Thus, the null hypothesis in McNemar’s test is that we have “marginal homogeneity”. That is, the row totals and column totals have the same distribution: P a +P b =P a +P c , and similarly that P c +P...Thus, the null hypothesis in McNemar’s test is that we have “marginal homogeneity”. That is, the row totals and column totals have the same distribution: P a +P b =P a +P c , and similarly that P c +P d =P b +P d . Notice that this means that the null hypothesis actually simplifies to P b =P c . In other words, as far as the McNemar test is concerned, it’s only the off-diagonal entries in this table (i.e., b and c) that matter!
Thus, the null hypothesis in McNemar’s test is that we have “marginal homogeneity”. That is, the row totals and column totals have the same distribution: P a +P b =P a +P c , and similarly that P c +P...Thus, the null hypothesis in McNemar’s test is that we have “marginal homogeneity”. That is, the row totals and column totals have the same distribution: P a +P b =P a +P c , and similarly that P c +P d =P b +P d . Notice that this means that the null hypothesis actually simplifies to P b =P c . In other words, as far as the McNemar test is concerned, it’s only the off-diagonal entries in this table (i.e., b and c) that matter!
