Dissemination is the phase of the data life cycle in which data collected and compiled are released to the public (UNSD, 2022). It is defined in UNECE’s 2018 Guidance on Modernising Statistical Legislation as: “the activity of making official statistics, statistical analyses, statistical services and metadata accessible to users”. Statistics should be sufficiently comprehensive and presented in such a way with appropriate metadata so that data can be understood without having specialised statistical knowledge.
Release schedules: NSOs are encouraged to make statistical data available to the public according to pre-defined release schedules. These schedules are important for preventing the manipulation of data for political purposes. For example, Ireland’s Central Statistics Office has a timeliness monitor to ensure that the publication of official statistics follows key release dates (CSO, 2024).
Metadata: these ‘data about the data’ provide information about the background, purpose, content, collecting, processing, quality and related information that a user needs to find, understand and manipulate statistical data.
Microdata privacy and protection: most NSOs release macro-data, which is aggregated microdata. In some cases, NSOs may release microdata containing information on individual persons, households or enterprises collected through survey interviews or from government administrative systems and registers. Dissemination of microdata, which may lead to direct or indirect identification of the data subjects should follow strict data privacy and protection regulations ensuring confidentiality and anonymity.
Data accessibility: data producers should strive to improve the accessibility of the data by designing dissemination platforms together with end-users, making the data available in formats that are widely used and do not require expensive software, providing translations in English and in all national languages, and facilitating use of the data by persons with disabilities or special needs.
Dissemination strategy: members of the NSS might draft a formal strategy for the dissemination of official statistics that outlines, among others, the use of different dissemination platforms and pricing policy, the procedures for dealing with errors and revisions, the contents of metadata, suggestions for improving users’ access and best practices for data tables and graphs.
Dissemination methods: data can be made available in PDF format, as downloadable datasets, or through interactive databases:
- Data tables in a PDF are useful for persons seeking a specific figure or a restricted set of data points, but not for persons seeking disaggregated data or large-N data for further statistical analyses.
Figure 1: Example of data table
(source: US Dept. for Homeland Security, 2023)
- Downloadable datasets are more useful for persons seeking disaggregated data as they allow for table customization. This can be done in several ways: different tables can be provided on separate excel sheets, users can download their preferred cross-tabulation from an online data browser, or differ cross-tabulations can be generated from .dta files.
Figure 2: Example of downloadable dataset
(source: Eurostat, 2024)
- Interactive datasets provide customizable visualizations which can improve knowledge and understanding of complex large-N datasets.
Figure 3: Example of interactive dataset
(source: IOM, 2024)
- Data portals: The most important tool used by an NSO to disseminate statistical data, whether in downloadable or interactive format, is the data portal, a web-based, interactive data and metadata platform, often composed of different databases relevant to the types of data collected, e.g., micro-, macro- and metadata (UNSD, 2022). Data portals can play a major role in reducing reporting burdens and improving the quality of reporting data and provide a single interface for accessing statistical data using standards of content and presentation. Today many developing countries lack fully functional platforms for data dissemination and reporting and rely on less sophisticated dissemination methods on their websites. A study by Paris21 and Open Data Watch (2021) demonstrated that a third of countries eligible for funding from the International Development Association did not have a data portal. The study recommends that data producers invest in improving metadata availability and multilingualism; integrating quantitative (e.g., Google Analytics) and qualitative (e.g., focus groups) feedback mechanisms to gain a better understanding of user needs; advancing national ownership as NSOs often rely on external partners for the design, implementation or management of data portals; improving data management practices by committing to a broader digitalisation process; and streamlining dissemination by combining multiple data outputs in an ‘all-in-one’ data portal.
- Social media: in order to stay relevant to the new ways in which people access data, NSOs have started reaching out to users directly through social media. Social media can help NSOs and data producers of the NSS promote the use of statistics, announce new releases, reach new audiences and receive user feedback.
- Paid versus open access: Official statistics are considered to be a public good, which is a reason for NSOs to provide free access to their data. However, charging for data might help to cover the costs of hiring experts and procuring resources for the aggregation, anonymization and linking of new statistics (or any costs related to additional processing requests). Data producers can choose between several paying models, such as only charging for more detailed or additional data, charging different rates for different uses, or only charging for more recent or relevant data.
Checklist for designing good data tables (UNECE, 2009):
- Tables should be able to stand alone, i.e., without the accompanying text
- Include the metadata explaining the variables.
- Include the title, column headers, row stubs and source
- Provide the data files in formats that are easily accessible, i.e., excel
- Shading encourages horizontal or vertical reading
- Use thousand separators and limit the number of decimal places
- Display the data in chronological order for time series
- Centre numbers around the decimal point
- Do not leave any data cell empty (missing cells should be identified)
Why should data producers disseminate microdata (IHSN, 2010)?
Data producers such as the NSO cannot be expected to identify and answer all of the potential research questions. Providing the microdata enables researchers to dig deeper, identifying any relationships between the variables recorded. By providing access, the producers allow for correction or replication which demonstrates confidence in the data, thereby increasing the actual and perceived reliability of the data. User feedback is key to improving the reliability of the data, as users ask questions or provide suggestions that the producers have overlooked. If researchers have access to reliable data, they will not need to collect their own data, which reduces the burden on respondents and the risk of inconsistent studies. The more the data are used, the greater the return on investment, thereby convincing funding agencies to spend more on supporting data production activities. In some countries, NSOs and partners may have a contractual obligation to provide microdata as data collection is funded by taxpayers and therefore considered a public good. In recent years, the ‘open-data’ movement has highlighted that all data collected by public funds are a public good. This has led to the development of new technologies that make use of open-data such as ‘mash-ups’ which bring together data from multiple sources in order to provide a new service.