Chapter 2 Summary

One way to get a research idea is to be skeptical. In this chapter, we show you a variety of ways to tap into your skeptical side. We also show you a variety of ways of directing that skepticism toward (1) conventional wisdom (common sense) and (2) published research.

When you first try to generate ideas, you should focus on quantity rather than quality. Try to come up with as many ideas as you can without judging them. Once you have a pool of ideas, then you can start judging them. Even the best of your ideas will probably have to be refined to make it a suitable research hypothesis.

In developing your research idea into a research hypothesis, you must be sure that your hypothesis is:

• testable
• reasonable: following logically from common sense, previous research, or theory
• relevant to theory, previous research, or real life

1. The problem may be that you need to improve your hypothesis. The book has many tips for doing so, but our favorite is to add a moderator variable (to learn what moderator variables are and how they can help you, see page 62).

2. The problem may not be with the hypothesis, but with your ability to explain why your hypothesis is interesting.

If testing your hypothesis seems impossible or unethical, don't abandon it. By following the advice on pages 63-67, you may be able to rescue your hypothesis. Although most hypotheses can, with a little tinkering, be developed into hypotheses you can test, there are two hypotheses to be cautious about:

1. null hypotheses: You can't prove that a treatment has no effect or that there is no difference between the effectiveness of two treatments. So, don't hypothesize that a treatment has no effect or that two treatments have the same effect.

2. cause-effect hypotheses: You can make cause-effect hypotheses. Indeed, your professor will probably require that you make a cause-effect hypothesis (In a cause-effect hypothesis, you state that changes in one variable cause a change in another variable. Rather than saying "causes a change," you might say "increases," "decreases," "improves," "affects," "makes," "changes," "helps," or "hurts.") However, to test a cause-effect statement, you must be able to manipulate/control/administer the cause (That is, you must have a treatment that you can administer to some participants and withhold from others). Therefore, you can't test the hypothesis that changes in employment status cause changes in self-esteem because you can't manipulate people's employment status.
   In short, we encourage you to make cause-effect hypotheses, but be sure that you can manipulate your suspected cause (if you are having trouble thinking about how you could manipulate your suspected cause, the tips on pages 65-66 may help. If you want more specifics on how to manipulate your treatment variable, skip ahead to Chapter 3).

Back to Research Design Explained Home Page