SOAN 2120 Lecture 9: Experimental Research
Document Summary
It is essential that the groups are equivalent: even with random assignment, the two or more groups will not be exactly equal in all respects (there will be random variations among the people in the different groups) If we were to let people choose their own group, we could compare pre-test scores (dependent variable measures) before the experiment begins. This is irrelevant, since it happens automatically with random assignment. Internal validity: there are 6 common threats to the validity of an experiment"s findings they can influence the results unfairly, history, maturation, testing. It also may be an issue if different researchers are used for the two tests since their judgments are different. If respondents are asked to fill out objective questionnaires, the tone of voice of the experimenter varies from one observation to another. Ideally, the experiment will be randomized, placebo-controlled, and double-blind. Post-test only vs. pre-test/post-test: consider a pre-test/post-test control group design.