Ingram Olkin: Mentor to Many
Amy Munice, ALM Communications
Crotona Park in the East Bronx of New York City was known as the home of the Detroit Tigers baseball icon Hank Greenberg. To Ingram Olkin, however, who moved near Crotona Park when he was 10 years old, the improbable but true fact is this area was the incubator of statisticians, and that he can name a dozen prominent statisticians who grew up within a mile or so of the park is what makes it of interest. Indeed!
What’s not so improbable is that Ingram Olkin came to know these Bronx boys—and they all were boys then—because he became their friend, colleague, or collaborator over his 65-years-and-counting career.
Whether these other Crotona Park sons wrote in their high-school yearbook that they wanted to be a statistician is unknown, but Ingram did. Ingram notes this with a tinge of marvel today as he marks his 90th year. Barely a week has gone by since 1947 when he has not been engaged in statistical work of one kind or another.
Ingram’s accomplishments are legion in the world of statistics—from being instrumental in launching prominent statistical journals (Journal of Educational Statistics, Statistical Science, etc.) to oft-cited work in multivariate analysis, inequalities, reliability, and meta-analysis. In some capacity, Ingram has mentored many of today’s statisticians—possibly more than many other statisticians.
It could be that Ingram finds it especially amusing that he wrote of his statistical career aspirations in high school because he knows it was in no way a foregone conclusion that his life would work out as it has. He doesn’t hail from an academic family. Quite the contrary, he was the only son of Lithuanian and Polish immigrants and the first in his line to go to college.
World War II created the only other career temptation Ingram has experienced, but it also gave him a leg up on completing his undergraduate mathematical studies and moving on to graduate work in statistics. Ingram explains:
I was born in 1924 and, when I was 18, it was World War II. I had started City College of New York in 1941 as a math major, which generally meant you were deferred from the draft. But then the Army said they needed people in radar and meteorology, asking us to enlist to get into one of these programs. A sophomore, I was inducted in 1943 to study meteorology and became a forecaster in the Army’s Air Force until 1946, when I was discharged.
The Army had sent me to MIT, which had many programs to train soldiers in specialties. I took a lot of math courses there and was able to get notes from my instructors that I had passed those courses so that, when I got back to City College, I received many credits and was able to graduate in one year, then going to Columbia for graduate school.
As a meteorologist, I was an officer at different airports. One station was in Stout Field, Indiana; another was LaGuardia, New York, and later I was in California. The key task for a meteorologist was to draw a weather map for a pilot who wanted to go from point A to point B. I did think about continuing in meteorology, but, by then, I was already immersed in statistics.
Ingram quips, “There still is an affinity for meteorology here—my wife (Anita) likes to watch The Weather Channel morning, noon, and night.”
A mentor to many, there were mentors in Ingram’s life who no doubt would have been disappointed if he had decided on a different life course. Selby Robinson, a math professor, was an early influence. Jules Joskow at the City College of New York also spent a lot of time with Ingram, spotting him as a student with deeper interest than others. It was at his urging that Ingram went to graduate school at Columbia.
When pressed, Ingram will tell you his ultimate mentor was Harold Hotelling, a name well known in all statistical and economics circles and also to automobile drivers in Chapel Hill, North Carolina, where a street is named for him. Ingram followed Hotelling from Columbia to The University of North Carolina at Chapel Hill and was one of the beneficiaries of Hotelling’s firm belief in paving the way for others.
Ingram reflects, “When you grow up, a lot of what you learn is by osmosis. You see what others do and you grow up with it, many times adopting similar positions. But when you ask me who the real mentor was for me, I think, ‘It was Hotelling.’ I always think of him that way.”
Osmosis of Hotelling’s ways then might help explain how Ingram came to be a mentor to so many during a career that included stints at Michigan State University and the University of Minnesota, followed by his years at Stanford dating back to 1961 and continuing to this day. He supervised 35 PhD students and served on many dozens of other doctoral thesis committees. Ingram is especially proud to report that many of his mentees later became chairs of other departments of statistics in the country.
Many may know that Ingram has played a particularly strong role in mentoring female statisticians, but few know to credit the CIA for feeding his drive to make this happen. No, not the Central Intelligence Agency, but rather the Culinary Institute of America, which his eldest daughter left Stanford to attend in hopes of becoming a master chef. She later found herself unemployable in the high-end culinary field. That the culinary world is rife with sexism and ubiquitous sexual harassment in the kitchens was something the Olkin family had to grapple with firsthand. Ingram has not one, but three, daughters—Vivian (now 64), Rhoda (a polio survivor and expert in psychology and disability, now aged 61), and Julia (now 55 and the only one of his daughters to become a mathematician).
All of my daughters faced little obstacles here and there because they were female. It was clear that there was discrimination against women in small ways or bigger ways.
Years later—at The University of North Carolina, Stanford, and Columbia—it became apparent there were no women on the faculty, even though a lot of the students were female. At one point, it occurred to me that women were often alone—the only woman—in math and statistics departments in both small and big schools and they were having difficulty getting tenure.
At that time, I was working closely with the National Science Foundation (NSF) and was able to get them to sponsor bringing three to five women to Stanford during the summer for one to three years. We encouraged these women to interact with Stanford faculty. When they later came up for tenure in their institutions, they could often get letters of recommendation from the Stanford faculty with whom they had worked. These weren’t necessarily women from the biggest name schools, but rather from places like Kansas, Texas, or Delaware. It was a very successful program, and the NSF created programs to entice women into the mathematical sciences. The NSF itself was changing, with many female program directors within their own organization.
In fact, I was just asked to write a letter of recommendation by one of the women I had invited to Stanford 35 years ago who is now being considered for a distinguished professorship. Overall, this was one of the most successful initiatives in my career and I consider it a major personal achievement.
How Times Have Changed
Ingram reflects upon this and other ways in which the profession has changed a great deal. He says:
In 1948, when I started my graduate studies, I was able to go to the math library and read every math or statistics journal. Today, that’s an impossibility, and this is really a problem to some degree. For example, in 1940, there were approximately 2,500 biomedical journals and today there are about 80,000.
There’s been a 15–20-fold increase in almost every field of applied or theoretical statistics. Keeping up is now a problem. It’s difficult to browse. Now there are search mechanisms to help you find something, but that isn’t the same as browsing where you would serendipitously find materials that are of interest.
When I was in graduate school, one of the faculty leaders—Gertrude Cox—had a vision of forming an Institute of Statistics, which would include theoretical and applied statistics as well as psychometrics and biostatistics. Her advice, which I didn’t follow but now see as being a very good idea, was for statisticians to find an application or field to concentrate on. This makes sense because every field has unique statistical problems. In psychometrics, for example, the problems are related to tests and measurement. In forestry, they have to estimate the amount of timber, which presents a different problem to solve than assessing if a drug works. Each field has its unique aspects.
If I were to give advice to students working towards their PhDs, I would urge them to be closer to the faculty and find out the interests of each faculty member before they decide what to work on. The interaction between students and faculty is a very important part of the learning process.
Yet Ingram’s career path was anything but one of specialization. He recounts:
I would get into new areas every 10 or 15 years … and, during my career, I was always passionate about the area I was working on at the time. Today, my focus is on meta-analyses, which is a particularly hot topic in medicine.
… Generally, I’m a theoretician, so I don’t get involved in specific medical applications, but rather with how to analyze that kind of data. My papers are theoretical papers that came out of actual applications. Like other theoreticians who work on methodology, I hope that the methods will be used by others for their own data.
… One of the plusses of being in statistics is that it applies to all kinds of fields and it tends to make you more of a generalist in your thinking. You get an interest in medicine, sociology, psychology, and so on. Personally, this has meant that I know people from many fields, as does my family. Perhaps being a generalist in this way has helped make me more open to the paths each of my daughters has chosen for her career.
The students of today are different, and the faculty of today are different. Today, there are many more two-career families, and that wasn’t the case 40 years ago. Then, the university was more central to the lives of faculty than it is now. Today, many commute to the university they teach in and have far more interests outside the university.
The biggest change impacting students is technology. Statistics is now coupled to computing and students tend to go into both fields jointly now. Years ago, computer science abilities didn’t figure large in a statistical life.
Looking back, Ingram expresses no regrets. He says, “I was blessed in my career. I had good mentors and colleagues and super students. It’s not clear to me what I could have done differently to improve on any of those because it is your teachers, colleagues, and students—your friends—who become important in your life. I’ve been fortunate to have a super group in each of those categories.
When I first started at Stanford, we all lived on the Stanford campus and all our neighbors were faculty and staff or somehow university connected. It was very good—great in fact. The bad part is that I have now survived many of them. We now have new friends of our vintage that come from other walks of life. In my department, however, the next-oldest faculty member is 75. That’s a big gap in age. This is just one of the problems of aging.
Then again, not many—any?—90-year-olds are invited to their students’ weddings or asked for letters of recommendation. Ingram doesn’t quite know which relationships he has had with students that would qualify as mentor-mentee. Does the fact that his work continues to be widely cited count? Is it only PhD students, or undergraduates as well? What about younger colleagues with whom he has collaborated and then written letters of recommendation for? However you define it, Ingram Olkin has mentored many.