EXECUTIVE SUMMARY
Background
1. Researchers in both the UK and the US have noted a gap in 'school readiness', i.e. how well prepared children are to start school, between children from less advantaged and more advantaged backgrounds. However, various research studies have also found that parenting and children's activities in the early years can make a difference to children's outcomes.
2. A composite measure of a child's home learning environment has been developed in the Effective Provision of Pre-school Education ( EPPE) project, the National Evaluation of Sure Start ( NESS) and the Millennium Cohort Study ( MCS). This composite measure of the Home Learning Environment was then found to be very effective in predicting future development for the EPPE, NESS and MCS longitudinal studies.
Aim and objectives
3. This study aims to investigate whether interview-based measures of children's activities are associated with cognitive ability at age 34 months, and whether they have independent effects once socio-demographic factors have been taken into account.
4. The objectives of this report are:
- to replicate, using GUS data, methods previously used with EPPE data to determine which variables should be included in a composite Home Learning Environment index
- to calculate the index, using the variables identified
- to examine the ability of the index to independently predict cognitive outcomes at age 34 months
Data
5. Growing Up in Scotland is a longitudinal study aimed at tracking the lives of a cohort of Scottish children, from the early years through to childhood and beyond. In 2005, data were collected on a cohort of 5,217 children aged approximately 10 months, and they have been followed up annually since then.
6. In this report, eight activity measures from the Year 1 interview, 18 from the Year 2 interview and 25 from the Year 3 interview ( i.e. 51 activities in total) were considered for their relationship with cognitive development. A small number of these measures had been examined in the original EPPE work on this topic (Melhuish et al., 2001; 2008), but the majority had not been tested in this way in previous research.
7. Cognitive and language ability in the birth cohort was measured at the 3 year contact by two assessments: the naming vocabulary and picture similarities subtests of the British Ability Scales ( BAS).
Statistical analysis
8. Firstly, multilevel models were constructed for each of the cognitive outcomes ( BAS naming vocabulary and picture similarities at age 3 years), with the following tested as predictors: age, child gender, birth weight, number of siblings, respondent and partner (parental) education, household socio-economic status, lone parent status, child developmental status in the first year, child health status in the first, second and third years, amount of centre-based care/education, and home area characteristics (deprivation and urban/rural).
9. From the baseline demographic multilevel models for naming vocabulary and picture similarities measures of relative over- and under-achievement were calculated for each outcome. Three categories of relative achievement (performance) were used (unexpected over-achievers, average i.e. as expected, and unexpected under-achievers) for both child outcomes.
10. Each of 51 home activity items derived from the GUS interviews were analysed separately with a multinomial logistic regression for any significant association with the individual categorical variables of over- or underachievement. On the basis of these results activities were selected for inclusion in a Home Learning Environment index.
11. From consideration of the analyses of over- and under-achievement and analyses of internal consistency, nine activity items were chosen to form a Home Learning Environment index.
12. The correlations of this HLE index with Household socio-economic status was 0.22, with respondent's (mother's) education was 0.20, and with partner's (father's) education was 0.10. This indicates that this measure of learning opportunities provided in the home is only slightly associated with parental socio-economic and educational status and that it can be regarded as a measure that is relatively independent of family demographics.
13. To further examine the effects of HLE on the prediction of achievement over that provided by family and background characteristics for children, new multilevel models for Naming Vocabulary and Picture Similarities were created that added the HLE (see Table 2). Comparison of models indicates a significant contribution of the HLE to children's attainment. Adding the HLE to the demographic model, the child level variance explained increased by 30% for Naming Vocabulary and 21% for Picture Similarities.
Discussion and conclusions
14. In the analyses presented in this report, it is clear that a measure of the Home Learning Environment ( HLE) added to the understanding of the influences that might affect a child's cognitive development. The influence of the HLE was over and above that of standard measures of family socio-demographic factors such as parental education, socio-economic status and income.
15. Of the nine activity measures selected in the this report for inclusion in the Home Learning Environment measure, eight are similar to those identified in the original EPPE research in England (Melhuish et al., 2001; 2008a) and the ninth item (number of children's books in home) is similar to an item (number of books in home) identified in several previous studies ( e.g. Kirsch et al., 2002) as being related to children's cognitive development or educational achievement. Hence, the findings in this report are supported by research on other populations. It is interesting to note that none of the other items in the 51 tested added further to a measure of the Home Learning Environment.
16. Research findings such as those reported here suggest that policies that encourage active parenting strategies (including for disadvantaged parents) can help to promote young children's cognitive development and educational achievement both early and later in development.