Variable Analytic Tradition

In the simplest sense, a variable analytic approach to research means that researchers build scientific explanations by considering the effects of a variable or set of variables on another variable or variables. One application of this approach is the simple scientific experiment in which the researcher manipulates one variable known as the independent variable to observe the effect on another variable called the dependent variable. For example, researchers might expose two audiences to the same persuasive presentation, but manipulate credibility of the message source by leading one audience to believe the speaker is highly qualified while leading the other audience to believe that the speaker is a low-credible source. In this way, the researcher can observe the effect of credibility on dependent variables such as attitude toward the topic of the presentation.

Experiments designed to determine cause and effect relationships represent the purest form of science. The variable analytic tradition is important because it was the fulcrum, or pivot, for the shift to the scientific study of communication. The variable analytic approach continues to serve an important [Page 995]role in contemporary communication research. Although the variable analytic approach is important to the progress of scientific communication theory, it depends on a set of central assumptions.

The central assumptions underlying the variable analytic approach to communication research include the following: (a) social interaction can be reduced to distinct, measurable variables, and (b) these variables are related in an organized and systematic way rather than in a random pattern. Although it is convenient to describe the variable analytic approach in terms of simple cause-effect relationships between two variables, the overwhelming majority of experiments in communication research are far more complex. Most manipulations of independent variables involve more levels of the factor. For example, rather than merely presenting high- and low-credible presentations, researchers might present high-, moderate-, and low-credible versions. Furthermore, researchers typically manipulate more than one independent variable in experiments known as multiple factor designs. In an attitude-change study, for instance, both source credibility and language intensity might be manipulated so that all combinations of levels of the independent variables are presented to audiences. Multivariate designs allow researchers to investigate the effects of one or more independent variables on multiple dependent variables at the same time. Richard Rudner described research that is conducted for the purpose of theory testing as operating in the context of validation.

However, in the 1970s, a selection of variables included in published scientific studies of communication often relied very little on formal theory in the selection of the variables. That is not uncommon, as Rudner observed, early in the evolution of a discipline or field. At that stage, researchers are most interested in the basic relationships among variables. Rudner describes this research phase as the context of discovery. The simultaneous emergence of the variables analytic approach and a scientific approach to the study of communication led to the erroneous depiction of the variable analytic approach as atheoretical. In contemporary research, however, choices regarding which independent and dependent variables should be included in a study and how those variables should be measured or manipulated depend heavily on existing theory. In fact, the vast majority of experiments published in scholarly communication journals in the past 20 years are explicitly designed to test specific theories.

...