< Back
Page 4 of 4
Like this article? We recommend
A Recap
A Recap
Before going on to that article, it's helpful to review what this paper has discussed. The important points are:
- A simple experiment takes two realities into account, one in which there is no post-treatment difference in the populations represented by your samples (the null hypothesis), and one in which a difference exists (the alternative hypothesis).
- You establish a critical value based on your desire not to reject a true null hypothesis. If your post-treatment outcome measure is within the critical value, you will conclude that the null hypothesis is true and reject the alternative hypothesis.
- If your post-treatment outcome measure is beyond the critical value, you will reject the null hypothesis and conclude that the alternative hypothesis is true.
- The portion of the distribution which represents the alternative hypothesis, and which is beyond the critical value, is the probability that you will reject a false null hypothesis. That distribution represents reality if the alternative hypothesis is true. It is the power of the statistical test: the probability that you will reject the null hypothesis when it is false.
Even more concisely: Establish a critical value that will be your criterion for rejecting the null hypothesis. Determine the percent of the distribution representing the alternative hypothesis that is beyond the critical value. That percentage is the test's statistical power.
< Back
Page 4 of 4