Conferences

Quality in Statistical Surveys is a complex (multidimensional) concept, which incorporates attributes and characteristics such as timeliness, relevance, accuracy, coherence, comparability, clarity and accessibility addressed to statistical data which is the main product of the various statistical systems and surveys. The quality of these products is defined and measured in accordance with the customer (data user) needs. In a similar way, we consider a statistical frame (or a register) as a product and we define quality with its major attributes and characteristics, focusing on the user, which is the statistical survey itself, needs. In this paper we attempt to define quality components addressed to any statistical frame, using the same attributes and characteristics redefined in context of statistical frames and registers. These components can be expressed in a descriptive and measurable manner through a metadata repository, organized in a SDMX (Statistical Data and Metadata eXchange) standard aiming at the harmonization and evaluation of quality in large surveys repeated in time and space (Official Statistics Surveys, multinational social surveys etc.). SDMX is a standard that facilitates the statistical data and metadata exchange. It offers artefacts for representing structures for reference metadata that can be used to produces metadata sets in XML (i.e. SDMX-ML the XML implementation of the SDMX standard). The artefact for creating such structures is called Metadata Structure Definitions (MSD) and can define all metadata and their definitions. This artifact has been used for representing the metadata for the statistical frames presented in this work. Therefore these metadata values can be represented with SDMX-ML metadata sets and processed by any SDMX tool (e.g. related with storage, transmission, dissemination). This opens a broad area of application providing registers and frames with structural metadata repositories.

Keywords: Metadata; Statistical frames; Quality; SDMX standard; XML