ABSTRACT
The random assignment allows the experimenter to be fairly certain that the participants in all treatments are approximately equal as to subject variables. Randomization procedures used to consist of a table of random numbers or a coin flip to assign participants to conditions (or treatments). Similar procedures allow even rather complex assignment problems to be handled in an unbiased fashion. For example, a group problem-solving experiment was conducted, in which participants met in groups of three to solve a particular human relations problem. Matching techniques can be very powerful in eliminating bias due to subject variables if the experimenter knows what subject variables are highly related to the experimental task and if they can get scores on these variables. The within-subject design ensures that subject variables are equal in all treatments, but this design brings up a problem with regard to the order in which treatments or stimuli will be presented to the participant.
