Question

The mayor of a large city will run for governor if he believes that more than 30 percent of the voters in the state already support him. He will have a survey firm ask a random sample of n voters whether or not they support him. He will use a large sample test for proportions to test the null hypothesis that the proportion of all voters who support him is 30 percent or less against the alternative that the percentage is higher than 30 percent. Suppose that 35 percent of all voters in the state actually support him. In which of the following situations would the power for this test be highest?

a. The mayor uses a significance level of 0.01 and n = 250 voters.
b. The mayor uses a significance level of 0.01 and n = 500 voters.
c. The mayor uses a significance level of 0.01 and n = 1,000 voters.
d. The mayor uses a significance level of 0.05 and n = 500 voters.
e. The mayor uses a significance level of 0.05 and n = 1,000 voters.

Answer 1

e. The mayor uses a significance level of 0.05 and n = 1,000 voters.

Explanation:

Power of a test:

For a high power of a test, two conditions are needed:

Large significance level.

Large sample.

In this question:

A significance level of 0.05 will lead to a higher power than a significance level of 0.01, which means that the answer is between options d and e.

A test with a sample of 1000 has a higher power than a test with a sample of 500, and thus, the answer is given by option e.