A cost analysis technique for research management and design

A classic research approach is to optimize experimental design and available resources. However, that approach does not guarantee that results will be useful. This article takes the view that research should not be undertaken unless there is assurance that final results will be significant to a level specified before any work is begun. Given that assurance, the relationship between cost and statistical significance is explored, a new approach to optimization s developed, and the use of models in research design is considered.