Sample Size Determination in Hypothesis Testing and Estimation

Sample data may be collected with different objectives in different scenarios. In some cases, one may wish to collect enough observations that would guarantee minimal prespecified reliability for the conclusions derived from the data. On other occasions, in addition to specified constraints regarding reliability, an optimal value is desired of some objective function that would guarantee that the sampling procedure is optimal in some sense. In all these cases, a major decision variable is the required sample size. In this article, sample-size determination is addressed, relating, in particular, to four often-encountered scenarios: hypothesis testing, parameter estimation, relational modeling, and optimal sampling