But every educational experiment is not a success. And that's a good thing. It is imperative that every educational experiment be objective and verifiable, else all of them will be suspect. What better way to demonstrate this than to take note of cases where one of our great ideas failed to improve student achievement? An experiment that is not a success (whether it is carefully planned and controlled or informal and dependent on our personal observations and intuition) lends credence to those experiments that do succeed. It also provides us with data, ideas, and intuition that may guide our next foray into the very dense intellectual jungle of how to help students learn efficiently, effectively, and with long-term retention. If we can figure out why what we did and what the students did was unsuccessful, maybe that will help ensure success the next time around.

I have been thinking about this a lot lately, because almost everything new that I have tried this semester seems not to be working. On exams similar to those given a year ago, or several years ago, students are performing less well. The final exam is a week off, so I don't know for sure. However, all the evidence so far shows that my students this fall are not doing as well as a similar group did last fall. I really don't think the students can have changed that much from one year to the next, so I need to try to figure out which of the changes I made in this course is responsible, or whether there is a negative synergy in the several changes I have made.

There are a number of ways in which this year's course differs from last year's. Because of the way we are giving quizzes, my teaching assistants are spending less time interacting with their students. The group of teaching assistants is completely different, and they have less teaching experience. Students are taking computer-generated quizzes online in a secure testing facility, rather than pencil-and-paper quizzes. I am spending more time interacting with some of the students, because there is an hour-long, optional problem-solving session each week. As a result of time constraints, the students are working less often in groups, and they are not working on challenging group problems. The rhythm of work each week is different because quizzes (which spur many students to study) are given at a different time. Assuming that my perception is real and students are learning less well this semester, any or all of these differences may be important. Or there may be other differences that I am unaware of or have not identified.

What next? For part of my final exam I am using the ACS First-Term General Chemistry Special Exam (available from the ACS Exams Institute at Clemson University). I have used this exam for the past few years. It contains both algorithmic and conceptual questions and will allow me to make a more objective comparison of this year's class with those in previous years. If students do as well or better on the ACS exam compared to last year's students, I will be quite pleased. Nevertheless, I am certain that I can improve on what I have done this semester and I will try some variations in the second semester of this two-semester sequence--as I have every year I have taught these two courses. An even bigger improvement in student performance on the ACS Second-Term General Chemistry Special Exam (which I have also used for several years) will be my goal for next semester.

If my suspicions are confirmed and my students do not do as well as last year's students did on the ACS exam, then I need to examine the course components and organization even more carefully and perhaps to make even more changes. However, I don't plan to go back to doing exactly the same thing I did a year ago. I am sure that I can improve on that, but obviously I have gone off in a somewhat misguided direction, or, at best, implemented good ideas badly. I still expect to approach teaching the next course in a different way, but what I do will be tempered by the negative results I have obtained this semester. I certainly will not teach this course the way I planned last summer, when I was thinking about my courses, organization, and teaching methods for this academic year. I intend to include more group work on challenging problems, and to get back to a weekly rhythm that I hope will help students to budget their time and to study steadily and effectively throughout the semester.

Any experiment has considerable value, provided that we can learn from it how to create new, successful approaches to teaching. Even if we are completely satisfied with a course, there is almost certainly room for improvement, and there is certainly room for variation and experimentation. Without such innovations, no matter how small, teaching would be much less interesting and fun. A failed experiment is far better than no experiment at all!