From my experience, the concept of directional hypothesis testing is a challenge to learners who are somewhat new to using statistical concepts to prove or disprove a hypothesis. One important guideline to keep in mind is to be sure to state the direction of the relationship between the independent and dependent variables. Are you simply trying to prove that the independent variable does affect the dependent variable in some way? Or, are you trying to prove that it affects it in a specific way, for example the outcome of the dependent variable is either greater, or less, than its previous state because of the influence of the independent variable. For example, yesterday I read a new American Medical Association study that showed that vitamin E supplements actually increase the risk for a particular type of cancer. Previous studies, that were not as rigorous, had demonstrated the exact opposite. In this case, the researcher is clearly trying to show that there is a one-way relationship between the independent and dependent variables - that more vitamin E means a lower risk of cancer. The original researchers were correct in the fact that vitamin E does affect the frequency of cancer, but their hypothesis stated it going the wrong way!

However, there are some studies that might require a less specific approach, as the researcher only wants to determine if there is an effect in some way - not the direction of the effect. For example, what is the effect of after-school sports programs on standardized test scores? Here, the researcher is not stating that after-school sports affects test  scores in a particular way, so this is a two-sided hypothesis test.

In summary, in order to conduct useful research, be sure to state the direction of your hypothesis statement!

More about a hypothesis statement in a practitioner research study:

http://www.businessweek.com/news/2011-10-11/vitamin-e-supplements-boost-prostate-cancer-risk-study-says.html

More about hypothesis statements:

http://www.public.asu.edu/~kroel/coursesb/coursesa/www500/HYPOTHESIS%20Fri.pdf

Dr. Daniel J. Smith, MA, MBA, PhD

480-861-8851 (cell)

[email protected]