2. Read the Attached Article and Answer the Following Questions

HON 180 2010
Assignment 1 Brief Answers

1. Do Review Exercises Chapter 2 pp. 24 -26 1, 2, 4, 7, 8, 9, 11

2. Read the attached article and answer the following questions:

a. Was the study described in the article an observational or a randomized controlled study? Briefly explain.

b. Carefully describe the treatment group and the control group in the main study reported in this article.

c. The article mentions “previous research based on self-reporting of physical activity has produced similar results.” Why did researchers have concerns about research studies involving self-reporting? How did the researchers address the concern in this study?

d. Briefly explain why Amy Zlot, a genetics researcher, said that more research is needed to see if the results hold in other ethnic groups.

e. Describe at least one source of potential confounding in this study and explain why it is a source of potential confounding.

Remember the definition of a confounding variable: a factor that is relevant to the outcome of the study that the treatment and control groups differ by, other than the treatment itself. For example, consider the true fact that “smokers have shorter life expectancies than non-smokers.” A confounding variable is gender. Here is why. Gender is relevant to the outcome of the study (longevity), because men in general have shorter life-spans than women. Furthermore, the treatment and control groups (smokers and non-smokers) differ in their gender makeup: More smokers are male than female.
Note on explanations relating to confounding variables. When the authors of our textbook and I ask questions such as “Why did the investigators adjust for age? education? marital status?” it is incomplete to say “Because age, education, and marital status are potential confounding variables.” It is completely obvious that they are potential confounding variables; to just say that they are potential confounding variable is superficial. To receive full credit, you must explain how each variable could affect the outcome of the study.

37 points total

1. (a) (2 points) The investigator is wrong. He is not taking into account the size of the population of the state. Michigan in fact has twice the population of Minnesota. So in fact Michigan has lower crime rate than Minnesota!

(b) (2 points) This statement is definitely true; the population of the US increased during this time period, so the crime rate in fact did go down. People in the US abided the law more. Note as in 1(a), it would be better to take the population into account by calculating a rate from the start. Because of the increase in population the decrease is even more dramatic than the comparison between 28000 and 22000 suggests.

2. (a) (2 points) No, we cannot conclude that American thieves prefer American cars. What is missing from the equation is that there are far more Corvettes on the road then Q45’s. For example in 2002, there were 33586 Corvettes sold compared to 8065 Q45’s. Other confounding variables could be the ease with which one can steal the cars can differ, but that is most likely a smaller effect than the availability. Both cars are rather expensive so the difference in the value of the cars cannot account for the difference either.

(b) (2 points) Similar to (a), there are far more 3-series BMW’s on the road than 7-series.

(c) (2 points) This statement is false. For example, a rate could be high even with a large denominator. The rate is low because of the relationship between the numerator and the denominator: the numerator is small compared to the denominator. However with a fixed numerator, the number with a larger denominator will be smaller.

4. (a) (2 points) They were controlling for age and gender as possible confounders. Men and women differ in their overall health levels and may have different responses to smoking and stopping smoking. Similarly different age groups differ in health and could have different responses to smoking (e.g. smoking may not affect young people as much older people).

(b) (2 points) Many smokers give up smoking because they are sick. So recent ex-smokers include a lot of sick people! A number of you mentioned the fact that ex-smokers sometimes suffer from withdrawal symptoms, so they might be less healthy because of these withdrawal symptoms. This fact may also play a role, but the dominant one is the first one mentioned.

7. (a) (1 point) It is an observational study.

(b) (2 points) You should address all three. 1) Age. Rates of cervical cancer go up with age while younger women use the pill more. 2) Education. Women of different education levels have different patterns of sexual activity (frequency, number of partners, protection, use of contraceptives) 3) Marital status. Women of different marital status also have different patterns of sexual activity (married women have fewer partners and have a lower rate of contraceptive use among other things).

(d) (1 point) No, the conclusions of the study were not justified because of the confounding discussed in part (c).

8. (2 points) There is a little over a quarter of the year between Memorial Day and Labor Day, so if burglars were continuously active at the same level, more than 25% of the burglaries would occur during this period. Some of you pointed out that the 25% figure makes no direct connection between people going on vacations and burglaries (e.g., people go on vacations other times of the year as well). This is true, but the authors of the book wanted you to note the fact that there is a little over a quarter of the year between Memorial Day and Labor Day.

9. (a) (1 point) False. They failed to confirm the observational studies.

(b) (2 points) True. People who eat lots of fruits and vegetables can differ in many ways other than the food they eat. They can have different lifestyles (e.g. different levels of exercise) and even live in different areas (areas with cleaner air and water, better health care).

(c) (1 point) False. Well-design randomized control studies should produce sound results by eliminating the possibility of confounding.

11. (a) (2 points) The treatment group consists of those who finished boot camp. The control group consists of other prisoners—including those who do not volunteer, or those who volunteer but do not complete the program.

(b) (1 point) This is observational. The prisoners decide whether to volunteer for boot camp and whether to stay in the program or drop out. That is the problem: those who volunteer and stay the course might be quite different from those who volunteer but drop out. They might differ in motivation, types of crimes committed, quality of the network of friends and family outside of prison…

(c) (1 point) False because it is an observational study and there is the possibility for confounding as described in part b.

2. (a) (1 point) This is observational study. The subjects themselves determined who was in the treatment group or the control group by leading an active lifestyle or not.

(b) (2 points) The treatment group was Amish people with the gene who had very physically active lifestyles. The control group was Amish people with the gene who did not have very physically active lifestyles. The treatment group was not Amish who have the gene.

(c) (1 point) They were concerned that self-reporters might overreport or underreport the amount of physical activity they do. It is also somewhat difficult to keep track of one’s physical activity. Researchers in this study addressed their concern by asking participants to wear electronic monitors.

(d) (1 point) All the participants in the study (treatment and control) were Amish, so the study is only valid (in so far as it is valid) for Amish only. There may other genes involved which may occur with higher prevalence in the Amish community.

(e) (2 points) Note that in this case, a confounding variable will be a factor that is relevant to begin overweight that the treatment group (high activity persons with the gene) and the control (low activity persons with the gene) differ in. Some possibilities are age, gender, diet. For example, the treatment and the control groups may differ in their average age. For example there might be more young people in the physically active group (the treatment). Young people tend by less overweight than people in their 40’s and 50’s. So age could be a confounding factor. There might be more males on one of the groups, and males might respond to the gene differently and have different patterns of overweight.