APPENDIX B. 

 Effect of the Area Sample on Census Comparability 



The 1982 and 1978 Censuses were conducted 

 primarily by mail. Despite every effort to assemble a 

 complete list, a small portion of the Nation's farms were 

 not included on the census mail list. To improve the 

 coverage of the 1978 Census, especially in counting the 

 number of small farms, the mail-out/mail-back 

 enumeration was supplemented by the direct 

 enumeration of all households in a sample of areas in 

 all States, except Alaska and Hawaii. Due to budget 

 reductions, the direct enumeration sample was 

 eliminated in the 1 982 Census. 



In late 1978, enumerators visited all households in 

 approximately 6,400 segments in rural areas (areas with 

 less than 2,500 population) and completed a census 

 form for each agricultural operation. These forms were 

 then matched to the census mail list. Data from those 

 cases which were not matched to the mail list were 

 used to estimate the number and characteristics of 

 farms not on the mail list at the State, regional, and 

 national levels. No county-level estimates were 

 developed because the area sample size was 

 insufficient for reliable estimates at that level. 



The U.S., region, and State data for 1978 shown in 

 the 1978 Census of Agriculture publications included 

 data for farms represented on the mail list plus 

 estimates from the area sample for farms not on the 

 mail list. To provide comparable data for 1982 and 

 1978, estimates from the area sample have been 

 subtracted from the 1978 data. Thus, all 1978 data 

 presented in the 1 982 Census of Agriculture publications 

 include data only for farms on the 1978 mail list. 



In 1978, the area sample farms accounted for 9 

 percent of all farms in the United States, but only 1 

 percent of the total value of agricultural products sold 

 and 1 percent of the land in farms. The area sample 

 farms represented 25 percent of all farms with sales of 

 less than $2,500. The contribution of the area sample 

 farms to the total farm count varied widely by State, 

 from a low of 2.0 percent in North Dakota to a high of 

 23.8 percent in New Hampshire. 



The following table shows the 1982 results and 

 adjusted 1 978 data for farms on the mail list, as well as 

 1978 data which include estimates for farms not on the 

 mail list. 



Effect of 1978 Area Sample on Census Comparability: 1982 and 1978 



tFor meaning of abbreviations and symbols, see introductory texX] 



