Notes on data infrastructure developments

As we go about selecting the indicators that are to be used in Social OMC, there is a wealth of survey-based and administrative datasets on which regular monitoring of child poverty and well-being across the Member States can be built. The major datasets considered as potential sources include the following.

The EU Statistics on Income and Living Conditions (EU-SILC), which is a valuable source of data on child well-being. A special, one-shot module asking parents about the well-being of their children has been added to the 2009 wave, further enhancing the potential of this dataset.¹²⁴ A limitation of EU-SILC, however, is that children themselves are not directly surveyed, since it is the parents who are interviewed. Another concern is about sample size, especially as regards the longitudinal data. A validation exercise has highlighted the fact that improvements in documentation (especially on country-level national datasets) are needed. In certain cases, the variability in poverty and inequality figures over time dictates caution.¹²⁵ The EU-SILC is the obvious source on which to base the educational deprivation index due to be developed or the social participation of children indicator. To this end,

122 EU Task-Force 2008: 82. To get this type of information, it is necessary to involve children (at least those who are in schools and of an age when they are capable of answering questionnaires).

123 See details in OECD (2009).

124 The battery of questions of the new module is listed in Annex 3.4 of this report.

125 In some countries, the relatively weak robustness of some basic indicators appears to be a recurring problem, so the sampling design and data quality might need to be addressed.

a workshop on evaluation of the 2009 module would seem advisable, with the involvement of the various stakeholders.

The Labour Force Survey (LFS) covers a larger sample than the EU-SILC, but suffers from a lack of data on income. Furthermore, the data relate almost exclusively to the situation of respondents at the time of survey, and this may differ from their circumstances in the near or distant past. This limits, for example, the use of the data to measure the work intensity of households. So far, there has been little effective use of the longitudinal data that the survey compiles, even though they could potentially provide insights into the ease or difficulty of movement between employment, unemployment and inactivity (which is an important determinant of the income and living standards of households).

The Programme for International Student Assessment (PISA) is undertaken by the OECD every three years and focuses mainly on educational attainment. It is based on large school-based samples of 15-year-olds. The most recently published results are from the 2006 survey. PISA’s main potential contribution is to the access-to-education aspect of well-being. The data might be exploited further, especially as regards the breakdowns by parental education and the migrant status of parents.

Progress in International Reading Literacy Study (PIRLS) is a worldwide comparative reading assessment that is carried out every five years. It is based on school-based samples of 9- and 10-year-olds (the fourth grade of elementary school).

The most recently published results are from 2006. The number of participating countries is growing: the most recent (2006) wave covers 40 countries. PIRLS provides important information on children’s reading literacy achievement, as well as on various influences (home, school, national) on how efficiently students learn to read. It is suggested that the coverage of the child-related indicator pools should be extended to the reading literacy of 10-year-olds, broken down by parental education level, using the PIRLS survey.

In addition to PIRLS, the Trends in International Mathematics and Science Study (TIMSS) is undertaken every four years, at the fourth and the eighth grades, to provide data on maths and science. The most recent available results are from the 2007 wave, which covers 62 countries. At this stage, there is no suggestion that science and maths indicators should be included, but TIMSS is a potential source of such data.

The Health Behaviour in School-aged Children Survey (HBSC), coordinated by the World Health Organization (WHO), is carried out every four years. It is based on school-based samples of 11-, 13- and 15-year-olds. Country coverage is being extended wave by wave: the most recent wave (2005/06) covers 41 countries. The HBSC is an important source of data on behaviour and risks, subjective well-being and health, relationships, and school well-being. The main drawback to the HBSC is that there is no direct access to the microdata – this would be necessary to carry out the required analysis in order to construct satisfactory indicators.

For a cross-check of HBSC data on risk-taking behaviour, the European School Survey Project on Alcohol and Other Drugs (ESPAD) is also available. This survey, which is undertaken every four years, is based on samples of 16-year-olds.

The most recently published data are from the 2007 wave, which covered 35 European countries. The basic aim of the survey is to collect comparable data on the use of alcohol, tobacco and other drugs among students throughout Europe, so ESPAD could be an important source of data on behaviour and risks. Should the ISG decide, as a next step in the indicator-development process, to monitor smoking, alcohol and drug use among 16-year-olds, contacting the ESPAD research team would be useful to clarify the preconditions for data access.

Some of these surveys already cover children aged 11 and over, and there is scope for the others (in particular the EU-SILC) to lower the age of respondents and start interviewing children about their well-being. This could be carried out within the limits of the numerous methodological, legal and ethical issues recommended by the EU Task-Force report. Related to the issue of data infrastructure, it should be mentioned that an OECD project is already planned to evaluate the above-mentioned sources with respect to their comparability across countries.

In addition to more in-depth utilisation of survey-based datasets, the potential of administrative records at the national level has to be explored further: most importantly, indicators of crime and violence, of the extent and coverage of institutionalised care of children, guardianship and other forms of extra-family care.

Table 3.1: Overview of child well-being indicators, OMC and suggested new breakdowns

Dimension Indicator with 0–17 age breakdown Breakdown A1: Income A1.1: At-risk-of-poverty rate Child age, work intensity,

household type, migrant status A1.2: Relative median poverty risk

gap

A1.3: Persistent at-risk-of-poverty rate A1.4: Dispersion around the poverty threshold

A2: Material deprivation A2.1: Primary indicator of material

deprivation Child age, work intensity,

household type, migrant status A2.2: Secondary indicator of material

deprivation

A3: Housing A3.1: Housing costs Child age

A3.2: Overcrowding Child age

A4: Labour-market

attachment A4.1: Children living in jobless

households Child age

B1: Education B1.1: Low reading literacy

performance of pupils aged 15 Average performance by socio-economic status, migrant status

B1.2: Early school-leavers B2: Health B2.1: Life expectancy at birth

B2.2: Life expectancy at birth by SES B2.3: Infant mortality

B2.4: Infant mortality by SES B2.5: Perinatal mortality B2.6: Vaccination in children B3: Exposure to risk-taking behaviour

B4: Social participation and relationships, family environment B5: Local environment