Worry over the future of a small Statistics Department probably ceased when Harvard offered Donald B. Rubin a senior professorship. Boasting a string Ivy credits and a ten-year stint at the Educational Testing Service (ETS) in Princeton, N.J., Rubin adds a practical knowledge of statistics and a willingness to share his experience.
The newest addition in a department of four senior and three junior professors, Rubin stresses the importance of "recon-structing the bridges" that Professors C. Frederick Mosteller and William G. Kochran built at the department's inception 25 years ago.
"I have know him since he was a graduate student," said Mosteller of Rubin, "and the areas that he is interested in, the social sciences, interest me as well. I expect his joining us to increase the relationship between the Statistics department and the social sciences."
Once Statistics department chairmen, Mosteller is now primarily active in the department of Public Health, Rubin said. Kochran retired in 1976.
"There is an enormous amount of statistics being taught at Harvard and very little by Harvard statisticians," said Rubin. Though "not necessarily bad," he continues, "it suggests a market" for those trained in statistics. It's a profession he describes as "nice for someone who has a broad interest in the sciences."
Teaching at the University of Chicago, from 1981 to 1983, Rubin says he was part of a "very tight" campus. It's geographical 'centralization, he continues, "makes it easier" to build bridges. At Harvard, he says one "wastes time in travel."
He said that Harvard's statistic department's small size and poor geographical location prevents bond-building with other departments and schools.
Half the size of University of Chicago's, Harvard's Statistics department distributes more administrative and department responsibilities per person. It's another reason why he describes Chicago's department as "more fluid."
Besides administrative projects, Rubin will keep busy with teaching duties and personal projects.
He is teaching a course in survey methods and participating in a consulting class that meets once a week over lunch. It's a chance for grad students, faculty members and undergrads to "bring in problems" of a statistical nature and discuss them, he says.
Although he describes it as a "friendly department," Rubin said that one of the "issues in rebuilding is to do a better job of undergraduate teaching." He added that he is "pretty accessible" to students.
His most recent research at ETS considered the efficacy of coaching on exam results.
The Federal Trade Commission (FTC) in 1980 considered filing suit for false advertising against Stanley H. Kaplan Educational Center, Inc., Kaplan claimed a test score increase of 50 points for those students who took advantage of their coaching.
Rubin's research validated the effects of coaching but credited Kaplan with only a 10 to 20 point increase. He attributes the increase left to inevitable "maturation."
Praising the SATs as exams that are not elitist and that help equalize the college admissions process, Rubin says. "It's easier to be prejudiced and discriminatory on the basis of recommendations than among anonymously graded tests."
Consulting for government agencies such as the Internal Revenue Service, the Census Bureau, the Department of Labor and the National Institute of Health. Rubin is also researching nonresponse surveys and causual inference.
One of the problem existing in nonresponse surveys, he says, occurs in polls like the Census which "undercount minorities" because of nonresponses.
"I've got a technique called multiple imputation, I think is very useful in practice," he says. "One fills in a series of things which represent your uncertainly about what the right thing is"
A graduate student in statistics at Harvard in 1970. Rubin likes working in Cambridge, which he describe as "an exciting and intellectually active area." Currently he is tearing down the inside new house in Newton and caring for his children which, he says "absorb the time that used to be for hobbies"