While a well designed research question is essential, it is the research hypothesis that will actually shape a study.  The following considerations should be taken into account during the process of transforming a research question into a research hypothesis.

• Most hypotheses include all four of the aspects mentioned in “Formulating Clinical Questions”  (an experimental group, a treatment or exposure investigated, a control group, and an outcome to be measured).  While some hypotheses will not need all four of these, it is important that the omission of any one of them be by design rather than negligence.

• While it may seem like the easiest part of designing a hypothesis, selecting a control or comparison group warrants special attention.  The control/comparison group should be as similar to the experimental group as possible.  Consulting with a research scientist is highly recommended to reduce biases and to increase validity.  There are many options apart from the standard randomized trial.

• A hypothesis will include a prediction of the results.  A two-sided hypothesis will not state in which direction the experimental group is expected to deviate from the control group.  A one-sided hypothesis will state an expected direction of deviation