My rationale to choose teach statistics with simulation-based methods

Dave Klanderman , Trinity Christian College

It would be accurate to say that I was a skeptic when I showed up at a week-long MAA workshop at Dordt College in June 2013.  At the urging of my departmental colleague, who is now serving as our Provost, I signed up to learn more about a new statistics textbook, a new paradigm for teaching and learning statistics, and a chance to connect with both friends and family in Sioux Center, Iowa.

Additional sessions convinced me that this approach had merit and the comparison data using the CAOS provided the final piece of assessment evidence.

At the dinner preceding the first full day of workshop activities, textbook co-author George Cobb challenged us to rethink our assumptions about a standard first course in statistics.  In particular, he noted that theory-based models are accompanied by validity conditions, some of which have been “tweaked” over the years.  A good example of this phenomenon is the “Plus 4” method for creating confidence intervals for a single proportion in cases where there are too few successes or failures to satisfy the typical model.  George then claimed that simulation-based methods could achieve long-run approximations for p-values that would essentially match the results of the traditional theory-based tests.  Even better, in cases where the approximations did not match theory-based methods, the validity conditions were usually the culprit and made a simulation-based approach more accurate.

This initial discussion piqued my curiosity for the 4-day workshop that followed.  Even in the initial sessions, I appreciated the seamless transition from small sample simulations using coins, cards, dices, and spinners to the larger scale simulations using applets developed by Beth Chance, one of the other authors of the new textbook.  Additional sessions convinced me that this approach had merit and the comparison data using the CAOS provided the final piece of assessment evidence.

I was now convinced to try simulation-based methods.  But how to make such a significant change to my teaching paradigm?  Initially, I thought about simply dipping my toe in the water by using the “Buzz and Doris” exploration from Chapter 1.  However, the more I thought about it, the more I was convinced that it was time to give the entire course approach a try. 

My students also seemed to pick up on the increased engagement in class sessions and the advantage of doing inferential statistics from the very beginning of the course, as reflected in their end-of-semester course evaluations.

As I had already put in my book order for the fall semester, I agreed to serve as a class tester for the new book.  I was part of the control group in the fall with my more standard course in statistics and switched to the new textbook in the spring.  I also had a student assistant that helped me provide feedback to students working in teams on explorations during class sessions.  Her reaction matched mine:  the simulation-based approach to teaching and learning statistics is inherently more accessible and more engaging to students.  Theory-based methods are included in the text and I do make reference to them, but the focus remains on what students can conclude from large scale simulations.  My students also seemed to pick up on the increased engagement in class sessions and the advantage of doing inferential statistics from the very beginning of the course, as reflected in their end-of-semester course evaluations.

I was sufficiently convinced of this approach to teaching and learning statistics that I asked a subgroup of the author team to lead a two-day workshop at my college immediately following the end of the spring semester.  All of our full-time, half-time, and adjunct faculty members in the department joined a dozen colleagues from other local colleges and universities during the workshop.  We have made the switch for all of our learning environments, traditional, adult studies, and online.  I hope that you will consider a similar transition.

Dave Klanderman  (dave.klanderman@trnty.edu) is Chairperson of the Math and Computer Science Department, Trinity Christian College

Leave a Reply

Your email address will not be published. Required fields are marked *