Macro-Integration - Main Module (Theme)


Macro-integration is the process to integrate data from different sources on an aggregate level, to enable a coherent analysis of the data. When there are two or even more independent data sources, inconsistencies will inevitably occur. Discrepancies are caused by various kind of errors, like: sampling error, nonresponse error, coverage error, measurement error and processing error (Eurostat, 2009). Macro-integration can be divided in two stages. In the first stage the source data are adapted to comply with the correct definitions. The second part is called data reconciliation. This is the process in which the remaining discrepancies are resolved or at least reduced at the aggregated level. Data reconciliation is often called balancing in the literature. As mentioned in United Nations (1999, page 23) and Rassier et al. (2007) data reconciliation improves the accuracy.

Data reconciliation can be divided into adjusting for major errors and for the remaining (sampling) noise. Large errors are often corrected manually, by using subject matter knowledge. As correction methods for (large) errors are difficult to formalise, these are not covered well in the literature. A reference for such methods is Bloem et al. (2001, Chapter V). Manual integration is discussed in a theme module in this handbook (“Macro-Integration – Manual Integration”).

Working with statistical data based on samples and questionnaires and influenced by non-response etc. means working with margins of error. Even when samples are perfect and response is 100% there will be inconsistencies. The cause is then a statistical one. In such a case balancing could be done automatically. Compared to methods for the correction of large errors, methods for the correction of noise are much more explored in the literature. In this handbook, there are a modules on four of these methods: “Macro-Integration – RAS”, “Macro-Integration – Stone’s Method”, “Macro-Integration – Denton’s Method”, and “Macro-Integration – Chow-Lin Method for Temporal Disaggregation”.


To read the entire document, please access the pdf file (link under "Related Documents" on the right-hand-side of this page).


Your feedback is appreciated. Please send your remarks, suggestions for improvement, etc. to memobust@cbs.nl.