Example of Power and Sample Size for 2 Proportions

A university financial aid officer wants to determine whether male or female students are more likely to get a job in the summer. Results from a previous study suggest that 60% of students get a job in the summer. Before collecting the data for a 2 proportions test, the officer uses a power and sample size calculation to determine how small of a difference the test can detect when the sample size is 1,000 and the power is 0.9.

  1. Choose Stat > Power and Sample Size > 2 Proportions.
  2. In Sample sizes, enter 1000.
  3. In Power values, enter 0.9.
  4. In Baseline proportion (p2), enter 0.6.
  5. Click OK.

Interpret the results

With a sample size of 1,000 and a power value of 0.9, the officer can detect a difference between proportions of approximately 7% in either direction. This difference is adequate, so the officer collects the data for the 2 proportions analysis.

Power and Sample Size

Test for Two Proportions Testing comparison p = baseline p (versus ≠) Calculating power for baseline p = 0.6 α = 0.05
Results Sample Size Power Comparison p 1000 0.9 0.669724 1000 0.9 0.528190 The sample size is for each group.

Power Curve for Two Proportions

By using this site you agree to the use of cookies for analytics and personalized content.  Read our policy