Simulation, Randomization, and/or Resampling: How Much Is Really Necessary?

• Activity Number: 651
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Education
• Author(s): Christopher J. Malone*+ and Tisha Hooks
• Companies: Winona State University and Winona State University
• Keywords: Introductory ; Inference ; Simulation ; Randomization

Abstract

There has been a recent movement towards the use of simulation and randomization procedures in the teaching of statistical inference. Wood (2005), Tintle et al (2011), Pfannkuch and Wild (2012), and Watson and Chance (2012) present outcomes from the use of a randomization and/or resampling curriculum. Garfield et al (2012) presented outcomes from the CATALST curriculum which is centered on a more general concept of simulation and modeling. This body of work provides evidence of improved learning and gives the impression that wholesale change in the curriculum may be necessary. The authors have investigated how students progress through the learning of statistical inference. Our process begins with the use of simulations to intuitively introduce the concepts of inference followed by an extension of these methods to exact tests. Traditional methods are then introduced as needed. This work suggests that outcomes similar to those presented by the aforementioned authors can be achieved by devoting only a limited amount of curriculum to simulation, randomization, and/or resampling methods.