CHAPTER 17

1. Answer briefly.

1. Suppose you see Billiard Ball A approaching, then coming in contact with, stationary Billiard Ball B; when they have come- In contact,. B moves away. State the three conditions that had to be true, according to Hume, when X causes Y. Show how they apply to this case.

2. What's a "stochastic cause"? Why Is that different from a "deterministic cause"?

CHAPTER 18


I. In each case circle S (if the relation Is symmetrical) or N (if It's non-symmetrical)

1. S N Ann is Barney's sibling
2. S N Arthur Is Belinda's brother
3. S N Fire causes smoke
4. S N Exercise Is negatively correlated with heart disease

Il. Give a clear example of:

1. A case In which A and B are mutual positive. causal factors:

A:

B:

2. A case in which A Is positively correlated with B, but In which A Is not a positive causal factor for B.

A:

B:

3. A negative feedback loop Involving A and B.

A:

B:


III. Suppose there is a positive correlation discovered among university students between drinking lots of beer and failing one or more university classes. There are three basic sorts of causal structure that might be behind this positive correlation. Give an arrow-diagram, and an explanation of what's going on In a sentence or two, for each of the three.

CHAPTER 19

1. Answer briefly.

1. Describe the case of Fred and the Birth Control Pills.

2. Show how this Is a case of
(a) correlation

but (b) not of cause.

3. What is "screening off"? Show how the case of Fred and the Birth Control Pills is a case of screening- off.

CHAPTER 20.

I. Answer briefly.

1. Suppose that you're a vet, and In the course of your practice you've noticed that several of the dogs who are brought to you with bad cases of fleas have been fed exclusively on dry dog food. You check, your records, and determine that, of your doggie patients, 35% of the ones who are fed exclusively dry food have flea Infestations every summer. Explain why you cannot conclude from this that there's a positive correlation among your patients between eating exclusively dry food and getting fleas In the summer.

2. What additional information would you need to conclude this?

3. Suppose you had this additional Information, and could conclude that there was a positive correlation in your sample of dogs (your patients). What additional Information would you need to conclude that there's a positive correlation in the population of dogs in general?

4. Suppose now that you could conclude that there was a positive correlation in the general population of dogs. State and give arrow diagrams for FOUR different causal structures which might explain this positive correlation.

5. Now imagine that you have received a research grant and the use of a big laboratory to carry out experiments to determine which of your causal hypotheses Is correct. Explain In detail how you would run these experiments. Make sure to say what various results of each of the experiments would establish.

OMIT PART 5 ABOVE FOR NOW

CHAPTER 21

I. Define each term briefly:

1. Blind testing

2. Double-blind testing

3. A linear cause-effect relationship

II. Answer briefly.

1. Explain why the results of the fetal-alcohol-syndrome study discussed in the test appear to show the "'threshold effect."

2. The study does show that alcohol consumption causes damage to the fetus. Explain why the study does not show that a pregnant woman who wants to make sure not to harm her fetus should stop drinking altogether.

3. In the (imagined) study discussed in the book, kumquat consumption as low as one per month is shown to be a positive causal factor for cancer of the toenail, increasing the risk by tenfold. Explain why it does not necessarily follow that you should stop eating kumquats. How is this different from the fetal-alcohol example, above?

4. What is a placebo? Explain how use of a placebo helps to accomplish sample matching of the experimental and control groups.

CHAPTER 22

Answer each question briefly.

1. What's the difference between an experimental study (what is sometimes called a randomized comparative experiment) and an observational study (such as a prospective or retrospective study)?

2. Explain why experimental studies, when practical, are preferable to observational studies. Cite several reasons.

3. What's the difference between a prospective and retrospective study?

4. What's the difference between a "real time" prospective study, and an "after-the-fact prospective study?

5. Why is a real-time prospective study more likely to provide better information than an after-the- fact prospective study?