The Post Enumeration Program was conducted in two parts. The first, designated as the P-Sample, was designed to measure gross undercoverage in the 1980 census. As will be explained later, because of methodological problems the P-sample actually overestimates gross undercoverage and an adjustment must be made in the estimation process. The second part, designated as the E-Sample, was designed to estimate gross overcoverage in the census; this would include duplicated and erroneous enumerations.
-Cowan, Charles D. and Fay, Robert E.