Our purpose is to consider the efficiency of the county as the primary sampling unit for making state estimates of some socio-economic and agricultural characteristics of the farm population in North Carolina. The main emphasis is on the general purpose type of survey for the State as a whole—one in which both socio-economic and agricultural characteristics are to be estimated; the study is not directed at obtaining ‘best’ estimates for specialized items. Calculations herein presented indicate that the county is not a very efficient primary sampling unit for state estimates. It is expected that future investigation will show that the minor civil division, or some better defined statistical area of comparable size, will produce better results. Variances and relative costs need to be calculated, however, before it is possible to evaluate the relative gains of the smaller over the larger unit.