An Assessment of the Current State and Uses of Data Linkage in Household Surveys

Linking survey and administrative data is a practice that many survey agencies worldwide have been implementing for at least a decade now. Although it is believed that data linkage may reduce survey costs and ease respondent and interviewer burden, there is little evidence to support this claim. In addition, linking survey and administrative data is a complex procedure that, in some cases, poses specific methodological challenges. For example, negotiations with data holders are complex and time consuming; consenters are usually not a representative sample of respondents; the actual linkage process is not a straightforward procedures.  The aim of this paper is to critically assess the use of data linkage in household surveys, with a view towards the total survey error paradigm. First, we provide an overview of the ways data linkage is implemented in the different household surveys (i. e., the type of administrative data linked to the survey data, the consent procedures, the linkage process etc), then we discuss the uses of the linked data with respect to their impact on the original survey design (are the linked data used to supplment or substitute the survey data?) and we evaluate whether the linked data are used to enhance the quality of the survey data, for example, by carrying out validation studies