Stats 110A, Spring 2002 Name:
April 11 ID

Quiz 2 Solution

The Public Health Service studied the effects of smoking on health, in a large sample of representative households. For men and women in each age group, those who had never smoked were on average somewhat healthier than the current smokers. But the current smokers were on average much healthier than those who had recently stopped smoking.

a) Why did they study men and women and the different age groups separately?
The basic principle being followed here is to minimize “nuisance” variation by putting subjects into homogeneous blocks. “Age” is quite possibly a confounder, because health deteriorates with age and also older people are more likely to have smoked longer. “Controlling” for age then accounts for this. As for “gender”, since we know that genders react biologically differently, and also are subject to different societal pressures, we need to be prepared for the possibility that different genders will smoke differently and have different health effects.


b) The lesson seems to be that you shouldn’t start smoking, but once you’ve started, don’t stop. Do you agree? Explain why or why not.

Even if you want to agree, you should admit that this study is not evidence in support of following this prescription. First, this was an observational study (how could it be otherwise?) and therefore you should never rush to conclude causation. The causal claim you’re being asked to evaluate is this: does quitting smoking negatively affect health?

The data seem to say “yes”, but the experimental design doesn’t support this conclusion. There is at least one fairly obvious confounding factor: people in poor health are likely to quit smoking, either because the doctor tells them to, or because they simply are physically not capable of smoking any longer. (At least one of you pointed out that if you included those who died from smoking, the quitters would seem to be in even worse health than the smokers.)

Remember that the definition of a confounding factor is that it has to affect both assignment to treatment group and the response variable. Here, “poor health” determines which group they belong to (smoking or quitting) and the response variable (quality of health.)

Many people got distracted by the “recently” in the problem statement. There is an argument to be made that quitting smoking might cause short-term health problems. But if so, then you are agreeing with the premise of the argument that quitting does cause poor health (even if for the short term). And a problem with this is that it is hard to see how, in every age group, even among the elderly, quitting made health problems worse. Some people pointed out that there are, in fact, known side effects to quitting (most noticably a weight gain). But realistically speaking, the phrase “ current smokers were on average much healthier than those who had recently stopped smoking” must be interpreted to mean more than the fact that the quitters were slightly heavier than the smokers.

This answer (that the time period was too short to see the positive health effects of quitting) does a good job of explaining the results, but is not as convincing and full an explanation as describing a confounding factor that provides a more plausible explanation. This answer did not, therefore, receive full credit.