# What is the causal hypothesis that was tested by this experiment or study?

The hypotheses in these exercises are taken from Stephen Carey’s A Beginner’s Guide to Scientific Method.1. Consider the following hypothesis:Many dairy farmers claim that their cows produce more milk when they are listening to calm, soothing music, the sort of music we often hear in elevators and shopping malls.If you were to design a prospective observational study to test this hypothesis:a. What kind of cows would be in the experimental group?b. What kind of cows would be in the control group?c. What kind of results would confirm (strongly or weakly) the hypothesis?d. Is there anything that would need to be defined in order to make it easier to sort the groups?e. What are the most important confounding factors would we need to worry about (what else might be true of cows that are played calm, soothing music by their owners?)f. Explain how you would use matching to try to avoid doubts raised by at least one of the confounding factors you identified above.2. Briefly explain how your prospective study above differs from a randomized, controlled experiment.3. Consider the following hypothesis:By wearing seat belts, safety experts claim, we reduce the risk of serious injury or death in auto accidents.If you were to design a retrospective observational study to test this hypothesis:a. What kind of people would be in the experimental group?b. What kind of people would be in the control group?c. What kind of results would confirm (strongly or weakly) the hypothesis?d. Is there anything that would need to be defined in order to make it easier to sort the groups?e. What are the most important confounding factors would we need to worry about (what else might be more likely to be true of people not wearing seat belts, for example?)f. Explain how you would use matching to try to avoid doubts raised by at least one of the confounding factors you identified above.In the following exercises (all taken (but I’ve modified the questions) from Stephen Carey), you’ll find news reports of studies. Read the reports and answer the questions about them.4. A little exercise can help older people sleep better, researchers reported today in a new study. The study is being published on Wednesday in JAMA: The Journal of the American Medical Association.The study, undertaken by researchers at Stanford University, involved 43 sedentary, healthy adults, 50 to 76 years old, with mild to moderate sleep problems, such as taking longer than 25 minutes to fall asleep, and averaging only six hours of sleep a night.Half of those in the study participated in 16 weeks of aerobics, with two hour-long low-impact classes and two 40-minute sessions of brisk walking or stationary cycling each week. The other half did nothing.At the end of the study, the subjects who exercised reported that they fell asleep about 15 minutes faster and slept about 45 minutes longer than before. Those who did no exercise showed little or no improvement.a. What is the causal hypothesis that was tested by this experiment or study?b. Was this a randomized, prospective, or retrospective study?c. Could there have been a placebo effect or experimenter bias at work in this study? Briefly explain if so.d. If it’s not a randomized study, are there any important confounding factors that should make us doubt the study?5. LONDON—Willy Wonka would be horrified. Children who eat too much candy may be more likely to be arrested for violent behavior as adults, research suggests. British experts studied more than 17,000 children born in 1970 for about four decades. Of the children who ate candy or chocolate daily at age 10, 69% were later arrested for a violent offense by age 34. Of those who didn’t have violent clashes, 42% ate sweets daily.The researchers said the results were interesting but more studies were needed to confirm the link. “It’s not that the sweets themselves are bad, it’s more about interpreting how kids make decisions,” said Simon Moore of the University of Cardiff, one of the paper’s authors. Moore said parents who consistently bribe their children into good behavior with candies and chocolates could be doing harm that might prevent kids from learning how to deter gratification, leading to impulsive behavior and violence.a. What is the causal hypothesis that was tested by this experiment or study?b. Was this a randomized, prospective, or retrospective study?c. Could there have been a placebo effect or experimenter bias at work in this study? Briefly explain if so.d. If it’s not a randomized study, are there any important confounding factors that should make us doubt the study?