Conduct your experiment and obtain a list of values.
Find the mean of the values by taking the total of the numbers and dividing by the number of values. Find the standard deviation by squaring the difference of each value from the mean, summing the total, dividing by the number of values and then taking the square root of the new number. You can also enter the values into a graphing calculator and solve for standard deviation. For example:
Values: 2, 4, 5, 5, 8
Mean: 24/5 = 4.8
Standard Deviation = 2.16
Find an outlying value for which you are interested to know whether the power is accurate or not. For example, assume you want to evaluate the power of a test with a 9 value. Subtract the outlying value from the mean and divide by the standard deviation:
(9 - 4.8)/2.16 = 1.94
The result is the statistical z score.
Look-up the z-score to find the error rate. In this case, 1.94 equates to 2.6 percent error rate.
Subtract the error rate from 100 to find the power. In this case, the power is over 97 percent. Use this power for your longitudinal design.