Suppose a survey is taken of two groups of people where each person is asked a yes/no question and the proportion of people who answer yes is calculated for each group. Which of the following is true about a hypothesis test of the difference in the two proportions?
A) Normality can be assumed if the sample size for each population is at least 30.
B) The t-distribution should be used if the standard deviations are unknown.
C) The standard deviation must be assumed equal.
D) Normality can be assumed if, in each group, at least 5 people say yes, and at 5 people say no.

Question

Quizplus · Accepted Answer

Normality for the difference in proportions can be assumed if both groups have at least 5 successes (yes) and 5 failures (no), according to the success-failure condition. This is not about the overall sample size but about the distribution of responses.

Suppose a Survey Is Taken of Two Groups of People