Random sampling issues in a federal court case, a case study

Abstract

The authors of this paper, statistics professors at Bryant University, recently had the opportunity to act as expert witnesses in a case involving the Internal Revenue Service (IRS) as the plaintiff. The IRS used some sampling techniques in selecting a random sample that did not appropriately represent the population at hand. This action led the IRS to draw inferences about the population that were not likely conclusions on their part. The purpose of this paper is to highlight the fact that fundamental mistakes are made in the business and legal world regarding sampling. This legal case is a good case study to present to any statistics class, revealing both the pitfalls of inappropriate statistical sampling techniques, and incorrect inferences made based on an inappropriate sample. Basic random sampling techniques are developed in elementary and intermediate statistics classes, and the following paper highlights an example on random sampling that can be used in class, and is understandable by all students.