a. Component 3: Presumably the babies were all infants born during a certain time in one hospital. There should be nothing particularly different about these infants.
b. Component 4: The measurment of interest was weight gain or loss. That seems like a straightforward, objective measurement, so there should not be any problems with it.
c. Component 5: The setting in which the measurements were taken may have been somewhat of a problem, because all of the infants were in a common nursery. Therefore, if one infant cried excessively it may have impacted the entire group. It would have been ideal to test each infant in a separate setting, but that was not practical.
d. Component 6: There should have been no extraneous difference in the two groups of infants, since they were born at the same hospital and therefore presumably treated equally except for the heartbeat sound. There may be slight differences if they were born at different times of the year, but that information is not provided.
Problem 12.
A statistical difference may not be a meaningful one. Exercise 10b provides a good example. Learning that twice as many nicotine patch uses quit smoking than nonusers would mean more if the differnce was 50% versus 25% than if it were 2% versus 1%.
Problem 13.
Components 3 to 7 would all qualify. Probably most important would be 7, knowing by how much it raised cholesterol. Component 3 would be importantin order to know if results extend to people of your age, gender, and medical type; from component 4 you would learn how "coffee drinking" was defined; from component 5 you might learn about other changes in the environment of the participants (were they part of a larger study about healthier diets); from component 6 you would learn whether or not the non-coffee drinkers were different from the coffee drinkers in ways that could affect cholesterol, like amount of exercise or fat in their diets. A discussion of any three of these is sufficient
Problem 14.
Components 3 and 6 would reveal the problem. In comparing test results across countries, states, years, etc., we should learn about exactly who was included each timem and what extraneous differences there might be among th groups taking the tests. Often differences in results can be explained by the fact that the groups taking them are different on other measures. In this example, we would have learned that we were trying to compare the top 50% of students in one country with the top 3% in another country, hardly a fair comparison.
Problem 16.
a. The participants should have been randomly selected from among typical NordicTrack users.
b. The participants should have been weighed by the researchers rather than reporting weight loss themselves.
c.The NordicTrack should have been used in a naaatural setting, at home, rather than in a lab where someone else would be making sure participants used it regularly.
d. The advertisement would be misleading it there were major differences between the group tested and the typical user, i.e. of the group startedout grossly overweight or if they were on a lean diet at the same time.