Beschreibung

"While there is little doubt that linked survey and administrative records offer many scientific research opportunities, most users of linked data are unaware of the linkage methods used to combine the different data sources and the quality of the final linked product. Linkage is typically performed by technicians whose primary job is not research and who may not know for which purposes the linked data will be used. The linkage procedure itself involves several steps, some of which may involve subjective judgments and tradeoffs in which no optimal solution exists. Consequently, errors may be introduced through the linkage, potentially impacting the inferences and conclusions drawn from analysis of the combined data. The purpose of this chapter is to examine various error sources that can be introduced during the linkage process. The remainder of this chapter is organized as follows. Section 25.2 provides a conceptual framework for the possible errors that can be introduced by linking survey and administrative records. Section 25.3 describes the linkage consent process, summarizes the literature on linkage consent issues, and reviews factors that may influence linkage consent rates and biases. Section 25.4 provides an overview of linkage using unique identifiers and summarizes the different errors that can occur. Section 253 examines linkage procedures that rely on nonunique identifiers and reviews the potential errors that can arise. Section 25.6 discusses applications of record linkage and offers some practical guidance. Section 25.7 concludes with a Brief summary of the main take-home points from this chapter." (Text excerpt, IAB-Doku) ((en))