We introduce the principles of a sound design of experiments, so that the collected data can be deemed suitable for a rigorous statistical analysis.