A McNemar test indicates whether the distribution of values appears to differ across two category columns when the values in a particular row are related. For example, it can be used to assess whether voting patterns significantly differ from one year to the next, where each row represents an individual voter.
To perform a McNemar test on two category columns:
Note that the test will only be performed if the selected columns take on the same set of values, and have the same category labels (if any). If the columns take 3 or more values, the result is known as a Stuart-Maxwell test.