ABSTRACT

Quasi-experimental is the term for some non-experimental designs used to study the impact of an independent variable when subjects cannot be randomly assigned to conditions. One alternative explanation for the results of the shoplifting study has to do with the behavior of people other than the shop-lifters. It is possible that store owners and police became more reluctant to charge a person with shoplifting because they knew the person’s name would appear in the newspaper. The essence of the quasi-experimental design, however, lies in the various control groups we include to eliminate alternative explanations. In most cases these alternative explanations arise from the researcher’s inability to assign people to conditions randomly. Quasi-experimental designs are essentially a patchwork. The task for the researcher who uses a quasi-experimental design is to think of as many alternative explanations as possible. The researcher must then think of ways of collecting the data needed to minimize or rule out the alternative explanations.