91Ó°ÊÓ

Experiments versus observational studies When either type of study is feasible, an experiment is usually preferred over an observational study. Explain why, using an example to illustrate. Also explain why it is not always possible for researchers to carry out a study in an experimental framework. Give an example of such a situation.

Seat belt anecdote Andy once heard about a car crash victim who died because he was pinned in the wreckage by a seat belt he could not undo. As a result, Andy refuses to wear a seat belt when he rides in a car. How would you explain to Andy the fallacy behind relying on this anecdotal evidence?

Types of bias Give an example of a survey that would suffer from a. Sampling bias due to the sampling design b. Sampling bias due to undercoverage c. Response bias d. Nonresponse bias

Vitamin B A New York Times article (March 12, 2006) described two studies in which subjects who had recently had a heart attack were randomly assigned to one of four treatments: placebo and three different doses of vitamin B. In each study, after years of study, the differences among the proportions having a heart attack were judged to be not statistically significant. Identify the (a) response variable, (b) explanatory variable, (c) experimental units, (d) treatments, and (e) explain what it means to say that differences "were judged to be not statistically significant."

Colds and vitamin \(\mathbf{C}\) For some time there has been debate about whether regular large doses of vitamin \(\mathrm{C}\) reduce the chance of getting a common cold. a. Explain how you could design an experiment to test this. Describe all parts of the experiment, including (i) what the treatments are, (ii) how you assign subjects to the treatments, and (iii) how you could make the study double-blind. b. An observational study indicates that people who take vitamin \(\mathrm{C}\) regularly get fewer colds, on the average. Explain why these results could be misleading.

Issues in clinical trials A randomized clinical trial is planned for AIDS patients to investigate whether a new treatment provides improved survival over the current standard treatment. It is not known whether it will be better or worse. a. Why do researchers use randomization in such experiments, rather than letting the subjects choose which treatment they will receive? b. When patients enrolling in the study are told the purpose of the study, explain why they may be reluctant to be randomly assigned to one of the treatments.