-80-4 also want to know when you expect that are complementary to your List $113.50 · ASA/SIAM Member $79.45 · Code SA15 to finish the manuscript, so we can book and which could be mar- budget the book for the appropriate keted with it. Recurrent Events Data Analysis publication year. The book proposal form for Product Repairs, Disease encapsulates critical information Recurrences, and Other Content about your book project and Applications Now, it’s time to talk about the tech- conveys it clearly to publishing Wayne B. Nelson nical content of the book. Briefly staff, editorial board members, 2002 · xii + 151 pages · Hard · ISBN 978-0-898715-22-4 describe the focus of the book and and reviewers. It also can help List $105.50 · ASA/SIAM Member $73.85 · Code SA10 your motivation for writing it. How you, as the author, to think more do you expect people will use it, and deeply about the specifics of the Multivariate Statistical Process book you’d like to write. Taking Control with Industrial why is there a need for it? What specific topics have you chosen to address? the time to answer the questions Applications If there are related topics that you’ve thoroughly and thoughtfully is Robert L. Mason and John C. Young purposely omitted, explain why. an important step in finding the 2001 · xiii + 263 pages · Hard · ISBN 978-0-898714-96-8 List $87.00 · ASA/SIAM Member $60.90 · Code SA09 Be prepared to talk succinctly about right home for your project. Includes CD-ROM how your book will differ from other If you’re interested in submit- *A portion of the royalties from the sale of this book are books already on the market in the ting a proposal to the ASA-SIAM contributed to the SIAM student travel fund. same topic area. Series on Statistics and Applied www.siam.org/catalog For example, have you included Probability, please contact series www.siam.org / www.amstat.org full proofs, case studies, extensive acquisitions editor Sara Murphy 1-800-447-SIAM at [email protected]. n +1-215-382-9800 11/11 august 2012 amstat news 5 SAMSI Announces 2013–14 Programs structures that approximate or encapsulate given high-dimensional data. Cutting-edge methods of dimension reduction will be brought together from probability and statistics, geometry, topology, and computer science. These techniques include vari- able selection, graphical modeling, classification, dimension reduction in matrix estimation, empirical processes, and manifold learning. Working groups will include theoretical discussions of these ichard Smith, director of the Statistical tools and applications to image and signal analy- and Applied Mathematical Sciences sis, graphs and networks, genetics and genomics, Institute (SAMSI), recently announced dynamical systems, and machine learning. Rthat the two programs for the 2013–2014 year Program leaders for LDHD include Florentina will be Computational Methods in Social Science Bunea of Cornell, Peter Hoff of the University of and Low-Dimensional Structure in High- Washington, Chris Holmes of Oxford University, Dimensional Systems. Peter Kim of Guelph, Vladimir Koltchinskii of Like many areas of research, the social sciences Georgia Tech, John Lafferty of The University have experienced a data explosion. Social scien- of Chicago, Gilad Lerman of the University of tists are examining statistical and computational Minnesota, Sara van de Geer of ETH Zurich, methodology more these days for handling social Marten Wegkamp of Cornell, and Bin Yu of science data sets. Many statisticians and applied Berkeley. Ezra Miller is the directorate liaison. mathematicians also are focusing on social scienc- There are several opportunities for people to es in applications of their work, including looking participate in either of these programs. Financial at social networks and causal inference. support is available for visiting researchers to be Computational Methods in Social Science residents at SAMSI for one month to one year. will focus on social networks, agent-based mod- Young researchers have special opportunities to els, and new methodology for censuses and sur- participate that typically have a one-year appoint- veys. Program leaders include Robert Axtell of ment. Several postdoctoral positions also will be George Mason University, Elena Erosheva of funded for each program. the University of Washington, Doyne Farmer of Workshops and working groups give many the Oxford University and the Santa Fe Institute, Steve opportunity to collaborate with others on research Fienberg of Carnegie Mellon University, Krista Gile projects and network with their peers. Dedicated of the University of Massachusetts-Amherst, Mark workshops will allow graduate and upper-level Handcock of the University of California at Los undergraduate students to learn about the latest Angeles, and Tian Zheng of Columbia University. research and applications in the statistical and Smith is the directorate liaison. mathematical sciences. All involved researchers Low-Dimensional Structure in High- will receive chances to broaden their interests and Dimensional Systems (LDHD) is devoted to the skill sets, participate in cutting-edge interdisci- development of methodological, theoretical, and plinary projects, and make new connections. New computational treatment of high-dimensional researchers and members of under-represented mathematical and statistical models. Possibly lim- groups are especially encouraged to participate. ited amounts of available data pose added challeng- For more information or to apply, visit www. es in high dimensions. This program will address samsi.info. n these challenges by focusing on low-dimensional MAKE THE MOST OF YOUR ASA MEMBERSHIP Visit the ASA Members Only site: www.amstat.org/membersonly

6 amstat news august 2012 STAFF SPOTLIGHT Naomi Friedman

ello, my name is Naomi Mason University with a degree Friedman, and I am the in communications and recently ASA’s newest meetings earned my Certified Meeting Hplanner. As a member of the meet- Professional designation from the ings department, I will be working Convention Industry Council. on the annual JSM and planning I’ve been in the meetings additional conferences throughout industry for several years, start- the year. I am excited to be part of ing my career on the hotel side the team and wanted to take a before making the shift to plan- moment to introduce myself. ning. Most recently, I was with Although I grew up in the the Hearth, Patio, and Barbecue Washington, DC, area, I original- Association, where I worked on ly hail from New York (City) and meetings such as the HPBExpo— consider myself a New Yorker at an annual trade show—and with heart. I graduated from George their education program. Let’s face it, how can you go wrong with fireplaces and grills? When not at work, I can usually be found spending time with Statistics Without Borders my two crazy dogs, wandering Participates in Science around town, or planning my next trip with friends. In addi- Diplomacy Program tion to making my way around the United States, I love collecting stamps on my passport. tatistics Without Borders (SWB), the ASA all-volunteer So far, they include the United group, will provide pro-bono instructors for the Pyongyang Kingdom, France, Greece, Israel, Summer Institute (PSI) in Survey Science and Quantitative Turkey, and Canada. I can’t wait MethodologyS in North Korea this summer. The PSI is an inten- to see where I will go next! sive, international teaching program at the Pyongyang University I look forward to meet- of Science and Technology (PUST), the first and only private and ing many of you at JSM in San international university, which was launched in North Korea in Diego and this fall during H2R the fall of 2010. 2012 in New Orleans! If you have PSI is the first program of its kind in North Korea, which any questions or need assistance, has now approved visas for 15 visiting faculty. PSI students are please feel free to contact me at expected to be upper-level undergraduates and graduate students [email protected]. n whose academic credentials are comparable to those students in Ivy plus schools. The summer institute is jointly administered by the International Strategy and Reconciliation Foundation (ISRF), PUST, and SWB. The ISRF is a nonprofit organization recognized by the U.S. Treasury and Commerce departments for its humanitarian and educational programs in the Democratic People’s Republic of Korea. The PSI will begin as a three-year The managing editor of Amstat News is searching for pilot program that will take place from July 2012 to July 2014. ASA members who are willing to put themselves in the For four weeks, SWB members will provide both fundamen- spotlight and write a brief article about their life, to be tal and advanced courses in sampling, statistics, survey methods, published in an upcoming issue. census methods, questionnaire design, and computer-assisted The article should be 1,200 or fewer words and contain data collection and analysis. The courses were modeled after professional and personal information. Please include a those of the 65-year-old Michigan Summer Institute, a renowned photo of yourself and email it to Amstat News Managing international survey training program. n Editor Megan Murphy at [email protected].

august 2012 amstat news 7 The ASA Fellow Award—Revisited Robert Starbuck

he 2012 ASA Fellow awards were presented The percentages of Fellows awarded by employ- recently at JSM in San Diego, California. ment sector relative to the percentages of ASA Here, I present a brief update to previous membership by sector are shown in Figure 2. Tarticles about this award appearing in Amstat News. Employment Sector In the range of years shown, the percentages of ASA members by employment sector have remained relatively stable: 42% Academe, 47% Business/ Industry, and 11% Government. The counts of ASA Fellow awards given by employment sector since 2004 are presented in Table 1 and Figure 1.

Table 1—Counts and Percentages of ASA Fellow Awards by Employment Sector

Employment Sector Year Academe Business/Industry Government Total Figure 2. Percentages of Fellows awarded by employment sector relative to the percentages of ASA mem- 2004 36 (64.3) 11 (19.6) 9 (16.1) 56 bership by sector 2005 38 (67.9) 8 (14.3) 10 (17.9) 56 2006 50 (83.3) 5 ( 8.3) 5 ( 8.3) 60 2007 37 (62.7) 11 (18.6) 11 (18.6) 59 2008 32 (60.4) 13 (24.5) 8 (15.1) 53 The counts and percentages of Fellow nomina- 2009 36 (63.2) 15 (26.3) 6 (10.5) 57 tions by employment sector are shown in Table 2. 2010 43 (81.1) 5 ( 9.4) 5 ( 9.4) 53 The total number of nominations was the lowest 2011 45 (77.6) 8 (13.8) 5 ( 8.6) 58 this year since 2004 and was noticeably lower than 2012 37 (77.1) 7 (14.6) 4 ( 8.3) 48 last year in all employment sectors. The number of nominations from government was considerably lower this year than in the other eight years included in this report, and the number of nominations from business/industry was the second-lowest in the nine years included in this report.

Table 2—Counts and Percentages of ASA Fellows Nominations by Employment Sector

Employment Sector Year Academe Business/Industry Government Total 2004 44 (58.7) 16 (21.3) 15 (20.0) 75 2005 51 (57.3) 22 (24.7) 16 (18.0) 89 2006 81 (73.0) 19 (17.1) 11 ( 9.9) 111 2007 79 (65.8) 22 (18.3) 19 (15.8) 120 2008 60 (64.5) 18 (19.4) 15 (16.1) 93 Figure 1. Counts of ASA Fellow awards given by 2009 59 (62.1) 23 (24.2) 13 (13.7) 95 employment sector since 2004 2010 71 (72.4) 13 (13.3) 14 (14.3) 98 2011 76 (72.4) 18 (17.1) 11 (10.5) 105 2012 62 (75.6) 14 (17.1) 6 ( 7.3) 82

8 amstat news august 2012 So, how have the nominations fared in each of Gender the employment sectors? As shown in Table 3 and The membership of the ASA has changed signifi- Figure 3, nominations submitted this year from the cantly in the percentages of females and males, as business/industry and government sectors fared illustrated in Table 4 and Figure 4. This table looks better than those in the previous two years, and at the current ASA membership and subsets that those from academe were similar to those submit- joined the ASA in ranges of previous years. ted in the previous three years. Assuming the number of departures from ASA membership has been proportional to gender (i.e., females and males are equally likely to dis- continue or retain ASA membership), there has Table 3—Percentages of Successful ASA Fellows been a noticeable increase in the percentage of Nominations by Employment Sector female members. Employment Sector Year Academe Business/Industry Government Table 4—Percentages of ASA Membership by Gender 2004 81.8 68.8 60.0 Current ASA Members Female Male 2005 74.5 36.4 62.5 Joined ASA 1985 17 83 2006 61.7 26.3 45.5 ≤ Joined ASA 1990 20 80 2007 46.8 50.0 57.9 ≤ Joined ASA 1995 22 78 2008 53.3 72.2 53.3 ≤ Joined ASA 2000 23 77 2009 61.0 65.2 46.2 ≤ All 32 68 2010 60.6 38.5 35.7 2011 59.2 44.4 45.5 2012 59.7 50.0 66.7

Figure 4. Current ASA members by gender

The ASA Fellow award is almost always given to Figure 3. Percentages of successful ASA Fellows nominations by employment sector ASA members who have reached mid-career, and as such, the gender percentages of ASA membership that are appropriate reference points for years 2004–2012 are those reflecting members who joined the ASA on or before 1994–2000. Thus, the relevant reference percentages for females for Fellow nominations and awards are in the 20% to 23% range.

august 2012 amstat news 9 The counts and percentages by gender of ASA The percentages of female Fellow awardees have Fellow nominations in 2004–2012 are presented in been generally reflective of the reference percent- Table 5. ages. As shown in Table 7 and Figure 6, the decline in the percentage of female Fellow awardees in 2009 Table 5—Counts and Percentages of ASA Fellow was due simply to the lower percentage of female Nominations by Gender Fellow nominations. For nominees, the average chance of success is somewhat higher for females. Year Female Male Total 2004 14 (18.7) 61 (81.3) 75 Table 7—Percent of Successful ASA Fellow Nominations by Gender 2005 24 (27.0) 65 (73.0) 89

2006 25 (22.5) 86 (77.5) 111 Year Female Male 2007 22 (18.3) 98 (81.7) 120 2004 92.9 70.5 2008 16 (17.2) 77 (82.8) 93 2005 62.5 63.1 2009 12 (12.6) 83 (87.4) 95 2006 44.0 57.0 2010 24 (24.5) 74 (75.5) 98 2007 63.6 45.9 2011 19 (18.1) 86 (81.9) 105 2008 68.8 54.5 2012 19 (23.2) 63 (76.8) 82 2009 58.3 60.2 2010 70.8 48.6 2011 63.2 53.5 The percentage of female nominees in 2012 was consistent with the reference percentages. 2012 68.4 55.6 The counts and percentages by gender of ASA Mean 65.8 56.5 Fellow awards in 2004–2012 are presented in Table 6, and the percentages are presented in Figure 5.

Table 6—Counts and Percentages of ASA Fellow Awards by Gender

Year Female Male Total 2004 13 (23.2) 43 (76.8) 56 2005 15 (26.8) 41 (73.2) 56 2006 11 (18.3) 49 (81.7) 60 2007 14 (23.7) 45 (76.3) 59 2008 11 (20.8) 42 (79.2) 53 2009 7 (12.3) 50 (87.7) 57 2010 17 (32.1) 36 (67.9) 53 Figure 6. Percent of successful ASA Fellows nomina- 2011 12 (20.7) 46 (79.3) 58 tions by gender 2012 13 (27.1) 35 (72.9) 48 Conclusion The number of Fellow nominations from a given employment sector or gender is a key factor in deter- mining the number of awards from that sector. The other obvious key factor is the quality of the nominations. To increase the number of Fellow awards in an employment sector or gender to achieve parity in the percentages of awards relative to the percentages of ASA membership (see figures 2 and 5), the number of nominations from that sector or gender needs to increase, and these nominations need to be of good quality. The ASA Fellow award is a significant recognition of contributions to the statistics profession, and one that should reflect the constituency of the ASA membership. If you or others you know are deserving of Figure 5. Percentages by gender of ASA Fellow awards in 2004–2012 this award, please participate in and encourage others to participate in the award nomination process. n 10 amstat news august 2012 JASA HIGHLIGHTS New Approach to Analysis of Cancer Clinical Trials in June JASA

he June 2012 issue of the Journal of the Mid-study changes of a different type are consid- American Statistical Association features an ered in a second JASA A&CS paper. Lane Burgette article describing novel strategies for the analy- and Jerome Reiter consider the case when the tech- Tsis of sequentially randomized clinical trials in the nique used to measure important study variables is Applications and Case Studies (A&CS) section, along changed in the middle of the study. The Healthy with comments from experts in that research area. Pregnancy Healthy Baby Study is an observational Other A&CS articles include contributions in ecolo- study focused on identifying causes of adverse birth gy, economics, and brain imaging. The Theory and outcomes. Blood samples from expectant mothers Methods section contains important contributions on were sent to a laboratory to measure exposure to model selection in high-dimensional problems and several pollutants, including lead. Midway through the theory behind portfolio selection in finance, as the study, the investigators changed from one lab to well as many other contributions. another because the second could provide more fine- ly quantitated exposures. When investigators began Applications and Case Studies to analyze the data, they noticed there were differ- Cancer therapy is usually conducted in stages. A physi- ences in the distribution of the measurements from cian chooses an initial treatment based on the patient’s the two labs; the differences were larger than could disease severity and perhaps other patient characteris- be explained by chance. Unfortunately, no samples tics. That treatment will be continued if successful, were analyzed at both labs, which makes calibrating but may be altered if the initial treatment does not the two sets of measurements challenging. lead to a favorable response or if it has intolerable side In “Nonparametric Bayesian Multiple effects. Multistage strategies are known as dynamic Imputation for Missing Data Due to Mid-Study treatment regimes, or adaptive treatment strategies. In Switching of Measurement Methods,” Burgette and “Evaluation of Viable Dynamic Treatment Regimes in Reiter develop a set of methods that allow imput- a Sequentially Randomized Trial of Advanced Prostate ing the later measurements for those measured Cancer,” authors Lu Wang, Andre Rotnitzky, only using the earlier protocol. They treat the later Xihong Lin, Randall Milliken, and Peter Thall pres- measurements as “missing” data. Their imputation ent a new statistical analysis of data from a two-stage approach for these missing data is based on the clinical trial of advanced prostate cancer treatments. reasonable assumption that the two labs accurately The initial analysis of the data in 2007 identified the rank samples, although they score them on differ- best treatments in stage one and the best treatments ent scales. The rank preservation assumption (i.e., in stage two separately, but did not directly address the same ranking would be obtained in each lab) the question of the best two-stage dynamic treat- leads to three imputation strategies. The strategies ment regimes. This reanalysis carefully defines the work well in simulation and are then used to esti- full set of viable treatment regimes, constructs a new mate quantile regressions relating the distribution of compound endpoint incorporating both treatment infant birth weights to the level of maternal expo- effectiveness and treatment toxicity, and addresses sure to environmental contaminants. the patients who did not comply with their assigned viable treatment regime. Theory and Methods Estimation is carried out using an inverse prob- Model selection in linear regression models is an ability of censoring weighted estimator. Two com- important problem with an extensive literature. ments provide additional insight into the analysis The problem has become increasingly important in of sequential randomized trials. Daniel Almirall, modern high-dimensional problems in which there Daniel Lizotte, and Susan Murphy contribute are many predictor variables from which to choose. important points regarding the design of multistage Many Bayesian and frequentist approaches have trials and alternative approaches to accommodat- been proposed. The standard assumptions incorpo- ing multiple outcomes in the data analysis. Paul rated in Bayesian model selection procedures have Chaffee and Mark van der Laan describe an alter- resulted in procedures that are not competitive with native inference approach for such studies: targeted commonly used penalized likelihood methods. minimum-loss–based estimation. The article and In “Bayesian Model Selection in High-Dimensional comments shed light on an important class of trials Settings,” Valen Johnson and David Rossell propose for evaluating cancer treatments.

august 2012 amstat news 11 modifications of the Bayesian methods by relying on nonlocal prior densities on model parameters to improve performance. The authors show that model selection procedures based on nonlocal prior densities reliably find the correct model in linear model settings when the number of possible covariates is bounded by the number of observations. The key to the improved performance is that Bayesian approaches based on standard local prior densities (which are nonzero for null values) assign positive probabilities to incorrect models that include some coefficients that are truly zero. The nonlocal prior densities avoid this behavior. In addition to consistently identifying the true model, the proposed Bayesian procedures based on nonlocal prior distributions provide accurate estimates of the posterior probability that each identified model is correct. Simulation studies demonstrate that the new model selection procedures perform as well as or better than commonly used penalized likelihood methods in a range of simulation settings. A different type of “selection” problem is the focus of another Theory and Methods paper, “Portfolio Selection with Gross-Exposure Constraints” by Jianqing Fan, Jinjjin Zhang, and Ke Yu. Portfolio selection and optimization have been fundamental problems in finance ever since the development of Markowitz’s portfolio theory, which proposes choosing one’s portfolio (the amount of each asset to pur- chase/sell) to maximize the expected return of the portfolio subject to a constraint on the variability of returns. Though Markowitz’s proposal represented a significant theoretical breakthrough, it turns out there are practical problems in applying the theory because the optimal Markowitz portfolio is sensitive to errors in estimating/forecasting the expected returns and the covariance matrix of the returns. The problems are especially important in situations where there are a large number of assets from which to choose. Fan, Zhang, and Yu develop an extension of the usual Markowitz formulation by introducing a constraint on the total amount of investment at risk (which includes money used to buy assets and also the risk in short sales of assets). The so-called “gross exposure constraint” allows the authors to investigate the circumstances under which constrained portfolio selection has good properties relative to the uncon- strained solution. They show that for a range of values of the constraint parameter, the constrained approach leads to optimal portfolios that have smaller actual risk than the globally optimal portfolios while at the same time allowing for more accurate estimation of the risk. There are many other informative articles in both sections of the June issue, as well as a set of book reviews. The full list of articles and a list of the books under review can be viewed at http://amstat. tandfonline.com. n

12 amstat news august 2012 TECHNOMETRICS HIGHLIGHTS Computer Experiments Motivate New Approach to Bayesian Computation Hugh Chipman, Technometrics Editor

imulation-based techniques for Bayesian com- to motivate the problem of identification of change- putation have seen widespread application over points in a time series. In that problem, the removal the last two decades. Although flexible, they can of systematic shifts due to changes in measurement beS time consuming in some problems. This is espe- systems is an important step before the series can be cially true when the likelihood may be expensive to used. Prior information is used in a Bayesian analysis, evaluate. Approximation methods, such as variational with a closed-form form of a BIC-like expression for Bayesian inference, have shown promise, but may not identification of the change-points. provide sufficient accuracy. In “Bayesian Computation The remainder of the issue is devoted to papers Using Design of Experiments-Based Interpolation involving process monitoring. In “Monitoring Technique,” V. Roshan Joseph develops a new Warranty Claims with Cusums,” by Jerald F. Lawless, approximation method for posterior inference that is Martin Crowder, and Ker-Ai Lee, the monitoring of quick, accurate, and adaptive. reliability is considered. Using data from warranty By leveraging ideas from the design and analy- claims on North American vehicles, the paper devel- sis of computer experiments, a kriging model is fit ops practical monitoring methods designed to allow to the posterior distribution, giving highly accu- changes in claim rates to be detected in as timely a rate approximations in a variety of applications. manner as possible. The techniques are general, easy to implement, and In “Posterior Distribution Charts: A Bayesian applicable to many complex Bayesian problems. By Approach for Graphically Exploring a Process Mean,” using sequential design of experiments, the approxi- Daniel W. Apley develops a Bayesian approach for mation can be further improved though adaptive monitoring and graphically exploring a process mean addition of basis functions. and informing decisions related to process adjustment. The article is accompanied by several excellent dis- Observations are represented as a process mean plus a cussions and a rejoinder by the authors, which explore random error term, and the mean process can follow a number of extensions of the method. The discus- any Markov model. This includes a mean that wan- sants are Björn Bornkamp, Tirthankar Dasgupta, ders slowly, that is constant over periods of time with Xiao-Li Meng, Herbert K. H. Lee, John T. occasional random jumps, or combinations thereof. Ormerod, M. P. Wand, David M. Steinberg, and This exploratory approach is illustrated using an Bradley Jones. The directions explored in the dis- example from automobile body assembly. cussions suggest DoIT will provide fertile ground for Matthias Tan and Jianjun Shi also develop further research into the use of tools from computer Bayesian methods for process monitoring in their experiments for Bayesian computation. paper, “A Bayesian Approach for Interpreting Mean The remainder of the issue includes articles about Shifts in Multivariate Quality Control.” The focus here reliability, process monitoring, time series, and is on identification of the important factors that led to compliance testing. This latter topic is studied in an out-of-control signal when monitoring multivariate “Compliance Testing for Random Effects Models with quality characteristics. An adaptation of Bayesian vari- Joint Acceptance Criteria,” by Crystal D. Linkletter, able selection methods from linear regression provides Pritam Ranjan, C. Devon Lin, Derek R. Bingham, both a probabilistic framework for diagnosis and a Richard A. Lockhart, Thomas M. Loughin, and mechanism for incorporating expert knowledge. William A. Brenneman. For consumer protection, The issue closes with “Outlier Detection in many governments perform random inspections on Functional Observations with Applications to Profile goods sold by weight or volume to ensure consistency Monitoring,” by Guan Yu, Changliang Zou, and between actual and labeled net contents. Motivated Zhaojun Wang. Monitoring of profile data (also by a problem from a real manufacturing process, the known as functional data) is becoming increasingly paper provides an approximation for the probability of common with available sensor technology. To use such sample acceptance that is applicable for processes with data for profile monitoring, outliers must be removed one or more known sources of variation via a random first. This paper proposes a new testing procedure effects model. This approach also allows the assess- based on functional principal component analysis. ment of the sampling scheme of the items. After deriving the appropriate null distributions, the In “An improved Bayesian Information Criterion test statistic is demonstrated in a real-data example for Multiple Change-Point Models,” Alexis Hannart from a manufacturing process. n and Philippe Naveau use a meteorological application august 2012 amstat news 13 Maintaining Quality in the Face of Rapid Program Expansion Rebecca Nugent, Carnegie Mellon University

y, how times have changed. Many of today’s experimental design, and statistical computing senior faculty and leading industry statisti- help students gain programming skills and “data cians probably didn’t take nonintroductory sense” before the advanced methodology classes. statisticsM courses until graduate school. Now, incoming Advanced electives are offered in stochastic process- freshmen have not only placed out of introductory es and “special topics.” The special topics courses statistics and calculus, but they’ve already selected sta- are offered every semester with rotating topics that tistics as their major, applied for their first data analyt- include statistical methods in epidemiology, sta- ics summer internship, and declared their undying tistical learning, multivariate analysis, multilevel allegiance to big data. What on Earth do we do with and hierarchical models, and data mining. Given this new generation? our flexible—but stable—rotation design, most of In the department of statistics at Carnegie Mellon our majors take at least two or three special topics University, we’ve been asking ourselves this very ques- courses. Weekly seminars are occasionally offered tion. CMU has an undergraduate enrollment of on topics such as statistics in sports or statistics and nearly 6,000 students; recently, the department of the law. Qualified students also are allowed to take statistics has been teaching about 1,000 undergradu- first-year graduate courses, most commonly statisti- ates per semester from across the campus. Even with cal inference or methodological courses like parallel our high service demand, one pressing problem is the computing or machine learning. rapid expansion of the undergraduate statistics program. In the last five years, the number of statistics Undergraduate Research and economics-statistics majors has quadrupled to Carnegie Mellon University is fully committed to around 150 students. Given the department’s focus undergraduate research; the statistics program is no on groundbreaking research and strong commitment exception. In response to the overwhelming growth to high-quality vertically integrated education, we of our capstone methodology courses, we created are faced with redesigning our curriculum to satisfy an invitation-only research course that pairs small demand while maintaining our high standard of sta- groups of students with faculty clients across cam- tistical training for both industry and graduate school. pus (and occasionally industry) for a semester-long Not an easy task. research project. Accepted students learn the basic principles of research, including literature reviews, Program Overview methodology comparisons and critiques, and pre- Our program combines a solid theoretical back- sentation of results. Each team presents their final ground with thorough exposure to methodology, project in an oral defense, a poster session, and a both traditional and modern. After introductory written report. courses and mathematics prerequisites, majors take While enrollment tends to favor the more quali- a year sequence in probability and statistical infer- fied seniors, we encourage promising junior statistics ence. Doing well prepares them for the subsequent majors to apply, knowing that participation in this methodology sequence—a thorough treatment of course will prepare them for a senior honors thesis the linear regression followed by a semester of advanced following year. Our thesis students meet as a group data analysis methods including bootstrap simula- during the year to give updates on their research and tion, kernel smoothing, splines, generalized linear receive feedback from faculty and peers. We have and additive models, causality, and Markov mod- found that, rather than isolating the student and els. The enrollment in this sequence has increased adviser, the group dynamic has increased the quality by 400% over the past five years, not solely due to of the completed theses. the majors. As other programs such as mathemati- Research projects are also available to other cal sciences, business administration, and computer interested students. The department has a long science embrace the importance of understanding history of supporting undergraduate research modern statistical methods, more and more stu- internally and with outside grant support. Each dents are showing up in our classrooms. semester, there are four or five students working Lower-level electives in statistical graphics on National Science Foundation (NSF)–sponsored and visualization, sampling and survey methods, research projects, typically as members of research

14 amstat news august 2012 groups that include faculty and graduate students. to integrate the statistical analysis and the research Our summer programs have included participa- problem at hand. tion in Morehouse College’s Project IMHOTEP, Given the large number of students in both the a program designed to increase minority partici- program and our classes, building a sense of com- pation in statistics (also a feature of our current munity has never been more important. We pride NSF Research Training Group (RTG) grant), and ourselves on providing personal feedback and a this year’s group of 13 undergraduates sponsored welcoming, nurturing environment. However, by both the RTG and our NSF Census Research as numbers grow, the feasibility of this approach Node grant. If students do not qualify for NSF fades. Students who struggle often feel isolated in funding (due to citizenship), they are offered large classes; students who chafe under the necessar- course credit or department-funded stipends. ily stricter large class framework feel stifled. Oddly Independent studies are also available. In all cases, enough, we have found that fostering friendly com- students are expected to present their research to petition has been enormously helpful in building the department and at the annual Carnegie Mellon a community of creative, motivated undergradu- undergraduate research symposium. This year, we ate statisticians. For example, students might be had around 50 students competing in the statistics given a set of training data and a week to design research competition—a record number. an algorithm that optimizes a given error criterion. The department also has funds earmarked for Students can work in teams or on their own; they undergraduate support in the form of two endow- then present their approach and test their algo- ments, the DeGroot Memorial and Frederick rithms live in class on a set of test data. The winners Sorensen Memorial in Statistics funds. These contrib- might receive extra credit, but also bragging rights. ute to awards and travel subsidies for undergraduates We have held these types of competitions in several to attend and/or present their work at conferences. classes and they have yet to disappoint. One surprising result is that the winners are not Program Challenges always the top students in class. Some struggling One of our primary challenges has been the simul- students will completely devote themselves to the taneous preparation of students for both industry competition, learning additional advanced material and graduate school. The general CMU population on their own. The day of the competition is easily skews toward the quantitative analysis job market the most energetic day of the semester; the class- and, as such, needs to be well versed in methods, room is filled with cheers, groans, high-fives, and but also oral and written communication. These even dancing. Even in our statistical inference class, skills are, of course, part of a successful graduate our version of “Theory Jeopardy” had more than school experience, as well. Every upper-level course 80 students in a large auditorium frantically try- in our department includes writing and presenta- ing to solve inference problems in teams while tion skills. For example, in linear regression, stu- judges checked their work. When faced with dents are required to analyze large data sets from the “Daily Doubles,” the cheers from the teams real, interdisciplinary research problems and present able to double their scores were deafening. Who their findings in scientific reports. These projects knew statistics could be so cool? and reports build in complexity over the semester, We think the best way to manage our grow- culminating in something similar to a master’s-level ing program is to build a community in which qualifying exam experience. All students receive the statistics majors feel they are more than a individual report feedback (including grammar and name on a file. We advocate group collabora- spelling). Students are required to regularly defend tion and engagement in departmental activi- their work to faculty and other students in poster ties. Our seniors lead by example, both in the sessions, class presentations, or one-on-one conver- classroom and in the larger CMU community. sations. Students also are taught to critically analyze Anecdotally, we hear stories of upper-class sta- their work and suggest possible improvements with- tistics majors “adopting” newer majors in cam- out prompting. pus social organizations. The majors also have Almost every assignment, exam, or project is a student advisory committee to represent the grounded in a real, ongoing research problem. In departmental undergraduate interests; this com- upper-level courses, all data sets are large, messy, mittee sponsors both professional development and complicated. It is not uncommon for sopho- activities (e.g., interview preparation seminars) mores to be working with thousands of observa- and social events (e.g., a human histogram to tions with missing values. Students are presented celebrate World Statistics Day). This year’s with background material on a research problem CMU undergraduate statistics sweatshirt says and then asked to address specific scientific ques- “Statistics majors are always right … probably.” tions. They quickly become accustomed to the need Sounds about right to us. n

august 2012 amstat news 15 statistician's view

Misuse of Statistics a National Problem

was greatly interested and (http://finance.yahoo.com/news/ pleased to learn that ASA is inflation-life-cost-raising- now giving a course, “Stats child-145736881.html) is an forI Staffers.” I cannot think of a excellent recent example. more important group to get les- This media story says, “The sons in statistical literacy than cost of raising a child from birth those who are in the corridors of to age 17 has surged 25 percent power. The course seems excel- over the last 10 years, due large- lent, but I think that there are ly to the rising cost of groceries even more basic needs in the area and medical care, according to of statistical literacy, which I’m the Department of Agriculture, not sure are being met. I’m refer- which tracks annual expenditures ring to distortions that occur in on children by families.” Even the media. CNBC’s May 7 arti- more misleading is the head- cle, “The Inflation of Life - Cost line, which says, “The Inflation of Raising a Child Has Soared,” of Life - Cost of Raising a Child Has Soared” [bolding is mine]. The words “surged” and “soared” are not found in the Department of Agriculture report, which can be downloaded from www. cnpp.usda.gov/Publications/CRC/ crc2010.pdf. The words surged and soared are obviously misleading, because a 25% increase over 10 years represents an annual rate of inflation of 2.26% (compound- ed). This is just about the overall rate of inflation in the United States. What we see here is verbal distortion of statistical facts, which happens frequently in the public arena. I’m not sure whether the misuse of statistics is best addressed through education in statistical literacy, some watchdog function such as www.factcheck.org, or something else. But we do have a national problem, which I think should be of interest to ASA and its members.

Sincerely, Morris Olitsky Statistician, USDA, FNS, Mid-Atlantic Regional Office

16 amstat news august 2012 columns 175 Broadening and Deepening Statistical Thinking: Educating for 2014 and Beyond Richard Scheaffer

s I write this column, the debate about the standards can be viewed (and periodically critiqued) future of the American Community Survey is at www.smarterbalanced.org and www.parcconline.org. taking place in Congress, with one member How can colleges aid the CCSS project to improve Astating, “We’re spending $70 per person to fill this statistics education? The statistics and probability out. That’s just not cost effective, especially since, in standards of grades 6–12 are well beyond what one the end, this is not a scientific survey. It’s a random finds in current state standards, with few exceptions, survey.” How better to state the need for expanded and beyond what many mathematics teachers have statistics education for all? experienced in any deep way. Thus, a central task of The ASA has a history of leaders with a penchant colleges and universities is to improve the education for education, such as 1944 president Helen Walker, of prospective teachers. who, from at least a decade earlier, pushed for more Guidelines on what is needed and how it may and better statistics education at the college level and be provided can be found in a soon-to-be-released encouraged adding statistics to the school (K–12) report of the Conference Board of the Mathematical curriculum. Fred Mosteller, 1967 president and a Sciences, titled “The Mathematics Education of phenomenal educator, established the ASA/NCTM Teachers II.” This report challenges mathematics and Joint Committee on Curriculum in Statistics and statistics departments to work with colleges of educa- Probability, influential in expanding statistics in the tion to develop sound courses for future teachers that schools ever since. That committee developed the emphasize the topics they will actually teach. quantitative literacy projects of the 1980s, which led Typically, statistics departments are not engaged to the inclusion of statistics in NCTM mathemat- with education departments in any productive way; ics standards and, later, provided support for the AP this will have to change if statistics education at the Statistics program. These efforts to build statistics into school level is to make any pronounced improvement. school mathematics have led to such programs as the Otherwise, the statistics standards will be sidestepped Meeting Within a Meeting Statistics Workshop for or, perhaps worse, taught poorly. Mathematics and Science Teachers and Beyond AP As expected, total enrollment in the typical col- Statistics Workshop for AP Statistics teachers. lege introductory statistics course is low in statistics The Journal of Statistics Education and Significance departments as compared to mathematics depart- provide valuable resources for both school and college ments across four-year and two-year colleges, with statistics education, while the GAISE report (www. the most rapid expansion taking place in the latter. amstat.org/education/gaise) provides guidelines for sta- So, statisticians (and the ASA) need to reach out to tistics teaching in both venues. The latter has become improve statistics teaching from this wider perspec- a standard reference wherever standards in statistics tive. More alarming, though, is that enrollments in teaching are being addressed and will continue to second statistics courses drop dramatically across the carry the mark of sound statistics education into the board. Greater emphasis must be placed on devel future. In light of its primacy, the ASA should con- oping useful second courses in statistics to serve the sider updating this report. needs of many students who need to think statistically Today, the future of statistics education in the at a deeper level, including future K–12 and college schools is intimately connected to the future of the teachers and, perhaps, even future politicians. Common Core State Standards (CCSS) in math- With the support of the ASA, statistics educators ematics, an effort coordinated by the National are making progress in developing sound curricula, Governors Association and the Council of Chief informative assessments, and innovative materials State School Officers. Even though the standards for both schools and colleges. This progress must were released some time ago, there are numerous continue and expand to include ever-widening support activities open to contributions from inter- domains of students. Such would provide a worthy ested parties such as the illustrative tasks project tribute to the first 175 years of the ASA and bode http://illustrativemathematics.org. Progress on the well for the second. n all-important assessments being developed for these

august 2012 amstat news 17

columns MASTER'S NOTEBOOK Experiences as a Biostatistician at an Academic Health Center Cindy Weng, University of Utah, and Jamie Grady, University of Connecticut

ake a glance at any published medical article and you will likely see that statistical methods, in terms of study design and analyses,T play a major role in its success. Medical research heavily relies on the application of appropriate statistical methods. In addition to the technical aspects of a statistical analysis lays another crucial element: the human element. An effective collaboration between clinical researchers and biostatisticians is an essential aspect of a successful statistical analysis. Here, Cindy Weng and James Grady share how they learned to collaborate, prioritize, and appreciate their work as consultants at academic health centers. Cindy Weng’s Perspective When I began working at the Study Design and Biostatistics Center (SDBC), I was fortunate to have mentors who were experienced biostatisticians in medical research. They collaborated effectively with medical researchers—by using innovative research techniques—in various aspects of medicine, basic science, study design, statistical methods, epidemiology, and causal inference in diversified medical areas. At that time, I thought I was well prepared to take on the responsibilities of my new how to approach this specific problem, but on job; however, I was quickly surprised by how little I research study methodology as well so I could knew about certain aspects of statistical methodol- further educate myself. I continue to learn and ogy and design, particularly sample size and power expand my knowledge on this topic where I can. analysis. We seemed to cover little of this in my In fact, I recently discovered a book titled Sample graduate statistics program. In addition, nonstatis- Size Calculations in Clinical Research—written by tical skills such as consulting, effective nonverbal Shein-Chung Chow, Jun Shao, and Hansheng and verbal communication, and business and man- Wang—that has proven to be useful. agement knowledge are essential to the employed I think continually expanding your skill set is biostatistician today. necessary in today’s competitive world. My first assigned project had to do with sam- Since earning my MPH degree with an empha- ple size and power determination for a tiny study, sis in biostatistics, I have developed expertise while around five patients. The study used a crossover assisting in statistical analyses of epidemiologic and design to compare the effect of noninvasive corti- observational studies. There are many rewards that cal stimulation among children with hemiplegic come from my career, including direct interaction cerebral palsy (CP) in an intervention to several with physician scientists, contributing statistical sham groups. A paired t-test to determine power concepts to studies, and seeing results from these of the study seemed to be an elementary statis- studies published in prestigious journals, regardless tical problem to solve for most biostatisticians, of my authorship status. I will continue developing but at that time, was difficult for me. Therefore, my expertise by taking additional education courses I sought out my mentors’ guidance on not only at JSM and my local university.

august 2012 amstat news 19 columns

In fact, I expect to earn a comparative effective- into any of them. However, in my experience, ness research certificate in the summer of 2013. what may seem like a quick consult is usually Comparative effectiveness research (CER) com- much appreciated by the client. pares existing health care interventions to determine the benefit and harm of each and the most effec- Write: Another challenge in academia is tive intervention to treat patients on an individual being asked to write a statistical section of basis. The promise of CER to biostatisticians is that a manuscript or grant proposal, sometimes it enables us to best assist researchers in designing without extensive background informa- studies that can support patients, physicians, poli- tion or knowledge of the subject area. These cymakers, and resource planners to make practical investigators might only use statisticians spo- medical decisions. Through this certificate, I hope radically, every few years when they need to to continually develop professionally and would like submit a new grant. This is the reality of the to fulfill my goal of becoming a CER expert. consulting statistician. Of course, it is not the To broaden my experience as a biostatistician, preferred arrangement for an ongoing col- I am an active member of the ASA Section on laboration, but it can be the beginning of one. Statistical Consulting. I also have been enjoy- Consulting statisticians can provide assistance ing volunteering for Statistics without Borders when planning studies, an exercise that is (http://community.amstat.org/statisticswithoutborders/ often quite rewarding. home). My first project with this organization considered the effect of humanitarian aid given by Prioritize: Experience will give you the ability the International Rescue Committee at Pakistan to prioritize. You might need to go the extra refugee camps. The objective was to compare U5 mile for the highly funded senior investigator’s mortality of children born to women ages 15–49. grant; maybe not for the fellow who needs a With the number of volunteers and amount of sta- ‘quick analysis’ for a poster submission to a tistical expertise at SWB, the SWB New Projects conference. Experience on the job gives you Committee would like to contact nonprofit orga- confidence in making these decisions and pri- nizations in need of statistical consulting. I am a oritizing your effort. member of the SWB New Projects Committee and am assigned to the CharityFactors organization. Reward: An ongoing challenge is getting Through these activities, I hope to gain more experi- credit for consultations and collaborations on ence in communicating, collaborating, and working grants. Some universities are changing their with nonprofit organizations. tenure and promotion criteria to reflect the value of collaborators. Be sure to understand Jamie Grady’s Perspective how your institution views and/or rewards col- Beginning with graduate school, I have 25 years laborations when planning your career. of experience consulting at academic health cen- Statistical consulting and collaboration will con- ters. Projects have been both big and small, and tinue to play a major role for some statisticians, and the variety of clients fit my personality, as I enjoy may be the basis of their jobs. Having good men- juggling many activities at once. Success in this tors when you enter the field of statistical consult- environment requires flexibility—client investiga- ing can help foster a successful career. Becoming a tors don’t always plan things perfectly—and you person with strong statistical training who also can need to be accommodating. communicate effectively to nonstatisticians will be Interact: One rewarding aspect for biostatisti- a strong asset to any business, industry, or academic cians at academic centers is being involved in workplace. Last but not least, continuing education the educational mission of a university and and attending conferences help tie this all together interacting with other medical faculty, stu- by allowing you to stay current, develop collegial dents, and postdocs. The many projects that relationships, and improve your skill set. come about from basic science and clinical For more information about consulting and the research studies at a health science center allow statistician-client relationship, please visit www. you to use many of the statistical methods amstat.org/sections/cnsl/brochures.cfm for a brochure published by the Section on Statistical Consulting. learned in graduate school. Sometimes, I feel You can give a copy to your clients as well. n like a generalist on the job; I am on the surface of many statistical applications, but not deep

20 amstat news august 2012 columns STATtr@k Preparing for a Career as a Sports Statistician: Two Interviews with People in the Field Jim Albert, Bowling Green State University

any students are fascinated with statistics and sports and ask aboutM a possible career working as a sports statistician. To help them understand this career, Jim Albert contacted Ben Alamar and Keith Woolner, who are actively working as statisticians with professional sports teams. Each was asked a similar set of questions, and their answers shed light on the background, academic skills, statistical methods, and people skills that are important for a person in this vocation. JA: When were you first inter- JA: Is your position as a sports ested in statistics applications statistician a full-time or part- Interview in sports? time position? with Ben Alamar BA: I wrote my first paper on BA: I work part time with the modeling the probability that Oklahoma City Thunder and an NFL team would make the do a variety of other consulting playoffs during my third quarter work in the field, but I am also a in graduate school. I liked the professor of sports management Ben Alamar earned a BS in eco- paper, but never got it published. at Menlo College. nomics from the University I didn’t return to stats in sports of Minnesota and an MA and until I was a post doc at UCSF. JA: What skills and academic PhD in economics from the I did some consulting work for training (e.g., college courses) University of California at Santa a start-up company in the field are valuable to sports statisti- Barbara. Currently, he is a pro- and decided that it was the area I cians? fessor of sports management at wanted to concentrate on. Menlo College and works part time as a sports statistician for BA: High-level statistics courses the Oklahoma City Thunder. JA: What are your professional of all types are valuable, as are Alamar founded the Journal of duties as a sports statistician? acquiring advanced data manage- Quantitative Analysis of Sports— ment skills such as SQL. an ASA journal dedicated to BA: I provide analysis on play- the statistical analysis of sports ers and team strategy, as well as data—in 2005. tackle larger research projects.

august 2012 amstat news 21 columns

JA: Are there specific statistical any new tool when you have had Interview tools or topics that you find success previously without it. As with Keith especially helpful in your work? decisionmakers gain more expo- Woolner If a person had to take, say, sure to information that analysis three courses in statistics to can provide, they tend to become help them in your work, what more interested. courses would they be? JA: Do you think there will be Keith Woolner earned two bach- an increasing demand for elor’s degrees from MIT—one BA: Some important tools statisticians in your particular in mathematics with computer include basic regression analysis, sport? science and one in management logistic regression, Monte Carlo from the MIT Sloan School of simulation, classification, and Management. He then earned a hierarchical regression. Just as BA: Yes, I do. Data sets are master’s degree in decision anal- important as the technical tools becoming more complex ysis from Stanford University. though is the skill of effectively (motion capture technology is While he was employed as a communicating the analysis to being used to track everything software developer, he regularly nontechnical audiences. that moves on the court 25 times contributed to Baseball Prospectus a second) and the general con- (www.baseballprospectus.com). JA: Is there specific statistics or cept of using statistics is gaining In 2007, Woolner left Baseball data-management software more acceptance; these factors Prospectus to join the front office that you find helpful? will lead to more teams employ- of the Cleveland Indians profes- ing large analytics groups. sional baseball team. BA: R and SQL are very useful JA: Are there particular websites JA: When were you first interested JA: What are the first steps in for the interested student to in statistics applications in sports? entering the sports industry as visit to learn about the current a statistician? work in basketball analytics? KW: I’ve always had an affinity for both baseball and math, and I memorized many of the statis- BA: There is no clear path. I rec- BA: Hoopdata.com, basket- tics on the backs of baseball cards ommend that aspiring sports ana- ballvalue.com, and apbr.org. as I child, but it wasn’t until I lysts try to answer a question they was an undergrad at MIT in the think a general manager or coach JA: Do you have other general late 1980s that I discovered the would find interesting, then find advice for high-school or col- Usenet newsgroup rec.sport.base- a way to get that work into the lege students who are inter- ball on the Internet and learned hands of people who might be ested in a career as a sports about the emerging field of saber- interested. There is no shortage metrics and the work researchers of people interested in working in statistician? like Pete Palmer and Bill James the field, but there is a shortage were doing. I was fascinated by of people who have actually done BA: The advice I give to anyone it, eventually started tinkering good work in the field. interested in basketball statistics is to read Basketball on Paper, by around with my own methods, JA: In “Moneyball,” there was Dean Oliver, try to start think- and began publishing my own some resistance to the use of ing like a GM/coach instead of invented stats in the mid to late statistical methods to learn a fan (think about how analysis 1990s. That led to my involve- about players, especially by can actually be used to inform ment with Baseball Prospectus and people who were not part of decisionmaking), and work on set me on the path to an eventual the baseball establishment. communicating complex analy- career in sports statistics. But, I Do you think there is a similar ses to people who do not know didn’t seriously think a career in resistance to the use of statisti- linear algebra. sports was an option for me until cal methods in basketball? just a couple of years before I joined the Indians.

BA: I would not classify it as JA: What are your professional resistance, but I think there is a duties as a sports statistician? natural skepticism of employing KW: My title is director of baseball analytics, and in that role, I manage a team of analysts and 22 amstat news august 2012 columns programmers who support base- with the technology you’re likely baseball analysis is to just start ball decisionmaking by organiz- to encounter will help tremen- doing it on your own. Build ing, analyzing, and presenting dously. Most of the information up a base of knowledge so you’re information. We help baseball you’d be dealing with in sports aware of the state of the art and operations put the best pos- statistics would be in a database, complement that with the techni- sible team on the field. In many so learning SQL or another query cal skills you need to answer your cases, that involves developing language is important. In addi- own questions well. Write, write, statistical models to measure tion, mastering advanced spread- and write some more. Your body player value, forecast future per- sheet skills such as pivot tables, of work is your résumé, and if you formance, and answer questions macros, scripting, and chart cus- demonstrate your capabilities and about game strategy and tactics. tomization would be useful. develop expertise in a particular Domain knowledge—truly area, teams will notice. The more JA: Is your position as a sports understanding the sport you want you can show that you both have statistician a full-time or part- to analyze professionally is criti- the quantitative skills and the base- time position? cal to being successful. Knowing ball knowledge to help a team, the the rules of the game; studying better off you will be. KW: My current role with how front offices operate; find- JA: What kinds of data do you the Indians is a full-time posi- ing out how players are recruited, tion. When I was with Baseball developed, and evaluated; and collect and analyze beyond team Prospectus, it was part-time and even just learning the jargon used and individual performance? mostly a hobby. within the industry will help you integrate into the organization. KW: There’s a tremendous amount JA: What skills and academic You’ll come to understand what of detailed data being collected training (e.g., college courses) problems are important to the during every game. The basic out- are valuable to sports statisti- GM and other decisionmakers, come data (how many hits, walks, cians? as well as what information is runs, errors, homers, etc.) are the available, how it’s collected, what most obvious example, but we also KW: I would say there are three it means, and what its limitations sets of skills you need to be a suc- are. Also, I recommend keeping cessful sports statistician: up with the discussions in your Quantitative skills—the sta- sport’s analytic community so tistical and mathematical tech- you know about the latest devel- niques you’ll use to make sense of opments and what’s considered the data. Most kinds of course- the state of the art in the public work you’d find in an applied sphere. One of the great things statistics program will be helpful. about being a sports statistician Regression methods, hypoth- is getting to follow your favorite esis testing, confidence intervals, websites and blogs as a legitimate inference, probability, ANOVA, part of your job! multivariate analysis, linear and JA: What are the first steps in logistic models, clustering, time entering the sports industry as series, and data mining/machine a statistician? learning would all be applicable. I’d include in this category designing charts, graphs, and KW: My path into sports was other data visualizations to help atypical, so it’s hard to use that present and communicate results. as a basis for a strategy to get Technical skills—learning into the industry. There are more one or more statistical software opportunities in sports now than systems such as R/S-PLUS, SAS, there were years ago, but also SPSS, Stata, Matlab, etc. will greater competition due to the give you the tools to apply quan- increased awareness of sports titative skills in practice. Beyond analysis as a career path from the that, the more self-reliant you popularity of “Moneyball” and are at extracting and manipulat- the like. ing your data directly, the more What I usually tell people quickly you can explore your data who ask me that question is and test ideas. So being adept that the best way to break into

august 2012 amstat news 23 columns

collect information about where JA: How have statisticians Because baseball was ahead on the field each ball was hit, what influenced the operations of of the curve in adopting statisti- kinds of pitches were thrown, and your team? cal analysis, we may be closer to what situations each batter and the saturation point than other sports. Eventually, there is an pitcher faced. KW: Although I can’t get into upper limit to how many jobs There are cameras installed in specifics, I think where we have there could be with major league every MLB ballpark that track had the most influence is on teams, as the number of clubs is the path of the pitch in flight improving the consistency of pretty small and it’s unlikely that to the plate, so we know how our decisionmaking process. The a team would need hundreds of fast each pitch was thrown, how clarity that comes from perform- analysts to meet their needs. The much each curve ball curved, ing the same analyses and run- number of opportunities may where each pitch was released ning the same models every time be even greater in other sports and crossed the plate, whether we evaluate a different scenario than in baseball. But, even now, the batter swung, and how hard provides some balance against baseball teams are still hiring and he hit it if he made contact. being swayed by emotional reac- expanding their analytic depart- Beyond the data collected tions, or unduly influenced by ments, so we’re not at that point in-game, we have many years’ the loudest voice in the room. At yet. There’s still room for growth. worth of scouting reports, phys- the same time, you will always iological tests, medical histories, have a need for the knowledge psychological profiles, coach- and intuition of experts like JA: Do you have some other ing assessments, contract data, coaches, scouts, and executives, service time records, arbitra- general advice for high-school because every situation is unique. or college students who are tion case histories, negotiation There’s no one-size-fits-all records. We also have thousands interested in a career as a approach that works every time, sports statistician? of hours of recorded video. All and no computer can evaluate these sources can be useful in every possibility. But having that analyzing and forecasting both consistent evaluation process as KW: Don’t neglect the soft skills. team and player performance. a starting point helps ground Being successful in sports analysis your thinking so you avoid mak- is not just about having the most ing mental mistakes as you work data or the best algorithm. Many through a decision. brilliant analysts have struggled because they couldn’t get oth- JA: Do you believe there will ers to listen and buy into their be an increasing demand for ideas. It’s not enough to be right; statisticians in your particular you also have to be persuasive. sport? Communication skills, both spo- ken and written, are important and under-rated. If you can’t KW: I think there will be an explain what you’ve analyzed to increasing demand for statisti- someone who doesn’t have the cians who both understand their same statistical training, you’ll sport’s unique characteristics and have a hard time influencing the can identify the right statistical decisions they make. methods to apply to a problem. Also realize that choosing a More and more, teams are career in sports means spend- even looking beyond their ing a lot of time and long hours sports operations department with your coworkers. How you and employing statistical tech- are perceived by your colleagues niques on their ticketing, determines how much they like marketing, ballpark/stadium you, respect you, and trust your operations, personnel schedul- judgment. Optimism, humility, ing, food and beverage, and open-mindedness, and a sense merchandising data. There are of humor go a long way toward many ways an analyst could building the foundations of good help a club in these areas, too, working relationships. n so they shouldn’t be overlooked.

24 amstat news august 2012 education

First eCOTS Claims Success

he first biennial Electronic Conference on Teaching Statistics (eCOTS) was hosted by the Consortium for the Advancement Tof Undergraduate Statistics (CAUSE) from May 13–18. The conference was presented over the Internet and showcased three themes: Debating the Big Ideas of Teaching Statistics, Statistics for the Modern Student, and Commercial Resources for Teaching Statistics. A total of 420 statistics educators and students registered to participate live in the eCOTS program offerings. The week began with a series of 30-minute breakout sessions. Sessions covered a variety of topics, including interactive teaching of probability distributions and theory, using advertisements to teach statistical literacy, the second course in statistics, teaching bootstrapping and randomization-based methods, data visualization on the iPad, teaching in the online environment, students. This was followed by a series of panel using games to teach statistics, statistical comput- presentations in which the leaders of breakout ing, simulation and audience response systems in sessions reconvened to respond to audience ques- the statistics classroom, and what students need to tions and discuss their work in more detail. The know about statistics in the 21st century. conference culminated in a keynote presentation Twenty-two virtual poster presentations were by Webster West, who focused on the effect of pre-recorded and available throughout the week. technology on the teaching of statistics. Posters provided five-minute audio-visual presen- Those interested in learning more about tations on teaching statistical bioethics online to eCOTS and viewing recordings from breakout involving undergraduates in statistical consulting sessions, posters, keynote talks, and panel pre- to understanding the beguiling coincidences seen sentations are encouraged to visit www.causeweb. in big data. Participants could view posters and org/ecots. leave feedback on a special discussion board set up for each poster. Upcoming Workshop A short workshop was held on May 17 about CAUSE is pleased to announce “Identifying and the CATALST project at the University of Addressing Difficult Concepts for Students in Minnesota (www.tc.umn.edu/~catalst). This work- the Introductory Statistics Course,” a one-day shop was led by three graduate students from the workshop geared toward instructors at two-year statistics education program at the University colleges or instructors who are new to teaching of Minnesota: Rebekah Isaak, Laura Le, and statistics. The workshop will be led by Marjorie Laura Ziegler. The fully subscribed workshop Bond of Monmouth College on September gave participants a two-hour introduction to the 15 in Monmouth, Illinois, and November 7 in CATALST curriculum. Jacksonville, Florida, as a pre-AMATYC confer- The final day of eCOTS began with a keynote ence workshop. There is no registration fee to presentation by Hans Rosling, who talked about attend. Details and registration information are using a fact-based worldview to engage statistics available at www.causeweb.org/workshop. n

august 2012 amstat news 25 ASA Announces Statistics Poster and Project 2012Competition Winners

he American Statistical Association is their advisers provided by Texas Instruments. pleased to announce the winners of the Second-place winners received $100 and a plaque; 2012 Poster Competition and Project third-place winners received $50 and a plaque; and TCompetition. First-place winners received $200, a honorable mentions received certificates. plaque for themselves and their school, and grade- The poster and project competitions are direct- appropriate graphing calculators for themselves and ed by the ASA/NCTM Joint Committee on Curriculum in Statistics and Probability, with Linda Quinn of Cleveland State University serving as the poster competition leader and Jamis Perrett of Texas A&M University serving as the project competition leader. K–12 posters are due every year on April 1. Projects (written reports) for grades 7–12 are due every year on June 1. Visit the competitions web page at www.amstat.org/education/posterprojects for information about previous winners and how to enter. Also access instructional webinars and a rubric of how the posters and projects are judged. 2012 National Project Competition Winners Each year, the statistical project competition attracts a wide variety of submissions in which students from grades 7–12 conduct creative studies. This year, the submission deadline for the project competition was changed to June 1 to enable participation from high-school students who may have been preparing for the AP Statistics exam administered in mid-May. The new deadline also made it possible for teachers who might otherwise be busy at the AP reading to assist with the competition judging. The statistical project competition was especially useful for these students because it provided them with opportunities to apply all the statistical skills they had acquired throughout the school year to solve real-world problems of interest to them. Motivation to participate in the competition included monetary awards, plaques, and new Texas Instrument calculators. Results of the project competition and a list of the judges can be found at http://magazine.amstat.org.

26 amstat news august 2012 2012 National Poster Competition Winners

Grades K–3

First Place Brandon Samuels “Having a Ball with Statistics” Manatee Bay Elementary Weston, FL

Second Place Adva Oshri and Sophia Zapata “Look Who’s Talking” Manatee Bay Elementary Weston, FL

Third Place (tie) Joyce Huang “Is the Three-Sided Die Fair?” Worthington Park Elementary Westerville, OH

august 2012 amstat news 27 Third Place (tie) Oren Aviad, Peter Berganross, Iesa Khanji, Kabir Khwaja, Victoria McNelis, Arnav Paliwal, Ava Pfannenbecker, and Cole Wissink “World Pollution” Beecher Road School Woodbridge, CT

Honorable Mention Joseph McClelland “What’s My Number? Blood Glucose Values” Ada Vista Elementary Ada, MI

Honorable Mention Brooke Shirley, Anthony Aronson, Gino Bartolini, and Madison Prenni “Oreo Obsession” Saltsburg Elementary School Saltsburg, PA

28 amstat news august 2012 Grades 4–6

First Place Andrew Wyatt “Show Me Your Colors” Lied Middle School Las Vegas, NV

Second Place (tie) Carter Gates “Does the Size of a Pumpkin Determine the Number of Seeds Inside?” Rydal Elementary Huntingdon Valley, PA

Second Place (tie) Jonathon Price “Now You See It, Now You Don’t” Edgewood Elementary Fruitport, MI august 2012 amstat news 29 Honorable Mention Tyler Pearson and Wyatt Hinkson “Favorite Pet?” Grace Christian School Bowie, MD

Honorable Mention Kayla Scott Grades 7–9 “State Rankings vs. Private Schools in the United States” Grace Christian School Bowie, MD

First Place Benjamin Townson “Which Storm Caused the Worst Power Outages in Connecticut?” Jockey Hollow Middle School Monroe, CT 30 amstat news august 2012 Second Place Jahnik Kurukulasuriya “Hacked” Campus School of Carlow University Pittsburg, PA

Third Place Sarah Schultheis, Taylor Carroll, and Lizzy St. Clair “Sizzlin’ Relaxin’ Summer Getaway” Palmyra Area Middle School Palmyra, PA

Honorable Mention Peter O’Neill “Do Gender and Exercise Affect Self-Esteem?” Clarkstown High School South New City, NY

august 2012 amstat news 31 Grades 10–12

First Place Julia Benyo, Kacie Moore, Erika Shores, Laura Sierko “Dangerous Debt: What Is the Perceived Impact of Credit Card Debt in the United States?” Cuyahoga Valley Christian Academy Cuyahoga Falls, OH

Second Place Sonya Tran and Bridget Kilbane “Which Social Networking Site Is More Popular? Facebook vs. Twitter” St. Joseph Academy Cleveland, OH

Third Place Layla Barati “Do Teens Really Care About Their Community?” East Career and Technical Academy Las Vegas, NV 32 amstat news august 2012 Honorable Mention Yan Ting Cheung and Erin Bator “Which Is Healthier? Fruits, Fruit Snacks, or Candy Bars?” St. Joseph Academy Cleveland, OH

2012 Regional Poster Other Region/ASA National Poster Competition Leader: Linda Quinn, Cleveland State University Competition Leaders Contact: Rebecca Nichols, ASA Director of Education, [email protected] Connecticut Chapter Statistical Poster Competition www.amstat.org/education/posterprojects Marianne E. Messina, Bristol-Myers-Squibb www.amstat.org/chapters/Connecticut/home/Poster/ Students outside the regional competition areas poster_index.htm submit their posters directly to the ASA office to be judged separately by the Washington Statistical Michigan Statistics Poster Competition Society as part of the Other Region. The best post- Dan Frobish, Grand Valley State University ers from each region are sent to the national judging. www.gvsu.edu/stat/mspc-home-30.htm Information about regional poster competitions and winners is available on the individual regional poster Nevada K–12 Statistics Poster Competition competition websites. David Thiel, Clark County School District www.amstat.org/chapters/nevada/ Get Involved k12postercompetition.html For information about how you can start a regional poster competition or mentor students in your Ohio Statistics Poster Competition area, see the article appearing in the July 2011 Jerry Moreno, John Carroll University issue of Amstat News at http://magazine.amstat. www.bio.ri.ccf.org/ASA/poster.html org/blog/2011/07/01/poster-comp-how-to. You can download a flyer about the ASA poster and project Pennsylvania Statistics Poster Competition competitions and other K–12 statistics education Pete Skoner, Saint Francis University programs and resources to share with your local www.francis.edu/paposter.html schools at the bottom of www.amstat.org/education. For additional information or questions regard- Washington Statistical Society Poster Competition ing how to get involved in the poster or project Barnali Das, Westat competitions, contact ASA Director of Education www.amstat.org/education/posterprojects Rebecca Nichols at [email protected]. n

august 2012 amstat news 33 education ASA Group Reviews Mathematical Education of Teachers II

ollowing the Because MET2 will influence the statistical prepa- Review Group Common Core ration of teachers and the teaching of statistics at the State Standards K–12 level, the American Statistical Association estab- Nancy Boyton Jerry Moreno F(CCSS) Initiative in lished a MET2 review group consisting of K–12 sta- Christine Franklin Rebecca Nichols Mathematics, the tistics teachers and university-level statistics educators Conference Board of with strong ties to K–12 statistics education. With Amy Froelich Chris Olsen the Mathematics statistics taking a more prominent role in the CCSS, Robert Gould Mary Sullivan Sciences (CBMS) especially at the middle- and high-school levels, the Bradley Hartlaub Roxy Peck released a draft of The review group generally focused on recommendations Mathematical regarding statistics education for pre-service teachers Patricia Humphrey Michael Posner Education of Teachers II and on the need for professional development oppor- Henry Kranendonk Iwan Praton (MET2) as an update tunities for in-service teachers. Successful implemen- to the original docu- tation of the CCSS will require statistics content that John McKenzie Amy Wagaman ment published in focuses on the conceptual understanding necessary for 2001. The report the development of sound statistical reasoning. focuses on the mathematics and statistics preparation The ASA review of MET2 was well received by of K–12 teachers with the primary audience intended CBMS. The MET2 draft report is available at www. for college and university mathematics and statistics cbmsweb.org. The ASA review of the MET2 draft departments and individual mathematicians and stat- report is available at www.amstat.org/outreach/pdfs/ isticians responsible for educating K–12 mathematics/ MET2ASAResponse.pdf. n statistics teachers.

34 amstat news august 2012 awards and deadlines

American Statistical Association Conference on C. R. and Bhargavi Rao Prize STATISTICAL PRACTICE embers of the Rao Prize Committee areM accepting nominations for the C. R. and SAVESAVE THETHE DATEDATE Bhargavi Rao Prize for The ASA announces the Outstanding Research in Statistics. The prize, Conference on Statistical Practice awarded by Penn State Innovations and Best Practices for the University’s department of statistics, was estab- Applied Statistician lished to recognize outstanding and influential February 21–23, 2013, New Orleans, LA C. R. Rao innovations in the theory and practice of mathematical statistics, international leadership in directing Statistical Practice 2013 brings together hundreds of statistical statistical research, and pioneering contributions by a practitioners—including data analysts, researchers, and scientists—who engage in the application of statistics to solve real- recognized leader in the field of statistics. world problems. The conference will provide an opportunity to The Rao Prize is awarded every two years learn about the latest statistical methodologies and best practices (odd-numbered years) to an individual work- in statistical design, analysis, programming, and consulting. ing in the United States. The awardee receives a medal, cash prize, and invitation to visit Penn State and give a talk. Key Dates: Nominations should include a letter describing the nominee’s outstanding contributions August 30, 2012 to leadership and research in statistics, a cur- • Practical Computing Expo Submissions Due rent curriculum vita, and two supporting letter. Submissions are due by December 3 and should September 13, 2012 be sent to Rao Prize Selection Committee Chair, 326 Thomas Building, Penn State University, • Poster Abstract Submissions Due University Park, PA 16802-2111. C. R. Rao held the Eberly Chair in Statistics at October 1, 2012 Penn State University from 1988–2001. He now • Registration Opens serves as Holder Emeritus of the Eberly Chair in Statistics. He was the founding director of the December 31, 2012 Center for Multivariate Analysis. A President’s • Early Registration Deadline National Medal of Science Laureate, Rao is recognized worldwide as a pioneer of modern statistical theory and one of the world’s top five January 25, 2013 statisticians, with multifaceted distinctions as a • Housing Deadline mathematician, researcher, scientist, and teacher. His contributions to mathematics and statistical theory and applications have become part of undergraduate and graduate courses in statistics, econometrics, and electrical engineering at universities throughout the world. Learn more about CSP 2013 at www.amstat.org/csp. For additional information, see www.stat.psu. edu/news/awards/raoprize.html. n American Statistical Association Conference on STATISTICAL PRACTICE New Orleans, LA February 21–23 2013

august 2012 amstat news 35 36 amstat news august 2012 people news

Harvard Statistics Department Presents First Dempster Prize Matthew McClellan, Harvard Staff Writer

Editor’s Note: This article originally appeared in the Harvard Gazette on May 23. We have reprinted it with permission.

’m not very good with num- enough if I cannot convince Art bers,” said Arthur Dempster, of its validity,” said Blyth. “ professor emeritus of theo- The practical applications reticalI statistics, while pondering of statistics have always driven the founding year of the Dempster, so it is no surprise American Statistical Association. that he spent a year working On May 11, generations of at Bell Telephone Laboratories students and faculty celebrated before he joined the department the inauguration of the Arthur in 1958. Or that he has inspired P. Dempster Award and the students who work not only in Xiao-Li Meng (left), Whipple V. N. Jones Professor of 55th anniversary of the Harvard academia, but in finance, tech- Statistics, speaks with Arthur P. Dempster before the event Statistics Department. Of course, nology, and elsewhere. that celebrated the inauguration of the Arthur P. Dempster the large turnout gave lie to Much of Dempster’s work Award and the 55th anniversary of the Harvard Statistics Dempster’s claim; he’s been quite has been on integrating logi- Department. good with numbers, shown by cal and probabilistic reasoning Photo courtesy of Harvard Gazette the many who attended. beyond the established Bayesian Stephen Blyth, professor of framework, which requires quan- the practice in statistics and tification of prior information. conclusions, but scientists want managing director at the Harvard Dempster’s rule of combination to focus on processes of interest Management Co., established established a method to combine without getting stuck in low- the Dempster fund to recognize evidence from different sources level details of observation. promising graduate students in and Dempster-Schafer Theory, Blocker’s work finds inspira- the department, especially those or the theory of belief functions, tion and application in biology. working in theoretical and foun- which provides a framework for Expression microarrays measure dational statistics. “I’m happy for quantifying ignorance. the level of gene expression in the award to support ‘deep think- “I am not pessimistic about cell samples, allowing analyses ing’ about uncertainty, broadly the future,” said Dempster, “but of thousands of genes in parallel. defined, rather than simply rote I am pessimistic about what we They can be used to study chang- or ‘procedural’ application of sta- can do with limited data.” es in gene expression in response tistical machinery,” said Blyth, And this is where inaugu- to pathogens in comparison with who hosted the event at the ral award winner Alexander uninfected samples. “Millions of Federal Reserve, where the fund Blocker picks up. His research dollars are now invested in build- is located. deals with the problems of pre- ing huge collections of genomic Blyth, who studied under processing data: To make reams data,” said Blocker. “For example, Dempster and earned his PhD in of data comprehensible to an some hospitals are now routinely 1992, sees the award as a com- investigator, the raw data is collecting microarray data on memoration of his experience typically calibrated, smoothed, incoming cancer patients.” as a student and recognition of or otherwise simplified before As hospitals build massive the support he received from the being analyzed. Information databases to catalog the genetic faculty and the rest of the depart- is necessarily reduced, or even profiles of tumors, improving ment. Dempster’s influence has destroyed. Retaining more preprocessing techniques could only grown with time. “I know information through prepro- allow researchers to make bet- my argument is not strong cessing would allow for stronger ter use of data they are already

august 2012 amstat news 37 people news

collecting. It also could identify relevant supplementary information that would allow inves- Two PhD Candidates tigators to reinterpret data later. “Handled improperly, preprocessing could compromise the Take Home Natrella utility of these endeavors, wast- ing effort and delaying scientific Scholarships progress,” said Blocker. Will Guthrie, Natrella Scholarship Selection Committee Chair Blocker’s ultimate goal is to establish principled guidelines for data preprocessing. The ambition was betrayed by a few nervous laughs and one sincere “good luck” from the crowd. From the example of Dempster and the founding faculty, this kind of ambition has become a departmental tradition. Blocker also enjoys the department’s tradition of strong advising. Even the title of his presentation betrayed Anis Megahed its influence: “The Potential and Perils of Preprocessing: A he Quality and Productivity Section awarded two Multiphase Investigation.” Mary G. and Joseph Natrella scholarships during the “By the amount of allit- 2012 Quality and Productivity Research Conference, eration,” said Blocker, “you all Twhich was held June 4–7 in Long Beach, California. know one thing: I am Xiao-Li’s The recipients for 2012 are Mohammed Zafar Anis, student.” a PhD candidate in the department of mathematics at Xiao-Li Meng, chair of the the Bengal Science and Engineering University, and department, said that when Fadel Mounir Megahed, a PhD candidate in the Grado Blocker first came to the depart- Department of Industrial and Systems Engineering at ment (after studying economics Virginia Tech. at Boston University), he quickly Anis was recommended for the award by Murari Mitra produced research results that and Tirthankar Dasgupta. His presentation at the confer- had eluded some of Meng’s exist- ence was titled “A Review of Process Capability Indices.” ing PhD students and contin- Megahed was recommended for the award by William ues to produce strong research Woodall and Jaime Camelio. The title of his presentation results. But, said Meng, Blocker was “From Fault Detection to Diagnosis: An Investigation and his research, very much like Using Statistical Process Control and Visual Analytics in the department, are still great High-Density Data Environments.” works in progress. The winners were chosen for their outstanding teach- “When I first started, Xiao- ing, community service, mentoring, leadership, scholar- Li would give me advice, and I ship, and commitment to the pursuit of quality improve- would figure out how right he was ment through the use of statistical methods. Each winner two months later,” said Blocker. gave a research presentation at the conference and received “Four years later, I’ve shortened a $3,500 scholarship, plus $500 for travel expenses and that cycle to two weeks. If I get it complimentary registration for the conference and pre- down to one week, I think I’ll be conference short course. n ready to graduate.” The scholarships are funded from the ASA Natrella Scholarship Fund and the Quality and Productivity Research Conference.

38 amstat news august 2012 people news

Bruce Lindsay, Willaman Professor of Statistics and head Royal Statistical Society Toasts Queen of the department of statistics at Penn State University, was recently appointed as holder of the Eberly Family Chair in Statistics, one of the highest honors awarded to faculty members in Penn State’s Eberly College of Science. The appointment was made by the Office of the President of the University based on the recommendations of colleagues and the dean in recognition of Lindsay’s national and international reputation for excellence in research and teaching. Lindsay also has played a leading role in the creation of the nation’s policy regarding statistical data. In 2002, he was Staff members of the Royal Statistical Society stand in front of their offices on Errol Street in chair of the National Science London, England, to celebrate the Queen’s Jubilee Anniversary. From left: Teresa Dewane, Foundation Workshop on the Nicola Emmerson, Jack Beeby, Mawreen Chapman, Toni Young, Anna Mair, Charlotte Future of Statistics and served Stovell, Andrew Garratt, Abdel Khairoun, Roeland Beerten, and Paul Gentry as one of the co-editors of the resulting advisory report to the National Science Foundation. at Berkeley and a founding co- associate editor for Technometrics, From 1995 to 1997, he served director of the Microsoft Statistics Sankhya, and Statistics Survey. She on the National Research and Information Technology has served as an associate editor Council Committee on Fish Laboratory of Peking University of the Journal of the American Stock Assessment Methods. in Beijing, China. Previously, she Statistical Association, Annals of Lindsay has supervised 30 was a professor at the University Statistics, and Statistica Sinica. PhD degree recipients during his of Wisconsin-Madison; visit- During the 8th World time at Penn State and was hon- ing professor at Yale, ETH, Congress in Probability and ored in 1998 by the university INRIA, and the Poincare Statistics (the quadrennial joint chapter of the scientific research Institute; and a member of the conference of the Bernoulli society, Sigma Xi. He was named technical staff in the math center Society and IMS) in July, Yu deliv- a Fellow of the Institute of at Bell Labs. Her major research ered the 2012 Tukey Memorial Mathematical Statistics in 1987 interests include statistical infer- Lecture in Statistics. The Tukey and the American Statistical ence, machine learning, informa- Lecturer is one of the highest Association in 1998. He was the tion theory, data problems from academic honors on a statistician 2010 Fisher Lecturer. remote sensing and atmospheric bestowed by the Bernoulli Society. science, networks, neuroscience, Yu earned her BS in math- and text documents. Jointly ematics from Peking University with others, she holds three U.S. in 1984 and her MA and PhD Bin Yu patents: time-varying network in statistics from UC Berkeley has been selected as tomography, combined LMS in 1987 and 1990, respectively. president-elect of the Institute of prediction for lossless com- She was a 2006 Guggenheim Mathematical Statistics (IMS). pression of audio signals, and Fellow and is a Fellow of the She will serve a three-year approximation lasso methods for ASA, IMS, IEEE, and the term—one year as president-elect language modeling. American Association for the (2012–2013), one year as presi- Yu has published more than 70 Advancement of Science. dent (2013–2014), and one year papers in refereed journals, as well For a complete list of IMS elec- as past-president (2014–2015). as more than 30 refereed confer- tion results, visit http://imstat.org/ Yu is a professor and chair ence papers and book chapters. Yu news/2012/06/19/1340124690475. of the department of statistics is the action editor for the Journal html. at the University of California of Machine Learning and an

august 2012 amstat news 39 people news High-School Students Show Poise, Maturity in Pittsburgh Competition

ore than 1,500 high-school students from about 70 countries presented projects to judges from a variety of disciplines during the Intel International Science and Engineering Fair in Pittsburgh,M Pennsylvania, May 13–18. The American Statistical Association sponsors special awards for the best use of statistics. The ASA Pittsburgh Chapter hosted a large team of local statisticians, who reviewed the statistical content and merit of all the presented projects. During the first day of review, the judges narrowed the field to around 50 projects that showed a sophisticated level of statistical analysis. Of those, 15 were selected for final interviews the following day with multiple teams of judges. The judges eventually selected three winners and four honorable mentions. In general, the judges were impressed with the quality and variety of the students’ research, as well as their poise and intellectual maturity during the interview process. The first-place award of $1,500 went to Shreya Mathur, 15, from Oxford High School in Oxford, Mississippi, for “Developing a Novel Test to Detect Cancer Genes from Microarray Data.” Mathur developed her own statistic and wrote code in R to implement her simula- tions and analysis. The second-place award of $500 went to Henry Lin, 16, from Caddo Parish Magnet High School in Shreveport, Louisiana, for “A Generalized Holographic Model of Cosmic Accelerated Expansion.” Lin demonstrated his MCMC/Metropolis-Hastings approach with dynamic visualization tools that the judges agreed could be used in any university-level astrostatistics course. A third-place award of $250 went to Nicholas Schiefer, 17, of Holy Trinity School in Richmond Hill, Ontario, Canada, for “Apodora: Markov Chain-Inspired Microsearch,” a text-mining tool with several convergence properties. Honorable mentions were awarded to the following: Mingsha Zhou, 18, from Marianopolis College in Westmount, Quebec, Canada for “Rapid Evolution of Brown Trout in the Kerguelen Islands” Travis Sigafoos, 18, from Champlin Park High School in Champlin, Minnesota, for “A Spectrum of Triangulation: ADHD, Circadian Rhythmicity, and Bipolar Symptoms” Emily Hu, 16, from Lexington High School in Lexington, Massachusetts, for “The Effects of Mindful Decision Making on Post-Decision Regret” Madison Chakoumakos and Zibo Zhuang, both 17, from Oak Ridge High School in Oak Ridge, Tennessee, for “Characterizing the Elements of Earth’s Radiative Budget: Applying Uncertainty Quantification to Climate Models” Pittsburgh Chapter President Rebecca Nugent, from the department of statistics at Carnegie Mellon University, presented the awards to the winners. All winners and honorable mention awardees also received one-year subscriptions to Significance and CHANCE magazines. n

40 amstat news august 2012 section • chapter • committee news sectionnews

Biometrics Han Liu of The Johns Hopkins University for Edited by Songthip Ounpraseuth, Biometrics “The Nonparanormal Skeptic” Section Publications Officer Jennifer Sinnott of Harvard University for “Omnibus Risk Assessment via Accelerated The Biometrics Section recently chose Yang Ning of Failure Time Kernel Machine Modeling” The Johns Hopkins University as the David P. Byar Young Investigator Award winner for “Reducing the The Byar award comes with a $1,500 prize, while Sensitivity to Nuisance Parameters in Nonstandard the travel awards include $800 to go toward the win- To view Likelihood.” ners’ travel to JSM so they can present their papers. section news in The David P. Byar Young Investigator Award is its entirety, visit given annually to a new researcher in the Biometrics Quality and Productivity http://magazine. Section who presents an original manuscript at the amstat.org. Joint Statistical Meetings. The award commemorates Members of the Quality and Productivity David Byar, a renowned biostatistician who made sig- Section (Q&P) invite you to attend this year’s nificant contributions to the development and appli- Fall Technical Conference (FTC) in St. Louis, cation of statistical methods during his career at the Missouri, from October 4–5. The conference is a National Cancer Institute. This year, the committee premier forum in which to discuss topics at the received a record-setting 62 submissions of high-qual- interface of statistics and quality, leading to a more ity papers and chose the following additional travel effective use of statistics to improve quality. award winners: The theme for this year is “Statistics and Quality: Expanding the Horizon.” The program Huaihou Chen of Columbia University for “A includes a range of talks about subjects such as Marginal Approach to Reduced-Rank Penalized experimental design, reliability analysis, data min- Spline Smoothing with Application to Multilevel ing, and statistical process control. Highlights Functional Data” include a plenary address by Connie M. Borror Shuo Chen of Emory University for “A Bayesian from Arizona State University and the W. J. Hierarchical Framework for Modeling Brain Youden Memorial Address by C. F. Jeff Wu from Connectivity of Neuroimaging Data” Georgia Tech. The Q&P invited session on computer experi- Jeff Goldsmith of The Johns Hopkins University ments features William Myers from the Proctor for “Corrected Confidence Bands for Functional and Gamble Company, who will give a talk Data Using Principal Components” titled “The Unique Challenges of Computer Experiments: An Industrial Perspective,” and Min Jin Ha of The University of North Roshan Joseph from Georgia Tech, who will give a Carolina for “Testing and Estimation of Partial talk titled “Composite Gaussian Process Models.” Correlation Networks” Q&P also is sponsoring a short course on Peisong Han of the University of Michigan for October 3, titled Methods and Applications of “Conditional Empirical Likelihood Inference for Generalized Linear Models, which will be taught Unbalanced Longitudinal Data” by Douglas Montgomery. This short course describes and illustrates the use of generalized Yen-Tsung Huang of Harvard University for linear models and shows how this approach is an “Joint Analysis of SNP and Gene Expression attractive alternative to more traditional analysis. Data in Genome-Wide Association Studies” For more information and the complete program, visit http://cba.ua.edu/ftc2012.

august 2012 amstat news 41 section • chapter • committee news

Survey Research Methods Small-Area Estimation (Partha Lahiri) John Finamore, SRMS Publications Officer Address-Based Sampling (Michael Link) The Survey Research Methods Section (SRMS) Sampling for Nonstatisticians (Safaa Amer) webinar series was developed in response to a desire to provide members with the opportunity Reconsidering Mail Survey Methods in an to obtain information about the latest trends in Internet World (Don Dillman) statistical methodology. Nonresponse Bias Analysis (Kristen Olson The inaugural SRMS webinar occurred in and Jill Montaquila) November of 2009, when Mike Brick gave a talk titled, “Dual Frame Theory Applied to Landline Modern Methods for Missing Data (Paul and Cell Phone Surveys.” Since that time, SRMS Allison) has sponsored 10 additional webinars on the fol- The Calibrated Bayes Approach to Sample lowing topics: Survey Inference (Rod Little) The Psychology of Survey Response (Roger Tourangeau) Paradata to Monitor and Analyze Survey Processes (Frauke Kreuter) Surveys, Data, and Indicators from the NSF If you have an idea for a webinar, email your (Nimmi Kannankutty) suggestions to Marilyn Seastrom, SRMS education officer, at [email protected] or John Finamore, SRMS publications officer, at [email protected]. More information about the SRMS webinar series, including registration Biopharmaceutical Applied Statistics details and presentation slides from previous Symposium webinars, is available at www.amstat.org/sections/ srms/webinar.cfm. The section also seeks applicants for the student The 19th annual biopharmaceutical applied statistics travel award in doctoral programs in statistics, sur- symposium, BASS XIX, will be held November 5–9 vey methodology, or allied disciplines. Support is at the Mulberry Inn Suites in historic Savannah, offered for students to attend the Joint Statistical Georgia. At least 16 one-hour tutorials on diverse Meetings, to be held in Montréal, Québec, from topics pertinent to the research, clinical development, August 3–8, 2013. and regulation of pharmaceuticals will be presented Preference will be given to students presenting from November 5–7 by speakers from academia, the a paper or poster at the conference. In addition, applications must be supported by a current pharmaceutical industry, and the U.S. Food and Drug member of SRMS. Approximately three awards Administration (FDA). Four parallel two-day short will be granted to cover conference expenses up courses will be presented November 7–9. Highlights to $600. Winners are expected to attend JSM of the symposium include the keynote address and sessions and the SRMS business meeting. reception on November 5 and the FDA biometrics Application forms are available at www. session on November 7. amstat.org/sections/SRMS/index.html. The deadline for applications is December 3. If you have BASS is a nonprofit entity, sponsored by any questions, contact Jill A. Dever at jdever@rti. the department of biostatistics at Virginia org. Previous student travel award winners and Commonwealth University and the Jiann-Ping JSM 2013 student paper competition winners Hsu College of Public Health at Georgia Southern are not eligible for this award. University. Its purpose is to raise funds for graduate fellowships in biostatistics. Registration is now open at www.bassconference.org. For more information, contact Karl Peace at [email protected] or [email protected].

42 amstat news august 2012 professional opportunities

California Professional Opportunity listings may not exceed 65 words, plus equal oppor- nLABioMed seeks a PhD statistician to tunity information. The deadline for their receipt is the 20th of the month two join UCLA CTSI to direct biostatistics months prior to when the ad is to be published (e.g., May 20 for the July issue). at a research institute in a teaching hos- Ads will be published in the next available issue following receipt. pital. The successful candidate will be Listings are shown alphabetically by state, followed by international listings. expected to provide support for grant Vacancy listings may include the institutional name and address or be identi- proposals, design protocols, perform fied by number, as desired. data analyses, review CTSI proposed Professional Opportunities vacancies also will be published on the ASA’s studies and participate/organize biosta- website (www.amstat.org). Vacancy listings will appear on the website for the tistical seminars for investigators and entire calendar month. Ads may not be placed for publication in the magazine trainees. Mentoring early investiga- only; all ads will be published both electronically and in print. tors and fellows is a priority. Faculty Rates: $320 for nonprofit organizations (with proof of nonprofit status), appointment possible at UCLA School $475 for all others. Member discounts are not given. For display and online of Medicine. Los Angeles Biomedical advertising rates, go to www.amstat.org/ads. Research Institute is proud to be an Listings will be invoiced following publication. All payments should be Equal Employment/Affirmative Action/ made to the American Statistical Association. All material should be sent Drug-Free Employer. Applicants will be to Amstat News, 732 North Washington Street, Alexandria, VA 22314- considered without regard of sex, race, 1943; fax (703) 684-2036; email [email protected]. color, religion, ancestry, national origin, Employers are expected to acknowledge all responses resulting from pub- pregnancy, age, sexual preference, mari- tal status, veteran status, medical condi- lication of their ads. Personnel advertising is accepted with the understanding tion or disability (reasonable accommo- that the advertiser does not discriminate among applicants on the basis of race, dations will be provided upon request). sex, religion, age, color, national origin, handicap, or sexual orientation. Also, look for job ads on the ASA website at www.amstat.org/jobweb. Connecticut nThe Babson College Mathematics and Science Division invites applicants for a tenure-track assistant professor position starting September 1, 2013. The ideal candidate will possess a terminal degree in statistics, analytics, or related field or expect to complete the requirements for the degree by September 1, 2013, a proven teaching record with the desire to deliver statistics and analytics courses, and a strong research agenda. www. click2apply.net/gqgy77n. EOE.

Looking for a Job? The ASA JobWeb is a targeted job database and résumé-posting service. www.amstat.org/jobweb

august 2012 amstat news 43 Join the Massachusetts Conversation nHarvard University Statistics It is now easier than ever to Department seeks a preceptor for become engaged in the ASA and 2012–2013. Applicants should have a its members through social media. minimum of a Master’s degree in statistics, biostatistics, or closely related field, ASA Community—Join and demonstrated excellence in teach- this online setting for ASA ing statistics at the introductory level. members to communicate, For details and link to the Harvard collaborate, and share. University academic positions website, go to https://academicpositions.harvard. Twitter—Follow Amstat News edu/postings/4105. Position will remain on Twitter @AmstatNews open until filled. Harvard University and keep up to date on is an Affirmative Action/Equal the profession. Opportunity Employer. Facebook—Check out Missouri the ASA’s page on Facebook nDepartment of statistics/division to keep up with the latest deadlines, news, and of biological sciences at University of activities and to share your Missouri invites applications for a joint tips and comments with mid-level tenured position in stochastic colleagues worldwide. modeling. We're interested in candidates who use Bayesian approaches to Engage your fellow statisticians and enhance your mind, education, and career at www.amstat.org.

44 amstat news august 2012 model biological problems. A PhD in statistics/biostatistics is required. Apply online at http://hrs.missouri.edu/find-a- job/academic.Include cover letter, vita, statement of research/teaching interests and contact information for three refer- ences. August 27, 2012 deadline. hrs. missouri.edu/find-a-job/academic. MU is an Equal Opportunity-Affirmative Action Employer. Nebraska nAssociate Professor or Professor. Collaborative research (including campus and inter-campus partnerships to pursue external grant funding), teaching, curriculum development and advising in graduate programs in public health. Commitment to scholarly growth is expected. Preferred areas of emphasis include: statistical analysis of high dimensional (genomics/pro- teomics) data and statistical methodology for clinical trials or observational studies. PhD in biostatistics/statistics or equivalent. We encourage applicants to visit jobs.unmc.edu. UNMC is an EO/ AA Employer. Ohio nThe department of quantitative health sciences at Cleveland Clinic is recruiting for faculty, postdoctoral, and master’s-level positions. Many areas are being sought, including biostatistics, data mining, health economics, health status measures, and analysis of population-based registries. Details for all positions, as well as application instructions, are on our website: www. lerner.ccf.org/qhs/jobs. Cleveland Clinic is a AA/EOE. International nFaculty position in mathematical statistics or stochastics. KU Leuven (University of Leuven, Belgium) invites applications for a full-time tenured academic position in statistics at the department of mathematics. We are looking for a dynamic and motivated individual with an excellent research

august 2012 amstat news 45 record in mathematical statistics or stochastics. Details and online application tool: www.kuleuven.be/personeel/jobsite/ vacatures/science.html#2013_2014 Applications accepted until September 27, 2012. Starting date October 1, 2013. EOE.

nWe are recruiting for senior and principal statisticians at Statistics New Zealand (www.stats.govt.nz), in its statistical methods division to strengthen technical leadership and expertise in several areas related to the production of official statistics. Fixed term positions will be considered, and financial assistance for relocation is available. For more information go to the following link: http://careers.stats.govt.nz/detail. aspx?jobId=106830&CoId=1379& rq=1. Questions, contact John Lopdell, Manager, Methodology Development Unit ([email protected]). EOE.

46 amstat news august 2012 nDemand The ASA’s Popular Web-Based Learning Program Is Now Available Anytime, Anywhere

Enjoy previously recorded web-based lectures on an on-demand, pay-per-view basis. Each two-hour webinar features top statisticians discussing their specialties. Available topics include: Bayesian clinical trials Stratiﬁ ed analysis Survey research methods

Please visit www.amstat.org/learnstatondemand for the full catalog and to register.

august 2012 amstat news 47 AUGUST 2012 • ISSUE #422

AMSTATNEWS ADVERTISING DIRECTORY

Listed below are our display advertisements only. If you are looking for job- placement ads, please see the professional opportunities section. For more job listings or more information about advertising, please visit www.amstat.org.

misc. products and services David Anderson...... p. 24 Wiley-Blackwell...... p. 36 The Pennsylvania State University...... p. 12 Social & Scientific Systems, Inc...... p. 18

professional opportunities FM Global ...... p. 46 Iowa State University...... p. 44 MD Anderson Cancer Center...... p. 47 NORC...... p. 44 North Carolina State University...... p. 43 U.S. Census Bureau...... p. 45 Westat...... p. 46 Worcester Polytechnic Institute...... p. 46

software Cytel Inc...... p. 34 JMP, a business unit of SAS...... cover 4 Minitab Inc...... centerfold NCSS...... p. 16 Salford Systems...... p. 23 SAS...... cover 3 StatSoft...... cover 2

48 amstat news august 2012