In Exercises 21–28,determine whether the study is an experiment or an observational study, and then identify a major problem with the study.

Driver Aggression: In testing a treatment designed to reduce driver aggression in the United States, the original plan was to use a sample of 500 drivers randomly selected throughout the country. The program managers know that they would get a biased sample if they limit their study to drivers in New York City, so they planned to compensate for that bias by using a larger sample of 3000 drivers in New York City.

A study is conducted to test a treatment’s effectiveness in reducing the aggression of American drivers by choosing a sample of 500 drivers of New York City and then increasing the sample size to 3000 drivers to remove the biasedness from the study.

Anobservational studyis conducted to draw inferences about a population of units. There is no intervention present from the experimenter’s side over any of the units that are a part of the study.

Anexperimental studyis carried out to verify inferences drawn about a population of units. The treatment or factor understudy is assigned to one group while the other group does not receive the treatment.

Since the study is done to “test” a treatment’s effectiveness to reduce aggression in American drivers, it implies that the study is an experimental study. Moreover, a cost factor is involved in acquiring additional units, equal to 2500, to increase the sample size to 3000 drivers, which ascertains that this is an experiment.

Increasing the sample size from 500 to 3000 will not eliminate the biases of the sample that have arisen because the sample represents only a city and not the entire population of the United States.

Thus, biases in the sample lead to wrong conclusions of the experiment and cannot be removed by an increase in the size of the sample.