The numbered problems are from Rosner, Fundamentals of Biostatistics, 5th Edition. Most of the problems can be solved on your calculator, some will also require statistical tables. Wherever possible, use R to check you answers and submit your R code with your solutions.
Suppose that n = 9 observations X1, ..., X9 each have the same variance s2 but are autocorrelated, with
Corr(Xi, Xj) = rij = r|j - i|
for some 0 < r < 1. Apply Equation 5.13 on page 138 to find the standard error of the sample mean. Show that the standard error increases from s /sqrt(n) to s as r increases from 0 to 1. What is the importance of this result for measurements replicated under laboratory conditions?
Suppose that apples from a certain orchard have a mean weight of 90 g with a 10% coefficient of variation. If they are packed 12 to a bag, what is the probability that the bag will weigh less than 1 kg? Given 100 bags, what is the probability that at least 1 will weigh less than 1 kg? Given 1000 bags, what is the probability that at least 10 will weigh less than 1 kg? State any assumptions you make.