ABSTRACT
Sampling frames for rural household surveys include residential and commercial directories of addresses and telephone numbers and directories of all possible telephone numbers. The choice of a sampling frame affects the efficiency of the sample design. Blanks and foreign units cause, at a minimum, increased sampling error. The sampling frame of all possible telephone numbers is built by adding four digit suffixes to a list of three digit area codes and three digit prefixes obtained from Bell Communications Research. The cost of selecting the sample from an area frame is also relatively more expensive than the cost of selecting a sample from a list frame. Dual frame methods offer increased strengths when evaluated against the four factors—coverage, coverage bias, sampling efficiency, and cost. The increase in sampling error is caused by a positive intracluster correlation coefficient. In short, the efficiency of design can be greatly affected by the choice of the sampling frame.
