Read through the theory and examine the step in logic between the event it describes and the way it describes the event. Determine whether the step in logic is natural and uses reasonably accepted assumptions to make the step. If the step is too large, esoteric or uses its own assumptions as a foundation, read through its data, knowing that it will need substantial empirical data to justify its use. For instance, if you were testing a theory that said that listening to pop music helps you think as a result of brain stimulation, then you would require substantial information to support the foundation that pop music helps you think, because this is counter to traditional logic.
Evaluate the realistic nature of your theory, as well as its general value. Ask yourself if you can test your theory and answer by determining if you can break down the theory into specific parts enough to test it. Assess its general value by determining if the theory would be interesting to the average person. As an example, your pop music theory is realistic because you can test whether listening to pop music can improve your cognitive abilities and even test brain activity while listening.
Consider the power of a theory by examining which questions it seeks to answer. Identify descriptive theories as ones that tell you what something is, explanation theories as ones that tell you how things work, prediction theories as ones that say something will happen and intervention theories as ones that tell you what can be done about it. For instance, your pop music theory is an explanation theory, since it explains how it produces the effect.
Evaluate the falsifiability of your theory by determining if the theory can be retested either through the original testing process or through another similar process. Examine the empirical data supporting your theory, looking for evidence to support the theory. As an example, your pop music theory could be tested and retested, so it is falsifiable through test, which strengthens it as a theory. Unfortunately, without a significant amount of empirical data, it will not survive as a workable theory.