The Extraordinary Power of Statistics

With more than 3,400 individual presentations arranged into approximately 181 invited sessions, 400 contributed sessions, and 500 poster and speed presentations, the 2016 Joint Statistical Meetings will be one of the largest statistical events in the world.

In addition to the 45 parallel sessions taking place during most of the meetings, there are other activities you can add to your program for a fee: Professional Development courses, roundtable discussions, the Career Service, and workshops.

This year, the exhibit hall will be the place to be. The Opening Mixer will take place there, and we’ll also have Spotlight Chicago, which will feature events throughout the week. Moreover, if you are looking for a way to help the local community while at JSM, you’ll want to visit IMPACT CHICAGO, also taking place in the exhibit hall. Finally, just outside the exhibit hall, we’ll have an art show featuring data artists.

Here are a few more highlights to let you know what to expect. We hope to see you there.

Featured Speakers

Monday, August 1

10:30 a.m. – 12:20 p.m.
Medallion Lecture
Tracing Pathways of Dependence: How Far Did We Get?
Nanny Wermuth, Johannes Gutenberg-University/Chalmers University of Technology

Tracing pathways of dependence to understand development was a main aim of geneticist Sewell Wright when he formulated—a century ago—linear generating processes, represented them by directed graphs, and evaluated the fit to his data. This approach started to be generalized with graphical Markov models in the 1970s, permitting variables of any type, by using the concept of conditional independence and combining directed with undirected graphs. We have now a most suitable subclass, named “traceable regressions,” to model development in ordered single and joint responses together with a set of context variables. A main difficulty was to find special, testable properties of the generated distributions needed to concentrate on conditional dependences in addition to Markov structure. Traceable regression includes linear regressions, generalized linear models, subclasses of structural equations for longitudinal studies, and models for planned and virtual interventions. Here, we use several examples of studies to illustrate and summarize the now available features of traceable regressions and to point at open research questions.

Tuesday, August 2

4:45 p.m. – 6:15 p.m.
ASA Deming Lecture
Profound Knowledge from a Knowledge Use Perspective
Vincent P. Barabba, Market Insight Corporation

This presentation will focus on the need to improve the manner in which knowledge developed from statistical practice is presented so that the statistically based knowledge will be effectively used to improve decisions.

8:00 p.m. – 9:30 p.m.
ASA Presidential Address and Founder & Fellows Recognition
Appreciating Statistics
Jessica Utts, University of California at Irvine

“Appreciating Statistics” as the title of this talk is meant to convey a two-fold meaning. First, as statisticians, we have much to offer, and we need to make sure the importance of what we do is apparent to a broad range of audiences. Historically, we have not held a visible place of prominence with policy makers, the media, the public, or the wide range of professionals who could benefit from understanding statistical information. We all can play a role in publicizing what statistics has to offer and fostering appreciation for what we do. The second meaning I hope to convey is that the value of a degree in statistics and related fields is appreciating in multiple ways. According to the Bureau of Labor Statistics, “statistician” is projected to have the ninth-highest growth rate from 2014–2024 among all professions and the third-highest growth rate among those requiring a college degree. U.S. News and World Report ranked “statistician” as the number-one best business job based on a combination of financial and work-life issues. We need to educate students and those who help them make career decisions about the many tangible and intangible benefits of a career in statistics and encourage a diverse workforce to meet the needs of our profession. The ASA is implementing multiple initiatives to encourage and advertise both types of appreciation of statistics, but as 2013 ASA President Marie Davidian noted, all of us can play a transformative role in promoting our discipline. Find out how you can help!

Wednesday, August 3

2:00 p.m. – 3:50 p.m.
Medallion Lecture
Model Averaging and Post-Model Selection
Gerda Claeskens, KU Leuven

Several choices have to be made such as “which and how many estimators to average over” and “which weights to use.” Data-driven weights can be chosen by minimizing an estimator of the mean squared error. In general, those weights are not unique. We prove there are multiple weight vectors that yield equal model-averaged estimators in linear regression. A restriction to singleton models results in a drastic reduction in the computational cost. If we take into account that the weights are random variables, rather than fixed during selection, we show the averaged estimator is biased, even when the original estimators are unbiased, and its variance is larger than in the fixed weights case. This relates to the “forecast combination puzzle”; there is no guarantee that the weighted averaged forecast will improve on the original forecasts. The distribution of model-averaged estimators is, in general, hard to obtain. We work out the special case of an estimator after model selection by the Akaike information criterion AIC. We exploit the overselection properties of AIC to construct valid confidence regions that take the model selection uncertainty into account.

4:45 p.m. – 6:15 p.m.
COPSS Fisher Lecture
Personalizing Disease Prevention: Statistical Challenges
Alice S. Whittemore, Stanford University School of Medicine

The recent presidential allocation of U.S. resources for precision medicine reflects a national focus on personalized health care. Patients and their doctors are increasingly basing such care on statistical risk models that use a person’s lifestyle and genetic covariates to assign him or her a probability of developing a disease or other adverse health outcome in a given future time period. The use of such personal risk models will increase as we learn more about the genetic and epigenetic causes of disease, and as the routine sequencing of peoples’ entire genomes becomes practical. In this talk, I will describe some of the statistical problems that arise when evaluating the accuracy and utility of these models. These problems would have interested Sir Ronald A. Fisher, who did much of his seminal statistical work while serving as the Arthur Balfour Professor of Genetics at the University of Cambridge.