McNemar test
Test a change in a binary measure on the same cases.
When to use it
Use McNemar when the same people are measured twice on a yes/no variable. Only the people who changed contribute to the test. For example: did more people pass after a workshop than before it?
Assumptions
Each row is one pair. Both columns are binary. Concordant pairs (no change) do not enter the statistic.
Running it in Tensr
Options
Prop
Type
Reading the output
Eighty people classified yes or no before and after. Discordant pairs were built in both directions in similar numbers.
McNemar test for before vs after, n = 80 pairs. The primary result is not significant (p = > .999). This is not large enough to treat the comparison this page is about as a reliable association. The result is non-significant: the data are still compatible with no effect. Metrics: Discordant a→b = 5; Discordant b→a = 6; p-value = > .999.
Reporting (APA 7)
McNemar test for before vs after, n = 80 pairs. This result is not significant (p = > .999). Report the estimate with that p, and do not describe the pattern as a reliable effect.
Coming from SPSS
Crosstabs → Statistics → McNemar. The menu label is still on the Analyze menu even though an older retired-label list still names McNemar.