The dean of students at a large college is interested in learning about their opinions regarding the percentage of first-year students who should be given parking privileges in the main lot. He sends out an email survey to all students about this issue. A large number of first-year students reply but very few sophomores, juniors, and seniors reply. Based on the responses he receives, he constructs a 90% confidence interval for the true proportion of students who believe first-year students should be given parking privileges in the main lot to be (0.71, 0.79). Which of the following may have an impact on the confidence interval, but is not accounted for by the margin of error?
Very few sophomores, juniors, and seniors responded, so the responses may overrepresent first-year students. This can create nonresponse bias if students who did not reply have different opinions from those who did. The margin of error accounts for random sampling variation, not this systematic difference; therefore, the answer is nonresponse bias.
AI-written and checked against the verified answer.