The McNemar test is used for before-and-after research designs. It is used with matched pairs or when a subject is his/her own control. The purpose is to determine the significance of the observed change. The requirements are listed below.
Nominal data
Two groups
Related groups
Expected frequencies five or more
Two-tailed test only (The chi-square table is used for the critical value.)
The formula for the McNemar test is:
where A = number of individuals changing in one direction, D = number of individuals changing in the opposite direction from A, and E = expected number of individuals under the null hypothesis 1/2 (A + D). Df = degrees of freedom is always 1 for the McNemar test.