5. Data description
5.2 Overview of the data
In general, the data can be divided into two main parts, namely a questionnaire part and a register part, where the information from both parts is combined. The following describe the questionnaire data and register data respectively.
The main source of new information is a questionnaire survey – in the following called the main survey – involving 3,615 completed interviews with non- Western immigrants and descendants from eight nations. These eight nations were selected among the biggest groups of immigrants and descendants in Den- mark, and include Ex-Yugoslavia, Iran, Lebanon, Pakistan, Poland, Somalia, Turkey and Vietnam, and account for 64% of all non-Western immigrants and descendants in Denmark. The total sample was 6,257 persons, giving a response rate of 57.8%, which is satisfactory compared with other Danish and foreign sur- veys (see below).
In order to compare analyses from the main survey with similar results for the Danish population as a whole, selected questions from the main survey were re- peated in an omnibus survey, which 961 persons answered. The original sample for the latter was 1,499 persons selected from the general population (i.e. in- cluding immigrants and descendants), giving a response rate of 64.1%.
The register data used here is based on administrative registers in Statistics Den- mark. The Danish population is divided into three groups: Non-Western immi-
grants and descendants, Western immigrants and descendants, and Danes21. The
primary register data part consists of the whole population of non-Western im- migrants and descendants who resided in Denmark on at least one of the dates 1/1-1984, 1/1-1985,...., 1/1-1998, 1/7-1998, and who on one of these dates were between 16 and 70 (both years included).
In addition, a sample was drawn of 25% of Western immigrants and descendants who resided in Denmark on at least one of the dates 1/1-1984, 1/1-1985,...., 1/1- 1998, 1/7-1998, and who also on one of these dates were between 16 and 70. As for the non-Western immigrants and descendants, children of the selected per- sons who were under 16 on 1/7-1998 were also included. This gives a total of 45,615 persons.
In order to compare the results for Western and non-Western immigrants and descendants with similar results for the Danish population as a whole, a sample of 2% of the whole population was selected (i.e. including immigrants and descendants). The sample was selected in exactly the same way as the two previous samples, i.e. 2% of all those who resided in Denmark on at least one of the dates 1/1-1984, 1/1-1985,...., 1/1-1998, 1/7-1998, together with the children of the selected persons who were under 16 on 1/7-1998 – 93,673 persons. It should be mentioned here that register data are included for all persons selected for the omnibus survey (original sample), i.e. 1,499 persons. For Olaf Ingers- lev’s analyses of health data, data on 1/7-1998 have been supplemented with children of the above persons, i.e. children who were under 16 on 1/7-1998 – in all 137,415 persons. In all, the primary data thus consists of 209,499 persons. Finally, the data are supplemented with information on so-called d-family mem- bers22 of the persons in the above-mentioned samples who actually resided in
Denmark on 1/7-1998 – 112,329 persons in all.
There is supplementary information from a wide variety of registers on all selected persons – if, that is, the person exists in the register concerned. The most important registers used and the type of variable they contain are shown below in figure 5.1.
Like the questionnaire surveys, the registers have their uncertainties and other problems, though these are not more serious than that they can be dealt with in the relevant places in the analyses.
21 The more idiomatic term "Danes" is used – in line with our neighbouring countries –
to describe the 90-95% of the population which Statistics Denmark designates "Other".
22 A d-family is defined as one or more persons who live at the same address and who
are either: Single, a married couple, registered partners (two persons of the same sex who are civilly married), cohabitants (with joint children) or living together (without joint children).
Figure 5.1. Overview over the registers and variables used in the analysis
Register Variable Population statistics
Integrated database for labour market research (IDA)
Overall social statistics
Crime statistics Labour market schemes (AMFORA)
Central register for Labour Market Statistics (CRAM)
Income statistics Rent subsidy register Housing and Building Register (BBR) National Hospital Discharge Register National Health Service Register
Education Classification Module (UKM)
Contains a large number of demographic variables, e.g. sex, age, marital status, date of marital status, etc. Also contains the variables country of origin and date of entry.
Consists of variables from different registers. Contains informa- tion about education, wages, occupation and other labour market variables.
Contains information abort various social benefits received, plus duration, e.g. old age pension, early retirement benefit, unemploy- ment benefit, cash benefit, etc.
Contains information about criminal records, case decisions, etc. Information on whether the person has been in a labour market scheme (e.g. job training, leave of absence, early retirement). In- cludes starting and finishing dates, and number of hours in the scheme.
Contains information on annual degree of unemployment and membership of unemployment insurance funds.
Contains information about practically all kinds of personal in- come and paid taxes.
Information on rent subsidy/rent subsidy to non-pensioners, rent, etc.
Information on the size and use of housing, whether it is owner- occupied or rented accommodation, etc.
Information about use of hospitals, including admissions and visits to the casualty ward. Also contains both administrative and medical information, e.g. diagnoses.
Contains information about services carried out in the primary health sector by doctors, specialists, dentists, etc.
Includes variables on examination averages at upper secondary schools/higher preparatory courses/adult education centres (VUC), year of completion, etc.
5.3 Design, response, representativeness and non-response in the main