An official website of the United States government
In evaluating the prospective integration of survey data with administrative record data, one generally needs to balance several factors related to cost and data quality. In practical applications, however, these factors are not known with perfect precision, and measurement of them often requires a substantial effort by the statistical organization. This paper evaluates some ways in which the balance between data quality and cost can be affected by errors in estimation of these quality and cost factors. Principal emphasis centers on a simulation study involving sample allocation in the presence of uncertainty regarding stratum-level variances and per-unit costs.