EDI: ELECTRONIC DATA INTERCHANGE
FOR STATISTICAL DATACOLLECTION AND DISSEMINATION
W.J. Keller, Statistics Netherlands, The Netherlands
In this paper, we will present some experiences in the Netherlands with EDI for statistical
datacollection and dissemination. We will consider the changes to be made for large scale EDI
datacollection. We will argue that EDI demands a dramatic redesign of the way we collect and
process statistical information, but the rewards in terms of response burden, quality and
efficiency might be well worth it.
We will also discuss some projects at Statistics Netherlands dealing with EDI for statistical
dissemination. We will cover Statline (our statistical database with a traditional on-line querytool
on a remote DOS client) and its new experimental version, with a so-called dynamic WEB on
Internet. We will argue that the Internet not only provides great opportunities but also great
challenges for statisticians.
We will focus, besides on the technological aspects, on the conceptual and organisational
implications EDI for Datacollection and dissemination.
Keywords: Official Statistics, Datacollection, Dissemination, EDI, Internet, Meta-Information
1. OFFICIAL STATISTICS
National Statistical Institutes (NSI’s) are confronted with several strategic issues resulting from
new demands from our customers as well as new developments in Information Technology (IT).
Efficiency and market-orientation are the key-words now. We need to produce at lower costs.
Furthermore we need to lower the costs we inflict upon our suppliers of data. The outcome
should be a product that, although not actually sold on a market, our clients eventually want.
Furthermore we are confronted with new developments in IT. They will give us the opportunities
to construct the necessary tools to meet the new demands. In a situation like this a NSI needs to
make the right strategic choices.
The statistical production process is influenced by the growing demands of our clients (output)
and of our respondents (input). Co