Develop Your Hypotheses

There may be a delay in getting access to the finalized data if you or your classmates have not already submitted your data to the instructor.  You do not need the finalized data for your section in order to develop your hypotheses.  In fact, you would normally develop your hypotheses BEFORE you collect any data at all!  In this case, you will be developing a study based on a subset of a larger study.  We are taking an existing dataset and asking questions of it.  To do such a study, all we need to know prior to writing hypotheses is what the variables are how the data was collected.

In writing your hypotheses, you need to ensure that they are “testable”.  In this course, “testable” means that your hypotheses can be tested using either a Spearman Correlation or U-Test, which are tests that you already have experience with from the Statistics, Figures, and Modeling Lab.  If you and your partner BOTH have experience with more advanced statistical techniques (learned in other courses), then you may adapt your hypotheses and methods to those techniques, but only with the permission of your instructor.

In order to have a cohesive study, your two hypotheses must share either a dependent variable or an independent variable.  Your study will also be stronger (easier to interpret) if both hypotheses are testable using the same statistical test.  These experimental design structures are summarized with examples on the next page.

Be deliberate here.  Don’t just choose the first (or second) variable that comes along.  You have been