Feinstein A.R. Clinical biostatistics

Saint Louis : C. V. Mosby, 1977. — 488 p.In the exchange that occurs when clinicians and statisticians collaborate to conduct a research project, the statistician is regularly asked for a great deal of mathematical activity. If the project is being planned prolectively, one of his assignments is to determine “sample size” by estimating how many patients will be needed for the research to attain stochastic significance.17 If the project is a clinical trial, the statistician is asked to prepare a suitable arrangement and schedule for the randomized allocation of the compared treatments. After the research data have been collected (or if he is first consulted after the data are already available), the statistician is customarily given the job of performing a suitable “statistical analysis”. For this job, he usually develops ways of sorting the data to display the results in various tables or graphs, and he then administers the statistical ritual used for testing the “significance” of what has been found.