McNemar's test is used to analyze paired nominal data, often in 'before and after' studies. It focuses specifically on 'discordant pairs'—subjects who changed their status—to determine if the shift in proportion is statistically significant.
💡
Pro Tips
Discordant cells (Yes/No and No/Yes) drive the entire result.
For small samples (discordant sum < 25), the Exact Binomial method is automatically used.
An Edwards continuity correction is applied for larger samples to improve accuracy.
!
Fun Facts
"Quinn McNemar developed this test in 1947 specifically for psychological testing where the same subjects are measured twice."
"In medical research, it's the gold standard for comparing the sensitivity of two different diagnostic tests on the same patient group."
"The test ignore 'concordant pairs' (people who stayed the same) because they provide no evidence for a change in proportion."
