FORT EVANS PROJECT

FAQ List

Fort Evans is a small city located in Silica County, California. The city is divided into 2 districts: Northeast and North Central. Imagine that you are the city Census Director and that you have been asked to take a sample of the people and families of Fort Evans in order to estimate the median family income.

Proceed as follows:

1. Your sample should be a stratified cluster sample, where the strata are the city districts and the clusters are families. The families should be selected proportional to the number of persons in the family. You may ignore the difference between sampling with and without replacement. One person is to be sampled from each family chosen for the sample. The sample size is 1000.

Provide the following:

1. A description of the method used to select the sample.
2. A subset of the random indices of the persons chosen for the sample.
2. Weight your sample to:
1. Account for the probability that the family is included in the sample. Describe these weights.
2. Adjust for nonresponse using race (white-nonhispanic, black non-hispanic, hispanic, other) and sex. Use the technique of raking.
Match the city totals:

Males	26100
Females	28638

White non-Hispanic	42790
Black non-Hispanic	5022
Hispanic	5043
Other	1883

Provide a table of the weights to be used.

3. Impute any missing values of education level. Use the hot deck technique with race and sex to determine the class. Provide a list of the elements that had their educational items imputed, and provide the elements that were used to impute the missing items. Estimate the proportion of adults with a college education. Compare your estimate to that obtained without imputation.
4. Graph the distribution function for family income for those in your sample. Be sure to use the weights. Use the distribution function to estimate the median family income.
5. Use the bootstrap to determine a 95% confidence interval for median family income. Provide a histogram of the bootstrap values for median family income.