CSUF Department of Psychology
Unit Banner


What Makes a Good Question/Hypothesis?

There are several components of a good research question. What is meant by "good" is one that is useful for conducting research. Here, the term 'research question' refers to a question form of a hypothesis. (Remember, a hypothesis is simply a testable explanation for a phenomenon. Really, there's very little difference between a research question and a hypothesis, so don't get confused by the terms.)


A good research question...

1. Answers something new! Replications, by themselves, are not good enough. The point of conducting research is to advanced knowledge--to test something in a unique way, to test the prediction of a theory, to explore new methodology, etc. It is extremely important to show that a scientific finding can be replicated. Any valid scientific finding needs to be replicable...or in other words, reliable. The best kinds of research studies both replicate previous research while incorporating something new. That's not always possible in a single study; sometimes it takes two or three experiments to accomplish this. (That's why most journal articles will contain a number of research studies rather than a single study.)

2. Is based on and builds upon previous research. A good research question tests the prediction of a theory. A research question that you generate off the top of your head can be useful, but almost always, research questions developed after an understanding of previous research and theory is much stronger and more relevant to science (and will be viewed more favorably by reviewers of your research).

3. Is practical (e.g., Can you really do this study this semester? Do you have the equipment/resources to sufficiently answer your question? Will you need a zillion subjects for the study?)

4. Is often simple. You don’t have to answer the “big” questions of psychology – just add a little piece to existing research. A good piece of advice when conducting research is "Keep it simple!" You don't want to be left with a pile of data that you don't know how to interpret. (You'd be surprised at how often that happens!)

5. Is based on primary (not secondary) sources. Primary source basically refers to the original source of research, usually from a publication. A secondary source, on the other hand, is a summary or description of the primary source. Newspapers, magazines, and topic books are secondary sources; journal articles are usually primary sources.

6. Is specific. If you're conducting an experiment or doing a correlational study, the research question should state a relationship between variables and say something about testing those relations. It must be specific without being too wordy. It's a good idea to incorporate some aspect of the operational definition of terms in one's research question. This means that you simply can't throw out terms like "love" in your question without specifying what you mean by it and how it will be measured.


"Does overlearning lead to decreased performance?" ["Overlearning" refers to continued learning once it is 100% learned.]

This says something about the relationship between variables: a) overlearning and b) decreased performance. This also says something about testing the relationship between those variables: "lead to" (or, in other words, does overlearning cause decreased performance?).

"Are females smarter than males?"

This research question is not that good. It does include variables: a) gender, and b) intelligence, and it does say something about the direction of the relationship (i.e., "smarter than") but it does not tell us what is meant by "smarter than." There are MANY different ways of operationally defining intelligence, one of which should be included in this hypothesis. A better research question would be "Do adult females score higher than males on the WAIS-III?" (The WAIS-III is a widely-used standardized intelligence test.