Section 1.4 Bias in Studies 29 7. Surveys with many nonresponders often provide misleading results. 8. Large samples usually give reasonably accurate results, no matter how they are drawn. Practicing the Skills In Exercises 9–16, specify the type of bias involved. 9. A bank sent out questionnaires to a simple random sample of 500 customers asking whether they would like the bank to extend its hours. Eighty percent of those returning the questionnaire said they would like the bank to extend its hours. Of the 500 questionnaires, 20 were returned. 10. To determine his constituents’ feelings about election reform, a politician sends a survey to people who have subscribed to his newsletter. More than 1000 responses are received. 11. An e-store that sells cell phone accessories reports that 98% of its customers are satisfied with the speed of delivery. 12. The e-store in Exercise 11 sends a survey out to a random sample of 1000 recent customers, asking whether they are satisfied with the products they received. Ninety percent of those returning questionnaires indicated that they were satisfied. A total of ten customers returned the questionnaire. 13. A television newscaster invites viewers to email their opinions about whether the U.S. Congress is doing a good job in handling the economy. More than 100,000 people send in an opinion. 14. A police department conducted a survey in which police officers interviewed members of their community to ask their opinions on the effectiveness of the police department. The police chief reported that 90% of the people interviewed said that they were satisfied with the performance of the police department. 15. In a study of the effectiveness of wearing seat belts, a group of people who had survived car accidents in which they had not worn seat belts reported that seat belts would not have helped them. 16. To estimate the prevalence of illegal drug use in a certain high school, the principal interviewed a simple random sample of 100 students and asked them about their drug use. Five percent of the students acknowledged using illegal drugs. Working with the Concepts 17. Nuclear power, anyone? In a survey conducted by representatives of the nuclear power industry, people were asked the question: “Do you favor the construction of nuclear power plants in order to reduce our dependence on foreign oil?” A group opposed to the use of nuclear power conducted a survey with the question: “Do you favor the construction of nuclear power plants that can kill thousands of people in an accident?” a. Do you think that the percentage of people favoring the construction of nuclear power plants would be about the same in both surveys? b. Would either of the two surveys produce reliable results? Explain. 18. Who’s calling, please? Random-digit dialing is a sampling method in which a computer generates phone numbers at random to call. In recent years, caller ID has become popular. Do you think that caller ID increases the bias in random digit dialing? Explain. 19. Who’s calling, please? Many polls are conducted over the telephone. Some polling organizations choose a sample of phone numbers to call from lists that include landline phone numbers only, and do not include cell phones. Do you think this increases the bias in phone polls? Explain. 20. Order of choices: When multiple-choice questions are asked, the order of the choices is usually changed each time the question is asked. For example, in the 2008 presidential election, a pollster would ask one person “Who do you prefer for president, Barack Obama or John McCain?” For the next person, the order of the names would be reversed: “John McCain or Barack Obama?” If the choices were given in the same order each time, do you think that might introduce bias? Explain. Extending the Concepts 21. Literary Digest poll: In the 1936 presidential election, Republican candidate Alf Landon challenged President Franklin Roosevelt. The Literary Digest magazine conducted a poll in which they mailed questionnaires to more than 10 million voters. The people who received the questionnaires were drawn from lists of automobile owners and people with telephones. The magazine received 2.3 million responses, and predicted that Landon would win the election in a landslide with 57% of the vote. In fact, Roosevelt won in a landslide with 62% of the vote. Soon afterward, the Literary Digest folded. a. In 1936 most people did not own automobiles, and many did not have telephones. Explain how this could have caused the results of the poll to be mistaken. b. What can be said about the response rate? Explain how this could have caused the results of the poll to be mistaken. c. The Literary Digest believed that its poll would be accurate, because it received 2.3 million responses, which is a very large number. Explain how the poll could be wrong, even with such a large sample. Answers to Check Your Understanding Exercises for Section 1.4 1. No. The sample taken at the football stadium is biased, because football fans are more likely to be sampled than others. The fact that the sample is big doesn’t make it any better. 2. No. There is a high degree of nonresponse bias in this sample.
navidi_monk_essential_statistics_1e_ch1_3
To see the actual publication please follow the link above