Because it is usually not possible to study an entire population of experimental units, the researcher must choose a sample and use the information obtained from the sample to derive conclusions appropriate for the population. For the conclusions to be valid, the sample should be representative of the entire population; otherwise, a bias may occur. The researcher must identify biases before the sample is chosen and should attempt to control for them by the proper choice of experimental units. If possible, the experimental units included in the study should be chosen using some form of proba-bility sampling. The most common forms of probability sampling are random, strat-ified, cluster, and systematic sampling. After the sample has been chosen, the re-searcher must assign the experimental factor levels to the experimental units. The appropriate randomization technique depends on whether each experimental unit is assigned one or more than one level of the experimental factor. The appropriate randomization technique will prevent extraneous factors from becoming confounded with factors of interest and thus will result in a valid comparison of the factor levels.
Ronald G. Marks (Thu,) studied this question.