• Central question(s).
  • If you have more than three, try thinking bigger.
  • Why is the question important? What will we be able to do only if we answer it?
  • Literature review should show that the question has not already been answered. This may involve debunking bogus claims that it has been answered, or only answered for a narrow subset of relevant species, conditions, etc.
  • If you find an important question that hasn’t even been asked, congratulations!
• Hypotheses (plausible answers to the questions above)
  • If you don’t have at least two, are you sure the question hasn’t been answered?
  • If your two hypotheses are “yes” and “no”, rethink your question.
  • Having more than one hypothesis makes it easier to focus on your scientific task, which is to disprove at least one previously-plausible hypothesis. If you only have one, it becomes “your” hypothesis and you may look for results consistent with the hypothesis (though also perhaps consistent with other hypotheses) rather than stringent tests.
  • Give each hypothesis a short name, not a number, to use throughout.
• Predictions
  • A hypothesis is a prediction generator, so each hypothesis should make at least two distinct predictions. If not, look for a more-fundamental hypothesis.
  • Example: spherical-earth hypothesis predicts return of ships from over horizon, circular shadow on moon during eclipse, differences in shadows at same date/time with latitude.
  • Ideally, the link between hypothesis and prediction is so solid that nobody will question rejecting the hypothesis if the prediction fails. This can be difficult in practice.
• Methods
  • Which hypotheses will you test? You may not have resources/methods to test all of them.
  • Which specific predictions will you test first? If you get one false prediction, you can reject the hypothesis without doing more tests (if the hypothesis really makes that prediction).
  • For those hypotheses that survive round 1, what followup tests are planned?

Kinraide, T.B., & Denison, R.F. (2003) Strong inference, the way of science. American Biology Teacher 65: 419-424.