Discussion: Methodology Architecture - an area under construction

Abstract

A number of statistical agencies have recently launched major transformation programmes, generally with a corporate enterprise architecture at its core; a large number of architectural initiatives are cited by Eltinge et al. [5]. However, Lorenc et al. [6] note that while international cooperation is flourishing in the field of conceptual architectural frameworks, progress has been slower on the methodology side. To remedy the gap between architecture and methodology, Eltinge et al. [5] outline a potential a framework in which methodology is one of the key components. Whereas the potential framework [5] is at a very schematic level, Tam and Gross [7] enter into concrete detail by suggesting that the enterprise architecture for statistical agencies should be extended to include a Methodology Architecture (MA) within the Business Architecture layer. They describe MA as laying out the vision for methodology, supported by an envisaged future inventory of statistical methods and uses, contrasting with the current inventory, and a transition plan for migrating from the current methodological state to the future state. Starting with the CBS [8] inventory of the “as is” state, and proceeding via the generic process repository for existing methods and methodology outlined in the joint European Statistical System (ESS) strategy [9] and the HLG-BAS [10] library of the “could be” state, with an inventory of various possible methods, the MA approach of setting out the “to be” state of the inventory of statistical methods could be viewed as the third logical step in this sequence. The papers which we will discuss here concern Methodology Architecture initiatives at the Australian Bureau of Statistics [1], Statistics New Zealand [2], the Office for National Statistics [3] and the National Agricultural Statistics Service in the U.S. Department of Agriculture [4]. We start off by presenting the key drivers for change (leading to the launch of methodology architecture initiatives) in Section 2. Thereafter, we proceed to present salient features of the four papers with respect to the five key success factors for a methodology architecture identified by Tam [1], i.e. innovation (Section 3), industrialisation (Section 4), contemporisation (Section 5), capacity building (Section 6) and the building of support (Section 7). We then present a few concluding remarks in Section 8.