When we are comparing proportions, we are testing whether or not there is difference between populations.
When we write down H0: p1 = p2, we are communicating that there is no difference between the proportions.
In other words, if I were to subtract these two proportions, I would get value of 0. So, algebraically, p1 - p2 = 0.
Hence, any differences that which are obtained in our sample values would be determined against 0.
What will the alternative hypothesis communicate if we write down the following:
p1 < p2
p1 > p2
p1 <> p2 (not equal)?
What will the null look like for each and what does that point to about the data