THE ZONE DEFINITION PROBLEM IN SURVEY-RESEARCH - AN EMPIRICAL EXAMPLEFROM NEW-YORK-STATE

This paper examines problems of zonal definition in the context of a r ecent empirical project on the geography of inter-firm linkages in New York State. It is argued that the results of a spatially structured s urvey of private companies can change significantly depending on the m anner in which the study regions of the analysis are defined. Variatio ns in the composition of study regions are a result of either changes in spatial scale, or spatial zoning at any one scale. This point is il lustrated with data from 472 New York State manufacturing firms aggreg ated into four different zonal systems. The results of the analysis su ggest that misleading interpretations of spatial data can emerge, even when logical boundaries are selected from the outset.