|
Abstract:
|
Many educators are using simulation/randomization-based methods to teach inference, and momentum for these approaches is growing. Textbooks that heavily integrate these methods are now available (Lock et al., 2013, Tintle et al., 2016). These books use simulation/randomization/bootstrap methods to cover inference for one proportion, two proportions, one mean, etc. Furthermore, Tintle et al. (2011, 2012) have demonstrated an improvement in understanding and retention, indicating that the revised curriculum is effective. This leads some to believe that a complete reform of the introductory course may be necessary. Several instructors at Winona State have used simulations to teach inference, but our approach differs. Simulation methods are used early on to introduce inference for proportions. This transitions to exact tests, and then inference for means is taught with traditional methods. This poster will focus on assessment items measuring student understanding at different points in the course. We will present our results and discuss whether an increase in understanding can be achieved by devoting only a limited amount of the curriculum to simulation/randomization methods.
|
