The proportion of births in England and Wales registered to married couples has been falling steadily for the last three decades, from over 90% in the late 1970s to 55% in 2008 (source: ONS). At the same time the proportion of births to cohabiting couples has been increasing over time, from around 10% in 1986 to 30% in 2008.

The impact that being born to non-married parents has on children's cognitive, social and emotional development is therefore an important question for academics and policy makers alike. The issue seems likely to be the subject of much debate at the forthcoming general election, since the Conservative party has proposed to "recognise marriage in the tax system".

Previous commentators have concluded that children born to cohabiting parents have worse outcomes than those born to married couples. It is also widely acknowledged, however, that cohabiting parents differ systematically from married parents in many ways aside from their formal marital status; typically they are less educated, younger and have a lower household income, than married parents. They may also differ in less easily observable ways, for example in their relationship quality, stability and commitment to their partner even before the birth of their child. Once these factors are accounted for, there may be smaller or no differences in their children's outcomes. There has been remarkably little systematic work on this issue from the UK to date. Further research in this area is therefore valuable, and our project has four key aims which will inform the debate:

1. To describe in a simple but comprehensive way how young people born to parents of different household types differ in the development of a range of key skills (including cognitive, non-cognitive, behavioural, social and emotional skills), and engagement in a range of risky behaviours.
2. To describe some of the possible pathways, such as relationship stability and parental mental health, through which children born into cohabitation compared with formal marriage may either gain, or suffer, developmentally.
3. To assess to a greater extent than has previously been possible how far such differences in developmental trajectories (and the possible pathways examined) can be given a causal interpretation or whether they simply indicate positive (or negative) selection into marriage by different types of parents.
4. To assess the policy implications of the above, including a review of what the impact of changing the tax incentives to marry might be.

The project will mainly use data from three datasets; the Millennium Cohort Study, the Avon Longitudinal Study of Parents and Children, and the Longitudinal Study of Young People in England. These three cohort survey datasets will allow us to compare outcomes across children at a range of ages, from 3 to 17. We will also use the British Cohort Study, which has detailed life and relationship histories of cohort members, to address the causal impact of marriage on children's outcomes more closely.