Answer
In a test of hypothesis about two proportions we have that $H_0:~p_1=p_2$. So we must find a point of estimate $p̂$ that will be the same for $p_1$ and $p_2$. The best point of estimate is the pooled estimate of the population proportion.
When constructing a confidence interval, we do not assume that $p_1=p_2$. We must find a point of estimate for $p_1$ and another for $p_2$. This way, we do not use a pooled estimate of the population proportion.
Work Step by Step
Given above.
