Theme 3: Project 2 – Comparative Analysis of Early Childhood Outcomes

Project Lead

Lucinda Platt

Team Members

Michela Franceschelli
Rachel Wilde

Description

Children’s early health, development and well-being vary widely both within and between countries. Early child outcomes are influenced by both individual and family characteristics and local, national and institutional contexts. However, we still have only limited understanding of the extent to which differences in national cultural and institutional (policy) differences drive variations between countries, and how far they stem from the different characteristics and circumstances of families bringing up young children in different countries.

Children’s early health, development and well-being vary widely both within and between countries. Early child outcomes are influenced by both individual and family characteristics and local, national and institutional contexts. However, we still have only limited understanding of the extent to which differences in national cultural and institutional (policy) differences drive variations between countries, and how far they stem from the different characteristics and circumstances of families bringing up young children in different countries.

The project explores patterns of early child outcomes, addressing these questions in a longitudinal framework.

Hence, this project addresses four main questions:

1) How do selected early child outcomes vary cross-nationally among a set of broadly comparable developed nations?

2) What are the key individual and family-level influences on early childhood outcomes across these countries – and is their influence consistent for each of the countries considered?

3) To what extent are cross-national differences in composition of families with young children (e.g. ethnic, socio-economic) implicated in cross-national variation in child outcomes?

4) What can cross-national comparisons reveal about the importance of national and institutional context in early child outcomes?

The project exploits the wealth of mature and developing birth cohort studies from UK, Ireland, Australia, New Zealand and France, involving collaboration with researchers from all these studies, to explore patterns of early child outcomes, addressing these questions in a longitudinal framework.

What the project did

This project engages with how cross-cohort comparisons can shed light on issues of child development and early child education and care (ECEC), by providing insight into institutional effects across a set of countries identified for both similarities and differences in early child outcomes and institutional context.

It has two main strands. First a study of breast-feeding and how it varies with socio-economic background. This is a ‘proof of concept study’ and involves collaboration and data harmonisation across 4 cohort studies from UK, Ireland, Australia and New Zealand. While these four countries share a number of institutional features which renders them appropriate for comparison, they also differ strongly in breast-feeding rates, with higher rates in Australia and New Zealand and low rates in the UK and Ireland. Given the recognised importance of breast-feeding for children’s development, this is a highly relevant issue to consider in relation to potential policy learning. Breast feeding follows a clear class / SES gradient, though this appears to be stronger in some contexts than others. We therefore ask two questions: first, is the SES gradient in breast-feeding equivalent across the four countries? Second: how far can compositional differences among mothers (including in class background) help to explain differences in breast feeding rates across the countries?

Second is a comparative study of the relationship between formal pre-school child care and children’s non-cognitive skills as they enter primary school.  For this we use the Growing up in Ireland compared with The Millennium Cohort Study. Non-cognitive skills are a source of substantial academic interest given their potential to support learning and to provide additional benefits over and above cognitive skills in adult outcomes. While substantial attention has been paid to the relationship between ECEC and cognitive development, rather less has been paid to the relationship between ECEC and non-cognitive skills, and what there is offers rather mixed results. By exploring the relationship in two distinct but relatively comparable contexts, with fully harmonised data sets, we aim to increase the confidence in the robustness of consistent findings, while highlight the extent to which any differences do not stem merely from different operationalisations in different country contexts, but represent clear institutional differences that therefore challenge generalisability. We ask does participation in formal childcare compared to other forms of institutional care lead to a greater command of non-cognitive skills on school entry?

Key features of these projects are the use of advanced quantitative methods for addressing the research questions: decomposition of binary outcomes and propensity score matching, respectively; and the careful harmonisation of cohort data from different contexts that have not been made explicitly comparative a priori. This post-hoc harmonisation, and the ways in which analyses have been replicated across the different country partners for the breast feeding study illustrates both the potential and the challenges in this type of comparative study, and is intended to be used as a model, potentially for a succeeding major project.

Both papers have required substantial data work and the harmonisation of the data sets and co-ordination across the partners has been non-trivial. This has introduced substantial delays as we have had to depend on others running data and dealing with queries and corrections at one remove and around others’ availability. There have also been numerous changes of personnel, which caused disruption. Hence the time-to-output has taken longer than in more straightforward scenarios. But as noted, this experience and the harmonisation itself has the potential to offer an output in itself and we are exploring how we could release files to benefit the wider community in harmonisation.

Findings

For paper 1, we find that there are some differences in the relationship between maternal characteristics and breast-feeding propensity across the four contexts, leading to some reconsideration of the received wisdom and highlight the potential for intervention. Secondly, we find that only a small part of the cross-country differences in rates can be explained by compositional differences, highlighting the significance of national context and policy.

For paper 2, we find that formal childcare has only a small and, if anything, a negative impact on non-cognitive skills at child entry, even conditioning on prior levels.

Potential Impact

We expect these papers to be of substantial policy interest and impact and will be working closely with our contacts already established within the DfE and DH to amplify them once the papers are published in suitable outlets. There has already been some interest in the presentations we have offered on our interim findings from policy audiences but we aim to use publication to focus the impact and direct it towards peer reviewed outputs. We will also accompany publication with the use of our media contacts, in particular we aim to follow up successful pieces for the BBC website with a further contribution, and we will also be writing blogs on both pieces. We are targeting high ranked journals for publication, such as Journal of Marriage and Family, Economics Letters or others of similar calibre.

Interim findings in the form of existing presentations can already be found at https://lucindaplatt.com/.

News

Blog

Events

Publications