Content
About the statistics
Definitions
-
Name and topic
-
Name: Child-care (discontinued)
Topic: Education
-
Responsible division
-
Division for Education and Culture Statistics
-
Definitions of the main concepts and variables
-
Place of residence Individuals are grouped according to whether they live in: 1) Oslo 2) Trondheim, Bergen, Stavanger 3) Municipalities with more than 10 000 inhabitants, and 4) Municipalities with less than 10 000 inhabitants.
Family type classification is made according to whether respondents are: 1) Single parent without income 2) Single parent with income 3) Married / cohabitant without income 4) Married / cohabitant with one income 5) Married / cohabitant with two income.
Mother's country of birth is defined by the following classification: 1) Norway 2) EU, USA, Canada, Australia and New Zealand 3) Asia, Africa, Latin America, Oceania excluding Australia and New Zealand, and Europe outside EU / EEA.
Region includes the following counties: 1) Akershus and Oslo 2) Hedmark og Oppland 3) Rest of Østlandet 4) Agder and Rogaland 4) Western Norway 5) Trøndelag 6) Northern Norway.
Age is measured in completed years at the end of the year.
-
Standard classifications
-
Low-income families are defined as both families with lower income than 50 (OECD-def) and 60 per cent (EU-def) of median income after taxes for all households.
Educatinon The variable parents’ level of education, mother's level of education and father's level of education on the basis file is coded according to Norwegian Standard of Education (NUS 2000).
Country of birth is coded according to Standard for land and citizenship in personal statistics.
Administrative information
-
Regional level
-
National, regional and residential area.
-
Frequency and timeliness
-
Not relevant
-
International reporting
-
No international reporting.
-
Microdata
-
Not relevant
Background
-
Background and purpose
-
Similar surveys about families' child-care preferences, labour force participation, and use of cash-for-care are conducted by Statistics Norway earlier in 1998, 1999 and 2002. In 2002 there were fewer questions about the cash-for-care, and several questions about kindergarten than in 1998 and 1999. The survey in 2010, builds on the survey from 2002 and there are even fewer issues related to the cash-for-care. Section of kindergarten is further extended. It asked about how satisfied parents are with day care services and what they believe is quality in kindergarten. Moreover, two new sections included in the study: one on day care facilities for schoolchildren (SFO), and the other for homework help at school .
-
Users and applications
-
Users of data will primarily be the Ministry of Education and Research, and Ministry of Children, Equality, and Social Inclusion, media and others interested in information about families' child-care preferences, labour force participation, and use of cash-for-care.
-
Coherence with other statistics
-
Parts of the survey can be viewed in the context of official statistics in kindergartens and primary schools.
-
Legal authority
-
The Statistics Act § 2 -1.
-
EEA reference
-
No regulation.
Production
-
Population
-
The population in this survey are households that had children aged 0-9 years at time of the interview (see section 3.3.).
-
Data sources and sampling
-
Not relevant.
The sample was drawn from households that had children aged 0 - 9 years, a sample of 4530 children aged 0 to 5 years and 1 500 children aged 6 to 9 years. The reason that the sample is divided into two age groups is that the first sample are children of kindergarten age, while the other group is in the SFO-age (see section 2.1.). To ensure representation from all groups in the population, it was decided that there should be more people with low education, low income and with backgrounds from Africa, Asia, Latin America and Europe outside EU / EEA.
-
Collection of data, editing and estimations
-
The data were collected by telephone interview. Interviews began in mid-September and ended in mid-December 2010. Sampling units were celected from children aged 0 - 9 years (see section 3.3.), but the interview was conducted with the main carer, that person who stated as main care-giver in the National Register. Only one child was drawn from each family within the relevant age. Meanwhile, the main care-giver also interviewed about all children living in the household aged 0 to 9 years.
Interviews are accomplished by a computer-based questionnaire. Control subroutines are employed in order to avoid erroneous responses and incorrect data registration. In some cases the interviewer is notified when doubtful data entries. In other cases variables have a limited range in order not to enter values surpassing reasonable values. There are built-in control subroutines for valid outcomes for the response options.
To calculate the results of this survey, the data should in most cases be weighted.
-
Confidentiality
-
The analysis unit is person.
-
Comparability over time and space
-
Parts of the survey can be compared with previous "families' child-care preferences" studies.
Accuracy and reliability
-
Sources of error and uncertainty
-
Both in total counts and sample surveys erratic responses may occur. Errors may arise both in the collection as well as in the data revision process. PC´s are used in the collection of data in this survey. The interviewer reads the questions from the screen, and registers the answers directly into the data programme. An important advantage by using PC-based registering is that pre-programmed skipping of questions is employed in order to avoid placing questions to respondents for whom certain questions are inappropriate.
PC assisted interviewing gives the opportunity to monitor response consistency between the different questions directly. For every question a range of proper values are defined. In addition, error messages are programmed in order to alert the interviewer when typing values that not are consistent with previous responses.
We avoid entering invalid input and we achieve reduced non-response on certain questions by reduced risk for skipping questions that should have been raised.
Errors may occur when respondents give wrong answers. One reason is difficulties for the respondent to remember circumstances far behind in time. Additionally, questions may be misunderstood. When questions relate to issues people find complicated, we must expect that erratic responses may be found. Data collection errors may also come from questions respondents find sensitive. In such cases, respondents may intentionally reply incorrectly. Responses may also be influenced by what the respondent consider socially desirable.
Processing errors take place when there are discrepancies between the values registered and the values reported out of the process. Such errors may occur for instance by recoding. By means of various examinations, such errors are attempted to be identified and corrected.
When all errors as far as possible are corrected, experience indicate that statistical outcomes in most cases to a relatively little extent are affected by collection and processing errors. However, the effect of such errors may have importance in some cases, and every error will not necessarily be detected.
Of the total sample of 6,015 was reached interviews with 3262 parents. This corresponds to an answer input on 54.2 per cent. In most cases, mothers were primary caretakers. Of the original sample, there were 209 cases where the father had primary custody and was interviewed. Total was conducted interviews with 99 fathers.
The uncertainty of the findings based only a part of the population is often called sampling variance. Standard deviation is a measure of this uncertainty. The size of standard deviation depends, among other factors, on the number of observation in the sample, and on the distribution of the current variable in the whole population.
Statistic Norway has not made exact calculations to compute standard deviation for the findings. However, in table 1, the approximate size of standard deviation is given for observed percentages.
To illustrate the uncertainty associated with a percentage, we can use an interval to give the level of the true value of an estimated quantity (the value obtained if making observation on the whole population instead of observation based on a part of the population). Such intervals are called confidence intervals if constructed in a special way. In this connection one can use the following method: let M be the estimated quantity, and S the estimate of standard deviation of M. The confidence interval will be an interval with limits (M - 2*S) and (M + 2*S).
This method will give, with approximately 95 per cent probability, an interval containing the true value.
The following example illustrates the use of table 1 for finding confidence intervals: The estimate of standard deviation of 70 percent is 3.2 when the estimate is based on 300 observations. The confidence interval for the true value has limits 70 ± 2*3.2, which means the interval, is from 63.6 to 76.4 per cent.
Table 1. Standard deviation in per cent
Antall observasjoner
5/95
10/90
20/80
30/70
40/60
50/50
25
4,4
6,1
8,2
9,4
1,0
10,2
50
3,1
4,3
5,7
6,5
7,0
7,1
100
2,2
3,0
4,0
4,6
4,9
5,0
200
1,5
2,1
2,8
3,2
3,5
3,5
400
1,1
1,5
2,0
2,3
2,5
2,5
800
0,8
1,1
1,4
1,6
1,7
1,8
1000
0,7
0,9
1,3
1,4
1,5
1,6
1200
0,6
0,9
1,2
1,3
1,4
1,4
1500
0,6
0,8
1,0
1,2
1,3
1,3
2000
0,5
0,7
0,9
1,0
1,1
1,1
2500
0,4
0,6
0,8
0,9
1,0
1,0
3000
0,4
0,5
0,7
0,8
0,9
0,9
3262
0,4
0,5
0,6
0,7
0,8
0,8
Not relevant.