There are two important situations in research when we need to think about statistical power:

1. when we are planning a study and
2. when we fail to reject the null hypothesis.

We always want to think about statistical power when we are planning a research study. It makes no sense to develop a strong research design, collect and analyze data, and not have enough statistical power to be able to reject the null hypothesis. Therefore, we should always think about how many participants we need in order to have enough statistical power to test the hypothesis of interest. What is enough statistical power? Many researchers think that a probability of .80 of rejecting the null hypothesis is good statistical power. How do we figure out how many participants we need? We estimate a likely population effect size from the current literature. We then calculate the sample size needed to detect an effect of that size with the desired level of power. We use the same procedures outlined in our power calculations but estimate N instead of estimating 1 - β.