Quasi-experimental research designs

In complex naturally occurring health-care situations, it is often difficult, and sometimes impossible, for ethical and/or practical reasons to meet the requirements of manipulation and control needed in true and ideal group experiments. If researchers wish to investigate causal relationships in such circumstances then quasi-experimental designs are recommended. The designs represent approximations of true experiments because they do not meet all the necessary requirements, such as random allocation of subjects to groups, control and manipulation. Consequently, quasi-experimental designs are not as powerful as true experimental designs in establishing causal relationships but nevertheless allow such relationships to be considered by a process of rejecting rival hypotheses. This article outlines the two main approaches to quasi-experimental designs and provides an example of each. Other methods related to the quasi-experiment are also highlighted.