I have been impressed by the range of experiences expressed in the last half-dozen posts. I will use some of them to prepare my students better for what they are going to go through. While I want to do a better job of teaching, and like everyone, want to be applauded and rewarded, this means doing a better job of using active learning techniques, not a return to passive techniques. I think I have learned important things about teaching and learning statistics, from TBL and elsewhere: 1) people can only learn math or stat by doing (listening, as to a lecture, is totally ineffective at changing any behavior), 2) while I know that statistics is about real life (see VIOXX and flu vaccine below), nearly everyone is convinced it's a hoop to jump through--it's only a mechanical step, 3) team learning has given me a way to establish realistic expectations of students and most students meet them. It is also worth remembering that with TBL, for students to do a good job requires a lot more work, especially at the beginning. TBL requires that students start working the first week. I have seen the results and students learn much more. Not everyone likes it. Regards, David Smith David W. Smith, Ph.D., M.P.H. Associate Professor, Biostatistics Fellow, Institute for Health Policy The University of Texas School of Public Health San Antonio Branch Campus voice: (210) 562-5512 e-mail: [log in to unmask] or [log in to unmask] ________________________________ Other Comments: I regularly get reamed on about 10% of my evaluations. At the same time, using TBL I get some of the best student ratings ever. One of the most common complaints I get is the lack of lecturing. I now give a mini-lecture before each chapter. I still get complaints about not lecturing. Another common complaint is that my class is disorganized. I've never been so organized in my life, even when I taught in a Nursing school, which has the most organized curriculum and classes I've ever seen. No possible criticism of my organization can be justified objectively but I still get a small number of low scores on this item. These complaints, based on item analysis of my evaluations, are always part of responses by students who give me the lowest rating across the board they think they can get away with. They just check off the worst scores. I would do what I can to give them a better experience, but only subject to my standards of acceptable learning (see 1 and 2 above). I a lways give a mini-lecture. I base it on the theory (with some truth) that no one remembers anything after 20 minutes and try to keep it to that length. (I rarely succeed.) That way, I can say I always give a lecture, regardless of the students' comments. My lectures have to cover the formulas and a sample problem. But I also spend time discussing the larger scale role of what we are doing. I talk about the flu vaccine crisis and the VIOXX withdrawal (see below), even when it has nothing to do with ANOVA. No one has ever complained about trying to link the local classroom activities to the larger problems of our world, though I know I don't always do a good job of this. Problem-orientation and active learning are the only way to teach mathematics and statistics. My students MUST learn to do statistical problems themselves. Not just the toy problems in the text book, but making sophisticated decisions that can't be verified in trivial ways (looking in the back of the book). The need to both see and do multiple approaches to statistical methods and data, including numbers, formulas, words, and pictures. They need to present statistical information to others in all these formats. Why? Because that is part of what they will be doing the rest of their professional lives. (If they won't be, then they don't belong in our program. We may not be appropriate for their goals and needs.) Moreover, they need to learn how to deal with statistical problems that will come up in the future, not merely jump a classroom hurdle. My students must also learn to address the real problems of life. How did we ever approve VIOXX for sale in the US? Both its approval and its removal were decisions to which statistics was central. The death rates of VIOXX vs the control were different and statistically significant, but neither rate was particularly high. It took a thousand cases in each treatment group to show a difference. Someone, namely people using statistics, must make these decisions. (No matter what decision is made, some people will die. Don't forget that people died who were taking the other drug.) This kind of problem--the one with no obvious, mechanical answer--is the meat of Introductory Statistics. The simple problems, with mechanical solutions and easily verifiable answers, are only the overture to the real stuff, though necessary. Why? Because it is the only statistics course my students will take before they are turned loose on the world. This doesn't just happen once in a blue moon, we also had a dustup with the flu vaccine last year, where one of the two major manufacturers had its production facility closed (by another country). The production process, while standard, is long and potentially hazardous. Each year, in about Feb or March, someone must decide how many fertilized chicken eggs to purchase in order to grow flu vaccine. (Another statistical decision.) These orders must be filled in a very short time in order to keep production on schedule. (Leading to very tired spring chickens.) Moreover, based on a limited number of laboratory results from October and November of the previous year, the three or so strains of flu that appear to be most likely to circulate next year must be selected. We all saw what happened when one of the two factories supplying the US was disapproved for production in August. Sorry, much too late to order more fertile eggs. Because of the seasonal nature of this particular production process the fertilized eggs simply aren't available in August. What decision do we make then? Do we let the market handle the whole thing or do we turn to an administered system? This is a political and economic decision with statistical components. (President Ford floundered on just this sort of decision regarding swine flu in 1974. Some think it made Carter president. It certainly led Carter to fire the director of CDC when he came into office.) The decisions required, each year, are political and economic and require statistical information and statistical tools. They are not easy and they do not have answers you can find in the back of the book. In any given year there are a lot more examples. None of the options we can choose as a society are happy ones or simple ones, either with VIOXX or flu vaccine. People will die regardless of which choices we make as a society or as individuals. With respect to flu vaccine, recent evidence indicates we might be doing it all wrong anyway, maybe we should be vaccinating all school children and skipping the adults. This approach led to fewer deaths in Japan. Statistics is about life (and of course death) with all its vicissitudes. This is the most important thing I can teach my students.