Reliability and Validity of Dependent Variables

Carefully considering definitions and measurement of your dependent variables is critical. However, even carefully considered dependent variable measurement is not foolproof. Thus, you should properly plan for and monitor data collection throughout your study to increase the likelihood that your dependent variable data are valid, accurate, and reliable. As discussed in the previous chapter, data collection within the context of a single case design occurs repeatedly over time, and almost always via observational recording. Humans observe and record behavior (usually based on researcher-devised systems) and, in turn, make decisions based on those observations. Mark Wolery, a single case researcher who considerably influenced the field of early childhood special education, has said, “humans are the worst data collectors but are often superior to all other options” (2011). Although difficulties are not specific to single case design or repeated observational measurement, the nature of measurement in single case design does pose some different problems than those generally faced by group design researchers. We describe potential issues and discuss the ways that single case researchers can prevent these problems and detect and minimize the effects when they do occur.