26 Chapter 1 Basic Ideas Extending the Concepts 29. The Salk Vaccine Trial: In 1954, the first vaccine against polio, known as the Salk vaccine, was tested in a large randomized double-blind study. Approximately 750,000 children were asked to enroll in the study. Of these, approximately 350,000 did not participate, because their parents refused permission. The children who did participate were randomly divided into two groups of about 200,000 each. One group, the treatment group, got the vaccine, while the other group, the control group, got a placebo. The rate of polio in the treatment group was less than half of that in the control group. a. Is it reasonable to conclude that the Salk vaccine was effective in reducing the rate of polio? b. Polio is sometimes difficult to diagnose, as its early symptoms are similar to those of the flu. Explain why it was important for the doctors in the study not to know which children were getting the vaccine. c. Perhaps surprisingly, polio was more common among upper-income and middle-income children than among lower-income children. The reason is that lower-income children tended to live in less hygienic surroundings. They would contract mild cases of polio in infancy while still protected by their mother’s antibodies, and thereby develop a resistance to the disease. The children who did not participate in the study were more likely to come from lower-income families. The rate of polio in this group was substantially lower than the rate in the placebo group. Does this prove that the placebo caused polio, or could this be due to confounding? Explain. 30. Another Salk Vaccine Trial: Another study of the Salk vaccine, conducted at the same time as the trial described in Exercise 29, used a different design. In this study, approximately 350,000 second graders were invited to participate. About 225,000 did so, and the other 125,000 refused. All of the participating second graders received the vaccine. The control group consisted of approximately 725,000 first and third graders. They were not given any placebo, so no consent was necessary. a. Was this a randomized experiment? b. Was it double-blind? c. The treatment group consisted of children who had consent to participate. The control group consisted of all first and third graders. It turned out that the results of this study seriously underestimated the effectiveness of the vaccine. Use the information provided in Exercise 29(c) to explain why. Answers to Check Your Understanding Exercises for Section 1.3 1. Observational study 2. People who smoke may be more likely to drink alcohol than people who do not smoke. Therefore, it might be possible for smokers to have higher rates of liver cancer without it being caused by smoking. 3. a. Cohort study b. Prospective 4. a. Case-control study b. Retrospective SECTION 1.4 Bias in Studies Objectives 1. Define bias 2. Identify sources of bias Objective 1 Define bias Defining Bias No study is perfect, and even a properly conducted study will generally not give results that are exactly correct. For example, imagine that you were to draw a simple random sample of students at a certain college to estimate the percentage of students who are Democrats. Your sample would probably contain a somewhat larger or smaller percentage of Democrats than the entire population of students, just by chance. However, imagine drawing many simple random samples. Some would have a greater percentage of Democrats than in the population, and some would have a smaller percentage of Democrats than in the population. But on the average, the percentage of Democrats in a simple random sample will be the same as the percentage in the population. A study conducted by a procedure that produces the correct result on the average is said to be unbiased. Now imagine that you tried to estimate the percentage of Democrats in the population by selecting students who attended a speech made by a Democratic politician. On the
navidi_monk_essential_statistics_1e_ch1_3
To see the actual publication please follow the link above