Journal of Official Statistics, Vol.8, No.3, 1992. pp. 351–375
Some Aspects of Measurement Error in the United States Objective Yield Survey
Mark Reiser, Ron Fecso, and Meng-Koon Chua
Abstract:In this paper we consider aspects of measurement error in the United States Objective Yield Survey for corn, which is conducted by the United States Department of Agriculture National Agriculture Statistics Service. Various models are used to assess measurement error in variables that appear in forecasting models for end-of-season yield. Two major components of yield are number of ears (per sample unit) and average grain weight per ear. Variables that are used in forecasting number of ears are simple counts, such as number of stalks, and have very high reliability ratios. The reliabilities for variables that are used in forecasting grain weight, although adequate, are substantially lower than those found for number of ears. We also consider a model for the reliability of yearly means, and discuss the effect of measurement error on end-of-season forecasts.
Keywords:Agricultural statistics; forecasting; reliability; structural equations; LISREL.
Copyright © Statistics Sweden 1996-2018. Open AccessISSN 0282-423XCreated and Maintained by OKS Group