Achieving statistical power through research design sensitivity

Abstract

The challenge for nurse researchers is to design their intervention studies with sufficient sensitivity to detect the treatment effects they are investigating. In order to meet this challenge, researchers must understand the factors that influence statistical power. Underpowered studies can result in a majority of null results in a research area when, in fact, the interventions are effective. The sensitivity of a research design is not a function of just one element of the design but of the entire research design: its plan, implementation and statistical analysis. When discussing factors that can increase a research design's statistical power, attention is most often focused on increasing sample size. This paper addresses a variety of factors and techniques, other than increasing sample size, that nurse researchers can use to enhance the sensitivity of a research design so that it can attain adequate power.