Previous research has demonstrated persistent links between living in social housing during childhood and subsequent, usuallyadult, disadvantage. Recent work using the 1958 and 1970 British birth cohorts has also shown that children in families that moved into social housing during childhood were often worse off than children who were born into social housing and whose families never changed housing status. This finding suggests that the circumstances that precipitated the move were likely to be important. Family change is likely to be one such precipitating factor. In this paper, I use data drawn from the Fragile Families and Child Wellbeing Study to explore the ways that family structure and housing tenure are co-determined and how they independently contribute to child health and cognitive development. The models will pay particular attention to differences by race/ethnicity. They also will examine whether comparisons of immigrant and non-immigrant families are informative insofar as immigrants families in the US have less recourse to housing assistance than their native born counterparts. Although social housing appears to be a mark of disadvantage, there may well also be countervailing effects because, ceteris paribus, social housing tenants have lower housing costs and greater disposable income. To explore this issue, the paper will also examine whether access to housing assistance is associated with a lower incidence of poverty (measured using indicators of deprivation).

Email: W.Sigle-Rushton@lse.ac.uk

The consequences of childbearing postponement: how are children of older mothers faring?
Alice Goisis, London School of Economics

While numerous medical studies have claimed that giving birth after age 35 puts the child at an excessive risk, the literature hasfocused almost exclusively on health outcomes at birth. Consequently we have only partial knowledge about the consequences of older childbearing on the well-being of children. A better understanding of this link is needed, not least because childbearing is being increasingly postponed in richer countries. As women with high socio-economic statutes are the most likely to postpone their first birth and other women are likely to have smaller families, the characteristics of older mothers has been changing in ways that may affect the relationship between maternal age and child wellbeing. It is possible that the characteristics of older mothers may be able to more than compensate the reproductive risks claimed by the medical literature and result in improved outcomes for children. Using data from the Millennium Cohort Study (U.K.), this paper investigates the consequences of childbearing postponement on child wellbeing by comparing first born children of older and younger mothers. The rich data allow a wide range of cognitive and behavioural outcomes to be examined. OLS models and an estimator based on the eliminant method will allow me to assess whether the developmental process of children of older mothers significantly differ from that of children of younger mothers and to what extent increased socio-economic wellbeing compensate for the biological risks involved with conceiving at an advanced age.

Email: A.Goisis@lse.ac.uk

Orphanood, schooling costs and child's time allocation in Zimbabwe
Rafael Novella, ISER, University of Essex

This paper explores the effect of orphanhood, a main source of children's vulnerability, on the allocation of children's time toschool and work activities. The deterioration of one of the best education systems in Africa and a considerable high incidence of orphanhood make Zimbabwe an interesting case of study. In particular, this paper explores the determinants of the child's time allocation for those able to attend lower secondary (O-level). After controlling for household wealth and a diverse set of covariates at individual, household and community fixed-effects, I find that orphans are less likely to attend school and more likely to perform work activities. In addition, I find that the marginal cost of investing in an orphan's human capital is more expensive than in a non-orphan, particularly for orphans who are girls and older children. Finally, I find that there exists discrimination within households against children living with the households' head when they are not biologically related.

Email: grnove@essex.ac.uk

Children's home lives in the context of parental migration in South Africa
Rachel Bennett, University of Southampton; Victoria Hosegood, University of Southampton, Africa Centre for Health and Population Studies, University of KwaZulu-Natal; Marie-Louise Newell, Africa Centre for Health and Population Studies, University of KwaZulu-Natal; University College London; Nuala McGrath, Africa Centre for Health and Population Studies, University of KwaZulu-Natal; London School of Hygiene & Tropical Medicine

The dispersal of parents and siblings between multiple households is a commonplace experience for many children in SouthAfrica. The conditions and experiences of children living apart from their migrant parents have long been a source of concern for researchers, government and non-governmental organisations. Few sources of empirical data are available that adequately allow us to situate childhood experiences within the context of changing parental circumstances including work, migration, cohabitation, relationship stability and parental death. This paper explores the social and residential dynamics of migrant parents and their children in KwaZulu-Natal, South Africa using detailed, longitudinal data about members of rural households in a demographic surveillance population, and a linked sample survey of adult migrants. We describe the migratory and residential experiences, and socio-demographic characteristics of 236 migrant parents and 460 of their biological children (less than 18 years) reported in surveillance or survey interviews. One quarter of the migrant parents reported at least one child as being a member of both their destination and origin households. The majority of parents (72%; 67% of mothers and 84% of fathers, p<.05) did not report any of their children as members of their destination household. Thirty percent of these 'left-behind' children were resident locally but were not living in the 'origin' household of their migrant parent. We explore factors associated with types of dispersed family arrangements and the extent of children's residential stability by age. These findings contribute to understanding the dynamic and multiple spaces in which children and their parents live.

Email: rjb2g09@soton.ac.uk

Fragile families in the UK and the USA
Kathleen Kiernan, University of York; Sara McLanahan, Princeton University; John Holmes, University of York; Melanie Wright, Princeton University

In both the UK and the US, dramatic increases in non-marital births over the past forty years have exposed growing numbers ofchildren to non-traditional family structures. The capabilities of unmarried parents, the evolution of their relationships over their children's lives, and the effects of family structure and relationship transitions on children are largely unknown. This study explores these topics using the first five years of the Millennium Cohort Study and the Fragile Families and Child Wellbeing Study, which follow birth cohorts in the United Kingdom and in the United States, respectively. In both countries, unmarried parents have lower capabilities and less stable relationships than married parents, and single mothers tend to be more disadvantaged than cohabiting mothers. Relationship instability is associated with a variety of negative outcomes for both mothers and children. One key difference between the two countries is greater instability and family complexity in the United States. In addition, while cohabiting mothers in the UK are only slightly more disadvantaged than their married counterparts, cohabiting mothers in the US tend to be far worse off and more closely resemble single mothers than married ones.

Email: kathleen.kiernan@york.ac.uk

Adolescent childbearing, union history, and women's health and well-being at midlife
Kristi Williams, The Ohio State University; Sharon Sassler, Cornell University

Despite recent declines in the adolescent birth rate in the United States, approximately 10% of all births occur to women age 19or younger—a higher rate than any other developed country. Moreover, the vast majority of teen births occur to unmarried women. Despite the prevalence of adolescent and nonmarital childbearing, little is known about their long-term consequences for women's mental and physical health over the life course or about the role that subsequent marriage and cohabitation unions play in ameliorating any negative effects. We use 29 years of panel data from the National Longitudinal Survey of Youth 1979 to examine differences in age-40 mental and physical health between women who had a teen birth and those whose first birth occurred at later ages, distinguishing those whose birth occurred in marriage from those with nonmarital first births. We then differentiate women who had a nonmarital first birth during adolescence by their subsequent union histories and estimate the effect of marrying or cohabiting versus remaining never-married on midlife self-assessed health and psychological well-being, paying particular attention to the paternity status of the mother's partner and the stability of marital unions. To partially address selection bias, we employ multivariate propensity score techniques. Preliminary results suggest that teen childbearing is negatively associated with midlife health and this effect is exacerbated among those who had a teen birth while unmarried. We find little evidence that these negative health consequences of nonmarital childbearing are mitigated by either subsequent marriage or cohabitation.

Email: williams.2339@osu.edu

Serial cohabitation among men in Britain: does work history matter?
Erzsebet Bukodi, Institute of Education

This paper examines the importance of men's serial cohabitation experience in the outcomes of their cohabitations asking threequestions: (1) Are serially cohabiting men less likely to marry their partners but more likely to dissolve their cohabitations than single-instance cohabitors? (2) If so, what part is played by features of their work histories in explaining these outcomes? (3) Do associations between serial cohabitation and cohabitation outcomes result from selection or causation? The analyses are based on data from two British birth cohort studies relating to people born in 1958 and 1970. I use discrete-time event-history techniques along with duration models with selection. I find that serial cohabitors are less likely to transit to marriage but are more likely to dissolve their partnerships. The analysis clearly shows that serial cohabitors form a selected group via their early work histories. However, the negative effect of serial cohabitation on marriage and the positive effect of serial cohabitation on separation remain significant and substantial even after I control for cohabiting men's before-cohabitation and within-cohabitation work histories and explicitly correct for possible selectivity of serial cohabitation. It thus appears that the experience of serial cohabitation itself affects the attitudes of men towards marriage.

Email: E.Bukodi@ioe.ac.uk

Long-term trends of men's co-residence with children in England and Wales
Ursula Henz, London School of Economics

Motivated by the observation of decreasing family involvement of men in the US, the study carries out corresponding analysesfor England and Wales by examining changes in men's (and women's) frequencies of co-residence with children using the National Statistics Longitudinal Study (LS). The aim of the research is establishing long-term trends of men's co-residence with children in England and Wales. The second aim is exploring possible differences in these trends compared to the US. Period comparisons are straightforward for the calendar years 1971, 1981, 1991, 2001. Establishing cohort trends for cohorts born between 1930 and 1970 is more difficult because of the ten-year gaps between the census waves. The paper will propose a method for approximating the cohort and age-specific co-residence rates and will show trends in living with at least one child under age 18 for all men and women as well as for subgroups (by education, ethnicity). Initial analyses show a decreasing family involvement of men (in the limited sense of co-residence) but not to the same degree as in the US in the same period. Other interesting results are the relative homogeneity of men's fatherhood experience in young adult years compared to women, and the considerable level of fatherhood involvement of African Caribbean men.

Email: U.Henz@lse.ac.uk

Parental separation and adult psychological distress: has the association changed over time?
Rebecca Lacey, Mel Bartley Hynek Pikhart, Mai Stafford and Noriko Cable, University College London

Background: The relationship between parental separation and increased psychological distress in adulthood is well established, however little is known about how this may have changed over time and by gender. The "reduced effect" hypothesis suggests that as societal separation rates increase there will be less effect upon the children involved. This is thought to be mainly due to reduced stigma following the 'normalization' of divorce, increased availability of services and greater awareness of maintaining contact between both parents and children. This study tests the "reduced effect" hypothesis with regards to the effect on psychological distress in adulthood using three British birth cohorts.
Methods: Multiply-imputed data on 5,362 participants of the 1946 National Survey of Health and Development (NSHD), 10,923 participants of the 1958 National Child Development Study (NCDS) and 10,714 participants of the 1970 British Cohort Study (BCS) were analysed. Parental separation was measured from 0-16 years and psychological distress measured at age 30 yrs (BCS), 33 yrs (NCDS) and 36 yrs (NSHD). Logistic regression analyses were conducted testing the association (controlling for mother's education, psychological health, age and father's social class) and Wald tests conducted to test gender-separation and cohort-separation interactions.
Results: In all studies parental separation is associated with increased odds of reporting psychological distress. There was no evidence that this varied by gender or cohort.
Conclusions: Despite increasing parental separation across these three cohorts there was no evidence to support the "reduced effect" hypothesis.

Email: rebecca.lacey@ucl.ac.uk

The association between father absence, age at puberty, and reproductive timing in British men
Paula Sheppard, London School of Economics; Rebecca Sear, Durham University

The role of fathers is not yet well understood. Many studies have reported mixed results with regard to the relationship betweenfather absence and child survival in the developing world. Among Western populations, researchers have generally found that father absence is correlated with earlier age at puberty and first reproduction, particularly for girls. Little has been done, however, to test the effects of father absence on the timing of boys' life history transitions within this context. Drawing upon data from the National Child Development Study, a large British cohort, this study tested whether father absence had an effect on the timing of male puberty and reproductive events. A further aim was to determine whether the effects of father absence differed according to the different childhood stages at which the boys lost their father so as to test whether father absence is merely a single indicator of a sub-optimal childhood environment or if it is an independent effect. The results suggest that father absence before age seven is significantly associated with early reproduction, while father absence between ages 11 and 16 only is associated with delayed voice-breaking (a proxy for puberty), even after adjusting for other factors denoting a stressed childhood environment. We conclude that fathers exert an influence on male reproductive outcomes and that these effects may be independent of other factors indicating an adverse early environment.

Email: P.J.Sheppard@lse.ac.uk

Marriage duration and divorce: the seven-year itch or a life-long itch?
Hill Kulu, University of Liverpool

Most studies show that the risk of divorce is low during the first months of marriage; it then increases, reaches a maximum andthereafter begins to decrease. The reason for the rising-falling pattern of divorce risk, however, is far from clear. Classical psychological literature considers this pattern consistent with the notion of a seven-year itch. Married couples experience a gradual decline in marital quality suggesting that the short 'honeymoon period' of passion is followed by a longer 'posthoneymoon period' of strife. Other researchers argue that that the rising-falling pattern of divorce risk is simply a consequence of misspecification of divorce models because of omitted covariates or unobserved heterogeneity. The aim of this study is to investigate whether the rising-falling pattern of divorce risk is caused by the seven-year itch or whether unobserved heterogeneity is the main reason. We use register data from Finland and apply multilevel hazard models. We first study the hazard of divorce over marriage duration with and without controlling for a set of demographic and socioeconomic characteristics of women and their partners. We then control for unobserved heterogeneity to find out any changes in the shape of baseline risk. We also examine the hazard of divorce over marriage duration separately for the first and for subsequent marriages and across cohorts to identify any changes over years. Our preliminary analysis shows that when controlling for unobserved heterogeneity the risk of divorce increases over marriage duration rather than follows the rising-falling pattern; this suggests that the marital itch may last

Email: hill.kulu@liverpool.ac.uk

Divorce in India: trends, patterns and household structure
Premchand Dommaraju, Nanyang Technical University, Singapore

Indian marriage system has been characterized by universal marriage with strong caste endogamy, relatively early marriage andby very low divorce rates. But with liberalization, both economic and media, and increasing participation of women in tertiary education and in the paid labour force, the traditional marriage system has come under pressure. If media reports are to be believed, divorce rates in India are on the upward march (See, New York Times, 2008, The great Indian wedding is succumbing to the great Indian divorce; The Times (UK), 2008, The divorce rate in Delhi has doubled). But these reports are based on anecdotal and selective data. The absence of good quality data on divorce has meant that there has not been any rigorous demographic study of divorce trends in India. Data on divorce are not compiled by the Indian government and major demographic surveys (DHS and NSS) also do not have questions on divorce. This paper will use data from two new nationally representative surveys, Indian Human Development Survey (IHDS, 2005-6) and District Level Health Survey (DLHS-3, 2007-8), which for the first time included questions on timing of divorce and remarriage, along with questions on household structure and living arrangements among other things. Using this two datasets, I will in this paper: a) present a comprehensive picture of divorce trends and divorce patterns for various groups (religious, rural/urban, regional, age patterns, etc.); b) I will analyse some the consequences of divorce, especially on living arrangements and household structure; c) situate the findings within the broader literature on changes in attitudes towards marriage and divorce in India, and changes in institutional and structural factors that might explain recent changes in divorce rates.

Email: premchand@ntu.edu.sg

Examining cohabitation trajectories after a first birth: findings from across Europe and the United States
Brienna Perelli-Harris, University of Southampton

The increasing percent of births within cohabitation across almost all of Europe and the United States indicates that cohabitationis becoming more common as a setting for childbearing. However, the higher percent of first births than second births in cohabitation in most countries suggests that cohabitation is not necessarily a setting for childrearing, or that cohabitors and married couples differ in their fertility behavior. Here we employ union and fertility histories from 12 countries to investigate what happens after a first birth. First, we employ event history analyses to examine whether second conception rates are higher in marriage or cohabitation. Second, we use competing risk hazard models to analyse whether cohabiting couples with a first birth are more likely to experience marriage, a second conception, or union dissolution. In countries where the risk of marriage after first birth is highest, cohabitation can be considered a stage in the family formation process, with marriage postponed to later in the life course. In countries where the risk of dissolution after first birth is highest, cohabitation can be considered an unstable state, with a higher risk of lone motherhood. In countries where the risk of second conceptions is highest, cohabitation can be considered more of an alternative to marriage, assuming that having additional children signifies union stability and a suitable setting for childrearing. Preliminary analyses show that countries differ considerably by outcome after first birth, suggesting considerable cross-country variation in the meaning of cohabitation with respect to childbearing.

Email: b.g.perelli-harris@soton.ac.uk

Work on families at the Office for National Statistics
Karen Gask, Office for National Statistics

The families team at the Office for National Statistics (ONS) is charged with meeting the needs of users of statistics on families and households. This presentation describes the work of the team over the past year, including a new set of statistics on families, and a description of civil partnership in the UK, five years on from its introduction in December 2005. Planned future work on families at ONS will also be described and users will be invited to comment on what they see as the priorities for work in this area.

In April 2011, the Office for National Statistics published a new set of statistics on families in the UK between 2001 and 2010. This presentation describes the key trends in the statistics on families by family type. Types of family include married couple families, cohabiting couple families and lone parent families. The increase in people living alone is discussed.

The presentation also briefly examines trends in civil partnership formations and dissolutions in the UK, five years after the Civil Partnership Act enabled same-sex couples to legally register their relationship for the first time.

The intended future work programme of the families team is also discussed, including meeting unmet user needs for specific statistics and planning analysis on families statistics from the 2011 Census.

Users of ONS' families statistics are encouraged to provide feedback on recent and future work.

Email: karen.gask@ons.gsi.gov.uk

The Changing Demography of Mid-life, from the 1980s to the 2000s
DH Demey, Ann Berrington, Maria Evandrou, Jane Falkingham, University of Southampton

This study examines changes between 1984 and 2007 in the demographic and socio-economic circumstances of British men and women in mid-life. Changing living arrangements in mid-life reflect historical changes in the occurrence and timing of life events such as marriage and parenthood, as well as increased longevity. In order to place mid-life in this wider demographic context, the paper first reviews changes over time in kin availability across the adult life course using the British Household Panel Survey (2001) and Understanding Society (2009). The paper goes on to use data from the General Household Survey (1984–2007) to document shifts over time in living arrangements for those aged 20–79. In the final part of the paper we focus specifically on those aged between 45 and 64 and examine how their characteristics in terms of marital status, educational attainment, activity status and housing tenure have changed over the past quarter century.

Email: d.demey@soton.ac.uk

Intergenerational exchange in the UK
Tak Wing Chan, University of Oxford

In this paper, we use data from the British Household Panel Survey to explore the pattern and dynamics of the exchange of instrumental support between adult children and their non-coresident parents. In contrast to previous research in this area which rely on cross- sectional data, we are able to analyse the dynamic of intergenerational exchange with panel data. We employ fixed effects model and transition models, and report an asymmetry in the ebb and flow of intergenerational exchange. We also relate our findings to the research of historians and historical demographers and those from British community studies.

Email; tw.chan@sociology.ox.ac.uk