The Danish Longitudinal Survey of Children (DALSC)
- Børneforløbsundersøgelsen (BFU)
 
Introduction
The Danish Longitudinal Survey of Children (DALSC) was initiated in 1995 in order to provide basic representative information about the family conditions and development of Danish children. It was designed by researchers from SFI, the Danish National Centre for Social Research, in cooperation with other research institutions. It is the first longitudinal study in Denmark which aims to monitor children from birth until adulthood, and which allows research into the relationship between living conditions in childhood and subsequent life as an adult.
 
All children in DALSC were born in 1995 and must be living in Denmark at the time of the survey.

Originally, DALSC consisted of a sample of children by mothers with Danish citizenship and a sample of children by mothers without Danish citizenship. In 2003, SFI began a survey of children placed in care and to ensure comparability it was decided that all children in this sample also had to be born in 1995. Thus, today DALSC is a collective name for three surveys:

1.      A survey of 6,011 children randomly sampled among all children living in Denmark who were born between September 15 and October 31, 1995, by mothers with Danish citizenship (regardless of their country of origin). This survey is also referred to as 'the Danish Survey'. See a detailed description of the
Danish survey.

2.      A survey of 611 children living in Denmark, who were born in Denmark between April 1 and December 31, 1995, by mothers with non-Danish citizenship who had, them selves, been living in Denmark for at least three years. Specifically, this sample contains all children of mothers with citizenship in Ex-Yugoslavia, Pakistan, Iraq, Sri Lanka and Somalia and a randomly drawn sample of children of mothers with Turkish citizenship. This survey is also referred to as ‘the ethnic survey’. See a detailed description of the ethnic survey.

3.      A survey of all children born in 1995 who are, or at one time have been, placed in care environments, e.g. foster homes or residential institutions. The survey of children in care environments, which is called Children in Care – a Danish longitudinal Study (CiC), began in 2003, with approximately 600 children. CiC grows over time as children are placed in care. The children in the CiC have special needs which is why not described here – for more information, contact Anne Dorthe Hestbæk (adh@sfi.dk). 

Longitudinal survey 
DALSC is a longitudinal survey, i.e. a survey with repeated data collections about the same group of persons. At this point in time, the data collections from 1996, 1999, 2003 and 2007 are complete.
 
The longitudinal design implies that the survey repeats questions from one wave of the survey to the next which allows analysis of development over time and of causal relationships. But because DALSC (in 2007) has followed the children from infancy (around 6 months) until the age of 11, some questions are no longer relevant (e.g. questions about breastfeeding) while others have become relevant (e.g. questions about behavior in school). 
 
Register data (Statistics Denmark)
Much of the information about the parents is drawn from administrative registers at Statistics Denmark. Consequently, this part of the data is a genuine longitudinal data set (or panel data), in which all information is repeated at least once a year. The registers are comprehensive and cover topics such as socioeconomic status, education history and education level, employment and experience in the labor market, income and transfer income, country of birth and persons living in the household etc.

As a rule of thumb, variables in the registers are available once a year in the period 1981 until present day minus two years. 
 
Selection
Selection into the Danish or the Ethnic survey is done without consideration for the parents’ country of origin or their ethnicity. The selection variable is the mother’s citizenship only.
 
Until 2007, when a questionnaire to the children was introduced, the questions in the questionnaires have been directed to the parents. That is, the children have been the selection unit; the parents have been the respondents. For twins, for example, this design allows for the possibility of having one or both twins in the sample. In case both twins are in the sample, the same parent acts as the respondent twice.  

Family types
The family types in DALSC include the nuclear family, i.e. two adults (married or cohabiting) with at least one child as well as the single parent family. In the first waves of the survey, a third family type – children placed in care environments – was allowed but they were so few (less than 10) that analysis was unfeasible.

 

In order to answer these questions, and more, the longitudinal design is necessary.

Transitions
Frequent waves of the survey would allow a more precise picture of the children’s development but would also demand many resources. Therefore, a design was chosen in which new waves of the survey take place at several years’ interval but around important periods or transitions in the children’s lives.

The first wave (1996) took place towards the end of the normal leave-of-absence (i.e. within 24 weeks of birth) in order to describe the children’s situation before the parents (primarily the mother) eventually returns to work. At the time of the second wave (1999) the children were around three years old and well into their ‘day-care career’. At this time, the day-care alternatives (e.g. full time day-care centre or part-time day-care centre and part time care in the home), combined with other life circumstances, are important in determining the children’s well-being. The third data collection (2003) followed the important transition from day-care to school when the children were around seven years old. In the fourth wave (2007) the children were around 11 years old and had attended school for a few years (most were in the third grade). The 2007 wave of the survey was the first wave in which the children were old enough to supply information themselves.
 

Sample
DALSC uses three samples (Danish, Ethnic, Children in Care - CiC). The categorization of the children into the three samples was done in 1995. The categorization into the Danish and ethnic sample was final, i.e. children do not move between samples (e.g. children whose mothers received Danish citizenship between waves did not move from the ethnic survey to the Danish survey) or enter the samples (e.g. immigrants who came to Denmark later than 1995 do not enter DALSC).

But as always, there are no rules without exceptions. The CiC data increases in size over time: when a child is placed in care, this child enters the CiC data; if the care arrangement is terminated, this child remains in the CiC data. If the child was already in the Danish or Ethnic survey, it exits these surveys.  
 
Exit from the sample
Exit from the sample can be both temporary and permanent.

Temporary exits take place when children move abroad. If, in any subsequent wave of the survey they return to live in Denmark, they re-enter the sample.

In general, permanent exits take place if the parents state in any wave of the survey that they do not wish to participate (in current and future waves). In this case the child is permanently removed from the sample. Because a high share of parents in the ethnic survey refused to participate already in 1996, it was decided not to respect this refusal and to try, extraordinarily, to obtain a response again in 1999. This is the only time when refusal to participate has not been respected in DALSC.