Chapter 4 focused on the properties of outcomes, with a strong emphasis on describing reliability. Reliability can take on many forms, but the most common form encountered in health care and exercise science is test-retest reliability. Relatedly, Chapter 4 discussed ways in which outcomes are used in the clinical setting and the value of knowing the clinimetric properties of outcomes. Finally, that chapter addressed the concepts of accuracy, bias, and validity. Without reliable and valid outcomes, the foundation of research is weak. Thus, the hallmark of any line of research is establishing that the outcomes to be used are reliable and valid and measure qualities that are important.
The concept of validity is not limited to outcomes. As you will see, validity also applies to experiments. This chapter addresses the structures and methods used in experimental research. It discusses different options a researcher has in designing an experiment, as well as the three types of validity associated with experiments: internal, construct, and external. Each of these is important in different ways, and each determines the degree to which research results can be trusted and applied to your own set of circumstances. The content of this chapter lays the groundwork for understanding how to categorize evidence into the levels previously discussed. It addresses the strengths and weaknesses of the typical research designs so you can distinguish good research designs from poor designs.
After reading this chapter, you should be able to answer these questions:
What is an experiment?
How is cause-and-effect established?
Do the research results apply to my situation?
What is meant by "experimental validity"?
How does the experiment fit into the levels of evidence?
Experimental research is one of the most common forms of research encountered in exercise science. It is often considered synonymous with laboratory research. Although much laboratory research is experimental, as you will see, it is frequently nonexperimental as well. The classic definition of an experiment is research with at least two separate groups. One of the groups must serve as the control group, which receives no treatment or a standard treatment. The other serves as the experimental group and must receive a treatment that is different from that of the control group. The groups must be formed by random assignment of subjects. Research designs that do not meet these criteria are not true experimental designs and are defined as nonexperimental or quasi-experimental, depending on the structure of the design. Examples of these types are given later.
Experimental research: Method of research in which ...