Innovations in teaching: Alison Gibbs
Statistical science associate professor is having too much fun to go back to traditional lecturing
No point in calling Alison Gibbs at home to congratulate her on being one of the 2016 winners of the President’s Teaching Award.
“I don’t know about your home line,” the associate professor (teaching stream) said in her office in Sidney Smith Hall. “But I rarely pick up mine.”
Indeed, many Canadians no longer have such lines, a fact that is probably related to the failure of any polling company to predict the size of the Liberal majority in the federal election of last October.
“We’re not getting a decent random sample anymore, because of the way technology is going,” Gibbs explained. “You need other sources of information.”
Which is only one of the reasons this is an interesting time for the department of statistical sciences and for the study of statistics in general. Others have to do with the rapidly expanding range of activities to which statistics are newly relevant.
“Every transaction in a business is now being recorded in a huge dataset,” Gibbs said. “Business people want to use that and leverage it to their advantage.”
Medical research more than ever is an enterprise driven by statistics. English professors can study an author’s evolving vocabulary over time with statistical science. All this while the new analytical methods have changed the way baseball teams are recruited and managed.
“Moneyball was brilliant,” said Gibbs, referring to the system adopted by the Oakland Athletics and documented in a 2011 movie and a book by Michael Lewis. “Figuring out measures other than the traditional ones, that really worked.”
New applications have resulted in a rise in enrolment in STA courses at U of T from fewer than 500 four years ago to more than 2,000 in 2015-16. Only human biology courses in the Faculty of Arts & Science have higher enrolment.
The increase has required changes to the curriculum. Gibbs is described in her nomination letter as “the leading innovator of statistics education renewal in the department.”
Gibbs is so enamoured of teaching that in 2002 she gave up a tenure-track position at another Canadian university to take a teaching job at U of T, where she had earned her MSc and PhD. Her present U of T appointments include associate chair of undergraduate studies in statistics and chair of the undergraduate curriculum committee.
Nor are her activities confined to campus. Gibbs organized the “census at school” initiative of the Statistics Society of Canada, an effort to put high-school students in touch with StatsCan. On a global scale she is vice-president on the International Association of Statistical Education.
In her department Gibbs has led the “inverted classroom” revolution that shifts basic information to online learning and uses class time for active problem-solving.
“It’s really fun,” she says of the inverted model as it is applied to STA 220: The Practice of Statistics I. “I don’t think I could go back to straight lecture teaching ever again.”
Peer instruction, as the technique is known, provides results in real time. If 80 percent of students get a question right, the other 20 percent are aware instantly of their need to revisit the material.
“The most interesting peer-instruction questions are those which nearly everybody gets wrong, or that are spread evenly between right and wrong answers,” Gibbs said. “Then you get students talking to each other.”
The tonic ingredient is the enthusiasm, for which Gibbs is widely praised in course evaluations.
“I initially dreaded taking this class but it turned out to be my favourite class of the semester,” said one undergraduate. “It was all due to the excitement Gibbs had for statistics; it was infectious.”
Ashley Cohen, a double U of T alumna with a BSc in actuarial science and statistics and an MSc in statistics, credits Gibbs in part for her success after school.
“Not only was her commitment to her teaching and her students exemplary,” Cohen says, “but many of the skills and techniques I learned in her classes have had a hugely positive and helpful impact on my professional career as a biostatistician.”
While computers might seem to be a clear aid to data collection, the steady increase of computing power has created conditions that are “a little messier” than they once were. The need for careful instruction in statistical science has never been keener.
“Computation is at an whole new scale now because the dataset is often so large that you can’t do it on your desktop,” Gibbs said. “A lot of people are looking at data where not a great deal of design has gone into the collection process.”
While the interpretation of data now plays a larger role in many industries, it also is important to individuals and their private lives.
“Even to be an informed consumer and an informed voter, we need to understand statistics,” Gibbs says. “To take ownership over your health care, you need to understand statistics.
“Health studies seem to contradict each other. Back in the 1980s, the question was: Should I be eating oat bran? More recently it was Vitamin D. An issue now is colonoscopy and colon-cancer screening. There is evidence that just the stool sample test is sufficient. Colonoscopies are expensive and unpleasant. What’s the evidence that tells us we should do them?
“There is a famous quotation by H.G. Wells about how the ability to compute and think in averages will be as important as the ability to read and write. This is becoming more and more true all the time. Because people are doing more and more with data.”