The Vehicle Inventory and Use Survey (VIUS) has been conducted by the U.S. government to provide information on the number of private and commercial trucks in each state. The stratified random sampling design is described in U.S. Census Bureau (2006b). For the 2002 survey, 255 strata were formed from the sampling frame of truck registrations using stratification variables state and truck type. The 50 states plus the District of Columbia formed 51 geographic classes; in each, the truck registrations were partitioned into one of five classes:
1. Pickups
2. Minivans, other light vans, and sport utility vehicles
3. Light single-unit trucks with gross vehicle weight less than 26,000 pounds
4. Heavy single-unit trucks with gross vehicle weight greater than or equal to 26,000 pounds
5. Truck-tractors
Consequently, the full data set has 51 × 5 = 255 strata. Selected variables from the data are in the data file vius.dat. For each question below, give a point estimate and a 95% CI.
a. The sampling weights are found in variable tab trucks and the stratification is given by variable stratum. Estimate the total number of trucks in the United States.
What should your response variable be? Why is the standard deviation of your estimator essentially zero?
b. Estimate the total number of truck miles driven in 2002 (variable miles_annl).
c. Estimate the total number of truck miles driven in each of the five truck type classes.
d. Estimate the average miles per gallon (MPG) for the trucks in the population.