<span>p1=44/88=.50; p2=57/85=.67.
Under the null hypothesis of no difference, we pool the data to estimate the
common p of (44+57)/(88+85)=.584.
The test statistic is (.67-.50)/sqrt[(.584)(1-.584)(1/88 + 1/85)]=2.268 (which is stat sig. at a .095 level).</span>
