The data is the result of a linkage between different data sources. The most important data sources are the Central Population
Register (CPR) in the Directorate of Taxes, and the Aliens Register (Utlendingsdatabasen –UDB) in the Norwegian Directorate
of Immigration.
The CPR is the most important data source in Statistics Norway's population statistics system. In addition, data from the
1960 and 1970 population censuses are included as a supplement. The same applies to older statistical files which originate
from population register data. The definition of first-time immigrations is based on CPR data alone.
In 2004, UDB replaced the Fremkon register (the old Aliens register) and the Refugee Register. Central information from these
two registers were adapted to new standards and included in UDB. As a result, UDB now contains data going back to 1991, several
years before UDB was introduced in the handling of cases.
The single most important variable in UDB for the statistics is reason for decision. If the information on that variable is
missing, other similar variables are used.
An important supplementary source is the 2004 edition of the file for reason for immigration produced in 1995 and updated
by use of data from the Refugee Register. It covers the persons that do not have satisfying information in UDB.
The population register data have been processed. The new focus in the processing leading to reason for immigration is the
task to achieve the best linkage possible between the ID number in UDB and Personal Identification Number, and subsequently
to find the first reason for immigration for as many foreign-born as possible.
The process of obtaining a linkage between the two ID number series is based on links that are found both in UDB and CPR.
These links are controlled, linked and compared, and improved considerably. The result is a common definition of the persons
in the two input data sets.
When this task is accomplished, the next step is to find the information in UDB that, as best as possible, states the reason
for the first immigration event registered in the CPR.
One challenge in this connection is that most persons are registered with several cases (even if most of them are just renewals),
and that the information may differ from case to case. In many cases the most specific information is found on some of the
later cases registered on a person. For some persons this comprehensive information is copied to the former cases.
Another challenge is to identify the case which most probably is related to the first immigration registered in the CPR. There
are no variables in UDB that make it possible to pick out the most relevant case in all cases.
The main principle is that the reason for immigration is taken from the last registration before the immigration was registered
in the CPR, but there are exceptions if the actual case does not provide sufficient and reliable information. In these cases
other registrations or sources are used as indication of the reason for immigration.
One of these supplementary sources is the file of reason for immigration produced in 2006. That year the task was resolved
by finding the most relevant case for each of the main reasons, and then make a choice between these reasons (for persons
with more than one main reason). The choice was based on the main reason which is seen as the most reliable in each case.
In reality, refugee as reason for immigration was chosen before the others.
In case of missing source of reason for immigration imputations are made, based on variables like citizenship and age at the
immigration. In general, there is a greater need for imputations in the older data (from the beginning of the 1990s and earlier)
than in data registered in the new UDB system (started in 2004).
In this process, some children without their own reason for immigration get a value from their parents.
The problems with varying reliability may have the effect that some persons are allocated e.g. Refugee as the reason for immigration
in stead of Family, as would have been the result if the registrations had been more complete.
The next challenge is to identify the most probable reference in Norway (the "resident person") in family immigration cases.
If this reference is not stated in the UDB data, CPR information on relationships is used to find the spouse, children, parents
or siblings that possibly may have been the reference at the first immigration. In this process spouse takes precedence over
the other categories. The method identifies persons that may have been the reference for foreign-born. However, the method
does not guarantee that the identified person actually and legally was the reference.
The construction of a variable called "type of family unification" (distinguishing between reunification, accompanying person
and formation/extension) also causes several challenges, both concerning missing data and conceptual difficulties. The classification
is mainly based on assessments of dates of immigration and marriage (when relevant) of both the immigrant and the reference
person, and on registrations of that variable in the UDB data. The determination of limit values used in the classification
is based on assessments.
Not relevant.
