The McNemar's test

Modified on Mon, 30 Aug, 2021 at 8:51 AM

Definition


The McNemar's test is an alternative to the Chi² test, that compares proportions between group, for paired data. Paired data are data where pairs are less different from each other than from another data of the dataset.


In practice


This test has to be used when you compare proportions between paired groups like cases and controles or before and after treatment. 
The odd ratio (OR) given is computed as:

OR =  proportion of patients having not the outcome before intervention & having the outcome after intevention
         proportion of patients having the outcome before intervention & having not the outcome after intevention


If the OR is superior to 1, it has to be understood as "the percentage of patients free from the outcome before treatment that have got the outcome after treatment is more important than the percentage of  patients not free from the outcome before treatment that haven't got the outcome after treatment". 

For example, if the outcome is "Healthy" = 1, an OR superior to 1 could be understood as "the percentage of sick patients that become healthy after treatment is more important than the percentage of healthy patients that become sick after treatment". 


 

Pre treatment 

Healthy= 1

Pre treatment 

Healthy = 0

Post treatment 

Healthy = 1

a

b

Post treatment 

Healthy = 0

c

d


How to use it on EasyMedStat ?


  1. Go to Statistics > Test 2 variables
  2. Select two Yes-No variables you would like to compare taking care of the following order: the first variable field on the left side of the page is for the pre-treatment variable, the second variable field on the right side of the page is for the post-treatment variable.
  3. Click on "Compare pre-post treatment proportions"


See also


Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select at least one of the reasons
CAPTCHA verification is required.

Feedback sent

We appreciate your effort and will try to fix the article