We Have Some Serious Explaining to Do
Rob Santos, Chief Methodologist, Urban Institute
A search of The New York Times for the word “statistician” during the past 30 days turned up three hits, whereas a search for “psychologist” resulted in 56 hits and “economist” a whopping 204 hits. Do psychologists and economists really have more to say that’s newsworthy than statisticians do? Probably not, yet clearly someone thinks they do. Statisticians have lots of important information to convey, and society would be well-served if we had a greater media presence. Two of my presidential initiatives involve media training for statisticians, and the 2016 JSM President’s Invited Speaker is an award-winning science journalist for National Public Radio, Joe Palca. For this month’s column, I have invited ASA Vice President Rob Santos, who is leading the implementation of one of the initiatives, to discuss these initiatives and provide his insightful thoughts on statisticians and the media. In addition to being an ASA vice president, Rob is the Urban Institute’s chief methodologist and a past president of the American Association for Public Opinion Research. – Jessica Utts
As statisticians, we have some serious explaining to do. Our association tagline proclaims promoting the practice and profession of statistics. It’s an elegant, simple statement with a big footprint that includes everything from teaching to advocacy. It also speaks to a statistician’s role as a public spokesperson. The media thirsts for experts who can explain and visualize quantitative data and insights, yet statisticians generally have not been offered the opportunity to provide their perspectives. Indeed, our association could greatly benefit from bolstering the ranks of statistical spokespeople.
Shout-Outs to Current Media Spokespeople: STATS.org and Others
All too often, we see misinterpretations of statistical results in the media and public discourse. Take the 2016 presidential primaries, for instance. More often than not, the media declared one candidate was “running ahead” of their opponent by a few percentage points when the margins of error did not support such inference. In fact, STATS.org—a partnership between the ASA and Sense About Science—published a blog post on this very issue last fall.
And what about our old friend, the p-value? ASA member Regina Nuzzo published an article in Nature expressing in easy-to-understand terms the over-reliance of statistical significance (alone) in scientific research. More recently, the ASA undertook the onerous task of developing a statement on p-values (download the PDF here) that has enjoyed wide distribution in the scientific community.
There are more examples, but it seems the plethora of misinterpretations and missed opportunities for statisticians to show leadership in a spokesperson role dwarfs the instances in which we have had the opportunity to step forward. The need to expand our visibility in the media is without question. I’ll share some good news on this later, but first let’s consider a statistician’s role as an ‘explainer.’
Target members to receive this training include statistical professionals at all stages of their careers, as well as teachers, academics, and students.
An essential part of being a statistician involves communicating with people about how statistics work, what they can and cannot tell us, and why there are always limitations due to inherent uncertainty. Indeed, our efficacy as statisticians often depends on our ability to digest the needs of a study, translate them into an effective design, and later communicate the results of analyses. The teaching of statistics is deeply rooted in theories of probability and mathematics, as well as applications to real-world problems in manufacturing, government, social science, health, teaching, hard sciences, etc.
Yet, students of statistics are not uniformly afforded the opportunity to acquire and practice communication skills, especially to nonstatistical audiences such as the general public. If we wish to ‘promote the practice and profession of statistics,’ we should be able to speak to different audiences in terms they understand—be they fellow colleagues, scientists from other fields, policy makers and government officials, the media, students, or even the public. But, what is the ASA’s role in and contribution to helping our membership acquire or hone their communication skills?
New Initiatives Inspiring Members to Media Engagement
The ASA is spearheading several initiatives to not only be more publically visible, but to help its members sharpen their media skills. First, the association has built-in infrastructure on its staff with Steve Pierson, ASA director of science policy, and Ron Wasserstein, ASA executive director. Moreover, Jill Talley, the ASA’s public relations manager, recently joined the staff. These individuals each have roles in promoting statistics to a wide range of audiences. But, as strong as the ASA infrastructure may be, that will not suffice. Our members can and should have leadership roles as spokespeople on behalf of the statistical community. And on that front, we have even more good news.
The ASA is launching two initiatives to strengthen the presence of statisticians in the media: the Media Training Program and Statistical Media Ambassador Program. The Media Training Program is designed to provide statisticians with the basics for interacting with, and communicating to, journalists and the public. This forum will cover print, audio, and visual media such as blog posts, op-eds, in-person and telephonic interviews with journalists, and social media. Target members to receive this training include statistical professionals at all stages of their careers, as well as teachers, academics, and students.
Being a media expert involves a long road of practice and preparation. Considering that, media training cannot be an instantaneous, transformative session, but rather a starting point from which one’s skills can more readily be sharpened with experience and reflection over time.
Spanning up to two full days of training and exercises, depending on the types of media training desired (e.g., blogging vs. interviews with journalists), the Media Training Program will kick off at the ASA office in Alexandria, Virginia, this November. As the chair of the work group responsible for developing this program, I encourage you to consider attending, as space is limited for this pilot program. An official announcement will be made in the coming months.
The Statistical Media Ambassador Program, on the other hand, is designed for senior-level statisticians with media experience who wish to take their game to the next level and become a recognized statistical spokesperson for the ASA. Ambassadors in this program will support STATS.org and pledge to be available when journalists engage the ASA for an expert statistical spokesperson. Terms of service are expected to be three years with about a dozen or so statisticians on board at any point in time.
This program consists of formal training, complete with videotaped mock interviews, and provides an exceptional opportunity for senior-level statisticians to stretch and challenge themselves, advance our profession, and just have fun. Wendy Lou, professor at the University of Toronto and a Council of Chapters Representative to the ASA Board of Directors, chairs the work group responsible for designing the program. Like the Media Training Program, the Statistical Media Ambassador Program is expected to formally launch in 2017.
BONUS—National Public Radio’s Science Correspondent Joe Palca to Speak at JSM
Going hand-in-hand with the Media Training and Statistical Media Ambassador programs as a means to promote our profession, grow the capacity of our members, and educate consumers, this year’s prestigious President’s Invited Address will feature National Public Radio’s (NPR) award-winning science correspondent, Joe Palca.
A PhD scientist and journalist, Palca covers an array of scientific breakthroughs and highly technical concepts over the airwaves, doing so in a way that embraces simplicity and verve to maximize audience comprehension and enjoyment. He will present his own experience and perspective on the unique challenges in the scientific news-making process, and his address will elucidate the importance of communicating our value to the media and the public. A can’t-miss session at JSM, Palca will address attendees on August 1 at 4:45 p.m.
My Media Story
Statisticians and data scientists, by the interdisciplinary and meaningful nature of our work, deserve to be trained in media interactions and need not feel overwhelmed in communicating with reporters. My own recent experience working with the media demonstrates the value of such interaction and emphasizes the importance of media training.
In March, I was invited to write an op-ed piece for the LA Times on the state of political polling in the U.S. presidential primaries. The request came on the heels of the New Hampshire and Michigan primaries, where voting outcomes defied pollsters’ predictions. While not having been involved in polling for years, I remained close to the polling community and followed their methods and performance. I also knew that just about everything there was to say from a statistical methods perspective had already been published in the media by polling scholars.
Rather than simply recounting technical details that already had been published, I used this platform to provide thoughts about why pre-election polling results can sometimes be wrong and why we should expect capricious discrepancies to continue. Instead of talking about noncoverage and nonresponse bias or the finer points of prediction modeling for likely voters, I addressed how new technology (including social media) fundamentally changed how people consume and react to information and how concerns about privacy and identity theft altered public behavior in polling and voting. I suggested that heavy use of social media had motivated those who otherwise would not vote (e.g., younger voters) to show up at the polls. Predicting if and when historical nonvoters will vote is difficult using traditional models that heavily rely on past voter behavior. Thus, I offered a framework to explain why past methods won’t work as well in a society that increasingly embraces and is motivated by new technology.
As an ambassador to our profession, I wanted to show the public that statisticians not only know our statistical theory and its applications, but we can also offer insights that help the public understand why unexpected things happen. The op-ed was published Easter Sunday.
Yes, we statisticians need to promote our profession. And it can be challenging, fun, and rewarding. We hope you will join us in the effort to train a cadre of statistical media ambassadors and a new crop of media-savvy statisticians. As I stated at the beginning of this column, we have some serious explaining to do. Let’s get ready to roll!