Chair: Sorina Vaju
Room: S3B Sukiennice
Time: 10:30 - 12:00
Date: 29 June
|Title: <<< Quality Guidelines for Multisource Statistics >>>
The Guidelines on the quality of multisource statistics are being produced in the framework of the ESS.VIP Admin project, as a work package of the Essnet Quality of multisource statistics. They are intended to be practical and applicable in the National Statistical Institutes of the ESS, when several administrative sources are used exclusively or in combination with survey data to produce statistics. The guidelines are addressed to process managers who can use them in the design and the implementation of multisource statistical processes, to identify the factors having higher impact on output quality and to monitor and assess quality. First of all a review of existing quality guidelines published at national level has been carried out, with the scope of identifying the elements relevant for the development of guidelines within the project. The structure and the content of the guidelines have been defined and reviewed taking into account the comments from Eurostat and the ADMIN network of contact points. A draft version of the guidelines is expected by the end of the current Specific Grant Agreement, i.e. in September 2018, however some areas, for which some methodological work is still required, are expected to be further finalised in the framework of the third specific grant agreement of the Essnet. The paper will describe the quality framework adopted in the guidelines, will present the structure of the guidelines and a sample of the content. The guidelines are structured according to Eurostat statistics quality dimensions and take into account: the different basic data configurations when using multisource data, the phases and sub-processes where the errors generate, examples of measures and quality indicators. This work is carried out in collaboration among the members of the Essnet.
|Title: <<< Administrative data and quality, guidelines towards better quality of administrative data >>>
Statistical authorities need to produce data faster in a cost effective way, to become more responsive to users´ demands, while at the same time providing high quality output. One way to fulfil this is to make more use of already available data sources, and in particular administrative sources, most typically used in combination with other sources. Depending on the use of the administrative sources and the data configuration different statistical tasks must be applied. Usually it is not only one task but a sequence of different tasks that have to be applied, for example, data integration, imputation and editing or tabulation. For these tasks different methods are available and depending on the input data quality and the data configuration the same method can have limited use or produce lower quality outputs. The use of administrative data sources risks impacting negatively quality on several dimensions, in particular accuracy and comparability. Surveys and administrative sources have both particular strengths and weaknesses. Combining them may overcome these weaknesses, provided that suitable methodology and tools are used. At the same time, harmonised measures of quality for outputs that combine administrative sources with other sources (surveys) are necessary to ensure that European Union official statistics are of sufficient quality and fit for their intended use. This paper looks at the most frequent methodological challenges faced when integrating administrative sources and provides, for typical situations, preferred methods to have the best quality of statistical output. It also introduces the work of ESSnet on the Quality of Multisource Statistics (KOMUSO) to develop quality measures and guidelines related to the use of administrative sources.
|Ton de Waal|
|Title: <<< Quality Measures and Indicators for Multisource Statistics >>>
The ESSnet on Quality of Multisource Statistics is part of the ESS.VIP Admin Project. The main objectives of that project are (I) to improve the use of administrative data sources and (II) to support the quality assurance of the output produced using administrative sources. The ultimate aim of the ESSnet is to produce quality guidelines for National Statistics Institutes (NSIs) that are specific enough to be used in statistical production at those NSIs. The guidelines are expected to take the entire production chain in to account (input, process, and output). They also aim to cover the diversity of situations in which NSIs work as well as restrictions on data availability. The guidelines will list a variety of potential indicators/measures, indicate for each of them their applicability and in what situation it is preferred or not, and provide an ample set of examples of specific cases and decision making processes. Work package 3 of the ESSnet focuses on developing and testing quantitative measures and indicators for measuring the quality of output based on multiple data sources and on methods to compute such measures and indicators. Examples of such quality measures and indicators are bias and variance of the estimated output. Methods for computing these and other quality measures and indicators often depend on the specific data sources. Therefore we have identified several basic data configurations for the use of administrative data sources in combination with other sources, for which we proposes, revise and test quantitative measures and indicators for the accuracy and coherence of the output. In the presentation we will discuss the identified basic data configurations, quality measures and indicators for these basic data configurations, and methods to compute those measures and indicators. We will also point out topics for future research.
|Title: <<< Quality Guidelines for Frames of Social Statistics >>>
Under the umbrella of the ESS.VIP Admin the ESSnet Kumuso dealing with quality aspects on multisource statistics was established. Within SGA 2 of this collaborative projecta work package was dedicated to draftingquality guidelines for frames in social statistics. The paper describes the motivation and the generation process of the guidelines and presents the document in its current state. After describing thestructure of the documentas an outcome of a stock taking actionbased onresults of a survey conducted among EU member states and investigations atNSIs outside the ESS regarding the construction, use and assessment of sampling frames within the NSIthe paper outlines the definition of frames in social statistics and the associated processes. In the second part the paper looks in more detail to the contents and substance of the guidelineextending the focus to other possible forms of using a frame then sampling such as input for processing and/or direct production of statistics. Another aspect which is considered in more detail is what kind of quality indicators and procedures are proposed to assess a frame in social statistics. Finally the possible present and future role and further development of the presented quality guidelines within the ESS and for NSIs is discussed.