Poster abstracts

BSPS Conference poster prize 2014

Each year, BSPS gives a prize for the best poster from the Conference poster display. In 2014, the prize was split between the best student and the best non-student poster. The best student poster award went to Heini Vaisanen (LSE) for her poster The association between education and induced abortion in Finland: a cohort perspective.

The best non-student poster award went to John Black (Cambridge Group for the History of Population & Social Structure) for his poster Burial practices in Manchester1750-1850.

Heini Vaisanen poster prize winner

Hampshire themed-posters

Hampshire County Council - Infographics and Data Visualisation of Forecast and Census data
Jack Cox, Hampshire County Council Demography Team

Presenting data to a range of audiences in formats that allow for easy digestion, understanding and interaction is of vital importance for Hampshire County Council. The statistics we produce must be easily accessible to a wide range of audiences including members of the public, council members and other more advanced statistical users in the council as a wider entity. This poster wishes to showcase a range of infographics and data visualisations brought together from our Small Area Population Forecasts and the 2011 Census. Furthermore, two of our most recent videos will present Hampshire’s vital statistics data in an engaging format whilst the other will showcase how valuable census data is for historical analysis by looking back to the mid 1800’s to observe the changes that have taken place in Hampshire’s age and sex structure.


Teaching population history through local demographic studies of Hampshire Andrew Hinde, University of Southampton

This poster will describe some examples of how local research projects on Hampshire's population history are being used to teach undergraduate students at the University of Southampton about the methods and materials of historical demography. Students are shown archival sources in the Hampshire Archives and Local Studies Office in Winchester, and asked to undertake a small project on a local area (for example a parish, or part of a town). Working in groups, they collect and analyse data which illuminate some aspect of the population history of that area. They then produce a report describing the project and its results, and comparing the experience of their local area with known regional and national trends. Over many years, these small projects have been undertaken on many parishes, and allow a general picture of several aspects of the demography of the county to be painted. The poster presents some illustrative results which deal with topics such as seasonality in marriages, the analysis of mortality 'crises', comparing rates of population growth among parishes, and occupational change during the nineteenth century. There has been wider interest in this approach to teaching local history: the course co-ordinator has been invited to made two presentations to the Higher Education Academy.


A profile of Hampshire - using census data
Gemma Quarendon, Hampshire County Council, Jack Cox Hampshire County Council

The 2011 census provides us with a wealth of data not available elsewhere or for the low level geographies we regularly make use of within Hampshire County Council. Using largely the initial results from the 2011 census this poster will present a profile of Hampshire County Council (and its districts) and investigate how it has changed since 2001. Using the latest info-graphic techniques this poster will show how Hampshire has grown, aged, diversified and adapted since 2001 to show what it means to be a resident of Hampshire today.


Using the ONS Longitudinal study for local demography: a case study of Hampshire Rachel Stuchbury, Chris Marshall, Nicola Shelton, CeLSIUS UCL

The ONS Longitudinal Study covers England and Wales: approximately 1% sample using four birthdates from 1971 onwards. The linked decennial census data is available for the whole household, not just sample member. The data includes vital registration data ("events"), including: birth of sample member, birth of children to sample member, widow(er)hood of sample member and cancer and death registrations. The poster will show data to answer a series of questions: Is there a strong association between housing tenure in childhood and housing tenure in young adulthood in Hampshire? Did the association changed between 1991 and 2011? Has the pattern of housing tenure in young adulthood in Hampshire changed between 1991 and 2011? Has the likelihood of attaining Householder (or joint Householder) status in young adulthood in Hampshire changed between 1991 and 2011? How does this vary by housing tenure? For more information about the ONS Longitudinal Study please see


Forecasting demand for Hampshire County Council Social care through whole system computer simulation modelling
Joe Viana, Sally Brailsford, University of Southampton, Kevin Andrews, Hampshire County Council

The aim of the project was to improve forecasting of Social Care needs within Hampshire over a 5 year period, in support of the Hampshire Joint Strategic Needs Assessment. The ageing population has the potential to place excess demand on the social care system and the aim of this project, part of the overall Care Life Cycle Project, is to evaluate the effect of this demographic change on the demand for social care utilising complexity science principles. Data from the SWIFT system as well as nationally representative survey data including the English Longitudinal Study of Ageing was incorporated in a system dynamics computer model developed to address the problem. The demand for social care in Hampshire is not limited to the social care provided by the council. Social care in Hampshire is provided by multiple stakeholders, which can be grouped into: i) formal (Hampshire County Council), ii) Informal (friends, family, privately purchased and charities) and iii) both a combination formal and informal. It is the interaction/feedback between the providers of social care and the likelihood of the person with social care needs requesting care that impact the demand for services.

Preliminary results from the simulation model have been validated against historical data. As well as the production of alternative scenarios to project outcomes forwards over the agreed 5 year time horizon. It is hoped that this work can illustrate the usefulness of the approach with respect for care demand and promote its use in other parts of the organisation.


Determinants of Undocumented Migration from North Africa to South Europe
Amany Hassan Abdel-Karim, Tanta University, Egypt

This study focused on determining the impact of demographic, socio-economic, and migration networks factors that affect migration type (with or without visa/work permit) from North Africa to South Europe. Since Egypt and Morocco are the grand population of North Africa, they can be considered as representative of this population. Logistic regression model is the most appropriate model to be constructed for this research since the response data (migration type) is binary. Data of this study were obtained from Netherlands Interdisciplinary Demographic Institute (NIDI) that were collected by local research teams. Questionnaires information was provided by Egyptians migrants in Italy and Moroccans migrants in Spain. These data were concurrently collected from migrants and comparable survey instruments were used such that questionnaires share the same basic modular design and layout. The study results showed that households with higher levels of education, had job in their countries before migration, had higher levels of financial situation before migration are more likely to migrate legally. Egyptians migrated to Italy are more likely to have documented migration. In addition, migrants who had no children before migration, migrated alone, and those who thought it is easier to find a job in the receiving countries are less likely to migrate without visa or work permit. The results of this study can be useful for legislation of migration in the receiving countries.


After the 'One Child Policy'? Fertility Intentions and Ideals in China
Stuart Basten, Department of Social Policy and Intervention, University of Oxford, Hou Jiawei, Baochang Gu, Renmin University of China

In recent years, scholars have increasingly called into question the claim that the fall to low fertility in China was primarily driven by family planning restrictions; instead that economic development, urbanisation, and the development of improved educational and employment opportunities. In this view, China – and urban areas in particular – share more in common with other low fertility settings in Pacific Asia than perhaps previously recognised. Fertility preferences, as measured through ideal or intended number of children have been employed by demographers in a variety of ways. In Europe, fertility preferences have invariably been higher than actual fertility, suggesting that under certain policy/economic conditions birth rates could be raised. Such preferences also give a broad impression of general attitudes towards family sizes and help to test whether there are particular societal ‘norms’ (such as a ‘two-child norm’). In this paper, we present the results of a new meta-analysis of fertility preferences in urban and rural China covering the period from the implementation of the one-child policy in 1980 through to 2009. While there are a number of clear limitations to both the review and the constituent surveys, we find indicative evidence of widespread below-replacement level fertility preferences. These concur with other national level surveys. Finally, we consider the extent to which we can ‘trust’ the responses given in surveys. We conclude that if China were to relax its’ ‘One Child Policy’, it is likely that only a relatively small number of people would take advantage of the change and have a second child. For those that do, there appears to be a strong bias towards having one boy and one girl.


Demographic factors in the works of the greatest minds in history
Vadim Bezverbny, Socio-Political Research Institute of Russian Academy of Sciences, Moscow

Research question An entwinement of social, economic and demographic facts has been noted as early as by the greatest minds in history (Confucius, Thomas Aquinas, Niccolo Machiavelli, Adam Smith etc.) as an effective measure of state consolidation and well-being. The first example of the use of demographic politics goes far back to the history of Middle East. Shumerian king Ur-Namme a code of laws more than two thousand years before Christ among which there were laws containing economic measures of promoting family among population. The king of Eshnunna named Bilalama also made a couple of laws in the middle of the twentieth century before Christ that regulated legal norms, property rights and family legislation. In the eighteenth century before Christ a well-known state leader, the king of Babylon Hammurapi combined previous ideas of Eastern state rulers and suggested a set of measures for strengthening marriage morals and regulation of family relations. Methodology The methodology is based on the methods of the content-analysis of historical literature of the scientists of ancient times, antiquity, Middle Ages and Renaissance epochs. Results Thus, the above mentioned historical and contemporary views of scientists on population increase reflect a fundamental value of the demographic factor, both for internal processes of social and economic development, and for strengthening of the geopolitical status of a state. Beside it vast majority of scientists who founded the school of geopolitics paid special attention to the demographic factor.


Burial Practices in Manchester, 1750-1850 
John Black, Cambridge Group for the History of Population and Social Structure

Within the historiography of the population studies of the rapidly expanding cities in England of the late eighteenth and early nineteenth centuries there has be noted a divergence in baptism and burial practices and registration. This paper will use the town of Manchester as a case study to discover the underlying causes of this disparity. The population of Manchester grew from less than 20,000 in the mid-eighteenth century to approximately 250,000 in 1851. Its collegiate church had a monopoly on fees for marriage, baptism and burial registration, leading to a high proportion of Anglican events being registered at the church during the eighteenth century. The Manchester Bills of Mortality were derived only from the parish registers of the collegiate church until the nineteenth century. Disparities between the substantial numbers of baptisms and marriages registered at the collegiate church, and subsequently recorded in the Bills of Mortality, as compared to the much lower numbers of burials registered both in the Bills and the church, caused a number of contemporaries to comment on the extraordinary fertility of the Manchester population in the late eighteenth and early nineteenth centuries. Through analysis of the baptism, marriage, and, most importantly, burial registers of the churches and chapels of all denominations proliferating in Manchester in this period, this paper will delineate burial practices and highlight the causes of the disparities in burial registrations in this rapidly growing city.


Model and National Forecasts for Ectopic Pregnancies in Great Britain and Ireland Patrick Carroll, PAPRI

Abstract of Paper to offer for BSPS2014: Fertility and Reproductive Health: Model and National Forecasts for Ectopic Pregnancies in Great Britain and Ireland Would you like a paper on this topic for BSPS2014? I would be happy to present the results of work done her as outlined below. ABSTRACT Ectopic pregnancies have become more numerous in Great Britain and Ireland so as to number 10,000 cases a year in England & Wales and 1,000 a year, which represents an even higher incidence rate, in the Irish Republic in some recent years. This modern epidemic is the subject of correlational analysis with consideration of risk factors including PIDs/STDs (Pelvic Inflammatory Diseases/Sexually Transmitted Diseases) such as Chlamydia and Gonorrhoea, IUCDs (IntraUterine Contraceptive Devices) and births by Caesarean section. It was found that use of IUCDs was highly correlated with incidence of Ectopic Pregnancies and there was also significant correlation with Chlamydia incidence. A linear regression model has been fitted to national data in England, Scotland and Ireland. This has been used to calculate forecasts that are estimated numbers of Ectopic Pregnancies in future years in these countries. In this model the rate of Ectopic pregnancies per thousand women in child bearing ages is the response variable and rates of Chlamydia and usage of IUCDs are the explanatory variables. This summer some further development work is planned so as to consider developing the model with the addition of Caesarean section births as a further explanatory variable. Patrick Carroll PAPRI(Pension and Population Research Institute) 35 Canonbury Road London N1 2DG, UK Tel +44 (0) 20 7354 5667 Fax +44 (0) 20 7226 6601 Web site


An Introduction to the Scottish Longitudinal Study (SLS)
Susan Carsley, National Records of Scotland

This poster will introduce the SLS and the datasets, the application process for researchers interested in using the SLS and outline research examples.

The Longitudinal Studies Centre – Scotland (LSCS) was established in 2001 and hosts the Scottish Longitudinal Study (SLS). This study links together routinely collected administrative data for a 5.3% representative sample of the Scottish population (about 270,000 people). It currently includes a wealth of information from the censuses starting in 1991, vital events registrations (births, deaths and marriages), Scottish education data, and with appropriate permissions can be linked to NHS health data including cancer registry and hospital admission data.

The size and scope of the SLS make it an unparalleled resource in Scotland for analysing a range of socio-economic, demographic and health questions. Additionally, the longitudinal nature of the SLS is particularly valuable, allowing an exploration of causality in a way that cross-sectional data collected at a single point in time does not. In this way, the SLS can provide insights into the health and social status of the Scottish population and, crucially, how it changes over time.

The 2014 BSPS conference presents a timely opportunity to highlight new data that will be available as a result of the inclusion of the 2011 Census data and will help researchers decide whether the SLS is an appropriate resource for their research.


Ethnicity, health and unemployment in the UK: the mediating effect of unemployment on mental health outcomes
Natasha M Crawford, Institute for Social and Economic Research, University of Essex

Ethnic differences have been identified for a range of health outcomes, with those from ethnic minority groups, on average, experiencing poorer health than the majority white population. While patterns of ethnic inequalities in health vary from one condition to another, on average, Pakistani, Bangladeshi and Caribbean groups have the poorest health in the UK (Nazroo, 1997; Sproston and Nazroo, 2001). Similarly, elevated rates of mental illness have been identified among some ethnic minority groups, with the rate of schizophrenia among the Caribbean group commonly reported at between three and five times that of the white population (Bagley, 1971). Further, ethnic penalties in employment have been documented with most ethnic minority groups consistently faring worse than the white British population in the labour market (Berthoud and Blekesaune, 2006; Bell and Casebourne, 2008; Nazroo and Kapadia, 2013). Despite a well established negative association between unemployment and mental health (Murphy & Athanasou, 1999; Steele, 2013) and an ethnic penalty in unemployment, few studies have considered the potential mediating effect of unemployment when exploring ethnic inequalities in health. Utilising data from waves 1 and 2 of Understanding Society, we identify differential health outcomes according to ethnicity, with those from Bangladeshi and Pakistani groups reporting, on average, poorer mental health compared to those who identify themselves as White British. While we find no mediating effect of unemployment it may be that it is not the occurrence of unemployment but rather unemployment duration that mediates this association.


An assessment of the input approach to estimate household childcare the case of Plymouth, UK
Hoayda Darkal, Plymouth University

This study tests the application of the input replacement cost approach to estimate household childcare and its unreported contribution to GDP in the UK. In addition, it aims to provide an in-depth exploration of childcare arrangements in Plymouth. A multi-method approach was used consisting of primary time-use diaries completion and in-depth interviews conducted in selected areas in Plymouth. This was supplemented with UK’s Time Use Survey 2000 data evaluated at present day values, to improve representativeness of the results. Sampled households are families of married or co-habitant couples with children aged <15 years. Results at the micro-level show that families spend an average of 7-9 hours per day on childcare activities, with longer time spent during weekend days and in households with more children. A multiple-childcare arrangement model is found to be adopted by the studied families influenced by factors such as: child’s interests, parents’ employment, parent’s opinions of paid care, the high cost of institutional care and the availability of grandparents’ help. In addition, the strong relationship between mother’s employment and childcare settings has been confirmed. At the macro level, the estimated contribution of the monetary value of household childcare is found to be 9.1-12.13% of the UK’s GDP. Policy implications include improvements in employment policies enhancing flexible working conditions, longer maternity and paternity leave, and part time jobs; improved recognition of unpaid child carers and the nationwide development of advisory services that reach wider numbers of parents


Repartnering fathers in the UK
Alessandro Di Nallo, London School of Economics

The rise in divorce and separation in the UK over the last forty years, in conjunction with high rates of extramarital births and the gradual replacement of marriage with cohabitation, has resulted in increasing opportunities for repartnering. In this article, I present an in-depth analysis of the characteristics and determinants of new union formation of fathers in the UK – a topic that has been largely overlooked in the literature – using data from two cohorts studies, the NCDS 1958 and the BCS 1970. I specifically focus on the influence of the residence of biological children, the nature of the union when men experienced their first fatherhood , paying attention also to individuals’ background and social variables. I expect to find a negative impact of the number of previous union children (especially if coresidential) on the transition to a new partnership. I also expect a less marked difference in repartnering between men experiencing fatherhood in an established relationship and those becoming fathers outside a relationship, across cohorts. I will innovate the literature on new partnership formation by using multilevel event history models that will allow for studying multiple events of repartnering on an individual’s lifetime.


The Northern Ireland Longitudinal Study
Stefanie Doebler, Gemma Catney, Michael Rosato, Dermot O’Reilly, Queen's University Belfast, Fiona Johnston, NISRA

The poster introduces the Northern Ireland Longitudinal Study (NILS) and explains the structure of the data, its elements and potential for research and possibilities of data linkage. This presentation is useful and informative to researchers interested in using British longitudinal data, and Census-based data for population research, research on ageing, health outcomes, inequalities and migration. The poster also briefly presents a related data source that can be linked to the NILS, the Northern Ireland Mortality Study (NIMS) and highlights possibilities for comparisons with the SLS- and the ONS-longitudinal study.


Is the risk of teenage motherhood influenced by area of residence?
Rachel Doherty, Public Health Agency, Belfast, Michael Rosato, Bamford Centre for Mental Health and Wellbeing, University of Ulster, Londonderry, David M Wright, Dermot O’Reilly, Centre for Public Health, Queen’s University Belfast

The UK has the highest rates of teenage motherhood (TM) in Western Europe and unintended pregnancy is costly for adolescents and society in general. The relationship to individual social and material disadvantage is established but the influence of area of residence is unclear. We tested whether there were additional risks of TM in deprived areas or in cities. The Northern Ireland Longitudinal Study was used to identify a cohort of females enumerated in the 2001 Census who would be aged 15-19 between 2001 and 2009. Risk of TM was measured using multilevel logistic regression, adjusting for established risk factors (health, religion, family structure and SES). Settlement bands and Index of Multiple Deprivation represented an urban/rural gradient and area deprivation respectively. All the individual and household attributes were related to TM in the expected direction, e.g. the risk of TM for women in the most expensive housing was considerably lower than for their peers in the least expensive housing (ORadj 0.23, 95%CI [0.14, 0.40]). Risk of TM was elevated in deprived areas and there was an urban/rural gradient, with higher risks among city dwellers than those in rural and intermediate areas (ORadj 1.24 [1.04, 1.47]). We conclude that teenage motherhood is independently associated with area of residence. The higher risk in deprived areas may be a response to poor employment prospects and the elevated risk in cities may be due to greater opportunity given the higher concentration of teenagers, though other factors such as access to alcohol may also be important.


The main features of leaving the parental home in Russia
Alina Dolgova 1, 2, Ekaterina Mitrofanova 1, 2, 3, 1 National Research University Higher School of Economics (NRU HSE), 2 Research and educational group for Fertility, Family formation and dissolution of NRU HSE, 3 Institute of Demography of NRU HSE

Leaving parents, a start of an independent life are interesting and changeable questions of demographic and sociological research. This study is aimed to follow the dynamics of leaving a parental home among generations in modern Russian society and to identify the main directions of changes that occur during the process of separation from parents’ family. On this basis, we estimate the magnitude of changes in perspective of gender and generations. The preliminary results show that the majority of the respondents (95%) have an experience of separate living from their parents. The share of people, who do not live with parents, is almost twice lower for the youngest generation in comparison to the previous one: 42% versus 81%. Moreover, in the perspective of generations, there is a downward trend of the age of leaving parental family (from 24 years for the cohort of the 1920s to 19 years for the 1980s). The real age of the beginning of an independent life is inconsistent with socially accepted stereotype of this age (18-20 years): the real age is higher than the ideal. There is also a tendency that the later people start their independent life, the more they support the stereotype about the ideal age. Event history analysis results demonstrate that the probability of separation from parents is higher for women and for the youngest generation. Furthermore, it was revealed, that people from village tend to live longer with their parents and start an independent life later.


The relative importance of parents’ and own education for the timing of first birth and the role of intergenerational educational mobility. A cross-national comparison Adriana Duta, University of Southampton

Understanding fertility differentials by parental education and intergenerational educational mobility, beyond individuals’ own education, can bring new insights into the interplay between two important sources of socialization representing ascription and achievement (i.e. parents’ education and respondents’ education). Despite acknowledging that parent’s education and own education are two important socialization sources and key factors for individuals’ timing of first birth, no study investigated whose education matters more, parents’ or respondents’? Especially in the context of increasing absolute intergenerational educational mobility associated with educational expansion, very little is known about the fertility behaviour of those who are mobile and therefore have been exposed to two different types of socialization, according to the different levels of education at origin and destination. Moreover, given that social mobility patterns differ by country, our relationships of interest might differ by mobility regime. Therefore, using discrete time hazard models and diagonal reference models and relying on data from the ‘Harmonised Histories’ this study compares different countries with different social mobility regimes and family policies (i.e. Norway, France, Italy and U.S.) asking: (1) What is the relative importance of parents’ education and own education for the timing of first birth in different countries with different social mobility regimes? (2) Does the relative importance of parents’ and own education vary by the intergenerational mobility status (i.e. if upward, downward or non-mobile), and if so, is the pattern the same in different countries? (3) Is there a general pattern for males and females within the countries and between the countries?


Spatiotemporal variations in Chronic Obstructive Pulmonary Disease mortality in China: a multilevel analysis of the China Mortality Surveillance System
Xiaoqi Feng1, Peng Yin2, Yunning Liu2, Thomas Astell-Burt3,4, Jiangmei Liu2, Andrew Page3, Limin Wang2, Yong Jiang2, Shiwei Liu2, Lijun Wang2, Maigeng Zhou2, 1 Centre for Health Research, School of Medicine, University of Western Sydney, 2 National Center for Chronic and Noncommunicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing, 3 School of Science and Health, University of Western Sydney, 4 School of Geography and Geosciences, University of St Andrews

Background: Although Chronic Obstructive Pulmonary Disease (COPD) mortality is on the decline in China, does this trend occur across all areas of this vast country, or are spatiotemporal variations manifesting? Methods and Findings: Annual COPD mortality counts (years 2006-2012) for 161 counties and districts (Disease Surveillance Points, ‘DSP’) were extracted from the China Mortality Surveillance System and stratified by 5-year age groups (aged>20), gender and time. These counts were linked to annually-adjusted denominator populations estimated from the 2010 census. Multilevel negative binomial regression with random intercepts and slopes were used to investigate spatiotemporal variation. Regional and urban/rural differences were explored via fixed effects, adjusting for age, gender, and DSP-level compositional risk factors. COPD mortality varied over two-fold between DSPs across China (Median Rate Ratio: 2.40). A random slope fitted on the time parameter (incidence rate ratio (IRR): 0.93, 95% Confidence Interval (95%CI): 0.92, 0.94) indicated that the rate of decline was steeper in some DSPs than others. Rates were higher in the west compared with the east (IRR 2.67, 95%CI 2.08, 3.43) and in rural compared with urban (IRR 2.04, 95%CI 1.68, 2.48). Urban and rural differences in COPD mortality narrowed over time. East-west differences were constant. 57% of the geographical variation was accounted for by regional and urban/rural differences, whereas only a further 18% was explained by compositional risk factors. Conclusion: Although the gap between urban and rural China appears to be narrowing, regional inequalities in COPD mortality between east and west remain salient and a priority for decision-makers.


CALLS Hub poster
Allan Findlay, Fiona Cox, University of St Andrews, Chris Dibben, University of Edinburgh, Oliver Duke-Williams, University College London

The Census & Administrative data LongitudinaL Studies Hub (CALLS Hub) has been commissioned by the ESRC to support, promote and harmonise the work of the three LS Research Support Units (CeLSIUS, NILS-RSU, SLS-DSU), with the aim of providing a more streamlined experience for users. The three UK census Longitudinal Studies provide a unique and powerful research resource for a range of academic disciplines. They also form a powerful source of research evidence for policy-makers, practitioners and third sector bodies. CALLS Hub exists to help researchers find the information and resources they need in a straightforward way, and to promote the work and impact of the RSU’s to a wider audience. By bringing together the three studies, we can also highlight the potential benefits and possibilities of using more than one LS, either to allow regional comparisons or to build a national population. The aims of CALLS Hub are:

 To enhance the research potential of the LSs by co-ordinating the development of new resources and methodologies.

 To enhance and streamline the user experience of obtaining information about the LSs and applying to use them for research.

 To increase academic impact by developing communication strategies to raise awareness of the LSs, promoting their outputs and facilitating their impact strategies.

 To increase the economic and societal impact of the LSs by working together with key external stakeholders to develop research projects meeting their evidence needs.

 To facilitate and encourage the use of multiple LSs for UK-wide research.


Long-term effects of migration on population: The Hungarian case Some results from the SEEMIG project
Erzsébet Földházi, Hungarian Demographic Research Institute

Population of Hungary has been decreasing from the beginning of 1980’s as a consequence of natural decrease, but its effects have been being moderated by positive net migration until the latest years. However, the most recent data show negative net migration, and the question emerges about the future volume of net migration and its pattern during long-time period. It is investigated in SEEMIG project – one of the main goals of this project to analyze the migration processes in South-East European countries. Utilizing the fertility and mortality assumptions of medium variant of latest national population projection, a “real migration” model was calculated to assess the most probable migration processes in the future. Furthermore, a “crisis migration” model was calculated assuming new model in migration: the net migration becomes negative similarly to the “real migration” scenario, but it remains negative until the end of the projected period, 2060. In this latter case Hungary becomes a sending country. The medium scenario of population projection shows about 2 millions decrease in population of Hungary until 2060. Comparing to this scenario, the “real migration” variant shows further 600 thousand persons loss, while the “crisis migration” variant gives further 1.2 million loss. Changes in the age structure will be similar for all three variants throughout the projection period: the proportion of young and middle-aged persons will decrease, while the proportion of older persons will increase.


Partnership status and mortality in England and Wales: The effect of living arrangements or health selection?
Sebastian Franke, Hill Kulu, University of Liverpool

Research on health and mortality by marital status shows lower mortality rates and better health for married persons in comparison to single and separated individuals. Those differences, usually stronger for men than for women, persist even when controlling for socio-demographic and economic characteristics of individuals. Changes in England and Wales over the last 40 years -- such as rise in cohabitation, divorce rates, lone parent families, and life expectancy – invite a re-evaluation of these trends by focusing on health and mortality by different living arrangements. Using data from the ONS Longitudinal Study (LS) and applying hazard models, the project investigates health and mortality by partnership and family status in England and Wales between 2001 and 2011and examines the causes of mortality differentials. This poster presents preliminary findings for both men and women with specific focus on people at older ages.


Deciding to Disclose: Pregnancy and Alcohol Misuse
Jonathan Gray, Jakub Bijak, Seth Bullock, University of Southampton

The health journey of a generation begins with its founders. Supporting the reproductive health of parents is a core part of present and future public health as defined by the WHO. Taking alcohol consumption during pregnancy as an exemplar, we apply a combination of agent based modelling and decision theory to explore a scenario where a population of pregnant women choose how far to disclose their drinking patterns to their midwives. This is explored through an agent based model of a set of stylised scenarios, where populations corresponding to pregnant women and midwives play a series of games. In each game, the woman chooses to claim a level of alcohol consumption, and their midwife chooses whether to refer them to a specialist. Both players employ a simple decision rule using information from previous rounds to choose their actions. Midwives attempt to refer only women who need treatment, and women endeavour to avoid being stigmatised for their drinking behaviour while still receiving appropriate treatment. There are two key questions addressed in this work: firstly how far a relatively abstract decision theoretic agent based model is able to capture the dynamics at play in a complicated, and opaque real world situation; secondly, to investigate how information sharing within the two populations affects overall behaviour. The simulation model is able to produce a number of qualitative trends described in the literature, in particular an increased tendency to disclose as women have more encounters, and greater under-reporting of consumption by heavy drinkers.


The Risky Business of Asking For Help
Jonathan Gray, Jakub Bijak, Seth Bullock. University of Southampton

The proportion of the UK who are elderly, and require support is increasing rapidly. Identifying those in need of support, and effectively managing the necessarily limited resources available to provide it is a substantial, and growing challenge. This is compounded by the social stigma attached to age, and the need for assistance, which together with a lack of belief that help will be forthcoming can make those most in need of support reluctant to request it. This poster will outline our present work on developing an agent based model grounded in theories of decision making to explore the help-seeking behaviour of elderly people requiring support with activities of daily living. The decision to request help is construed as a balance between severity of need, belief that support will actually be available, and the desire to maintain an individual's self image as a capable person. Paired with this is the challenging task of making the most effective use of available resources to deliver care to those most in need. This will allow the exploration of the process of help-seeking by and identification of, those in need. with the intention of validating model outputs against longitudinal survey data (e.g. the English Longitudinal Survey of Ageing). In future, the simulation will support investigation of the implications of policy interventions, for example public campaigns to reduce the stigma associated with age.


Quantifying uncertainty in agent-based models: From emulation to calibration
Jason Hilton, Jakub Bijak, University of Southampton

This paper shows how statistical emulators can be used to calibrate agent-based models (ABMs) to fit demographic data, in order to properly quantify the uncertainty of the resulting predictions. ABMs generally include several free parameters, which are often unobserved. This is principally the case if the model in question purports to represent some aspect of human decision making or interaction. It is assumed that such parameters have a ‘true’ value in reality, which we wish to learn about through matching model outputs to empirical data. However, the dimensionality of the space combined with the computational expense of the model can make identifying the best-fitting values of the parameters difficult. This paper utilises statistical emulators to simplify this process. More specifically, we calibrate age-specific hazard rates in the Wedding Ring model of partnership formation (Billari et al. 2007) against the UK data through Gaussian process emulation. A comparison of a number of different approaches to the calibration task is undertaken, each using an emulator to stand in for the more computationally expensive simulation. Particularly, Bayesian approach involving Gaussian model discrepancy function, which generates posterior distributions for the calibration parameters, is tested. Various optimisation approaches and sequential experimental designs are also examined. Finally, efficiency in terms of number of simulation runs required and magnitude of error is used to assess the performance of candidate methods. The ultimate aim of this work is to propose a methodology for quantifying uncertainty in demographic forecasting utilising ABMs as well as empirical data.


Scottish Small Area Population Estimates: 2002-2010 Rebased Estimates using 2011 Census
Michael Hunter, National Records of Scotland

This poster covers the methodology and some of the key results of the rebased Scottish population estimates published in May 2014, by the National Records of Scotland. These estimates are the revised 2002-2010 estimates using population estimates published by the 2011 Scottish Census. This poster concentrates on the population estimates published at data zone level. Data zones cover the whole of Scotland and nest within local authority boundaries. Data zones are groups of 2001 Census output areas which originally had populations of between 500 and 1,000 household residents, and some effort was made to respect physical boundaries. In addition, they have compact shape and contain households with similar social characteristics. The National Records of Scotland’s annual population estimates, rolled forward from the 2001 Census, were around 49,000 fewer in 2011 compared to the 2011 Census. The poster focuses on how we apportioned out these difference over the time from 2001 to 2011 and across Scotland. Although mainly concentrating on the methodology and results, I will also cover a few issues that came to light during the exercise along with the solutions.


Community Socio-economic Status is Driving HIV Infections in Uganda: Evidence from Multilevel Modelling of Population-based Surveys.
Patrick Igulot, City University London

Background Socio-economic status has been documented to influence the risk of HIV infection in Africa. However, available empirical evidence is limited to the influence of individual level social economic status on vulnerability to HIV infection. This research examines the relationship between community wealth status and educational attainment and vulnerability to HIV infection in Uganda. Methods Multilevel binary logistic regression is used to analyse 39766 individual cases with HIV test results and 887 clusters obtained from Uganda HIV/AIDS Indicators Survey conducted in 2004-05 and 2011. Results With the exception of urban areas, an increase in the proportion of wealthy people in a community is significantly associated with an increase in the prevalence of HIV among women, men, and rural residents. An increase in the proportion of people with higher educational attainment in a community is significantly associated with a decrease in the prevalence of HIV among women, men, rural, and urban residents. In the analysis of trends, similar observations were noted both in 2004-05 and 2011. Conclusion Living in a community with higher wealth status increases vulnerability to HIV infection for women, men, and rural residents. However, living in a wealthy urban area decreases vulnerability to HIV infection. Living in a community with higher educational attainment decreases the vulnerability of women, men, and residents of rural and urban areas. To prevent HIV infections, policies and strategies need to ensure access to quality higher education, and address gender-based and regional inequalities in wealth.


Childbearing of Polish migrants in the UK: trends, challenges and implications for policy-making
Barbara Janta, University of Warwick

Since 2004 there has been a substantial increase in births to Polish mothers in the UK, with nearly 3% of all children born in the UK in 2012 having a Polish-born mother. Research shows that majority of these children are born to two Polish parents (Janta 2013) and that Polish parents in the UK have on average more children than parents in Poland. With the Total Fertility Rate of Polish women being 2.13 in England and Wales in 2011 (Dormon 2014) compared with 1.3-1.4 in Poland over recent years (GUS data for various years), this raises questions about factors that influence fertility decisions of Polish parents. Childbearing decisions and experiences, in turn, have an impact on the permanence of migrants' settlement in the UK. The aim of this paper is to provide a quantitative overview of childbearing trends among Polish mothers in the UK since 2004. We analyse fertility trends of Polish migrants through an analysis of birth registration data and an online survey of Polish parents, and draw a profile of Polish families in the UK. We discuss whether living in the UK changes the fertility plans of Polish migrants. Analysing a broad range of socio-economic factors, we investigate which factors play the most important role in decisions to start a family for Polish migrants. Finally, we discuss the implications of Polish migrants' reproductive behaviour, considering how the experience of starting a family influences migrants' decision to settle in the UK or to return to Poland.


Employment after first birth among immigrant women in Belgium
Tine Kil, Karel Neels, University of Antwerp

Belgium belongs to the group of North - Western European countries characterized by a high level of female employment in combination with family formation. This position is largely due to family policy that supports people in the combination of work and family, by example through childcare and parental leave. However, research for Belgium indicates that there is a strong socio - economic gradient in the decline of labour market participation after parenting and the use of childcare. The ethnic gradient in this decline has not been studied yet, however patterns of family formation and female labour force participation appear to be highly differentiated by migration background. Therefore in this research we analyse how female labour force participation (full-time, part-time or not working) in Belgium varies according to family composition and the extent to which these women make use of parental leave. With data from a Belgian Administrative Socio-Demographic Panel (1999 - 2010) we compare four major groups of migrants (Southern Europeans, Eastern Europeans, Turks and Moroccans) to Belgian women. We will perform Event History Analysis to link the birth of a child to changes in labour market position of the mother. Preliminary results show that there is a relatively stronger decline in labour force participation after the birth of a child for women with a migrant background. Especially for women from a Turkish or Moroccan origin the difference is large. East-European, Turkish and especially Moroccan women furthermore make less use of parental leave.


Young People and the Great Recession in the UK: Impacts on achievement related attitudes and behaviours
Mark Lyons-Amos, Ingrid Schoon, Institute of Education, University of London

The teenage years are a critical formative period for later achievements where young people develop aspirations and motivations towards the future and take first steps in pursuing a career path. Against the backdrop of the current economic crisis this paper examines how the societal context affects youth’s responses to the economic crisis. We adopt a developmental-contextual perspective, assessing the role of multiple interlinked influences. Do local labour market conditions play a role in determining educational aspirations and choice over and above parental socio-economic resources? Previous research conducted between 1980 and 2005 suggests that local unemployment rates have a positive association with enrolment in post-compulsory education, although the evidence is not univocal. In this study we add to the literature by investigating the impact of the recession on educational aspirations as well as subsequent behaviour using successive cohorts who were aged 11 between the period 1991 and 2012. We draw on data collected for the British Household Panel Survey (BHPS) and Understanding Society, two nationally representative household based surveys which followed individuals in yearly waves since 1991, including a British Youth Panel of 11-15 year olds. In our analysis we focus on younger age cohorts born between 1980-84, 1985-1989, and post 1990 to take into account age, period and cohort effects and enabling us to separate effects from the current economic downturn from previous trends. Each of the age cohorts comprises about 1000 individuals.


Census 2011 in Scotland – Disseminating findings and promoting uses of census data
Cecilia Macintyre, Tom Wallace, National Records of Scotland

The poster will highlight some of the activity which has been carried out to promote uses of the data from Scotland’s Census and our plans for the future. This includes our ‘Census Data Explorer’ - National Records of Scotland's key online resource for making the data from Scotland's Census available to users.


Using sequence analysis to investigate call record data in longitudinaal context Olga Maslovskaya, Gabriele Durrant, Peter W.F. Smith, Southampton Statistical Sciences Research Institute, University of Southampton

For interviewer administered surveys many survey agencies nowadays routinely collect call record data. Examples of such data may be recordings of the day, time and outcome of each call attempt. Researchers have increasingly become interested in how best to use and analyse such data which can be large and may exhibit complex hierarchical and time-dependent data structures. It is hoped that a better understanding of the calling patterns and the mechanisms leading to particular call sequences will help improve data collection through improved interviewer calling practices. It might help identify more difficult cases and unusual interviewer behaviour earlier on in the data collection procedure and may provide strategies for improved nonresponse adjustment methods. Although survey researchers have become increasingly interested in understanding and improving the process of data collection, to date analysis of whole call histories is still limited. This paper extends sequence analysis as a tool for investigating call record data in longitudinal context to better understand and improve survey processes. Sequence analysis offers an elegant way of visualising, displaying and summarising the normally quite complex call record data. Here, the method is used to inform survey management for adaptive and responsive survey designs. Sequence analysis is combined here with clustering, optimal matching and multidimensional scaling. The sequence analysis method is applied to call record data from the UK Understanding Society survey. We use the first three waves of the survey for the analysis. Implications of the findings for survey practice are discussed in the paper.


Making sense of families and households in Botswana over time
Oleosi Ntshebe, Andrew Channon, Victoria Hosegood, University of Southampton

This study provides an overview of household membership, characteristics and their changing character in Botswana between 1988 and 2007. In addition, the demographic processes that have influenced household composition over time are examined. The contexts that have shaped households over time include living arrangements with presence and absence of children, education, marriage, and employment status. Data analysed comes from nationally representative surveys: the 1988 Botswana Standard Demographic Health Survey, the 2000 Multiple Indicator Survey (MICS) and the 2007 Botswana Family Health Surveys (BFHS). This analysis is important for understanding the change in household composition over time, the contexts that have shaped these households, and the implications of such household changes. Further, the analysis affords information on Tswana families and households that is generally not available in census reports and other population survey analyses.


Anxiety and depression among medical students: a nationwide survey in China Xiong-Fei Pan, Ying Wen, Yun Zhao, Si-Qi Li, Hong Chang, Qing-Ping Xue, Zhi-Mei Zhao, Chun-Xia Yang, Department of Epidemiology and Biostatistics, West China School of Public Health, Sichuan University, Sichuan, China

Objectives: The study aimed to explore the prevalence of depression and anxiety and their correlates among medical students in China. Methods: We conducted a nationwide survey among medical students based on a questionnaire including demographic information, life events in the last four weeks before survey, and validated Chinese versions of the revised Life Orientation Test (LOT-R), 21-item Beck's Depression Inventory (BDI), and Zung's Self-rating Anxiety Scale (SAS). Results: We surveyed 9011 medical students with a mean age of 20.7 (standard deviation: 1.6) years from 33 universities. 38.3% were males, 42.4% majored in clinical medicine, and 41.4% came from the urban areas. Screening psychometric scales indicated that 19.9% and 14.5%, respectively, had depression or anxiety. Male students were more likely to have depression (P<0.001), and those majoring in clinical medicine showed higher possibility of anxiety (P<0.001). The year of study was negatively correlated with both depression and anxiety (P=0.032 and 0.005). Location of household (rural or urban), parents' education background, or household income per capita was not associated with either. Higher pessimism rendered less likelihood of both mental problems (P<0.001). Life events such as worries for current study and future work, distress due to peer competition, financial difficulties, interpersonal relationship crisis, smoking, drinking, and lack of sleep were correlated with higher risk of both depression and anxiety, while physical exercises, with lower risk (all P<0.001). Conclusions: Depression and anxiety appear prevalent among Chinese medical students. The study indicates it is important to pilot targeted counseling services to improve their mental health.


An analysis of 2011 Census results on alternative population bases
Claire Pereira, ONS

New and existing questions asked in the 2011 Census for England and Wales have allowed outputs to be produced on different population bases other than the standard definition of usual residence that is used to count the population. These are the workday, workplace, short-term resident and out-of-term population bases. This presentation will discuss the development and derivation of these population bases, present the range of outputs that have been published, and provide an analysis of these outputs to describe the main characteristics of the different populations that are counted by each of the alternative bases.


Is type of grave a class indicator? An analysis of the burial records of Coimbra’s municipal cemetery, 1885-1910
Mafalda Moura Pereira, University of Cambridge

This paper turns to a novel source, cemetery records, to reconstruct class structure, basing inferences regarding the deceased’s class identities on their ‘type of grave’. The civil burial records of turn-of-the-century Portugal are fuller than other archives of this kind in stating the deceased’s name, civil status, occupation, age, cause-of-death, place of birth and grave type. My paper observes the breakdown between types of grave for a cohort of adults (aged 9-98), reflecting on how far this may correspond to social stratification. Cemetery records’ comprehensiveness means they are an apt indicator of social structure. Further, the study is interested in whether type-of-grave better corresponds to the deceased’s socio-professional or occupational status or to genuine class differentials. How nearly equivalent, for instance, are men’s and women’s representation in elite mausoleums? In making class categorisations, the study uses some 7,000 burials, stripping out non-residents and individuals beyond its age end-points from city deaths. Women only listed as doing some form of housework especially interest the study insofar as, at individual level, type-of-grave serves as a class proxy and facilitates a depiction of women’s distribution across classes. Results show remarkably consistent proportions in the take-up of grave-types. Some 11% of both men and women go into mausoleums and 61% into graves-with-headstones. The grave category showing the most marked gender imbalance is the common grave (housing 24% of the female against 20% of the male population). The inference is that more women than men came to Coimbra seeking work and died in poverty.


Is dietary and alcohol consumption associated with cognitive disorders in Central Africa? A study from the EPIDEMCA program
Sophie Pilleron, , Maëlenn Guerchet INSERM UMR1094, Tropical Neuroepidemiology, Limoges, France, Jean-Claude Desport, Pierre Jésus, Department of Nutrition, University Hospital of Limoges, France, Pascal Mbelesso, Department of Neurology, Amitié Hospital, Bangui, Central African Republic, Bébène Ndamba-Bandzouzi, Department of Neurology, University Hospital of Brazzaville, Republic of Congo, Jean-François Dartigues, INSERM U897, Bordeaux University, France, Jean-Pierre Clément, Hospital and University Federation of Adult and Geriatric Psychiatry, Limoges, France, Pierre-Marie Preux, University Hospital of Limoges, Centre of Epidemiology, Biostatistic, and Research Methodology, CEBIMER, Limoges, France

Background: To date, no study concerned the relationship between dietary consumption and cognitive disorders in Africa where the greatest increase of the number of cases of dementia is expected in next decades.

Objective: To investigate the association between dietary and alcohol intake and cognitive disorders in elderly in Central Africa.

Design: A cross-sectional multicenter population-based study using a two-phase design was carried out on 2001 subjects aged ≥65 from urban and rural areas in Central African Republic (CAR) and Republic of Congo (ROC). Elderly with low performance to the Community Screening Interview for Dementia were then clinically assessed by a neurologist and underwent further psychometrical tests. DSM-IV and Petersen’s criteria were required for dementia and mild cognitive impairment (MCI) diagnoses, respectively. A food frequency questionnaire assessed the intakes of dairy products, fruits, vegetables, starches, legumes, oleaginous foods, meats or fishes, eggs and sweet foods during the past 3 days. Besides, we collected data on alcohol intake. Sociodemographic, vascular, and psychological factors were also collected.

Results: In adjusted multinomial logistic regression models, a lower consumption of oleaginous foods was associated with MCI (OR=3.7 [1.4-9.9]) and dementia (OR=2.8 [1.0-7.7]) in rural area of CAR. Alcohol consumption was associated with reduced probability of dementia in CAR (OR=0.3 [0.1-0.8]). In ROC, no association was observed.

Conclusion: Our study provides new data about association between diet including alcohol consumption and cognitive disorders in Africa. Further studies should investigate this relationship by analyzing at foods level and no more at food groups level.


Timing of fertility among migrants to England and Wales, 2001-2011
James Robards, EPSRC Care Life Cycle, Social Sciences, University of Southampton, ESRC Centre for Population Change, Social Sciences, University of Southampton

The fertility of migrants to England and Wales has been of interest to demographers because of the possible contribution of this group to the increase in the Total Fertility Rate (TFR) from 1.6 in 2001 to 1.9 in 2011 (ONS, 2013; Robards, 2013). Migration and fertility have been considered as interrelated events and their interrelation of increasing importance in Europe (Sobotka, 2008). In other European contexts an elevated level of fertility has been observed among migrants to the host country (Toulemon, 2004; Andersson, 2004). There are three main ways in which migrants can impact on fertility in the destination country (Robards, 2013). In the UK research on the timing of fertility among migrants has been limited because of a lack of accurate data on the date of migration (Robards, 2012; Robards et al., 2013). The 2011 Census has provided a suitable denominator, enabling the Office for National Statistics to report the TFR for women in the UK by country of birth (Dormon, 2014). At the 2011 Census non-UK born respondents were asked the month and year of first residence in the UK, information not previously available for recent migrants (Robards et al., 2013). This research uses the new 2011 Census data in the Office for National Statistics (ONS) Longitudinal Study (LS) to present results on the fertility of migrants before and after migration to England and Wales (2001-2011). In so doing it identifies fertility in relation to the migration event for the first time using 2011 Census data.


Life course and life events in statistics: The ONS Longitudinal Study
Nicola Rogers, James Warren, Kevin Lynch, Office for National Statistics

The ONS Longitudinal Study (LS) contains linked census and life event data for 1 per cent of the population of England and Wales. 2011 Census data have now been added to the study meaning that the LS now holds information on intention to stay, passports held, visitors, second addresses, main language and civil partnerships for the first time. It also means that for the second consecutive decade the LS will have information on general health, caring and religion, as well as linked data from five successive censuses that will support studies on the life course. This poster will give an overview of the LS and key variables it contains and its potential use in furthering understanding of the life course. Results will be presented from an exemplar research project that follows a cohort aged 15-24 in 1971 and examines their outcomes over the next four decades. The 2014 BSPS conference presents a timely opportunity to highlight new data that are now available as the result of inclusion of 2011 Census data and will help researchers decide whether the LS is appropriate for their research.


Parents and Grandparents: How much does age-distance between generations matter for fertility?
Roberta Rutigliano, Universitat Pompeu Fabra

Since women have acquired a new role in society, they have to struggle to reconcile career with family life. The general aim of this paper is to explore the dynamics within the extended family, focusing on the role of grandparents in different contexts. I investigate the relationship of the age-distance between grandparents and the adult child at the moment of (first) birth on the probability of having the second child. The key idea is to consider the age distance as a proxy for expectations about the future childcare supply by grandparents. Further, this measure provides a quick indicator for the trade off between postponement and aging. Although individuals are living longer and, to some extents, better, they can be too old to look after grandchildren especially when the middle generation is made by “great postponers”. I expect to find a non linear relationship between age-distance and transition to second child. I expect also a cohort effect due to the evolution of wellbeing overtime. I use data from SHARE (Survey of Health Aging and Retirement in Europe). In particular, I will use wave1,2,4 and some information from the 3rd wave. Methodologically, I will combine survival analysis, namely duration models, with a Propensity Score Matching approach.


Healthy Immigrant Effect in the UK across ethnic groups
Dorothee Schneider, Institute for Social and Economic Research, University of Essex

The Healthy Immigrant Effect (HIE), whereby recent immigrants have better health than the native-born population, is well established, e.g. in Canada, as is a negative duration effect with increasing length of residence. Despite a sizeable immigrant population, these effects are not well researched for the UK. Research question This study examines whether there is a HIE and duration effect in the UK. Because immigrant and ethnic minority populations intersect, and health levels vary across ethnic groups, this is investigated for the overall population and within ethnic groups. Methods The study uses the first wave of the UK Household Longitudinal Study, including an ethnic minority boost sample (n= 33,466, ages 21-60). Two dichotomous outcomes are used: poor self-rated health (pSRH) and diagnosed chronic condition. Survey-weighted logit models estimate the effect of immigrant status and years since migration (YSM) on probability of poor health. Models control for age and socio-economic status and include gender interactions. Results There is a large HIE overall for pSRH (OR 0.22; 95% CI 0.12-0.37) and chronic condition (0.32; 0.20-0.47). Stratified models show similar results among White and Pakistani/Bangladeshi, smaller HIE among Caribbean and larger amongst African and Indian. Increasing YSM decreases immigrants’ advantage until their predicted probability of poor health converges at 20-25 years with that of UK-born. Conclusion Evidence for HIE and duration effect is found for immigrants in the UK. The HIE varies across ethnicities, suggesting that immigrant status should also be taken into account as potential confounder in ethnic health inequality analysis.


Comparing the abortion rates among young people in Britain and France - what is the role of deprivation?
Rachel Scott, Kaye Wellings, Emma Slaymaker, London School of Hygiene and Tropical Medicine, Nathalie Bajos, Inserm (France)

Abortion rates among young people are higher in Britain than they are in France. Young people in Britain are more like to become pregnant than those in France, and less likely to have an abortion if they do. The association between deprivation and conception and abortion rates in Britain is well established; conception and abortion rates are higher in more deprived areas. Among those who do become pregnant, those who live in more deprived areas are less likely to have an abortion. However, abortion rates still tend to be higher in more deprived areas because of the higher conception rates. Less is known on the association between young people’s conception and abortion rates and area level deprivation in France. This study aims to examine and describe the rates and trends in conception and abortion among young people over the last 30 years in Britain and France, and to explain the differences observed between the two countries. It will focus on the role of socioeconomic disadvantage, using residence as a proxy indicator of area level deprivation, and whether its effects are different in Britain compared to in France. The research will use routinely collected data on abortions and conceptions, including individual level data on abortions, provided by the Department of Health and the Office for National Statistics. It is expected that conceptions and abortions will be less socially stratified in France than in Great Britain, and that this may go some way to explaining the variation in rates between the two countries.


Using patient records to investigate mortality and deprivation in the UK
David Smith, Les Mayhew, Ben Rickayzen, Cass Business School

Using patient records to investigate mortality and deprivation It is generally assumed that deprivation has a major impact on mortality rates with mortality rates increasing as the level of deprivation increase. While this has been shown to hold, calculations are often carried out on historical, aggregated data. For this paper we have investigated the impact on mortality of deprivation by using patient data provided by the THIN (The Health Improvement Network) database of patient records from 2005 to 2010. The data base contains around 4 million anonymised patient records including their patient histories. The data are being used to investigate chronic disease pathways over the life course and the consequent burden on the health service. We will begin with a description of the structure of the data base and how it can be used in the specific investigation of mortality. By splitting up the population into deprivation quintiles based on location, we demonstrate that mortality rates support the hypothesis that they increase as the level of deprivation increases. While this general relationship holds for all ages we find that the differences as expressed as relative percentages varies considerable over the age range, with the difference in males peaking around age 45 and females showing a different pattern with two peaks at ages 42 and 65. We also find that the differences reduce at older ages with mortality rates converging when patients reach age 80. We suggest reasons for why these patterns have emerged from this population and hence possible changes in the future.


The Social Gradients of Autism Spectrum Disorders in England
Alex Upfill-Brown, Tak Wing Chan, University of Oxford

In this paper, we use administrative data to explore the social gradients of Autism Spectum Disorders (ASD) in England. We report some demographic correlates with ASD that are consistent with previous research. But we also report results that are quite different from those based on Californian data. In particular, we show that in England pupils from lower income households are more likely to be diagnosed with ASD. This raises some questions about how the health delivery and school systems of different countries might produce different social gradients of ASD, an hypothesis which we will explore with multilevel models in the full version of this paper.


The association between education and induced abortion in Finland: a cohort perspective
Heini Väisänen, Department of Social Policy, London School of Economics and Political Science

Low education and high opportunity costs have been shown to be associated with higher risk of abortion, but studies often suffer from underreporting of abortions, and lack longitudinal data. UN has called upon reducing induced abortions by providing universal access to family planning, but whether that alone reduces socioeconomic differences in abortion behaviour remains unknown. This study explores whether the association between level of education (basic, upper secondary, further, undergraduate or postgraduate) and abortion changed over time in Finland, where comprehensive family planning policies and sex education have taken place since 1970. A unique longitudinal set of Finnish register data of women aged 20 or more born in 1955-59, 1965-69 and 1975-79 was analysed using concentration curves and discrete-time event-history models. This topic has not been studied longitudinally with comprehensive and reliable data before. Women with basic education had a higher likelihood of abortion than other women and the association grew stronger for later cohorts despite universal access to family planning and chancing educational composition of the population. For instance, the upper secondary group aged 20-24 had 17%, 40% and 55% lower risk of abortion than the basic education group in the 1950s, 1960s and 1970s cohorts. Selection into education may explain the pattern: although it was still fairly common to have only completed basic education in the 1955-59 cohort, it became increasingly unusual in the later cohorts. Thus, even if universal access to family planning is available targeted policies are needed to reduce socioeconomic differentials in abortion.


Universal Health Coverage in the Context of Population ageing: Patterns of Health Care Utilization among Elderly in Ghana
Nele van der Wielen, University of Southampton

Universal health coverage is a concept that is currently widely discussed. The aim of universal health coverage is to ensure that everyone in a population has equal access to health care services without fearing financial hardship. Increasing research has been undertaken in the field of health care utilisation. However, in many low and middle income countries age-related effects on health care utilisation are under-researched. Low and middle income countries have experienced rapid population ageing, raising the question of what factors determine health care utilisation among the elderly. For this research the Study on Global Ageing and Adult Health (SAGE) is used to investigate the patterns of health care utilisation (i.e. hospital use, medication use) among older adults, using Ghana as a case study. Binary logistic regression is applied to analyse individual-level correlates of healthcare utilisation among elderly, while multinomial logistic regression is utilised to examine the different kinds of health care facilities used in more detail. Expected results suggest that health care costs (i.e. total health care costs, out-of-pocket costs) and a wide range of socio-demographic and cultural factors are related to how individuals actually use the health care services available to them. Understanding factors associated with barriers to universal access will inform the design of policy.


The working patterns of the older health and social care labour force in England Alison Wadey, Centre for Research on Ageing, University of Southampton

In England, the concept of retirement has developed over the last 50 years to become a cultural expectation. However, in light of increasing longevity and with the implementation of the Extending Working Lives agenda, older workers are now being encouraged to remain in work for longer. Alternate patterns of work, such as part-time and flexible working opportunities, may facilitate working longer in later life. Additionally, the element of choice is evident as a strong influence that affects decision–making around State Pension Age. This research focuses on the health and social care labour force in England aged 50 years and over and uses secondary analysis of the Labour Force Survey to investigate the characteristics of selected occupations within this labour force. This presentation reports the employment patterns of health and care professionals and health and care support staff in two age groups: 50 to 59 and 60 years and over. Preliminary results suggest two key points. Firstly, that the majority of health and social care workers aged 60 years and over work on a part-time basis, and secondly, that those who work part-time, do so out of choice. Understanding the working patterns of the older health and social care labour force facilitates the development of relevant workforce planning legislation.


Eurostat products from the 2011 Census
Paul Waruszynski, Office for National Statistics

The European Union (EU) requires that the Office for National Statistics (ONS) supplies data from the 2011 Census to Eurostat, the statistical office of the EU, with the aim of creating consistent and transparent data from across the EU. The data supplied by the ONS includes census data for the whole of the UK and is aggregated in a series of pre-defined table layouts, or ‘hypercubes’. These tables have been the largest that ONS has produced from the 2011 Census. They also include a range of variables, some of which are new to the 2011 Census and others are not generated as standard within the UK. This presentation gives an in-depth look at the production of census statistics for Eurostat. Incorporating a focus on how the 2011 Census has bridged the differences between Eurostat specifications, and what data was collected in the 2011 Census. Particular attention will be on the challenges of harmonising data from across the UK as well as generating data on topics not routinely collected in the UK, for example Country of Citizenship. We will also demonstrate how a new data explorer will help to increase value to users by allowing the comparison of data across the EU.


Progress and developments of the Digitising Scotland (DS) Project
Lee Williamson1, Chris Dibben1,2, 1 University of Edinburgh, 2 St Andrews

The Longitudinal Study Centre – Scotland received funding to create a multidisciplinary research database from historical vital events records. The Digitising Scotland (DS) project will digitise the 24 million Scottish vital events record images (births, marriages and deaths) from 1855. This will allow research access to individual-level information on some 18 million individuals - a large proportion of those who have lived in Scotland since 1855. At the moment these records are indexed images accessible from Scotland’s People (website or the centre search rooms), however to undertake research means manually transcribing all the information needed (eg cause of death, occupation, etc). This has made any large-scale research project impossible – a situation that DS will change. The poster introduces DS and the progress on the work involved. For example work package 1 on digitising the birth, marriages and death vital events records from 1855-1973. Work packages 2 and 3 are the on automation of standardising and classifying text: for job descriptions to the Historical International Standard Classification of Occupations (HISCO) and for deaths to the International Classification of Disease -10 (ICD-10). While the full transcription of vital events from work package 1 will not be complete by BSPS we will still be able report on the quality of the transcription. The poster will also give an update on the progress on the other work packages, given pilot work is already underway to automate the process of standardising and classifying in conjunction with Computer Scientists utilising existing digitised individual records - with interesting results.


Family formation and female employment over the life course in France, the Netherlands, and Hungary
Jonas Wood, University of Antwerp

The relation between family formation and female employment has been high on the demographic research agenda for decades. Research on the impact of motherhood on labour market participation, typically finds reduced activity when young children are present in the household, but also that this negative effect weakens as the age of the child increases, suggesting a return to the labour market. Relying on cross-sectional data, many contributions are subject to a life-course fallacy and fail to control for socio-economic position before parenthood which has been shown to affect maternal employment patterns. This paper investigates changes in female employment related to family formation using longitudinal microdata for France the Netherlands, and Hungary between 1970 and the mid-2000s. The first and foremost conclusion of this paper is that before becoming a mother great differences in employment are found between educational groups - with low educated women being characterized by more non-employment - whereas the change in employment after a first child indicates few differences between educational groups. Hence the preconditions to motherhood vary strongly between educational groups, and these differences explain a great part of the observed educational differences in maternal employment patterns. The transition to a second child is marked by educational differentials, especially concerning part-time work. A second major conclusion is that the differences between these three distinctly different countries studied in this paper indicate that contextual factors determine the strength of the negative relation between family formation and female employment.


What becomes of the youths Not in Education, Employment or training (NEETs) in the E&W Longitudinal Study (LS)
Wei Xun, Chris Marshall, Rebecca Lacy, Nicola Shelton, Department of Epidemiology and Public Health, University College London

The recent recession has highlighted the substantial scale of youth unemployment in the UK and worldwide. Although it is assumed that for most young people, being a NEET is a temporary hiatus and sometimes a necessary step in the transition from adolescence into adulthood and subsequent financial independence, for some the disadvantages of economic inertia at this stage in the life course can have long-term effects on their socio-economic outlook later on. The aim of this paper is to investigate if being a NEET in youth can affect their future economic participation; hence, to what extent does a “shadow” of NEETs exist in socio-economic terms in the UK. Using the data from the England and Wales Longitudinal Study (LS), we will identify NEETs from a representative sample of 1% of the E&W census population aged 16-29 years, and follow them for 10 and then 20 years to examine the differences in their destination employment status, occupation and social class compared with their non-NEET counterparts. Due to the changes in the economic conditions, labour market and other relevant factors within the last few decades, three cohorts will be drawn from 1971, 1981 and 1991 for comparison purposes.