Over the past five years, Statistics Netherlands has been working on an ambitious redesign program. The general ideas of this program are embedded in a comprehensive enterprise architecture. More recently, the architecture has been complemented with a series of standard methods and a box of standard tools or combinations of tools that can be used to further streamline the core production process.
Before standard methods or tools can be readily applied in the production of statistics, the statistical processes themselves need to be (partly) standardised. For this purpose, a conceptual business model for processing statistical data was developed. An important concept in this model is the standard process step. A standard process step is a functional component that can be (re)used in a statistical process to meet one or more specific statistical targets. Underlying these components are generic (statistical) methods. Several standard methods may usually provide the same functionality. On the other hand, a specific standard method can be used for different functions. The idea is to design processes in terms of standard process steps rather than methods, in order to take account of this complex multiple relationship. This facilitates the re-use of building blocks and increases the transparancy and flexibilty of the designed processes. By identifying these standard steps and providing guidelines for their use, the model also aims to close the gap between the high-level view taken in the business architecture and designing statistical processes in practice.
The model and its concepts will be illustrated by applying it to the field of data-editing. Based on an overall data-editing strategy, standard editing methods are further unravelled in a limited number of functions and process steps. Further research involves identifying all standard process steps involved in processing data, as well as applying these to complete statistical production chains.
Keywords: Standardising statistical processes; Business architecture; Standard statistical methods; Generic building blocks
Biography: The presenter has been working at Statistics Netherlands for 15 years. First, she worked as a methodologist, specialising in sampling theory and multivariate statistics. For the past 3 years she has worked in the field of business architecture. Her work includes developing new ideas and policies in this area as well as assisting in redesigning statistical processes in practice, in accordance with the architecture.