Hypothesis testing enables you to determine whether differences in samples indicate actual differences in populations or are merely due to sampling error. For instance, a study conducted on about 1400 American children in a variety of settings compared the behavior of the children who attended day care with the behavior of those who stayed home. Aggressive behavior such as stealing toys, pushing other children, and starting fights was measured in both groups. The study showed that children who attended day care for more than 30 hours per week were about three times more likely to be aggressive than those who stayed home. Although the aggressive behavior observed in the study was well within the normal range for healthy children, these statistics have been used to persuade parents to keep their children at home until they start school.

The U.S. study found that the results were the same regardless of quality of the day care center and income of the family. However, the overall quality of care experienced by most of the children studied could be the problem-a survey of American day care centers that measured aspects such as number and expertise of caregivers found that only 10 percent of American day care centers provided high-quality care.

A similar study of preschoolers and aggressive behavior in Norway, where day care centers are subject to strict standards and the ratio of adult caregivers to children is high, found that the link between day care attendance and aggressive behavior was minimal. Another Norwegian study included an additional variable, differences between siblings, and found no relationship between day care attendance and behavior problems. These additional variables that are often out of the researcher's control are known as confounding variables.

A series of studies was conducted on various methods for reducing the number of cigarettes that smokers smoke. The study compared smokers who were simply told to smoke less and those who tried methods such as nicotine replacement therapy, electronic cigarettes, and using reduced tar, carbon, or nicotine cigarettes. Some methods were shown to be effective in reducing the number of cigarettes smoked.

Some of the studies were funded by the tobacco industry, which could profit from promoting strategies other than quitting as beneficial to smokers' health. When dealing with statistics, it is always good to know who is paying for a study, and whether the researchers are unbiased.

EXERCISES 

1. A pharmaceutical company has applied for approval to market a new arthritis medication. The research involved a test group that was given the medication and another test group that was given a placebo. Describe some possible confounding variables that could influence the results of the study.

2. Medical research often involves blind and double-blind testing. Explain what these two terms mean.