|
Abstract:
|
The International Price Program (IPP) collects data on United States trade with foreign nations and publishes monthly indexes on the import and export prices of U.S. merchandise and services. The IPP employs a three stage PPS design in which establishments, then broad product categories traded within establishments, and finally items within a category, are selected. Certainty selections can occur in the first two stages. We present three variations of the bootstrap rescaling method adapted to the IPP sample design: 1) sampling at the first stage, treating certainty units as probability units, 2) sampling that allows for certainties, and 3) a procedure that extends the previous method by collapsing single item strata. Finally, we compare the precision and bias of the three approaches by simulating 1000 samples of a simulated universe using the IPP sampling methodology.
|
