Phone: 612-624-5551
unews@umn.edu
24-hr number: 612-293-0831

Advanced Search

Feature

Joan Garfield, statistics teacher

Thanks to the work of educational psychology professor Joan Garfield and other teachers, statistics courses have become student-friendly and fun as well as instructive.

Statistically speaking, they are ahead of the curve

Student raves lead to national award for U statistics teachers

By Deane Morrison

March 13, 2007

As the class begins, the students debate what "Let's Make a Deal" contestants should do if they've just chosen door number three and host Monty Hall shows them that a booby prize was behind door number two. Do the odds favor a switch to door number one or staying with number three? They could ask Joan Garfield, their statistics professor, but she won't tell. Instead, she has the students play the game in class until the numbers make the answer clear. By having students learn statistics by doing it, Garfield, a professor of educational psychology in the College of Education and Human Development, and her colleagues have earned rave reviews from students and a national award for teaching innovation from the American Psychological Association. Presented to the department in February, the Innovative Practices in Graduate Education in Psychology Award was in large part due to Garfield, who was instrumental in reshaping her department's statistics teaching philosophy and style nearly a decade ago. Besides the award, Garfield is also a principal investigator on two grants from the National Science Foundation, one to assess students' statistical reasoning and another to develop class activities and lesson plans. "I think of myself as a revolutionary, trying to overthrow traditional ways of teaching statistics," she says. "I've been given a lot of freedom to do that in educational psychology." Garfield developed a new way to teach the subject and trains graduate students to use similar methods when they teach. "Joan is one of the people who have done the most to bring people together" to revolutionize statistics teaching, says Garfield's colleague Robert delMas, an associate professor of educational psychology. "She's a catalyst." The key to her success lies in never presenting statistics as a frightening or bewildering package of formulas. Instead, students collect their own data and analyze it to learn concepts first hand, making abundant use of computer tools to solve problems. Also, students get together in small groups to discuss their conjectures and test them with real or simulated data.

"Statistics is a fascinating subject," says Garfield. "It's been compared to detective work, investigating data to see what you can learn about the world."

Since 1998 students have reported greater satisfaction with statistics courses in the department, and enrollments have been up. All seven doctoral students who have completed their Ph.D.'s and taken part in the department's statistics training program have found academic positions, and several were told that their unique statistics teaching was a factor in hiring them. Two Ph.D. graduates of the program, Michelle Everson and Andrew Zieffler, are now colleagues of Garfield who have adopted and expanded her methods. In particular, Everson, now in her fifth year in the position, has emerged as a highly respected teacher in the same mold as Garfield. Calming math phobia Whenever Garfield or any of her colleagues remarks to a stranger on a plane or in an elevator that they teach statistics, the normal response is a groan. "I think people fear statistics because they connect it with math," says Garfield. "But you can do very well in statistics and not be good at math."

Data reveals great statistics teachers abound

Joan Garfield has received three teaching awards: Horace T. Morse Award for Excellence in Undergraduate Teaching (1995), the College of Education and Human Development's Distinguished Teaching Award (2002), and the Postbaccalaureate, Graduate and Professional Teaching Award (2006). Other University statistics teachers have also been recognized. Among them are Deborah Levison of the Humphrey Institute, who received an Award for Outstanding Contributions to Postbaccalaureate, Graduate, and Professional Education in 2006, and Jon E. Anderson of the University of Minnesota, Morris, who received a Morse-Alumni Award for Contributions to Undergraduate Education in 2003.

To remove the math phobia, Garfield shows students that they can succeed in statistics by thinking critically, without spending time manually crunching numbers. Take the "Let's Make a Deal" dilemma. On the basis of chance alone, a contestant has a two-thirds chance of winning by switching doors after the contents of one has been revealed. Garfield explains that originally, each door has a one-third chance of hiding a new car. However, the revelation of one door as a dud changes the odds. That is, if the contestant has chosen number three and number two is shown to be a dud, number one's chances of hiding the prize rise to two-thirds. Everson, too, has helped students overcome their fears. One "math hater" student wrote of her: "Statistics is no longer the bitter enemy I avoided facing for so many years ... Michelle Everson is responsible for making this class worthwhile." It's the real thing To give students experience handling real data, Garfield has them measure body dimensions such as head circumference for everybody in the class. The students see that the variability in head sizes primarily reflects the diversity of the students' heads. But while the students all get measured once, the instructor's head is measured by every student in the class. The variability in measurements of the teacher's head size reveals another source of variability, the "error of measurement." When they put their tape measures away, the students have gained a grasp of two key aspects of variability, the most fundamental concept in statistics. In his classes, delMas has students work with the example of an actual company that laid off a number of workers, all over 55. What, he asks, are the chances that age played no part in the decision of whom to lay off? The answer has a strong bearing in the workers' age discrimination suit against the company. "We try to show our students that statistics is a way of thinking about information--for example, in ads or in making decisions about weather forecasts and how they know the chance of rain," says delMas. "Or in evaluating data in order to buy a car. They should ask questions about how the data was collected. Why does one poll say [one thing], and another poll says the exact opposite? We hope students will become critical consumers." The study of statistics also helps students overcome common misconceptions, such as the idea that correlation means causation. For example, says Garfield, a study of neighborhoods showed a correlation between income and the amount of recycling in a neighborhood. But that doesn't mean having more money makes people recycle more. "Statistics is a fascinating subject," says Garfield. "It's been compared to detective work, investigating data to see what you can learn about the world."