Monday, April 20, 2015

Randomization Tests: Using Technology to Guide Inferential Thinking

In my AP Statistics classes, I have been using simulations as a means to teach various topics in the curriculum. For the past year or so, I've introduced the use of StatKey to do randomization tests.

With StatKey, randomization tests are fast, easy, and results are graphically presented in a clear way. This year, I have been doing an activity in which I incorporate some sort of simulation in almost every unit. It was a great way to introduce the idea of inference early on in the course and has been a nice supplement to traditional inferential procedures as the content progressed. The use of technology here makes statistical inference approachable for a broad range of student levels.

Last December, I presented how I use this technology at the annual Math for America Master Teachers on Teaching. Watch my MT² talk here:


  1. Your introduction to StatKey looks really intriguing. I love the idea of "planting the seeds" of inferential thinking before formalizing it in the fourth topic of the AP Statistics curriculum. I was wondering how you use StatKey to develop the inferential thinking first with your students and then formalize it later with your AP Statistics students. You intimated about it in your talk but only one specific example. How might one use StatKey in a variety of ways early in the year? beginning to rethink some of my topics for next year and hope to put them together over the summer. Your advice would be greatly appreciated. Also curious, do your students find the traditional formalized approach tedious then, when they can get to the answer so much quicker through a randomization test? Love to hear about how you handle their questions around this.

    1. Lynn, I think what I like about randomization tests like this is that it emphasizes the thinking and deemphasizes the computation elements of the traditional inference procedure. When I start inference, especially, I take my time through the four steps. These tests with the help of the software mean I can have students informally do the first and last steps of any inference procedure, without knowing how to justify or do the calculations.

      As for whether or not the students find the traditional steps tedious afterward, maybe. But I also think they feel more at ease with the ideas. Additionally, most of my students tell me that they felt most prepared for the inference topics once they took the exam.