Frequently asked questions

What is microdata.no?

The microdata.no service was designed to simplify and expedite the use of register data in research.

Norway has many good sources of register data, and Norwegian law allows them to be used for research purposes. However, the application process for obtaining de-identified microdata from registers is long-winded, and adapting datasets for researchers is time-consuming and costly.

microdata.no gives researchers instant access to register data and enables them to adapt the data without going through the application process.

What is the relationship between microdata.no and RAIRD?

The microdata.no service is a result of the RAIRD project, and is based on technology, data and metadata developed in the project.

RAIRD stands for ‘Remote Access Infrastructure to Register Data’ and is a joint project between the Norwegian Centre for Research Data (NSD) and Statistics Norway. The project is funded by the Research Council of Norway under the National Financing Initiative for Research Infrastructure (INFRASTRUKTUR).

Which registries will be included in microdata.no?

Registries for which Statistics Norway has processing authority (under the provisions of the Statistics Act). This applies to the following:

  • Norwegian National Registry
  • National Education Database (NUDB)
  • Register for Personal Tax Payers
  • Labour market data
  • FD-Trygd (event history database)

When was microdata.no launched?

The first publicly available version of microdata.no was launched in March 2018.

Is it solely descriptive analyses that can be extracted from microdata.no?

No. The analytical platform also supports linear and logistic regression (logit, probit, mlogit), analysis of variance, panel regression (available soon), etc.

There are no limitations in the technological solution as to which analyses can be supported (see separate section below on which statistical libraries are used), but all functionality that is introduced is first subjected to testing, evaluation and, where relevant, measures to protect confidentiality.

More options for analysis are to be added.

The analytical platform also includes a data processing function, with support for population delimitations, linking, aggregation/disaggregation and the development of separate variables.

Does microdata.no contain anonymous data?

No. microdata.no contains complete, unmodified data that prevents the direct identification of an individual. Privacy is protected through an ‘anonymisation interface’ that enables researchers to process and analyse the data without being able to view or gain knowledge of the personal data.

Is microdata.no mainly aimed at master’s degree students?

No. The service is part of the national infrastructure for research using register data, and the aim is to simplify and improve access to such data for all groups of researchers who qualify for access.

Master’s degree students were often cited as an example in discussions about the RAIRD project because they are precluded from using register data in research (beyond a 20% sample from FD-Trygd) due to the costs and time involved.

microdata.no endeavours to be part of an infrastructure that improves access for all groups of researchers.

Will microdata.no lead to changes in the current practice for obtaining register data for research purposes?

No. The purpose of the project is to stimulate more research and the establishment of more research communities that use register data in their research.

In order for microdata.no to be a success, the system must be capable of handling or improving the bulk of Statistics Norway’s data dissemination in the long term. To this end, the aim is for as many researchers as possible to carry out more of their register data work via microdata.no. The existing arrangement will continue, but microdata.no will supplement this with extra capabilities and volume.

microdata.no can be used for examining/analysing the scope of data prior to or as part of the application process for obtaining data, thereby playing a useful role for research communities that still want to obtain microdata.

How can I apply for access to microdata.no?

Information about this is available here: How to get access.

See also the questions below.

Who can enter into an agreement on behalf of an institution?

Where no institutional agreement currently exists, anyone affiliated with an institution can initiate the agreement. An authorised signatory at the institution must sign a digital agreement before it can be approved.

Can anyone sign on behalf of the institution?

No. The person signing the agreement must be an authorised signatory at the institution.

How do we add more administrators?

The person who initiated the agreement will be assigned a role in the system, which we have called ‘procurator’. The procurator is responsible for adding and removing administrators by initiating an amendment to the agreement. If the institution needs to replace the procurator, it must contact us at microdata@ssb.no.

How does the signing process work?

After the agreement has been forwarded for signing, the person signing the agreement will receive an email from the Norwegian postal service (Posten) informing them that a document is ready to be signed. Posten uses the Agency for Public Management and eGovernment’s (DIFI) Contact and Reservation Register, which contains the digital contact details of the population. More information about this (in Norwegian) can be found here. Notification to sign the document may be sent to a private email address as opposed to a work-related email address.

Why can users not view the data directly in microdata.no?

The simplified access in microdata.no (which makes it less time-consuming and reduces costs) is based on a model in which microdata is neither downloaded nor available for direct visual inspection. Instead, the researcher must use documentation (metadata), as well as the processing, analysis and visualisation techniques that are available in order to understand the data.

If it is nevertheless necessary to view data, or to link the researcher’s own data to register data, the researcher can apply to obtain data in the normal way in accordance with applicable regulations. In such situations, researchers can use microdata.no to simplify the application process. They can do this by examining data and adapting their own datasets, then applying to obtain specific microdata in line with current guidelines. This will simplify Statistics Norway’s work on adapting datasets considerably, whilst also reducing the cost to the researcher.

This is not currently possible. Linking the researcher’s own data to register data leads to a higher risk of identifying individuals, and is not compatible with the service’s simplified access model. Additionally, as the body responsible for the alignment of the data, Statistics Norway does not have processing authority for users’ own data. The necessary permissions must therefore be sought for such matters.

Future development projects will explore and develop models and tools for researchers to link to their own data in microdata.no.

Can users use typical statistical packages in microdata.no?

No. Traditional statistical packages and tools such as Stata, SAS, SPSS, R, Python, etc., are not compatible with the anonymisation model used in the service, where processing and analysis are performed in a controlled environment that enables the confidentiality of data to be protected throughout the analysis process.

microdata.no offers an alternative interface for the importing, processing, analysing and visualisation of register data (both event history data and status data). This interface protects privacy whilst also supporting well-known analysis techniques and working methods. microdata.no is distinct from other statistical packages in that substantially more metadata (information, data documentation) is fully integrated into the tool.

Which statistical packages/libraries are used in the microdata.no platform?

The algorithms used are retrieved from open and tested software libraries and modules.

The following modules from the Python programming system are currently used:

Will microdata.no eventually support new statistical methods?

Yes. The service will be developed on an ongoing basis and more analysis methods will be introduced over time. Before an analysis technique can be included, its ability to protect confidentiality will we assessed, and analysis output checks may be carried out.

Users’ needs and wishes with regard to functionality can be forwarded to support@microdata.no

Can researchers use/install plug-ins in microdata.no?

In order to protect confidentiality, we need to review all types of analyses and functions before they are incorporated into the system. However, researchers may ask for new algorithms/methods to be added, and a facility will be established to deal with such requests.

Users’ needs and wishes with regard to functionality can be forwarded to support@microdata.no

Will sensitive data be disclosed when using such a system? What about privacy protection?

Through the RAIRD project, microdata.no has been developed with a holistic approach to privacy and confidentiality, and privacy protection is embedded in the technology along several axes:

  • The data cannot be inspected or downloaded
  • Analyses, graphics and other output from the system are anonymous
  • The system can only be used by researchers and students from approved research institutions
  • Use is regulated through agreements between Statistics Norway and the institutions
  • Strong authentication is used to log in (the ID port)
  • All usage in the system is logged for a period of five years

microdata.no facilitates the use of register data for research purposes, but the data itself remains protected from direct access since the user is not given identifiable or sensitive information.

How can I analyse data without seeing it?

Visual inspection of data is traditionally a commonly used technique for understanding data, and for verifying and controlling the stages in a data processing process. In microdata.no, it is not possible for users to see the data matrixes. Metadata solutions, descriptive techniques and other tools have therefore been developed which will help users to understand the data and control the data processing processes without direct visual access to data values.

Can anyone register as a user?

No. Only researchers and students at approved research institutions are given access.

Information about this is available here: How to get access.

Why does microdata.no differ from the register data solutions in, for example, Sweden and Denmark?

The solutions in Sweden and Denmark are based on a so-called ‘Remote Access’ approach (RA), where researchers are given remote access to microdata (which is adapted according to an order) via Citrix or Remote Desktop-type solutions, and where analysis results are subject to manual analysis output checks aimed at protecting privacy. Under Norwegian law, this type of solution will necessitate the same requirements for dispensations and licences that are currently applicable. The RA model was therefore found to be unsuitable for RAIRD/microdata.no.

The RAIRD model is distinct from the RA models in that an anonymisation interface prevents direct visual access to microdata. The model simplifies the access process and largely eliminates the need for datasets to be adapted to detailed orders in advance.

Instead, researchers themselves have the freedom to assemble and arrange datasets based on the register data sources. The long-term ambition of microdata.no is to cover more and more of researchers’ data processing and analysis needs.

The access model was chosen because it provides the following benefits compared to traditional models:

  • Using microdata.no, the researcher can explore data freely and arrange datasets based on the register sources
  • The access application process is simplified considerably by the very fact that the researcher can use microdata via an anonymisation interface
  • Confidentiality measures can be automated

The RAIRD model also has some disadvantages compared to the Swedish and Danish systems:

  • The interface has only recently been developed, and is inspired by, but not identical to existing statistical packages
  • The researcher cannot view the microdata directly
  • Permission must be sought before researchers can link their own data to microdata.no
  • It is more complicated for users to develop their own statistical methods and use plug-ins because all new methods must be tested and approved in accordance with the anonymisation criteria in the service

These capabilities were selected with a view to achieving the goal of an anonymisation platform for register-based research, and thereby speed up and simplify access.

In situations where it is necessary to view data, or to link the researcher’s own data to register data, the researcher can apply to obtain data in the normal way in accordance with applicable regulations. In such situations, researchers can use microdata.no to simplify the application process. They can do this by examining data and adapting their own datasets, then applying to obtain specific microdata in line with current guidelines. This will simplify Statistics Norway’s work on adapting datasets considerably, whilst also reducing the cost to the researcher.