James Cochran

Jon Kettenring joined Drew University in 2004 as a fellow in the Charles A. Dana Research Institute for Scientists Emeriti, known as RISE, and has served as its director since 2008. From 1969–2003, he worked in industrial research at Bell Laboratories, Bellcore, and Telcordia Technologies. His primary research focus has been on applied methods for analyzing multivariate statistical data. From 1984–2003, he managed research groups in statistics, economics, computer science, and information analysis at Bellcore and Telcordia. Kettenring also held visiting appointments at the University of Washington, the University of Minnesota, Stanford University, and the University of Michigan. He earned his BS and MS degrees from Stanford University in 1961 and 1962 and his PhD from The University of North Carolina in 1969, following two years in the U.S. Army.

Throughout his career, Kettenring has been especially interested in the infrastructure that supports the statistics profession. He served as president of the American Statistical Association in 1997 and vice president from 1990–1992, chair of the Board of Trustees of the National Institute of Statistical Sciences from 2000–2004, chair of the Committee on Applied and Theoretical Statistics at the National Research Council from 1993–1996, member of the Mathematics and Physical Sciences Advisory Committee at the National Science Foundation from 2003–2006, and chair of the ASA Accreditation Proposal Review Panel from 2008–2009.

In 2005, the National Institute of Statistical Science (NISS) organized The Future of Data Analysis: A Conference in Honor of Jon Kettenring in recognition of his multiple contributions to the profession and NISS. He is a fellow of the ASA and American Association for the Advancement of Science, and, in 2001, he received the ASA’s Founders Award.

You are one of the few statisticians I have met who earned all of her or his degrees in statistics. How did you find this discipline so early in life? Did you intend to major in statistics when you entered Stanford as a freshman, or did something happen during your undergraduate education that led you to major in statistics and study statistics as a graduate student?

I started out thinking I would major in a branch of engineering, but nothing excited me especially. My junior year, I had the opportunity to study for six months in the Stanford-in-Germany program—probably the most important experience I had as an undergraduate student. After returning, I still needed a major and a plan to graduate. I had dabbled a bit in statistics as I explored industrial engineering. I was intrigued with the idea of dealing with uncertainty. So, in short, I fell into statistics out of curiosity and necessity.

You were executive director of several departments at Bellcore and Telcordia Technologies. How did you, with your education and experience in statistics, wind up in these managerial positions? What aspects of your training and experience in statistics were particularly valuable to you in this position?

A tradition at Bell Labs that carried over to Bellcore and Telcordia was that managers in technical areas should have technical backgrounds. That sounds simplistic, but it’s really important from a cultural perspective. Also, I enjoyed working with people from different disciplines. So it was natural for me to take on such assignments. They involved managing research groups in economics, software engineering, and other aspects of computer science, as well as statistics. Half the battle in such positions is to listen well and ask good questions—something that is very natural for statisticians. In fact, I believe statisticians often make excellent managers because of the way we think about problems.

In 2002, you, Bruce Lindsay, and David Siegmund produced a report on a workshop about the future of statistics that was held at the National Science Foundation (NSF). One focus of this workshop was educational reform, and issues identified included developing adequately trained teachers for AP Statistics courses and ensuring statistical literacy of instructors in other subjects in grades K–12, the need for integrating statistics throughout the K–16 curriculum, and expansion of statistics programs and options at both the undergraduate and graduate levels. Do you think our discipline is making progress in these areas? What activities in these areas have impressed you, and what activities would you like to see undertaken?

The two students shown here with Kettenring—Richa Patel (left), an economics and math double major, and Liz Pemberton (right), a physics major and math minor—spent last semester together studying multivariate statistics.

The instigators of the workshop were Marianthi Markatou and John Stufken, who were serving as program officers at NSF. They recognized that a forward-looking study of statistics—a Vision 2020—was needed at the foundation. It was hinted that, without such a report, our field was at a disadvantage. So the timing seemed right, and fortunately a broad representation of the community was able to participate. Bruce served as chair of the program committee and David, along with several other distinguished statisticians, were members. Following the workshop, I worked with Bruce and David to produce a report of the proceedings. An abridged version of the full report appeared in the August 2004 issue of Statistical Science. Writing the report was a lot more difficult than it should have been, and it took considerable time. One of my roles was to pull together the education section you mentioned. As a friend likes to remind me, I had zero qualifications for the task—and he was correct! Nevertheless, the educational reforms we included still feel right to me. If anything, perhaps they are understated. Now that we are in the era of Big Data and data science, there seems to be greater urgency to the opportunities for reforming our entire educational effort.

I would especially underscore the need to invest more heavily in high-school statistics education. I recently suggested doing this as the leading element of a six-pronged plan for developing future leaders. I believe it would help introduce more of the best and brightest students to our field at a young age. You can find more details in Leadership and Women in Statistics, which was edited by Amanda Goldbeck, Ingram Olkin, and Yulia Gel and published by CRC Press in 2015.

You were associated with the National Institute of Statistical Sciences (NISS) for many years. What was the nature of your involvement? Why did you devote so much time to it?

I served on the board of trustees for about 10 years and as chair of the board for five. NISS is a very important part of the infrastructure of statistics and has been a leader in cross-disciplinary work since its inception. One of its nicest contributions has been its post-doctoral program. It is also a strong partner with SAMSI [Statistical and Mathematical Sciences Institute]—both are housed in the same building in Research Triangle Park. I was happy to help create the NISS Affiliates Program, which brings together academic, government, and industrial interests under the NISS umbrella. I also enjoyed working with the directors of NISS during that period, Jerry Sacks and Alan Karr, who did so much to nurture and grow the institute to its present stature.

You are a fellow and director of the RISE program at Drew University. What is RISE, and what is its mission?

RISE stands for the Research Institute for Scientists Emeriti. It’s a group of researchers from various scientific disciplines who have retired from industry. Our mission is to mentor undergraduates at Drew who want to get involved in research. Most of us worked in either the pharmaceutical or telecommunications sector, both of which have a strong legacy in New Jersey. One of our members, Bill Campbell, recently received the 2015 Nobel Prize in Physiology or Medicine for work he and others did at Merck. Most of our students head to graduate or medical schools after graduation. The RISE program began 35 years ago. More than 400 students have participated. We have a lot of fun together, and it is a great way for retirees to stay active!

You chaired the committee that developed the ASA report on voluntary individual accreditation of statisticians. What led to the establishment of this committee? How was the ASA’s current accreditation program ultimately developed? How did the ASA decide what would be required for accreditation, and what led to the establishment of two levels of accreditation—Accredited Professional Statistician, or PStat, and Graduate Statistician, or GStat?

Our committee was a follow-up to a previous one chaired by Mary Batcher that recommended the ASA start an optional accreditation program. We took a fresh look at the opportunity, surveyed about 1,000 members, and recommended that the ASA Board launch an optional PStat program. We made this suggestion to a large extent because more than 40% of the respondents indicated they would apply for such a program. The ASA now also offers the GStat program, which we had considered but deferred. We also made recommendations for PStat requirements that mirrored those in the Batcher report. After the board approved our proposal, there was a lot of follow-up work to be done to launch the program and its processes. Iain Johnstone chaired the committee that brought the accreditation concept into practice.

Today, there are more than 350 PStat- or GStat-accredited statisticians. While this is a lot less than 40%, it does emphasize what Sally Morton, ASA president in 2009, said was one of the program’s main attributes, namely to provide a way to serve some of the underserved groups in the association.

I would like to add that the committee I chaired was probably the most effective one of all the ASA committees I served on over the years. While the issue under study was controversial, we managed to approach it in an open-minded way through a series of conference calls. The other members of the committee were Mary Ellen Bock, Roger Hoerl, Nancy Kirkendall, Bob Mason, David Morganstein, Vijay Nair, Bob O’Neill, Len Oppenheimer, and Ron Wasserstein.

What were the biggest issues you faced as president of the ASA?

That’s a hard question, especially given it was so long ago—1997! As a board, we dealt with quite a range of issues. One priority that I would label big was to complete, refine, and operationalize a new strategic plan that had been initiated by Lynne Billard in 1996. Its main themes were to enhance the reputation and health of statistics, support professional statisticians, and improve the efficiency of the ASA. Each theme had specific sub-goals, such as getting involved in policy issues. An issue that cut across the plan was how best to deal with the challenges and opportunities posed by electronic technology. We were able to use the plan to guide decision making and set funding priorities throughout the year. David Hoaglin led this effort as chair of the Strategic Planning Committee, and many others were involved on the committee and various specialized task forces.

At a more personal level, a big issue I tried to emphasize, as many from our community already fully understood, was the growing intersection of computer science and statistics, the increasing size and complexity of data sets, and what these trends might mean for us. It was for this reason I chose Alfred Aho, a distinguished computer scientist, to be the president’s invited speaker at the JSM in Anaheim and why I reiterated the words of John Tucker of the National Research Council in my presidential address at the same meeting that massive data problems would be the grand challenge for statistics in the 21st century. Now, nearly 20 years later, I think we can agree he was right on the money—even though massive seems to have been downsized to big in the press. For more about these developments, see the 2013 National Research Council report, “Frontiers in Massive Data Analysis,” published by the National Academies Press.