On Sample Size Determination

Abstract

One of the questions most frequently asked of a statistician is: How big should the sample be? Managers are anxious to obtain an answer to this fundamental question during the planning phase of the survey since it impacts directly on operational considerations such as the number of interviewers required. There is no magical solution and no perfect recipe for determining sample size. It is rather a process of compromise in which the precision requirements of the estimates are weighed against various operational constraints such as available budget, resources and time. In this article, we revisit to the estimate of sample size for various project characteristics. Examples for each are supported numerically.