|
The ethnic Survey
The children in the ethnic survey were born in Denmark between April 1 and December 31, 1995, by mothers with non-Danish citizenship. This time span, which was longer than the one used in the Danish survey, was necessary in order to reach a satisfying sample size. 611 children were sampled for the ethnic survey.
All mothers in the ethnic survey came from three countries from which Denmark had (in 1995) received many refugees: Somalia, Sri Lanka and Iraq, or from three countries from which Denmark received many immigrants: Turkey, Ex-Yugoslavia and Pakistan.
The mothers were required to have lived in Denmark for a period of at least three years prior to the survey. The reason for this was that, most often, refugees who had lived in Denmark for more than three years, tended to stay here. This was important, because DALSC was not intended to follow the children if they moved abroad. To ensure comparability between refugees and immigrants, this three year rule applied to immigrants as well. Because the number of Turkish immigrants in Denmark is much higher than the number of people from the other groups, only 25 percent of children with Turkish mothers, and all children of the remaining nationalities, were included in the sample.
Due to a low response rate (and consequently a small data set) already in 2003, it was decided to increase the size of the ethnic survey for the fourth wave of the survey in 2007. This was done in two ways: firstly by relaxing some of the sampling criterion concerning the children of the six nationalities in the original sample, and secondly by including children of Lebanese origin in the sample.
Detailed information about the response status (the maternal questionnaire) in each wave of the survey is available in tabular form here.
As in the Danish sample, the fathers’ citizenship or country of origin was not taken into account when the sample was drawn.
The table below gives an overview over the questionnaires used in the ethnic survey. The number of questionnaires used in the ethnic survey is smaller then the number used in the Danish surveye, mostly due to potential language difficulties.
|
|
Questionnaires in the ethnic survey
The 2007-questionnaire to the mothers (8910) and the children (8855) were also used in the Danish survey.
|
|
1996
The data was collected in the period March-May, 1996, by the interview staff of SFI Survey. As in the Danish survey, the mothers were interviewed.
Prior to the visit by the interviewers, the families received a letter of introduction explaining the survey. This letter of introduction was translated into 9 languages: Tamil (Sri Lanka), Somali (Somalia), Arabic (Iraq), Kurdish with Arab letters (Iraq), Kurdish with Latin letters (Turkey), Turkish, Urdu (Pakistan), Serbo-Croatian/Bosnian (Ex-Yugoslavia) and Albanian (Ex-Yugoslavia) by translators from the Danish Refugee Council. The interviewers were instructed to make sure that the families received the letter of introduction in Danish as well as in one or two translations, depending on their country of origin (e.g. if the mother was from Iraq, then the family received three versions of the letter of introduction: Danish, Arabic and Kurdish with Arab letters).
The interviewers visited the families a few days after they had received the letter of introduction. If possible, the interview was completed in Danish. If this was not possible, or if the family preferred so, the interviewer could request the assistance of a translator. This was done in 9 percent of the interviews. In contrast to the Danish survey, where it was preferred that the mother was alone with the interviewer during the interview, when the mothers were interviewed in the ethnic survey, it was preferred that the father or another person from the household was present or within reach during the interview, for instance in order to clarify questions that the mother did not understand or to act as a translator. Although the father was allowed to translate, the mother was the respondent, which meant that questions like ‘How do you…’ had to be reworded into ‘How does your wife…’ If other persons were present during the interview, the interviewer would note down who those persons were in the questionnaire. In 55 percent of the interviews, the mothers answered most or all of the questions themselves.
Response (mother questionnaire) in 1996
| |
Pct. |
Total |
| Response |
78.9 |
482 |
| Non-response |
4.9 |
30 |
| Refused/exited from sample |
16.0 |
98 |
| Total |
100.0 |
611 |
Interviews were obtained with (the parents of) 482 (79 percent) of the 611 children in the sample, which was deemed satisfactory. 99 parents (16 percent) either refused to participate or were not in the sample at the time of the interview.
Supplementary data collections
The ethnic survey was preceded by 15 quantitative interviews with key persons and families from the selected nationalities in order to gain a better understanding of their background and life circumstances. Most of the questions to the mothers in the ethnic survey were identical to the questions to Danish mothers but there were some differences as some questions were shortened in the ethnic survey, some were removed and others were added. The added questions were mostly about the fathers, e.g. about their education and occupation. These were necessary due to the lack of a questionnaire to the fathers in the ethnic survey.
|
|
1999
The data was collected in the period April-May, 1999, by interviewers from SFI Survey
Response (mother questionnaire) in 1999
| |
Pct. |
Total |
| Response |
75.8 |
463 |
| Non-response |
5.9 |
36 |
| Refused/exited from sample |
15.4 |
94 |
| Total |
100.0 |
611 |
Interviews were completed with the mothers of 463 children (76 percent).
Between the first and the second wave of the survey, 62 mothers had received Danish citizenship which did not mean that the children left the ethnic sample (and entered the Danish sample, for instance).
The mothers were interviewed according to one of two questionnaires depending on the children’s housing situation. The first questionnaire was used when the children lived with one or both of their parents, if they were adopted or if they had been in the care of grandparents or other relatives for more than one year. The second questionnaire was used if the children were placed in care or in an institution by the department of social services or if they were in a care arrangement with relatives which had lasted for less than one year – this questionnaire was only used in three cases and was not used in later waves. Because of the possibility that some of the parents who had not participated in 1996 would do so in 1999, the 1999 questionnaire repeated some of the background questions.
Most of the questions used in the ethnic and the Danish survey were the same although some of the questions in the ethnic survey had been rephrased in order to make them easier to understand. Some questions – especially concerning daycare arrangements – were unique to the ethnic survey.
The interviewers mailed a letter of introduction to the families. Again, this letter was in Danish as well as in the nine languages spoken by the original nationalities. The mothers who had obtained Danish citizenship since 1996 received this letter in Danish only. Preferably, the interview was carried out in Danish but if this was not possible, the interviewer could send for a translator (9 percent of the interviews).
The questions about the children’s medical history (vaccinations, health examinations, height and weight) were not in a separate questionnaire as in the Danish survey. Instead, they were filled in by the interviewer on the basis of ‘The Child's Book’ ("Barnets Bog") and vaccination certificate. ‘The Child's Book’ is a medical/health record used in the communication between the family, the general practitioner and the visiting nurse. In case the mothers were unable to find these documents, the interviewers sought permission to retrieve the information from the children’s general practitioner.
|
|
|
2003
The data was collected in the period April-June 2003, by interviewers from SFI, among families who had not refused to participate in the survey in 1999 or moved abroad since 1999. Without these families the sample contained, in 2003, 450 children and their families and 378 interviews (84 percent) were completed. However, when all the children who have been part of the sample are taken into account, in 2003, the response rate was somewhat lower.
Response (mother questionnaire) in 2003
| |
Pct. |
Total |
| Response |
61.9 |
378 |
| Non-response |
6.2 |
38 |
| Refused/exited from sample |
5.6 |
34 |
| Total |
100.1 |
611 |
The 378 obtained interviews among parents to children in the original sample corresponded to a response rate of 62 percent. Most of the loss of participants was due to refusals in 1999. According to Christensen (2004), when comparing the families still in the survey with the ones that had left (according to vocational education, connection to the labor market, economic conditions, housing arrangements and parents’ knowledge of Danish) there were no pronounced differences.
The mothers were interviewed in the home according to a standardized questionnaire. A translator was used in 2 percent of the interviews as they could not be carried out in Danish.
As in 1996 and 1999, preceding the interview the families received a letter of introduction but in 2003 it was not possible to translate the letter of introduction into Kurdish with Arab letters because a translator could not be found.
Again, there were questions concerning the children’s medical history (vaccinations, health examinations, height and weight) from ‘The Child's Book’and vaccination certificates. If the mothers were unable to find these documents describing the children when they were 4 or 5 years old, the interviewers asked permission to retrieve the information from the children’s general practitioner.
|
|
2007
The data was collected between January and December 2007 by interviewers from SFI Survey. Because of the relatively low response rate in 2003 (62 percent of the original sample), it was decided to boost the sample through a broader definition of the sampling criterion.
The table below shows the response rates in 2007 in the original sample drawn in 1995
Response rates in 2007 (original sample)
| |
Mother questionnaire
|
Child questionnaire
|
|
Percent
|
Total
|
Percent
|
Total
|
| Response |
50.1
|
306
|
47.5
|
290
|
| Non-response |
49.9
|
305
|
52.5
|
321
|
| Total |
100.0
|
611
|
100.0
|
611
|
The response rates were quite low in 2007. An interview was obtained with only half of the mothers from the original sample. Among the children, the response rate was even lower: 47.5 percent
Auxiliary sample
The auxiliary sample was established mainly in order to increase the number of observations available for analysis.
The larger ethnic sample was accomplished in several ways. For the six nationalities in the original sample, the earlier requirements - that the children had to be born between April 1 and December 31 and that the mothers had to have lived in Denmark for three years by 1995 - were relaxed. Now, it was sufficient for the children to have been born any time during 1995. The requirement that the mother had to have been living in Denmark for three years was abandoned. Further, a new sub-sample (children of Lebanese origin) was included in the sample. Finally, the condition that the children had to have been born in Denmark was replaced by a new one saying that it was sufficient if the children had immigrated to Denmark before 1998. Relaxing the earlier requirements to the sample allowed an auxiliary sample of 510 children thus bringing the total ethnic sample to 1,121 children.
The response rates in the auxiliary sample was a little higher than in the original sample: 53.1 percent among the mothers and 52.5 percent among the children.
Same questionnaires
The questionnaires used in the ethnic survey was exactly the same as the ones used in the Danish survey. The questionnaires contained sets of background questions designed specifically for the auxiliary sample. Filters in the questionnaire ensured that the original population did not have to answer the background questions as they had already done so in previous questionnaires.
|
 |
|
|
|
|
|
|
|
|
|
|
|
