A multiple frame consisting of both a list and an area frame is used in most agricultural surveys conducted by the National Agricultural Statistics Service of the U.S. Department of Agriculture. The area frame sampling is used to account for the lack of coverage by the list frame sampling. Typically, the area frame estimate of the portion not on the list (NOL) has low precision. To improve upon its precision, certain alternative approaches were investigated. This study of the Agricultural Labor Survey involved post-stratification of all list sample respondents and all NOL sample respondents in California and Florida for the 1991-92 surveys based on some auxiliary data available from the 1991 June Enumerative Survey. A difference estimator was also developed for the NOL portion using the auxiliary variables. No improvement in the precision of the multiple frame estimate was realized under the straightforward application of post-stratification. On the other hand, the difference estimation in conjunction with post-stratification led to some gain in precision and hence proved to be a better approach.