make a hypothesis about somethinginteresting. It must be something easily measurable and numeric and the hypothesis must be based on either supporting or arguing against an outside source - book, tv, online, whatever.Nothing nominal. And nothing too personal that people are not going to be willing to answer, or answer honestly (cough, response bias, cough). Furthermore, it must be about a MEAN ONLY. And you MUST do this BEFORE the data collection.I also adviseto stick with "greater than" or "less than" hypotheses, as "exact" hypotheses are nearly impossible to support, unless you're dealing with very small data value ranges or an extremely small population.

Examples of good hypotheses:

- Blahblahblah.com claims that Americans own on average 10 pairs of pants. I believe the actual mean is less than that.

- Some tv personality said that college students are awake for 15 hours per day. I believe the actual mean is less than that.

- My psychology textbook indicated that math professors drink at least 5 shots of tequila per day. I believe the actual mean is greater than that.

Notice how each of these examples would have subjects answering my question with a NUMBER.

Note also that you might want to prepare yourself for questions about measures of variables. For example, "do you count shorts as pants," or "how many ounces is a shot?"

Examples of bad hypotheses:

- The average American likes the color blue.

- The average college student has at least 3 family members s/he is ashamed of because of their lengthy prison sentences.