Refer to Exercise 16.1. After collecting the data and testing for parallelism of the four lines relating the reduction in blood pressure to the age of the subject, the researcher finds that there is significant evidence that the lines are not parallel.

a. Describe how would you test for differences in the adjusted treatment means?

b. Provide the form of the estimated mean response for supplement 1 for subjects of age 45 years.

c. The researcher wants to evaluate the supplements for subjects of age 80 years or older. What problems may she encounter using the data in her current study, if any?

Exercise 16.1

A researcher designs a study to evaluate three dietary supplements that are reputed to lower the systolic blood pressure reading for people who have high blood pressure. A inert supplement is included to evaluate the placebo effect. Twenty subjects all having systolic readings higher than 160 mmHg are randomly assigned to each of the supplements and to the control. The researcher is concerned with the disparity in age of the 80 subjects (20–60 years old) and thus wants to include the effect of age in the model also. Write a general linear model in which the response variable y, the change in systolic blood pressure after 6 months of treatment, is linearly related to the age of the subject A for each of the three supplements and the placebo. From previous studies, the researcher determines that the relationship between the reduction in blood pressure readings and age may be substantially different for the three supplements and the placebo. Identify all the parameters in your model.