## Curious David Redux: Musings and Swan Song about Teaching Statistics

I draw upon this material in the introductory chapter of a book I am working on.

Dec 7, 2009

As I’ve documented elsewhere, I have now taught more than seventy times a course required for the Carroll undergraduate psychology major. Can it really be that it was more than 30 years ago that my first student assistant, Larry Jost, and I proudly announced in our “journal,” *Occasional Papers in Psychology*, our successful translation of some 40 BASIC statistical analysis programs (obtained from a psychologist in South Africa)? We were so thrilled that we were able to introduce them into my PSY205 class using my TRS80 Model I Level II microcomputer with 16K RAM!

There exist a number of excellent resources on the Web, today, for supplementing a course like “mine.” Among my favorites is the Rice Virtual Lab in Statistics. I also find very well done and useful for our students the online resource Research Randomizer.

I’m wondering, though, whether it is time (or past time) for me to pass along the course to someone else. Indeed, is it perhaps time for the course to “go away” (though parting is such sweet sorrow 🙂 )— either by assuming that its content is addressed by other courses (such as are offered by my colleagues in Mathematics) or by merging it with another course, say, a two semester psychology research course. Both ideas have been proposed and discussed with faculty colleagues in recent years.

Though I think I have been unusually successful at developing rapport with students in a course whose subject matter is often anxiety-arousing, anyone teaching the same course time-and-again runs the risk of losing touch with new ways of teaching, failing to understand changes in student needs and student abilities and neglecting to capitalize on new and perhaps better ways of teaching the subject matter.

Is time to teach the course in a way more in “sync” with students of today to help students avoid choosing the wrong statistic? Should I offer the course in a totally online format? In a hybrid format ? Is there continued value in teaching a “canned” statistical package such as SPSS? Might it make more sense to teach the data analysis capabilities of open source statistical software freely available online? Should I incorporate more applications or draw more upon the work I do for Schneider Consulting which I strongly believe enriches my teaching? Should I make room for the more systematic teaching of effect sizes and statistical power?

I’d love to incorporate material on ethical issues of data analysis, provide opportunities to discover inappropriate data analyses, create opportunities for action research, and help students realize the limitations of statistics. Might it be worth my following through on my idea of a number of years ago of creating a highly trained group of undergraduate, cross-disciplinary, student data analysis **consultants**? Do I delude myself in thinking that my “Statistics and Experimental Design course” is the most important class that I teach?

I invite former and present students of mine who have taken this class with me to share their insights for possible new directions for this course.