Demonstrating the gambler's fallacy in an introductory statistics class

In this article, we describe a classroom demonstration that uses the Gambler's Fallacy to illustrate misconceptions about random processes and how they affect statistical interpretation. The demonstration used a database collected from simulated gambling by students picking professional football games with the point spread (i.e., a real-life random process). The results of student picks illustrated that random processes are not self-correcting and reinforced the relation between sample size and variability. Formal and informal feedback from students indicated that the demonstration was well received and recommended for future classes.