Epidemiological studies based on small sample sizes--a statistician's point of view.

Abstract

We consider 3 basic steps in a study, which have relevance for the statistical analysis. They are: study design, data quality, and statistical analysis. While statistical analysis is often considered an important issue in the literature and the choice of statistical method receives much attention, less emphasis seems to be put on study design and necessary sample sizes. Finally, a very important step, namely assessment and validation of the quality of the data collected seems to be completely overlooked. Examples from veterinary epidemiological research and recommendations for each step are given together with relevant references to the literature.