CHAPTER 13 PART 2Observational Studies and Experiments

Experimental units are not always drawn

randomly. For example, medical experiments may

deal only with patients who have the disease being studied. However, treatments should be assigned to

subjects randomly.

A single-blind experiment occurs when either the subjects or the evaluators are unaware of which subjects are receiving which treatment.A double-blind experiments occurs when BOTH the subjects and the evaluators are unaware of which subjects are receiving which treatment.

A control group receives no treatment.A placebo may be used so it is unknown during the experiment that the control group is not receiving the treatment.

It is not unusual for 20% or more of subjects given a placebo to report improvements. This is called the

placebo effect.

An article in a local newspaper reported that dogs kept as pets tend to be overweight. Veterinarians say that diet and exercise will help these chubby dogs get in shape. The veterinarians propose two different diets (Diet A and Diet B) and two different exercise plans (Plan 1 and Plan 2). Sixty dog owners volunteer to take part in an experiment to help their chubby dogs lose weight.Diet A: owners control the portions of dog food and dog treatsDiet B: a mixture of fresh vegetables with the dog food and substitute regular treats with baby carrots Plan 1: three 30-minute walks a week Plan 2: 20-minute walks daily1)Identify the subjects 2)Identify the factors and the number of levels for each 3)Identify the number of treatments 4)Identify whether or not the experiment is blind, double-blind, or neither 5)Identify the response variable 6)Design an experiment to determine whether the diet and exercise

programs are effective in helping dogs to lose weight

1) 60 chubby dogs 2) Diet and exercise, 2 levels each 3) Four treatments: Diet A, Plan 1; Diet A, Plan 2;

Diet B, Plan 1; Diet B, Plan 2 4) The design is at best single-blind since the

owners know which plan their dogs are on but the evaluators don’t know

5) Weight loss

Answers

A group of people are concerned that the coach of a local high school men’s and women’s basketball teams alters the amount of air in the basketball to gain an unfair advantage over opponents during home games. The idea is that the basketballs are pumped up with one pound per square inch less air than required, and his teams practiced with these altered balls all week prior to home games. Since these under-pumped basketballs would react differently to being shot at a basket, the team that practiced with these balls would have an unfair advantage when it came to shooting free throws.1) Describe how to use a retrospective study to determine if the

home teams have an unfair advantage when shooting free-throws.

2) Describe how to use a prospective study to determine if the home teams have an unfair advantage when shooting free-throws.

Answers 1) Past games are reviewed to see how many

free-throws the team made when they had the altered basketballs vs. when they didn’t. The differences in the percentage of free-throws made would be compared between the two types of basketballs.

2) Teams are selected and then their future free-throws are recorded to see if they improve when they have the altered basketballs. The differences in the percentage of free throws made would be compared between the two types of basketballs.

Researchers plan to investigate a new medication that may reduce blood pressure for individuals with higher than average blood pressure. 90 volunteers with higher than average blood pressure are solicited. Volunteers are randomly assigned 100 mg of the medicine, 200 mg of the medicine, or a placebo. Blood pressure will be measured at the beginning and at the conclusion of the study.1) Identify the subjects, treatments, and response

variable 2) Describe an advantage to random assignment of

treatment 3) Describe an advantage of the placebo 4) Is this study blind?

Answers 1) Subjects: 90 volunteers with high blood

pressure; 3 treatments: 100 mg, 200 mg, placebo; response variable: change in blood pressure

2) Randomization will equalize variability for which we cannot control. We may be able to establish causation.

3) We have an extra comparison with a control group. Blood pressure may change due to other variables.

4) The subjects are blind, assuming the 100 mg, 200 mg, and placebo look the same.

Today’s Assignment:Make sure you read the book! HW: p.312 #6-8, 10, 22, 24, 26, 33 Quiz Friday