Enter your data for Power and Sample Size for 2-Sample Equivalence Test

Stat > Power and Sample Size > Equivalence Tests > 2-Sample

From Hypothesis about, indicate how you want to express the equivalence criteria.
- Test mean - reference mean (Difference)
  Define equivalence in terms of a difference between the mean of the test population and the mean of the reference population.
- Test mean / reference mean (Ratio, by log transformation)
  Define equivalence in terms of the ratio of the mean of the test population to the mean of the reference population, as modeled with a log transformation of the original data. For this option, all observations must be greater than 0.
From What do you want to determine? (Alternative hypothesis), select the alternative hypothesis that you are trying to prove or demonstrate.
- If your hypothesis is about Test mean - reference mean (Difference), select one of the following options.
  - Lower limit < test mean - reference mean < upper limit
    Test whether the difference between the population means is within the limits that you specify.
    
    For example, an analyst wants to determine whether the mean strength of a new generic drug is within ± 10 mg/ml of the mean strength of a brand-name drug.
  - Test mean > reference mean
    Test whether the mean of the test population is greater than the mean of the reference population.
    
    For example, a food analyst wants to determine whether an improved formulation of a dog food has more mean protein per 100 g than the current formulation.
  - Test mean < reference mean
    Test whether the mean of the test population is less than the mean of the reference population.
    
    For example, an analyst wants to demonstrate that a new medication takes effect in less time, on average, than the current medication.
  - Test mean - reference mean > lower limit
    Test whether the difference between the population means is greater than a lower limit.
    
    For example, a researcher wants to determine whether the mean reduction in diastolic blood pressure induced by an experimental drug is more than 3 mm Hg greater than the mean reduction induced by the current medication.
  - Test mean - reference mean < upper limit
    Test whether the difference between the population means is less than an upper limit.
    
    For example, an analyst wants to determine whether the mean waiting time in the emergency department of a hospital in one location does not exceed the mean waiting time of a hospital in another location by more than 5 minutes.
- If your hypothesis is about Test mean / reference mean (Ratio, by log transformation), select one of the following options.
  - Lower limit < test mean / reference mean < upper limit
    Test whether the ratio of the population means is within the limits that you specify. Both limits must be greater than zero. A ratio of 1 indicates that the two means are equal.
    
    For example, an analyst needs to demonstrate that the mean bioavailability of a test formulation is within 80% (0.8) and 125% (1.25) that of the reference formulation, using log transformed data.
  - Test mean / reference mean > lower limit
    Test whether the ratio of the population means is greater than a lower limit.
    
    For example, an analyst needs to demonstrate that the mean bioavailability of a test formulation is greater than 80% (0.8) that of the reference formulation, using log transformed data.
  - Test mean / reference mean < upper limit
    Test whether the ratio of the population means is less than an upper limit.
    
    For example, an analyst needs to demonstrate that the mean bioavailability of a test formulation is less than 125% (1.25) that of the reference formulation, using log transformed data.
Enter a value for each equivalence limit included in the alternative hypothesis.
- Lower limit
  Enter the lowest acceptable value for the difference or ratio. You want to demonstrate that the difference (or ratio) between the mean of the test population and the mean of the reference population is not lower than this value.
- Upper limit
  Enter the highest acceptable value for the difference or ratio. You want to demonstrate that the difference (or ratio) between the mean of the test population and the mean of the reference population does not exceed this value.
Specify values for two of the following power function variables. Leave the variable that you want to calculate blank.
Tip
If you enter multiple values into a field, separate the values with a space. You can also use shorthand notation to indicate multiple values. For example, you can enter 10:40/5 to indicate sample sizes from 10 to 40 in increments of 5.
- Sample sizes: Enter the number of observations for each group. For example, enter 50 if you plan to collect 50 observations for each of the 2 groups. To assess the effect of different sample sizes, enter multiple values. Larger sample sizes give the test more power to demonstrate equivalence.
- Differences (within the limits): Enter one or more values to specify the difference (or ratio) between the test mean and the reference mean. The values that you enter must be within the equivalence limits. Differences (or ratios) that are close to an equivalence limit require larger sample sizes to achieve adequate power.
- Power values: Enter one or more values to specify the probability that the test shows equivalence when the population difference (or ratio) is within the equivalence limits. Common values are 0.8 and 0.9. For example, an analyst enters 0.9 to indicate a 90% chance that the test will demonstrate equivalence between the mean of the test treatment and the mean of the reference treatment when the means are actually equivalent.
Specify the amount of expected variation in the data. Minitab assumes that both the test population and the reference population have the same standard deviation and coefficient of variation.
- For a test of the difference, in Standard deviation, enter an estimate of the standard deviation.
- For a test of the ratio by log transformation, in Coefficient of variation (CV), enter an estimate for the coefficient of variation of the raw data.
If you have already collected and analyzed data, you can use the sample estimates from the data. The coefficient of variation equals the standard deviation divided by the mean. If equal variances can be assumed, you can use the pooled standard deviation from a relevant 2-sample equivalence test for the standard deviation. If you do not have data, base your estimate on related research, design specifications, pilot studies, subject-matter knowledge, or similar information.