Mapping of a new data source and initial quality assessment of the source

There can be two ways to go about the identification of new administrative data sources.

First, the statistical producer can have a clear list of needed sources for a specific statistical product and
then engage with the data holders in trying to identify the needed sources.

Another approach is to engage in dialogue with a data holder to learn more about all their
administrative data and once there is a complete list find out internally, what could be used for which
statistical products.

In any case, dialogue between the data holder and the statistics producer is key to build a mutual
understanding and good relations. Also, generating an understanding of the benefits of sharing data
can be very helpful and necessary in that process to increase everyone’s willingness to cooperate and
share.

1. Knowing the user needs

Meetings with users are important to know which statistical products are needed and if what is being
produced at the moment meets the needs of those using the statistics (e.g., is the format, level of
complexity, access according to user needs).

Needs can change. Therefore, checking in with the users regularly ensures that the statistical outputs
remain relevant for the users and therewith “fit for purpose”. At the same time, regular engagement
with users ideally increases mutual trust and understanding. Also, often data holders are also users of
the statistical outputs, thus making the good relationship and satisfaction even more important.

In the end, good quality is to match the user needs as closely as possible with the resources that are
available (striking a balance between resources and needs).

Therefore, the very first step in the identification of new data sources starts with gathering information
on the needs of the users to find out what the final statistical product is supposed to look like. The users
provide input to which statistical products are needed, which in turn provides the specifications on the
administrative data sources that could be used to produce the statistics.

2. Engaging in dialogue between statistics producer and data holder

In order to make use of a new administrative data source it is very important to engage and build
relations with the institution that owns the data.

Meetings between the data holder and NSO ensure that everyone is informed about each other’s work
and changes in processes that could affect the statistical product are communicated to all in a timely
way (e.g., a change in pension age, a change in the age limits of who is classified as a child and an adult,
a change in what is reported as a violent crime and more).

Investing time and resources in building this relationship from the beginning will also positively impact
cooperation and data sharing later on, as all stakeholder will have a better understanding of each
other’s needs and practices.

As part of the first meetings, presentations on the potential benefits of using administrative data,
increasing the quality and similar may be a good idea to increase willingness and by in from the data
owner/holder.
 3. Mapping
After identifying what the users’ needs are, and having established a relationship with the data holder,
the first step in assessing the quality of administrative data is to identify which data sources exist.

Some administrative data sources are already being used for producing official statistics, while other
administrative data sources could be a potential source for producing statistics – either replacing a part
of, or an entire survey, or producing new information on society which is not yet published.

To identify available sources a mapping tool is recommended. Such a tool provides a template for
detailing information about the source such as information about the data holder, which variables the
data source contains, the population demarcation, process of data collection, data validation
procedures, which statistics it can produce etc.

The mapping can be done sector by sector and will thus give an overview of how different data sources
within the same sector can possibly be linked. This overview is useful as the different public institutions
within the same sector might not themselves have this overview.

Once the mapping is done within a sector, it is possible to identify those administrative data already
being used for statistics, and which administrative data sources could be further investigated for their
potential.

A template for the mapping of data sources can be found as an attachment to this document.

4. Initial quality assessment of a new source

The first time an administrative data source is assessed, it is recommended that the assessment is done
in collaboration between the NSO and the administrative data holder to establish a common
understanding of why the data has been collected, what it is normally being used for and also learn
about each other’s concepts and definitions. Afterwards, the assessment is recommended once a year
on those statistics of highest priority and importance, and less often for other statistics (agreement on
how often between NSO, data holder and users).

The self-assessment tool contains a number of questions that help gain a better understanding of the
potential new data source. However, the tool does not provide a score but can serve as a basis for
evaluation by the tool user. It is the tool user’s own interpretation of the answers—and the tool user’s
prioritization of which ones are especially important for the data at hand—that constitutes the tool
user’s own assessment of data quality.

Sometimes, a data source can maybe not immediately be used as a source for official statistics but can
be used to more efficiently design a survey frame or other issues related to the production of statistics.
When engaging with new administrative data sources it is recommended to think openly and creatively
and not discard a source too quickly.

However, in a situation of scarce resources, both in staff and budget, that many NSOs and also
Ministries, Departments and Agencies (MDAs) are faced with, prioritising the most needed actions
might be a necessity.