Completely randomized design

This is the simplest experimental design: one pool of units and one randomization. For example, write the names of 60 volunteers on identical slips, shuffle them, and deal 20 slips to each of three treatments. The treatment groups do not have to be the same size, though equal sizes usually give the cleanest comparison. The random assignment tends to balance extraneous variables across the groups and lets you argue for cause and effect.