You can get a research idea by questioning common sense. You can find common sense statements to question by looking at old sayings, editorials, self-help books, cartoons, songs, popular magazines, and television commercials. Box 3.1 (p. 75) shows you how to get research ideas by questioning common sense statements.
You can also use a theory to generate hypotheses. For example, you could try to apply the theory to solve a practical problem or to understand a real life situation. If you don't like the theory, you could look for moderator variables that the theory hasn't accounted for (e.g., "The theory doesn't hold when ..."), try to prove that the theory is wrong about its claims concerning mediating variables (e.g., "people don't really feel an unpleasant arousal called dissonance"), or try to show that another theory makes more accurate predictions in certain situations.
An easy way to generate research hypotheses is to look at studies that have already been published. Because no study is perfect, you can usually do something to improve the internal, external, or construct validity of a published study.
Note that you may start with a general research topic, but to refine that general topic into a hypothesis, you must refine your idea so that it is a testable prediction about the relationship between two or more variables (by testable, we mean that if it's wrong, you can do a study that would find out that it's wrong). Not only should your hypothesis be testable but you should also have at least one good reason for your prediction. If you can answer "yes" to the questions in Box 3.3 (page 105), your hypothesis is probably pretty good.