Linking data resources and data imputation

In modern surveys, scholars increasingly tend to link data from a variety of sources. Such links may be established on the basis of a person’s unique characteristic, such as a registration number. But it may also be done statistically on the basis of sex, age, income, and educational attainment level. This latter method of linking is in fact a kind of data imputation, supplementing missing data in a justified way.

CentERdata has applied this method in estimates for education and the labour market: we uniquely linked salary record data to National Civil Pension Fund (ABP) data on early retirement. Another example is a survey in which we statistically linked income data from the CentERpanel to a Housing Needs Survey (WBO).