Abstract

The Crop Production Survey (CrPS) is the main source of data on production, area harvested, and number of bearing trees/hills/vines for crops other than palay and corn at the national and sub-national levels in the Philippines. It is one of the statistical activities conducted quarterly by the Philippine Statistics Authority (PSA). On the current design, large farms are completely enumerated to at most 5 farms for the crop in a province. The rest of the farms are selected in two stages. The first stage of sample selection is made at the municipality level (PSU) while the second stage is at the farmer-producer level (SSU). The limitations observed in the current design is that it may not be able to include those barangays that are top producers of the commodity as samples because it uses city/municipality as PSU, and it combines into one sampling design the household-based and establishment-based farms. Given the situation, a simulation-based sampling experiment was used following a data-generating process based on models with parameters estimated from the time series data of official statistics on the production of other crops and used to evaluate the performance of the different sampling designs. The new sampling design is expected to generate more efficient estimates for the CrPS specifically for household-based farms and establishment-based farms. Further, the results of the pilot survey suggest that the new sampling design of CrPS Household be a two-stage stratified sampling with two-way stratification.