|
Abstract:
|
R. A. Fisher's pioneering book, The Design of Experiments," published in 1935, laid the foundations of the modern design of experiments. The basic principles were randomization, replication, and local control. The importance of randomization, as a way of reducing biases in carrying out studies on humans, was early recognized by Bradford Hill. Today the randomized clinical trial is regarded as the "gold standard" for carrying out clinical trials. This lecture will discuss models in which the patient populations and hospitals are random samples or simply collections, and the implied limitations on making statistical inferences. In reality, the randomization process may serve as the only basis for making statistical inferences as the patients and hospitals are not random samples.
|
