Types of research designs

An experimental design is necessary for making cause-and-effect judgments about study variables. Although most researchers would agree that the experimental design is desirable, such a design is not possible or appropriate for every study--and even without an experimental design for support of the hypothesis, enough evidence might exist for a researcher to feel somewhat comfortable that variable X causes Y. At the very least, it can be demonstrated that X is always related to Y in some number of settings, with multiple samples of subjects and with controls for as many variables as possible. Before deciding to change a practice based on research findings, however, the clinician must also determine if X is not related to Y in some cases. Without experimental findings, the amount and quality of evidence available must be considered. This column has described three major types of research: experimental, quasi-experimental, and nonexperimental. Each type of research can produce valuable scientific evidence if the study is designed and implemented carefully. The type of research design and its appropriateness for the specific research question is only one consideration in critiquing research findings and making decisions about the application of those findings in practice settings. Future columns will address other considerations.