1. Why can't you design a controlled experiment to determine whether smoking causes lung cancer?
A: You would have to assign people to a smoking group and others to a non-smoking group. There are obvious ethical problems with this, but even so, how would you be able to guarantee that no one in the non-smoking group touches a cigarette, or that the smoking group all smokes the right amount?
2. An economist wishes to determine if a particular advertising campaign
for cigarrettes affected children's decisions to smoke. To do this,
he looks
through a national database that counted the number of children
who said they smoked each year from 1980 through 1995. He then compared
these smoking rates before and after the advertising campaign. Is
this an observational or controlled study? Why?
A: This is an observational study. Actually, its often called a retrospective or historical study, because rather than recruit subjects it uses information collected in the past. Here the treatment is their exposure to the advertising campaign, and the outcome is whether or not a child smokes. Because the children "chose" whether or not to be exposed to the advertising, it is an observtional study. (Of course, the children didn't really choose and had little control over whether or not they would be exposed. But the researchers certainly did NOT choose.) Note that in this study there's no way of knowing if the children exposed to advertising were the same one's who took up smoking!
3. A report of a study in the LA Times said that neighborhoods with a high number of liquor stores per capita also have a high crime rate when compared to neighborhoods with a relatively low number of liquor stores per capita. Is this an observational or controlled study? Why? Can we conclude that removing the liquor stores from the high-crime neighborhoods would lower the crime rate?
A: This is an observational study. It is impossible for researchers (or anyone) to decide which neighborhoods get a certain number of liquor stores. In effect, the neighborhoods themselves "chose" whether or not to have liquor stores. And so we can't conclude whether the liquor stores cause the high crime rate. Certainly this is evidence in support of the claim, but is not in itself "proof." It's quite likely that there's another fact, say poverty, that causes both high crime and a high number of liquor stores to appear in a neighborhood. Removing the liquor stores, then, would have little affect on the crime rate.
4. The LA Times (November 1997) reports that "A new study in the May 1 New England Journal of Medicine provides some of the strongest evidence yet that regular exercise helps protect women from breast cancer. The research, conducted in Norway, found that women who exercise at least four hours a week have a breast cancer risk about one-third lower than usual." Can we conclude that exercise prevents breast cancer? Why or why not?
A: No we can't. This is an observational study. (It doesn't say so explicitly, but to monitor breast cancer rates takes a long time, and there's no way to enforce an exercise regime on people for a long period of time.) The LA Times itself says: "Women who work out a lot also usually have other healthy habits, so it's difficult to say whether it is exercise that wards off cancer."