Meet The Man Who Actually Likes Statistics

By Eve Jacobs

The world of statistics is pretty dry for those not enamored of numerical data. So, it may take a "master educator" to win over the frustrated, the skeptical and the extraordinarily busy. Since most medical students fall into one or more of those categories, a teacher like Bart Holland has no time to rest on his laurels. His students owe a wee bit of thanks to Bronx High School of Science, where Holland says the message conveyed at all levels was that "science and research are not worth much unless you can communicate what you're doing to someone else." The school had its own peer-reviewed and indexed journal, completely written and edited by students, and an excellent debating team and English courses, he says. The science curriculum was strong, but Holland also learned the lesson that would guide his professional life: "Technical skill alone does not a good scientist make."

Holland graduated from Columbia College in 1977 with a major in anthropology and went on to Columbia 's Mailman School of Public Health, looking to do "something scientifically important," but not quite sure what that would be. It was there that he encountered biostatistics. "I kind of fell into it," he says. "Looking back, I can see a reasonable path and a reasonable outcome," he states, although that path was not immediately clear. He earned an MPH from Columbia in '79. Luck played its part, according to the statistician, as did "very good mentoring." Despite Holland not having a math background, the head of biostatistics at Columbia recognized his talents and "encouraged me and gave me very good career advice. I also learned the importance of mentoring."

Part of that advice was pointing the way to Princeton University 's Office of Population Research, where Holland studied population-based statistics. Explaining what that means, he says: "It's learning a set of statistical techniques for examining phenomena that occur in large groups of people." "Whether you marry, whether you commit suicide or come down with a particular disease is a personal, individual matter," he continues. "But you can zoom back from the individual and look at the population." "The number of people marrying, committing suicide, contracting a disease, plus or minus a few, which can also be predicted and will stay the same year to year," he explains. "If the numbers change, there must be a cause." Statistical techniques can then be used to better understand why two populations differ in rates of a particular disease, he states.

After earning his doctorate, Holland worked for the pharmaceutical industry for two years. After this stint in corporate America , he set his sights on academia, where he could conduct his own research and teach, and started at NJMS in 1985. "My first experience teaching was a course in biostatistics and epidemiology for medical students," he remembers. "I realized I could have fun with it figuring out how to communicate statistical concepts in plain English." Holland has continued teaching this course. He uses such creative "teaching tools" as loaded dice and two-headed coins when he discusses probability. "If you toss a coin 10 times and each time it comes up heads, who is to say that it couldn't come up tails on your next 10 throws? There is some uncertainty," he says. He likens the coin-question to the testing of drugs. "Where do you draw the line on the number of people in the test group?" he asks.

Holland 's own research has focused on that question. How do you estimate the sample size that you'll need for certain tests to be valid? There is a "right size" for a sample group in a clinical trial, he explains, "large enough to establish a difference between the test group and the controls, but not too large. You have to specify the number of people before testing." Statistics is a course required of all medical students, so all NJMS medical students for the last two decades have encountered Holland and his "much valued co-instructor Marian Passannante, PhD. When a subject has a reputation for being dry, as statistics does, you need to spice it up," he says. Humor and energy are two of his trademarks in the classroom. "In our class, students get involved in the demonstrations, measuring, counting, giving verbal "bets" on an outcome, doing something," he explains. He also uses technology liberally to keep the course lively. "You have to show the students that the material has practical value," he continues. "They will be reading articles for their entire careers. Understanding statistics is valuable." Holland also teaches "study design" at the UMDNJ-School of Public Health. How to make an experiment statistically valid; sample size; sample selection; how to recruit people into a trial without biasing the selection; and what to do when people drop out are just some of the topics he covers.

So what does a "master educator" do when he's not teaching? Write books, of course! After he published an article promoting rigorous clinical trials of claims of herbal medicine in Nature, Holland was approached by the Johns Hopkins University Press to write a book. Probability Without Equations was the result. The book is a tool to help physicians understand biostatistics, he says. If there's any question in your mind about just how successful a writer he is, visit Amazon.com and scan their "Top 10 Popular Math Books for 2002." Holland 's What are the Chances? Voodoo Deaths, Office Gossip and Other Adventures in Probability was the number 9 seller of the year in that category. Apparently, he has quite a knack for explaining probability and its effect on everyday life this time to a general audience in a lively and non-threatening way. Not only was it positively reviewed in the New York Review of Books, but it was translated into Japanese in 2004. The Korean version is due out soon.

His most recent ventures into the world of publishing have tested his language fluency. In 2004 his English language translation of Debunked! ESP, Telekinesis and Other Pseudoscience by French author and Nobel laureate Georges Charpak hit the bookstores, and in April 2006, his translation of a French history of epidemics and how they've been controlled over time, will be released. So what was the probability that Bart K. Holland , born, bred and educated in New York and now an associate professor of biostatistics and epidemiology in the NJMS Department of Preventive Medicine and Community Health, would become a Master Educator, statistics guru, and popular author? Only "the Master" can tell us for sure!