Question 208:

Asked on May 27, 2008

Tags: Hypothesis Test , 1-Sided Test

1

Answer:

No answer provided yet.

 The usual warning with 1-sided hypothesis testing is that you should decide ahead of data-collection if you will use a two-sided or one-sided. The other warning is that you should be rather certain that the treatment won't in fact increase the incidence of the disease.

You've stated that the new quality-of-care project is expected to reduce the proportion of a disease, assuming you are still in the planning stages, this seems like an acceptable use.  Of course, once you do collect the data, you can always also perform the more conservative 2-sided test and also see if those results are significant.