Calculating Power

• 1). Determine the type of power analysis you need. You may need either a post hoc analysis, where you have collected your data already and want to know the power of your analysis, or an a priori analysis, where you want to know how much data to collect to achieve a certain power.
  
  
• 2). Download and run G*Power, a free program for Mac and PC developed at the University of Duesseldorf (see Resources).
  
  
• 3). Select the type of analysis you are conducting from the "Test family" and "Statistical test" drop down menus.
  
  
• 4). Select "A priori" or "Post Hoc" as appropriate from the "Type of power analysis" drop down menu.
  
  
• 5). Type the relevant information into the fields in the "Input parameters" section.
  
  For a priori tests, you need to input the effect size, given as Pearson's r, the power you wish to achieve and the alpha level you use. This alpha is the cut off point for statistical significance used in your analysis; for example, .05 in psychology.
  
  For post hoc tests, you need to input the sample size of your experiment, the effect size you achieved, given as Pearson's r and the alpha level you used.
  
  For both a priori and post hoc tests, use the drop down menu to indicate whether your test was one-tailed or two-tailed.
  
  
• 6). Click "Calculate" in the bottom right.
  
  
• 7). Read the output in the bottom right of the screen. For an a priori test, the software presents the total sample size you need to achieve your desired level of power. For a post hoc test, you learn the observed power of the analysis you have conducted.