next up previous contents
Next: Two-tailed test Up: Proportions Previous: Required Inputs   Contents

Computational Details

Because the null hypothesis assumes equal proportions, $ p_1=p_2$ , we compute a pooled estimate $ \hat{p}$ as follows:

$\displaystyle \hat{p}\;=\;\frac{n_1\cdot\hat{p}_1+n_2\cdot\hat{p}_2}{n_1+n_2}
$

(This value appears in cell B14)

The $ z$ -statistic for testing equality of the population proportions $ p_1$ and $ p_2$ is:

$\displaystyle Z\;=\;\frac{\hat{p}_1-\hat{p}_2}{\sqrt{\hat{p}(1-\hat{p}}\sqrt{\frac{1}{n_1}+\frac{1}{n_2}}}
$

(This value appears in cell B17)

gene quinn 2006-12-04