Challenging the various methods for theory testing is as necessary as challenging theory in itself. In order to assess these methods it is first useful to explicitly define what the concept of a theory signifies and map the general framework for theory testing in practice. After having done this, the following discussion will contrast the experimental methods of theory testing against the nonexperimental methods in an attempt to distinguish; in what situations one can be preferred over the other, what precautions that can be taken to compensate for any of their shortcomings and to what extent each method contributes to causal inference. This paper will adopt the classifications of Kerlinger and Lee (2000), and categorize experimental research methods into (1) Laboratory experiments and (2) Field experiments, and nonexperimental research methods into (1) Field studies and (2) Survey research. From the following comparison it is concluded that experimental research methods are superior to the nonexperimental alternatives as they allow for greater level of control throughout the systematic enquiry that scientific research aims to uphold, but also that nonexperimental research has its merits if conducted under strict conditions.
By definition a theory is comprised of many elements out of which one is most frequently targeted when a theory is under scrutiny-the theoryâ€™s assumed relations among variables. A theory is built onto general assumptions of relationships between a number of different variables. Therefore, by breaking up these relationships and forming hypothesis from them, it becomes convenient to draw causal inferences in the form, â€œif x, then yâ€. In this way, testing hypothesises can aid researchers in their persuit to test theory.
The logical validity is virtually the same in experimental as it is in non-experimental research: â€if x, then yâ€(Ibid). However, two defining characteristic o