An experimenter working in the urban area wants to determine the average value of a variable highly correlated with race. She is considering the use cluster sampling , with city blocks as clusters and adults within blocks as elements (observation units). Explain why you would, or would not, use cluster sampling in each of the following situations.
a) Most of the adults in certain blocks are white, and most in other blocks are nonwhite.
b) The proportion of nonwhites is the same ine very block and is not clsoe to 0 or 1.
c) The proportion of nonwhites differs from block to block in the manner that would be expected if the clusters were made up by randomly assigning adults in the population to clusters.