Chair: Tasos Christofides
Room: S4B Lajkonik
Time: 17:00 - 18:30
Date: 27 June
Session 31 - papers & presentations
| Presenting Author | Abstract |
|---|---|
| Steven Thomas e-mail: steven.thomas@canada.ca | Title: <<< A new approach for Disclosure control -- Random Tabular Adjustment >>> The Government of Canada is investing in making more data available to Canadians. Statistics Canada is also investing in this initiative with the way that it assesses and treats disclosure risks. Data has historically been withheld from the Canadian public through a process of cell suppression whenever a disclosure risk has been identified. An alternative risk assessment is being proposed that will rely on ensuring that undesirable statistical inferences are prevented while useful statistical inferences can still be made through tabular perturbation. The Random Tabular Adjustment (RTA) process involves adding random noise to estimates where disclosure risks are apparent. At the same time, we are reassessing what we are comfortable with in terms of risk. The move to releasing more data comes at the expense of taking higher risks. In terms of quality, the new approach entails balancing various aspect of quality, such as pertinence, accessibility and accuracy. This presentation will highlight the challenges with disclosure control, the RTA method will be described and the appetite for risk will be discussed. |
| Arijana Amina Ramic e-mail: amina.arijana@gmail.com | Title: <<< Statistical Disclosure Control and Quality Reporting >>> The role of quality reporting is to demonstrate that high quality standards have been applied and achieved through in the statistical production processes, and to guarantee correct interpretation and use of the produced statistics. Statistical Disclosure Control (SDC) refers to the measures taken to protect data in accordance with confidentiality requirements, ensuring at same time that the usefulness of the data outputs is preserved to the greatest extent possible. Disclosure control measures reduce data quality (i.e. by suppressing data or changing detail levels), can affect the accuracy of information released (i.e.by data perturbation) or produce bias (i.e. using value rounding or noise addition), and limit access to certain groups (such as researchers). The degree and method of disclosure control may vary for different types of outputs as well as different statistics producers. The peer reviews in the European Statistical System (ESS) show that the highest standards for protection of the statistical confidentiality are applied across the ESS. Some quality indicators, as specific and measurable elements of statistical practice used to characterise the quality of statistics when possible, are also proposed for the information about the statistical confidentiality and security. Nevertheless, for the time being there is no information in the quality reports in the ESS about any disclosure measures applied or their influences at the quality of statistics produced and published. This paper focuses on showing the importance of the presentation of the SDC methods applied and their impacts at the statistics produced, in the quality reports. It provides a small analysis of the benefits for different kinds of the statistics users and of some possible and relevant quality indicators informing about SDC. Some information on the SDC measures applied and/or planned for the Population Censuses 2011 and 2021 will serve as example of good practices of reporting about SDC. |
| Fabian Bach e-mail: Fabian.BACH@ec.europa.eu | Title: <<< Statistical confidentiality: New initiatives in the European Statistical System >>> The protection of confidential information has a huge impact on how statistical data can be published and used for analysis, which makes it a key aspect of data quality. This paper presents new methods and tools currently being investigated in the ESS in order to publish more – and more useful – data without compromising statistical confidentiality. It covers new methodological and IT developments, where concrete use cases demonstrate their impact on data quality. For instance, a promising methodological direction is random noise: several ESS use cases at different maturity stages are presented, including recommendations for the harmonised protection of 2021 EU Census data. We also show how all these developments will enhance the user experience at various levels. |
| Kaija Ruotsalainen e-mail: kaija.ruotsalainen@stat.fi | Title: <<< Data exchange between the Nordic countries – supplementing education registers by qualifications completed in another Nordic country >>> High-quality and precise statistics production concerning the educational structure of the population usually requires a comprehensive register of qualifications that includes data on all degrees and qualifications completed by the population. An education register is the basis for producing education statistics and also important source data, for example, for population census and various survey-based statistics. Both Finland and other Nordic countries have a statistical register of education that contains data on degrees and qualifications completed by the population. The quality of the education registers is high as a rule because educational institutions are obliged to deliver data to the education registers annually. A big problem is caused, however, by the lack of data concerning qualifications attained abroad. No comprehensive register-based data source exists with data on the qualifications of immigrants attained in their home country nor on qualifications attained abroad by the original population. As part of the project ”Nordic Mobility” funded by the Nordic Council of Ministers, the coverage of education registers is improved by exchanging unit-level data between Nordic statistical institutes. This is enabled by the EU Statistics Act, based on which confidential data can be released from one ESS authority to another in order to develop and produce European statistics and improve their quality. This type of data exchange is likely to be the first of its kind at least in the history of Nordic statistical institutes. It is assumed that the data exchange will improve the coverage of the education registers of Nordic countries. At least in Finland this is likely to be visible especially in Åland where many of the population have completed their qualifications in Sweden, which means that data on completed qualifications degrees have not been included in Statistics Finland's education register. |
| Anne Berthomieu e-mail: anne.berthomieu-cristallo@ec.europa.eu | Title: <<< Quality assurance in micro-data exchange – The international trade in goods statistics as concrete example >>> The objective pursued when exchanging micro-data is to benefit from an additional data source with no extra burden for the providers of the statistical information (e.g. enterprises). The backbone of a micro-data exchange is a performant, robust and secure transmission system, embedding comprehensive guidelines and binding rules for both the sender and the receiver of the information. Rules binding the receiver are mainly targeting the respect of the data confidentiality while the sender is mainly bound to quality-related targets covering the data completeness, accuracy and punctuality. The quality monitoring is based either on information directly derived from the exchanged datasets or on related metadata.The quality assurance is built on five pillars: - a requirement system composed of binding rules on who should do what, how and when; - a guidance system aiming at promoting good practices and improving data harmonisation; - a reporting system composed of all types of metadata to be attached to the data flow or to be provided as an additional component; - a monitoring system gathering all information (e.g. error reports) on the data flows and the quality of the exchanged data; and - an assessment system producing compliance reports pointing out where actions are needed. Such quality framework is on the way to be implemented for the exchange of micro-data relating to intra-EU exports of goods. It takes its inspiration from the framework already in place to ensure the quality of the monthly trade in goods statistics issued from those micro-data and disseminated by Eurostat. The new aspects to be taken into account are inherent to the high sensitivity of micro versus macro data as well as to the challenge to make the exchanged data as useful as possible for the receiving partner countries. |