Building the Statistical Work Force: Increasing the Pipeline of U.S. Students
At the November 2012 ASA Board of Directors meeting, past-president Bob Rodriguez and I were pulled aside by ASA Public Relations Coordinator Jeff Myers. A Boston Globe reporter was hoping to talk to someone for a story on “the growing interest in statistics.” Her timing was fortuitous, and Bob and I arranged a phone interview with her.
As is often the case, most of what we said did not make it into the final piece. Luckily, however, our plugs for the International Year of Statistics and the 2011 McKinsey Global Institute report, projecting 140,000–190,000 new positions for individuals with data analytic expertise by 2018, did.
The Globe article is just another in a series on our field over the past few years. In The New York Times alone, careers in statistics have been getting some great press: “For Today’s Graduate, Just One Word: Statistics,” (2009); “What Are the Odds That Statistics Would Be This Popular?” (2012); and “The Age of Big Data,” (2012).
This press might be contributing to growing interest in statistics among U.S. students. As Bob’s President’s Corner columns in the August and September 2012 issues of Amstat News report, the number of high-school students taking Advanced Placement (AP) Statistics has soared, increasing 7% in the last year alone, and record numbers of undergraduates are majoring in statistics in several large departments nationwide.
This is fantastic news! However, a persistent concern for more than a decade has been that U.S. students are not pursuing advanced degrees in statistics, especially PhDs, in sufficient numbers to meet the demand. Even if the surge in interest at the lower levels is leading to more U.S. students going on to graduate study of statistics, reports continue to suggest it is not enough.
A recent Nature article cites a National Institutes of Health (NIH) survey of statistical geneticists who report considerable difficulty recruiting qualified candidates for research positions. Directors at our U.S. federal statistical agencies repeatedly cite the dearth of U.S. citizens with the requisite skills for positions there. Reports by the NIH Advisory Committee to the Director, Biomedical Workforce Working Group, and Data and Informatics Working Group call for more PhD biostatisticians and other quantitative scientists in health sciences research and greater NIH investment in relevant training programs. Leaders of biostatistical units in research institutions nationwide report their struggles to fill positions.
A decade ago, this concern was taken up in two NIH workshops to “examine the need to train more biostatisticians in the U.S.,” summarized in a 2006 Statistics in Medicine article documenting the flat production of PhDs in biostatistics and the heightening demand. Workshop leaders called for efforts to enhance awareness of careers in biostatistics for those with advanced training among U.S. undergraduates.
In 2003, responding to this challenge, the National Heart, Lung, and Blood Institute (NHLBI) issued a request for applications (RFA) for a Summer Institute for Training in Biostatistics (SIBS), which would host undergraduates majoring in quantitative areas for six weeks, expose them to statistical principles and career opportunities in biostatistics, and encourage them to pursue graduate training. Three SIBS programs were awarded to Boston University, the University of Wisconsin, and a joint North Carolina State University and Duke Clinical Research Institute (NCSU-DCRI) team.
For the past nine summers, I have co-directed the NCSU-DCRI program and witnessed how exposing undergraduates majoring in not only statistics, but also mathematics, engineering, and a host of disciplines, to the opportunities for PhD biostatisticians can inspire them to pursue graduate training. Through lectures on statistical methods and case studies presented by practicing biostatisticians and clinicians; field trips to sites such as DCRI, the U.S. headquarters of GlaxoSmithKline, and SAS Institute; and data analysis projects based on cardiovascular disease studies coordinated at DCRI, we showcase the excitement and rewards of a career in the field. Our sister programs exploit their own local resources to do the same.
The results have been inspiring. From 2004–2009, 409 students attended these three programs and more than 65% went on to graduate programs in biostatistics or statistics. Some, no doubt, would have regardless, but program evaluations and follow-up communications tell an encouraging story. I could fill pages with comments like these: “I’m leaving with a whole new appreciation for statistics. I never realized there was so much out there!” “I wasn’t sure of my future plans before this program, but after participating in SIBS, I am undoubtedly going to pursue a PhD in biostatistics.” “Without SIBS, I would certainly not be on the path I am now.” Numerous SIBS participants are now in graduate programs in statistics and biostatistics across the United States, and several have earned PhDs and hold positions in industry, government, and academia.
This success led to a new SIBS RFA in 2009, resulting in eight programs in 2010–2012 at the original three sites plus Emory University; the Universities of Iowa, Pittsburgh, and South Florida; and Washington University in St. Louis. In 2010–2011, there were 329 participants; among those eligible, more than 60% had gone on to graduate school by fall 2011. Six programs have been funded for 2013–2015; sites offering programs in 2013 are listed on the NHLBI SIBS website.
SIBS has exposed students to the exciting possibilities presented by advanced study of statistics work. So much so that past-president Nancy Geller proposed as one of her presidential initiatives that the ASA design a solicitation and identify funding sources for a SIBS-like program for statistics more generally.
My SIBS experience has led me to follow up on Nancy’s proposal with my own initiative. It has been more than 10 years since the NIH workshops, which focused on biostatistics. While the anecdotal reports of an insufficient number of U.S. students pursuing advanced statistics degrees continue, there has been no recent, systematic effort to document current and future work force needs, assess the adequacy of the pipeline to meet them, identify fruitful recruitment strategies (like SIBS), and describe training experiences that will best prepare these students for the expectations of employers in business, academia, and government.
With so much positive coverage of our discipline, the time is ripe. By JSM 2013, a workgroup chaired by Lance Waller (also a SIBS director) with members in academia, industry, and government plans to finalize a white paper similar to the 2006 article addressing these issues, including identifying potential funding sources for recruitment and training programs. The white paper will serve as a tool for engaging stakeholders and forging innovative cross-sector partnerships focused on achieving the needed statistical work force. An initial meeting of stakeholders will take place in late 2013, possibly at JSM.
Please contact Lance if you have ideas or suggestions. I look forward to reporting to you on the workgroup’s progress.
If you are interested in learning more about SIBS, look for an article in CHANCE magazine, Volume 26, Number 1, or visit my website for slides from presentations I’ve given. And if you mentor undergraduates, encourage them to apply!