Area frame sampling in agriculture

CONCEPTS AND GENERAL PRINCIPLES OF AREA SAMPLING, RATHER THAN SPECIFIC APPLICATIONS, ARE DISCUSSED. SAMPLING FRAMES SHOULD BE CONSTRUCTED IN RECOGNITION OF THE FACT THAT AGRICULTURE IS COMPOSED OF NUMEROUS SUBPOPULATIONS THAT MUST BE SAMPLED. A SAMPLE DESIGNED EFFICIENTLY FOR ONE SUBPOPULATION MIGHT BE OF LITTLE VALUE FOR ANOTHER. THUS SEVERAL SAMPLING FRAMES MIGHT BE REQUIRED OR IF A SINGLE SAMPLING FRAME IS TO BE CONSTRUCTED, IT PROBABLY SHOULD BE MULTI-PURPOSE. THE DEVELOPMENT OF IMPROVED SAMPLING FRAMES IS CALLED FOR BY: 1) THE TREND TOWARD LARGER, MORE SPECIALIZED FARMS, 2) THE GENERAL DEMAND FOR MORE ACCURATE STATISTICS, AND 3) THE NEED TO KEEP SAMPLE SIZES AND COSTS AS LOW AS POSSIBLE. ALSO, TO SOME DEGREE, SAMPLE SIZE IS RELATED INVERSELY TO CAPABILITY OF CONTROLLING NON-SAMPLE ERROR, WHICH ALSO IS IN FAVOR OF EFFICIENT SAMPLING TO KEEP SAMPLE SIZES AS SMALL AS POSSIBLE. SURVEY PLANS SHOULD INCLUDE PROVISION FOR STUDIES OF SAMPLING VARIANCE, RESPONSE ERRORS, COVERAGE ERRORS, AND COSTS. SUCH STUDIES SHOULD PROVIDE A CONTINUING BASIS FOR ADJUSTING THE ALLOCATION OF RESOURCES IN AN EFFORT TO ACHIEVE MAXIMUM ACCURACY AT A GIVEN COST.