Sample size for McNemar's test
Effect size is the difference in proportions between two observations of the outcome
In order to run an a priori sample size calculation for McNemar's test, researchers will need to seek out evidence that provides the magnitude of the treatment effect and the proportion of observations that are expected to change as a result of the treatment. The odds ratio is the effect size in this particular design. The odds ratio relates to the magnitude of the change from one level of the outcome variable to the other level of the outcome variable as a result of the intervention or treatment. Researchers must also specify what proportion of the observations will change as a result of the intervention.
The values to seek out in the literature are 1) the odds ratio or magnitude of the intervention on changing from one level of the categorical variable to the other and 2) the proportion of the population that will change as a result of the intervention. The articles should be theoretically or conceptually similar to the study of interest. Use the reported values in a sample size calculation for McNemar's test. This is called using an evidence-based measure of effect size.
For example, let's say researchers find high quality evidence that people are twice as likely to change group membership as a result of an intervention and that 75% of participants are expected to change. The odds ratio is 2.0 and the proportion of discordant group change (people that will change from one level of the categorical variable to the other level) is 75%.
The values to seek out in the literature are 1) the odds ratio or magnitude of the intervention on changing from one level of the categorical variable to the other and 2) the proportion of the population that will change as a result of the intervention. The articles should be theoretically or conceptually similar to the study of interest. Use the reported values in a sample size calculation for McNemar's test. This is called using an evidence-based measure of effect size.
For example, let's say researchers find high quality evidence that people are twice as likely to change group membership as a result of an intervention and that 75% of participants are expected to change. The odds ratio is 2.0 and the proportion of discordant group change (people that will change from one level of the categorical variable to the other level) is 75%.
The steps for calculating sample size for a McNemar's test in G*Power
1. Start G*Power.
2. Under the Test family drop-down menu, select Exact.
3. Under the Statistical test drop-down menu, select Proportions: Inequality, two dependent groups (McNemar).
4. Under the Type of power analysis drop-down menu, select A priori: Compute required sample size - given alpha, power, and effect size.
5. If there is a directional hypothesis, under the Tail(s) drop-down menu, select One.
6. If there is a non-directional hypothesis, under the Tail(s) drop-down menu, select Two.
7. In the Odds ratio box, enter the magnitude of the expected change in the form of an odds ratio. Example: "2.0"
8. Leave the alpha value at 0.05, unless researchers want to change the alpha value according to the current empirical or clinical context.
9. Enter ".80" into the Power (1-beta err prob) box, unless researchers want to change the power according to the current empirical or clinical context.
10. In the Prop discordant pairs box, enter the proportion of the population that will change as a result of the intervention. Example: ".75"
2. Under the Test family drop-down menu, select Exact.
3. Under the Statistical test drop-down menu, select Proportions: Inequality, two dependent groups (McNemar).
4. Under the Type of power analysis drop-down menu, select A priori: Compute required sample size - given alpha, power, and effect size.
5. If there is a directional hypothesis, under the Tail(s) drop-down menu, select One.
6. If there is a non-directional hypothesis, under the Tail(s) drop-down menu, select Two.
7. In the Odds ratio box, enter the magnitude of the expected change in the form of an odds ratio. Example: "2.0"
8. Leave the alpha value at 0.05, unless researchers want to change the alpha value according to the current empirical or clinical context.
9. Enter ".80" into the Power (1-beta err prob) box, unless researchers want to change the power according to the current empirical or clinical context.
10. In the Prop discordant pairs box, enter the proportion of the population that will change as a result of the intervention. Example: ".75"
Using the odds ratio and the proportion expected to change as result of the intervention and the 11 steps above, a two-tailed, 2.0 effect size at an alpha of .05 and 80% power and 75% of the population changing as a result of the intervention, 96 participants would be needed to detect the effect.
Click on the Statistics button to continue.
Hire A Statistician
DO YOU NEED TO HIRE A STATISTICIAN?
Eric Heidel, Ph.D., PStat will provide you with statistical consultation services for your research project at $100/hour. Secure checkout is available with Stripe, Venmo, Zelle, or PayPal.
- Statistical Analysis on any kind of project
- Dissertation and Thesis Projects
- DNP Capstone Projects
- Clinical Trials
- Analysis of Survey Data