This data collection includes syntax needed to: (1) merge data and commands used for weighting data; (2) derive the sample for analysis 1 (same father-mother households s1-5); (3) derive father involvement (dependent) variables; (4) derive Independent variables; (5) derive mother’s annual pay; (6) derive the relationship breakdown variables for analysis 2. All variable manipulations were derived to explore i) paternal involvement in childcare and housework, and ii) the association between paternal involvement in childcare and relationship stability.One of the root causes of persistent gender inequalities in economic and political life is that women carry a heavier workload in the domestic domain where they still do most of the work involved in looking after children and other family members. Women's engagement in employment has risen over the last four decades but men's contribution to childcare and housework has grown more slowly. Sen's (1992) 'capability framework' elaborates how state and organisational policies, social norms, and household economic and demographic circumstances shape men and women's options, decisions and behaviours. This framework suggests there are many social, economic, demographic and cultural factors, which exert logistic pressures on the arrangement of the domestic division of labour in households. Yet the relative importance of these factors in shaping men's involvement in childcare remains under-researched and largely based on small-scale qualitative studies or cross-sectional survey data (Norman 2010; also see O'Brien 2005). Our earlier research (Norman 2010; Norman et al 2014; Norman and Elliot 2015) used the Millennium Cohort Study (MCS) to develop measures of paternal involvement in childcare when the child was aged nine months and three years old. We found the mothers' employment hours had the strongest association with paternal involvement: if the mother worked full-time both nine months and three years after the child's birth then the father was more likely to be an involved parent when the child was aged three. Fathers were also more likely to be involved when the child was aged three if (i) they worked shorter hours in employment and (ii) if they were involved in childcare nine months after the child's birth; but the effect of both these variables was significantly weaker than that of the mothers' employment hours. This research project will build on this analysis, using the MCS, by developing more measures of paternal involvement in childcare to establish which employment and socio-demographic characteristics shape paternal involvement as children age from nine months to eleven years old. Part of the analysis will focus on intact households to remove the confounding impact of relationship breakdown. We will also analyse the relationship between paternal involvement and the probability of households remaining intact given previous research has found a correlation between paternal involvement and the quality of a couple relationship (e.g. Poole et al. 2014). The research questions to be addressed are: 1. How can we develop measures of paternal involvement over time as the child develops? 2. What are the key employment, socio-demographic, and attitudinal characteristics of fathers in the UK who report involved parenting behaviour when their child is aged 9 months, 3, 5, 7 and 11 years old? 3. Do trajectories of paternal involvement over the child's lifecourse vary between fathers and if so, what are the predictors? 4. Does paternal involvement when the child is aged nine months predict whether a household is still intact when the child reaches age eleven? In examining these questions, the project aims to contribute to scholarly and policy debates about what encourages or impedes fathers' involvement in providing care for their children. It will make an original contribution to the literature on parental involvement by using a representative sample of fathers to develop measures of paternal involvement, identify differences among fathers and explore how their involvement develops as the child grows older. This is particularly relevant in light of the growing attention to fathers within policy debates about work-family issues across Europe (e.g. European Union 2013), including UK policy, where the introduction of shared parental leave is the most recent reform designed to provide better support for fathers and their involvement in childcare (BIS 2014).

The UK’s Millennium Cohort Study - managed by the Centre for Longitudinal Studies - is a nationally representative survey following a cohort of children born around the year 2000 using a clustered, disproportionately stratified sampling design. The first sweep (2001/02), covers a cohort of 18,819 babies aged nine months (raised in 18,552 families). The second sweep (2003/04) followed the same cohort of children, plus 692 newly recruited families resulting in an overall, combined sample size of 19,244 households - although only 15,590 households were productive. Sweep 3 (2006) includes 15,246 productive households, sweep 4 (2008) includes 13,857 productive households and sweep 5 (2012) includes 13,287 productive households (see Hansen 2014: A Guide to the Datasets, 8th edition - available via www.cls.ioe.ac.uk). For the first part of our analysis, we filtered the sample to include only the same, two-parent households - containing a mother and a father - over the five sweeps of data. This reduced the sample size to 5,882 households. This allowed us to explore paternal involvement trajectories over the five sweeps of data. For the second part of our analysis, we filtered the sample to include all two-parent households - containing a mother and a father - in sweep one, which reduced the sample size to 13,411. This part of the analysis focused on exploring relationship trajectories between the parents, in particular relationship breakdown, over the four sweeps of data.