86 MONTANA AGRICULTURAL STATISTICS 



STATISTICAL METHODS 



GENERAL 



The Montana Crop and Livestock Reporting Service has two primary objectives: 



1. To develop accurate, reliable agricultural statistics 



2. To move those statistics to users as rapidly as possible 



Data collection to achieve these objectives, in most cases, involves sample surveys of farmers and ranchers, 

 grain companies, feed mills, seed dealers, retail stores, and many other agricultural enterprises. 



From data reported by these many sources, statisticians, using all available statistical tools, generate State 

 totals forming official USDA estimates. Cooperation of USDA and the Montana Department of Agriculture 

 provides resources for breaking State totals into crop reporting districts and counties. 



Montana agricultural statistics are available in two dimensions — State estimates and county or crop report- 

 ing district estimates. 



STATE ESTIMATES 



The U.S. Department of Agriculture is required by law to provide National crop and livestock statistics and 

 statistics for each major producing State. 



For many years and to a considerable extent today, mailed non-probability surveys served as the backbone 

 of data gathering. However, probability sampling, a more sophisticated sampling procedure, is being 

 incorporated as funds and applied technology allow. Some types of probability sampling used in Montana are 

 Area Frame Sampling, List Frame Sampling, and Multiple Frame Sampling. Sampling farm operators in a 

 specifically outlined geographic area is known as Area Frame Sampling, while sampling a list of all known 

 producers of a specific commodity is known as List Frame Sampling. A technique called Multiple Frame 

 Sampling takes advantage of the attributes of both the Area and List samples. Area Frame Sampling is used in 

 Montana for a number of acreage and livestock items and Multiple Frame Sampling is used for estimating the 

 cattle inventory and calf crop. 



The concept of probabilit>' sampling is maintained in all three types of sampling. Data collection therefore 

 involves personal enumeration and telephone or personal foUowup for all non-respondents to mail 

 questionnaires. These probability surveys are exjjenslve on a per sampling unit basis but yield State results that 

 have greater precision and allow measurement of the sampling error. 



The Numl)er 1 grain crop in Montana, wheat, is measured by objective yield procedures. Forecasts before 

 harvest incorporate plant counts, kernel and head counts, and weight. Sample fields are entered with permission 

 of growers, by enumerators in May, June, July, August, and September to take counts and measurements. After 

 harvest, gleaning provides estimated harvesting loss. 



COUNTY ESTIMATES 



Federal funding and regulations support State estimates for specific agricultural items that are important in 

 National totals. When additional items or geographic descriptions within States are needed, local funding is 

 required. To provide estimates by county for example, requires a much larger sample because each count>- must 

 have enough reports to provide a statistically sound basis for estimates. Without going into details of statistical 

 requirements and concepts, it is sufficient to point out that if reporting units display the same variation within 

 county as they do within the entire State, then a sample will be required in each county nearly as large as the State 

 sample. 



Large mailings are needed to place enough questionnaires in the hands of producers in each count>'. If 

 response is low in some counties, telephone followup is needed to secure additional reports. Also, Information for 

 minor crop acres Is secured from informed people in each county — such as county agents, ASCS personnel, etc. 

 County assessments and their change from year to year are a reliable indicator for livestock estimates by 

 counties. 



The Montana Department of Agriculture and Wheat Research and Marketing Committee provide funds for 

 local agricultural data. 



CHECK DATA 



In addition to producer surveys, check data are collected from the marketing channels to verify or 

 supplement the grower survey data. These data become available — usually long after harvest and are used to 

 establish final production and marketing figures. 



The final estimate is arrived at when all grower survey data and marketing data (check data such as: 

 handler or processor receipts, slaughter, and market shipments) have been analyzed. These data then become 

 the statistical history against which current forecasts and estimates are viewed. 



