Too often, experiments in atmospheric science are conducted without careful experimental design, and experimental design usually receives little formal attention in graduate programs. It is difficult to conduct controlled experiments in atmospheric science, and as a result many field projects produce inconclusive results, only suggesting new interpretations but seldom confirming them. The aim of this chapter is to discuss some aspects of a more formal approach to experimentation. This approach provides a valuable model even in those cases where measuring systems are inadequate or weather systems are too unpredictable to permit full use of these approaches.
A distinction is sometimes made between experimental research, involving use of these principles, and observational research, characterized by collecting a comprehensive set of measurements and then using them in exploratory analyses to learn about new phenomena. This does not seem to be a very useful distinction, because most experiments are strengthened by attention to these principles, so far as possible, and only differ in the degree to which hypotheses, critical measurements, statistical tests, etc., can be specified in advance.