Survey Methods 



For surveys conducted through spring 1984, USDA's 

 National Agricultural Statistics Service (NASS, formerly 

 the Statistical Reporting Service) drew a random sample 

 from all producers who responded to acreage and 

 production surveys. NASS sampled farms from the list of 

 respondents with a probability proportional to size so that 

 each unit of production (an acre or an animal) in the 

 region had an equal chance of being included in the survey. 

 Thus, these survey data tend to represent acreages of 

 surveyed crops on the larger commercial farms because 

 such farms have more acres. For example, a 1,000-acre 

 farm with 400 acres of a crop had twice the probability of 

 being sampled as a 1,000-acre farm with 200 acres of the 

 same crop. With the spring 1985 survey, NASS changed to 

 a full probability, multiframe sample. This change meant 

 choosing a sample from both a list frame (a list of 

 producers in each State) and an area frame (a random set 

 of land segments). By using a full probability survey, ERS 

 can statistically test future estimates for accuracy. 



The current survey is referred to as the Farm Costs and 

 Returns Survey (FCRS). FCRS is an annual survey of 

 farm costs, returns, and production practices. It is 

 conducted by ERS and NASS each winter in 48 States. 

 About half of the approximately 24,000 respondents are 

 asked to complete a detailed expenditures and receipts 

 questionnaire that asks for indepth information about the 

 farming operation's costs and returns. Other respondents 

 are asked for less detailed whole-farm costs and returns 

 data and for COP information for one of several specific 

 commodities. For example, Eastern and Western cow/calf 

 operations and those involved in dairy, hog, and fall potato 

 production were surveyed in 1986 for 1985 COP data. In 

 1985, data on sugar beet, pasture/forage, rice, and burley 

 tobacco crops were gathered. Because of limited sample 

 sizes and high enumeration costs, USDA cannot survey 

 each COP commodity every year, so each crop or livestock 

 enterprise is surveyed on a 4-year rotation. ERS updates 

 these costs between survey years, using recent yields, 

 prices, and price indexes provided by NASS. 



procedure is impractical, however, because few farmers 

 keep records by individual enterprise. Even those who 

 keep enterprise records generally use different accounting 

 methods, making it difficult to obtain comparable costs. 



Rather than solicit per-acre or per-unit costs directly from 

 farmers, ERS uses more general methods. COP 

 questionnaires from the FCRS obtain such information as 

 field operations, size and type of equipment, fertilization 

 rates, seeding rates, custom operations, and labor use. 

 This information complements price and quantity data 

 available from other surveys, including crop acreages, crop 

 yields, prices received for agricultural commodities, prices 

 paid for agricultural inputs, and fertilizer use on some 

 commodities. Other data include land values and rental 

 rates. COP surveys ask for the amount of hand labor for 

 crops, as well as labor used for livestock production. 



ERS estimated costs in Economic Indicators of the Fann 

 Sector: Costs of Production, 19S5 (5)^ by using data from 

 the following surveys: 



J/The 1982 survey was conducted in the Southern States 

 and the Pacific Northwest; the 1983 survey was conducted 

 in the Northern States. For sugar, partial results from 

 the 198A survey have been incorporated in the 1984 and 

 1985 estimates. 



The simplest procedure for determining the cost of 

 producing a commodity might appear to be to ask farmers 

 to report their costs for producing that commodity. This 



i/Italicized numbers in parentheses refer to items in the annotated 

 bibliography at the end of this volume. 



