ABSTRACT
The purpose of experiments is to make inferences, as unambiguously as possible, about the effects of treatments applied to experimental units. The units are taken as representative of the population of units to which the treatments may in the future be applied. Randomization is used to remove biases in comparing different treatments and to allow a measure to be obtained of the effects of uncontrolled factors that may contribute to the differences observed between treatment means. Blocking is used to control the effects of nontreatment factors that are known to affect the response measured.
