Hypothesis Testing for Two Population Proportions
In this section, we will step through how to perform a hypothesis tests to determine if there is a difference between two population proportions.
Distribution Used
We will continued to use the Normal Distributions and z-scores.

Assumptions
In order to be able to perform the analysis in this section, the following assumptions must hold true:
- the samples are random and independent of one another
- the sample sizes and proportions from each group are large enough such that:
- np>5np>5
- n(1−p)>5n(1−p)>5
Note: When the sample size and proportion are large enough, the discrete distribution approaches a normal/bell shaped curve:

The Difference Between the Two Curves Above
For the left-most curve:
- np=50×0.5=25>5np=50×0.5=25>5
- n(1−p)=50×(1−0.5)=25>5n(1−p)=50×(1−0.5)=25>5
- the curve closely resembles a bell-shaped curve
For the right-most curve:
- np=50×0.9=45>5np=50×0.9=45>5
- n(1−p)=50×(1−0.9)=5≯5
- the curve is skewed left and therefore not bell-shaped
Note: The right-most curve is at the ‘limit’ of acceptable. If the value of n was slightly larger or the value of p slightly smaller (to increase the size of 1−p), we could perform the analysis in this section on this data.