Explaining Sampling Distributions*Mark C. Fulcomer, Ph.D., UMDNJ/Restat Systems, Inc.
S. David Kriska, Ph.D., Restat Systems, Inc.
Marcia M. Sass, Sc.D., UMDNJ
Keywords: illustrating statistical concepts, sampling distributions, computational issues
Although they are important to statistical applications in general and to hypothesis testing in particular, sampling distributions remain difficult for practicing statisticians to explain to beginning students and clients alike. Despite the availability of simulation software, some communications problems with key aspects of sampling distributions are direct consequences of the enormous and complex computational work their construction entails (Fulcomer et. al, 2007 and 2008). In addition, other complications stem from the large number of statistical topics requiring mastery (e.g., probability theory) before a fundamental understanding of sampling distributions can be effectively grasped. Beginning with spreadsheet illustrations of the well-known binomial and Poisson distributions, this presentation explores treatments of this concept, from older “classic” sources (e.g., Hays, 1963) to current textbooks, emphasizing the sequencing of preliminary learning tasks and their mastery. Some ideas for illustrating other examples of sampling distributions to consumers of statistical methods are also presented.