Question

A current area of research interest is in the familial aggregation of cardiovascular risk factors. Suppose it is known that, overall, the average cholesterol level of children in the US is 175 mg/dl. Further assume that, as a researcher studying cardiovascular disease, you have access to a large number of hospital records of US mean who have died during the past year due to heart disease. You wish to conduct a test to determine if the mean cholesterol level of the children of these men is the same as children overall. You recall from the wonderful experience you had in your college biostatistics course that you can estimate the sample size required to conduct a test at a certain level and power. Suppose you wish to conduct a test at the 5% level to detect a difference between a true mean value and null hypothesis mean value of 15mg/dl with 90% power. Assume a population standard deviation of σ=50. Answer the following questions. Briefly describe whether you would conduct a one-sided test or two-sided test and why? According to your answer in part 6a, state the null and alternative hypotheses using notation established in class. Is a sample of size 90 large enough to conduct the test with the desired level and power? Show your work in determining this answer to receive credit.

Answer 1

This would be 2-tailed test

as she wish to conduct a test to determine if the mean cholesterol level of the children of these men is the same as children overall.

sigma = 50, alpah = 0.05, difference = 15 ,power = 90%

Using Minitab

stat -> power and sample size -> 1-sample z

n = 117

n = 90 is not large enough