Abstract:
|
Reducing duplication and matching records lacking unique identifiers are common practices associated with the construction and maintenance of a list sampling frame. The U.S. Department of Agriculture's National Agricultural Statistics Service (NASS) employs a record linkage system built using AutoStan and AutoMatch (originally developed by MatchWare Technologies) to maintain its list frame of farm operators and agribusinesses. The overall process consists of four steps: 1) reformatting, 2) standardizing (AutoStan), 3) matching (AutoMatch) and 4) review. Because the current matching engine is no longer supported and becoming increasingly obsolete, NASS has recently begun to explore alternative software options, such as Statistics Canada's G-Link package. In this paper, we describe the results of a preliminary study comparing G-Link with AutoMatch using list frame data from a national survey of organic farmers and discuss issues associated with upgrading the agency's record linkage system.
|