This paper describes a distribution fitting exercise that has been used in an undergraduate introductory simulation class. The intent is for students to collect data describing a customer arrival process, with the goal of determining whether the exponential distribution is a good fit. The paper briefly reviews the memoryless property, and presents some teaching tips directed toward students' understanding of the concept. The data collection exercise is then presented, with attention to common pitfalls students have encountered in the past. A presentation of the variety of distributions that emerge as "best fitting", when samples are actually drawn from an exponential distribution, serves as a warning against over-reliance on goodness of fit measures.
Timothy S. Vaughan