Vol. 38, No. 1,  Spring/Summer 2000- "The Pennsylvania Geographer"

SMALL GEOGRAPHIC AREA ESTIMATION

CONTRIBUTIONS TO FEDERAL AGRICULTURAL DATA:

A PENNSYLVANIA CASE STUDY

Daniel A. Griffith Department of Geography, Syracuse University

Abstract

This report describes a flexible methodology for calculating spatial autoregressive model-based small geographic area estimates, illustrating its utility with substitution computations for the suppression code (D) in order to allow more complete data tabulations to be released. Post-stratification figures reported here are those for the 1997 Census of Agriculture, which was conducted by the National Agricultural Statistics Service (NABS). The imputation equation was developed and evaluated in a previous study, using principal agricultural commodities for Michigan and Tennessee. Both major and minor commodities for Pennsylvania are treated in this study. The text describes applications of the estimation methodology, evaluating differences between two operational definitions of county densities. Measures of uncertainty are presented using oats production data, based upon Monte Carlo simulations, the jackknife technique, and asymptotic standard errors. The need for constrained maximum likelihood estimates is implied by the analysis of minor crops, regardless of whether the cultivation of such crops is thinly spread across most counties, or concentrated in relatively few counties. One important finding is that reasonably accurate model-based small geographic area estimates can be obtained even for minor crops.