Home > Department of Social Policy > British Society for Population Studies > Annual conference > 2012 - Nottingham > Abstracts - methods including event history & qualitative & mixed methods research sessions


Abstracts - methods including event history & qualitative & mixed methods research sessions

Strand organiser: Dr. Paul Norman, University of Leeds

The Dutch Regional Demographic Forecast with the Model PEARL
AH De Jong, PBL The Netherlands Environmental Assessment Agency

In 2004 the PBL Netherlands Environmental Assessment Agency (PBL) and Statistics Netherlands (SN) started a cooperation, with the intention to produce a regional forecast on population and households. In 2006 the PBL and SN have published their first regional demographic forecast, followed up by three updates in 2008, 2009 and 2011 respectively. The principal output of the regional forecast consists of the population and households for each municipality of the Netherlands, distinguished by a number of background characteristics. In addition, also numbers on several demographic events and household transitions are produced. The paper deals with the crucial decisions which had to be taken in the development of a regional forecasting model, the main algorithms of the model PEARL, the assumptions/input for the model, and the outcomes of the regional forecast of 2011.

Email: Dr. Andries de Jong: andries.dejong@pbl.nl|

Rethinking households: Using administrative data to count and classify households and to assess need and service planning
Gill Harper, Mayhew Harper Associates; Les Mayhew, Cass Business School & Mayhew Harper Associates

Households rather than individuals are being increasingly used for research and to target and evaluate social and economic policy in fields ranging from housing, social investment, utility consumption, health inequalities, and the design and delivery of services in education, health or social services. As a result accurate and timely household level statistics have become an increasing necessity across a range of uses in both the private and public sectors especially at local level. However, present sources of information on households are fragmented with significant gaps and inaccuracies that limit their usefulness. In this paper, we critique present statistical arrangements, pointing to the poor quality of much of the underpinning population data especially in areas of high population churn. We then describe a new approach to data collection and household classification based on local administrative sources. The result is a more integrated system that can not only re-create the existing official DCLG household typology but also enrich it, as well as provide much greater flexibility in terms of geography. The utility and advantages are demonstrated using examples based on work for the six Olympic London Boroughs during 2011. This work includes a comprehensive enumeration of households broken down into a new household typology which is also described in the paper. The result is a data set with a much greater range of household variables that can be linked to other administrative data sources as required as is demonstrated by way of examples.

Email: Gill Harper: harpergill@gmail.com|

Estimating Comparable Global Foreign Born Stock Tables using the EM Algorithm
Guy J. Abel, Wittgenstein Centre (IIASA, VID/ÖAW, WU) & Vienna Institute of Demography/Austrian Academy of Sciences

The World Bank recently released bilateral tables of global migrant stock tables from the 1960, up to the 2000, round of censuses (Özden et al., 2011). As with most collections of multinational migration data, the comparability of results is hindered by multiple problems. These include timing issues (from censuses in each round recorded on different dates), measurement issues (migrants can be defined from place of birth, nationality or ethnicity characteristics depending on census format) or missing values (where questions on the aforementioned characteristics are not included in a countries census or there was no census). In this paper, these issues are addressed using a model based imputations which have long been recommended for missing migration data (Willekens, 1994). Using the EM algorithm, parameters in a predicative model, which includes both demographic and geographic covariates, are estimated. The EM algorithm is applied, and results evaluated, twice. Firstly, data which are originally missing are imputed. These are compared to equivalent World Bank estimates, derived using non-model based methods. Secondly, data for which no foreign born stock measure is available are imputed. This is compared to estimates of stocks from nationality or ethnicity characteristics. The results represent a complete and comparable set of global bilateral foreign born migrant stock tables.

Email: Dr. Guy Abel: guy.abel@oeaw.ac.at|

Event history (survival analysis) session

Session organisers: Valeria Cetorelli, Ben Wilson, London School of Economics

This methods session will include four short presentations and a detailed discussion. Our aim is to share knowledge and experience of event history (a.k.a. survival analysis). Presentations will therefore be short (no more than 10 minutes), and the second half of the session will be reserved for debate, questions and discussion. Presenters will focus on how (and why) they have used event history, and to reflect on lessons learnt. The presentations will then be used to structure a chaired discussion about the merits of event history techniques. It will also be a chance to meet fellow researchers (and PhD students) who are using similar methods. Please contact Valeria or Ben with any questions (v.cetorelli@lse.ac.uk|  or b.m.wilson@lse.ac.uk| .

The recent rise in Palestinian mortality: Evidence from recent surveys
Marwan Khawaja, UN-ESCWA, Beirut

Recent data from several household surveys conducted in 1995, 2000, 2004 and 2006 indicate that Palestinian infant mortality rate stalled at a rate above 25 per 1000 birth since the onset of the second Intifada in 2000, increasing slightly more recently. This paper seeks to examine the causes of these trends, using available birth history data and both descriptive analysis and proportional hazard statistical models. The focus will be on the relative impact of conflict (and associated closures) and poverty on mortality change. Preliminary examination of the data reveals that the stagnation of infant mortality was due to a consistent increase in neonatal mortality during the past decade. Assessment of data quality of all the surveys showed consistency in reporting, and the trends cannot therefore be attributed to data collection problems in one or more of these surveys. Findings from proportional hazard models and data from the 2006 survey showed that infant mortality declined significantly during the 1990s, stalled during the early 2000s and increased significantly in 2004-2005. Such a reversal in mortality trends remained significant after adjusting for common risk factors and other family characteristics. However, variations across periods disappeared after controlling for conflict intensity. These findings hold for both the West Bank and Gaza Strip, although Gaza showed a significant rise in mortality throughout the early 2000s. Cousin marriage, short birth spacing, mother\'s education, and water supply remained significant predictors of infant death after adjusting for other factors and unobserved heterogeneity. Policy implications of the findings are discussed.

Email: Dr. Marwan Khawaja: marwan.khawaja@gmail.com|

The relationship between informal care-giving and mortality: an analysis using the ONS Longitudinal Study
S. Ramsay, London School of Hygiene & Tropical Medicine; E. Grundy, University of Cambridge; D. O'Reilly, Queen's University Belfast

Most care for people who need help with everyday activities is provided by family members and other 'informal' caregivers. Care-giving can be stressful and carers in England and Wales have a right to have their needs assessed, as well as those of the person they are providing care for. Numerous studies have suggested that care-giving may be associated with various health problems, including a higher risk of depression. However a recent analysis of the Northern Ireland Longitudinal Study found that caregivers had lower mortality than non-caregivers. This finding challenges common stereotypes that depict care-givers as comparatively vulnerable. In this paper we replicate the Northern Irish analyses using the ONS-Longitudinal Study for England and Wales. In doing so, this work helps to assess the broader significance of the Northern Irish experience and also affords some clearer conclusions about the mechanisms that might account for this comparative survivorship advantage.

Email: Dr. Susan Ramsay: susan.ramsay@lshtm.ac.uk|

Postponement of first births and revival of educational differences in Hungarian fertility after 1989
Jonas Woods, Karel Neel, University of Antwerp

During the 1990s, Hungary witnessed dramatic economic and social transformations in the transition from a planned to a market economy. In the turn to capitalism, the demand for highly skilled employees rose and unemployment levels appeared, especially affecting youth. For this group the unfavourable economic conditions and higher returns on education are incentives to prolong education and delay childbearing. In this paper we study the decline of first childbearing in Hungary after 1990 in the light of increasing enrolment in tertiary education and economic downturns. Analyses use individual-level data from the Hungarian Generations and Gender Survey conducted in 2003/2004, containing fertility histories for 13540 respondents. Multilevel discrete time event history analyses model the impact of individual-level educational characteristics and macro-level economic context on first birth hazards. Since literature shows that GDP does not provide sufficient links with micro-level living conditions and unemployment rates are not available pre-1990, the chosen economic indicator is wholesale price index closely mirroring consumer prices. Results show significant declines in first births for younger age groups, especially after 1989. This declining trend is closely linked with the educational expansion after 1989. Both constraints on childbearing due to educational enrolment and postponing effects of educational attainment prove stronger after 1989. Negative effects of rising prices, i.e. procyclical childbearing behaviour, have been found among younger ages, especially among higher educated and men. In conclusion the paper finds evidence for re-emerging educational childbearing differentials and parallels between the emergence of lowest-low fertility in Hungary and the emergence of subreplacement fertility in North Western European countries after 1970.

Email: Jonas Woods: jonas.wood@ua.ac.be|

A life course perspective on family formation in Bangladesh
Nahid Kamal. MEASURE Evaluation, UNC-Chapel Hill, International Centre for Diarrhoeal Diseases Research, Bangladesh (ICDDR,B); Gabriela Mejia-Pailles, London School of Economics

Early female marriage is still a concern in Bangladesh. Despite investments in female education and employment since the early 1990s, median age at first marriage among women aged 20-24 remains 16.6 years when the legal minimum age is 18. Applying Survival Analysis techniques to data from the 2007 DHS, we attempt to shed light on the circumstances surrounding early family formation (marriage and childbearing) in the country by examining the life course of Bangladeshi women. Our results reveal that among women with no education, 68% of those aged 20-24 and 84% aged 45-49 were married by age 16. In contrast, among women who had some schooling, the proportions married by age 16 among the age groups 20-24 and 45-49 were much smaller - 47% and 71%, respectively. Hence, our results suggest that birth cohort, followed by educational attainment, are more important determinants in the life course of Bangladeshi women\'s family formation than socio-economic status, religion, or place of residence. Contrary to the preconception of a fixed life course trajectory, results reveal that Bangladeshi women follow diverse patterns. Whereas, women with little or no education leave education prematurely (if ever in education) and enter marriage and childbearing at early ages, the trajectory for more educated women is not as clear cut. Some get married later upon completing education; others commence family formation while still in school. The paper discusses the plausible causes and consequences of the various trajectories of family formation followed by Bangladeshi women.

Email: Dr. Nahid Kamal: nahid.kamal@icddrb.org|; nahidkml@gmail.com|

Qualitative and mixed methods research in population studies

Session organisers: Dr. Elspeth Graham, University of St. Andrews; Dr. Joanna Sage, University of Southampton

Although population studies remains a largely quantitative research area, the value of qualitative and mixed methodologies for extending and deepening analyses of populations and population issues have been increasingly recognised in recent years. This session seeks to invigorate the use of mixed and qualitative methods in population studies by providing a forum for researchers to share both their findings and their insights on the application of qualitative and mixed methods to population research questions. The substantive theme of the session is 'understanding linked lives'.

Mixed methods as a long-term research strategy: investigating continuities and change in French Caribbean migration and transnational experience
Stephanie Condon, INED (Institut National d'Etudes Démographiques)

This paper intends to demonstrate how a mixed methods approach can be fruitful as a long-term research strategy. Combining quantitative and qualitative data has become almost standard practice in migration research in the field of population studies, quantitative data being used to describe migration trends and the socio-demographic characteristics of migrant populations whilst qualitative data is used to understand the dynamics of the migration process. Yet as one or other of the data sources are analyzed, separately or in conjunction, new questions constantly emerge. This has been the case over the last two decades for the author of this paper, in seeking to trace changes and continuities in the dynamics of migration from the French Caribbean. More recently, issues surrounding the existence of alternative migration options for this population have been central to my investigations. For most of the period from 1920s to 1980s, migration from Guadeloupe and Martinique was directed to the French métropole (or 'mainland France'). This was the result of the French state's policies of assimilation and organised emigration from its former colonies. Since the 1990s increasing numbers of 'French Caribbeans', either born in the Caribbean or in mainland France (as descendants of migrants), have sought other destinations, among which London. They have thus contributed to the much-publicized movement of young French skilled migrants to the UK capital. Bringing together qualitative and quantitative findings, we will discuss how these young French Caribbeans locate this migration as an extension - or a break with - their transatlantic social field.

Email: Dr. Stephanie Condon: condon@ined.fr|

A lifecourse perspective on experiences of parental marital disruption and changing family relations through the use of life narrative and life history calendar methods
Jo Sage; Maria Evandrou; Jane Falkingham, ESRC Centre for Population Change, University of Southampton

This paper outlines a qualitative lifecourse study involving 42 participants, which examined the effects of parental marital disruption occurring at different stages of the lifecourse on the quality of participants' relationships with their parents, and their willingness to provide care for them as they grow older. The longitudinal perspective adopted in this study responds to the lack of research tracing the outcomes of divorce for children across their lifecourse (particularly beyond young adulthood), and our findings have revealed that this was a significant gap in the literature. A qualitative lifecourse research framework such as the one applied in this study necessarily involves the retrospective collection of information about events unfolding over long periods of time, and memory recall can compromise the robustness of these data. We describe in this paper how we overcame this through an innovative approach involving the use of life history calendar, life narrative and semi-structured interview methods. This enabled us to accurately record the timing of marital disruption events in participants' lives, while eliciting narrative data about the meanings and effects of these experiences across their lifecourses. The research highlights how such an approach can instil rigour in population studies seeking to elicit retrospective data on the intersections of events, changing family relations, and intergenerational exchange frameworks across the lifecourse.

Email: Dr. Jo. Sage: j.sage@soton.ac.uk|

Family matters: migration and childbearing decisions of 'new' Poles in the UK
Paulina Trevena, ESRC Centre for Population Change, University of Southampton; Derek McGhee, ESRC Centre for Population Change, University of Southampton; Sue Heath, Morgan Centre for the Study of Relationships and Personal Life, University of Manchester

Following the rapid migration wave from Poland to the UK triggered by the 2004 EU Enlargement, the Polish community has become the single largest foreign-born national group in the country. Although Polish migration was initially seen as a transient phenomenon dominated by young and single males with no dependants, the significant increase in numbers of Polish pupils in British schools in the recent years and the current \'Polish baby boom\' taking place in the UK clearly indicate a shift towards family settlement and family formation. This presentation aims to explain the reasons behind these processes. Our analysis is based on the results of a large-scale qualitative study ('International migration and its impact on family and household formation among Polish migrants living in England and Scotland') involving 83 in-depth interviews with Polish migrants living in four different locations across the UK, both urban and rural. We have found that in the case of couples, especially those with children, family issues play a crucial role in the process of migration and settlement in the UK. Polish migrants who left their families behind are generally keen to bring over their children and, in the case of childless couples, to start a family in the UK because of three major reasons: work opportunities, relatively higher living standards, and the security of living in a country with a well-developed welfare system.

Email: Dr. Paulina Trevena: p.m.trevena@soton.ac.uk|

Transnational groups of reference and their role in justifications of intentions for the second offspring among Polish men and women living in Poland and in the UK
Joanna Marczak, London School of Economics

This paper has the overall aim to explore and compare the rationales behind, and justifications for, intentions about whether to have a second child among Polish men and women living in the UK and Poland. Low levels of progression to the second and higher parities have contributed to a very low Total Fertility Rate in Poland (TFR 1.39 in 2009). Despite low TFR the childbearing intentions of people in Poland have been shown by various surveys to be at around 2 children per person and there exists a gap between intended and actual fertility. Although there is no research on fertility intentions of Polish migrants in the UK, their TFR at 1,6 is higher than in Poland. Qualitative in-depth interviews (n=42) have been carried out with men and women living in London and Krakow who already have one child and the results suggest that top in the list of factors related to having the second child is the ability to provide adequate living standard for the family and child[ren] in the local socio-economic setting. However, what is considered as adequate living conditions and institutional setting for having another child depends largely on the frame of reference employed to define adequacy. Respondents\' narratives illustrate that they are not constrained by national boundaries and individuals draw on transnational comparisons and transnational groups of reference to identify acceptable standards which is visible against the background of intensified migration, and economic and political integration within the enlarged EU.

Email: Joanna Marczak: j.marczak@lse.ac.uk|

Qualitative and mixed methods research panel session

The purpose of this panel session is to encourage and stimulate rigorous and robust qualitative and mixed methods research within population studies by identifying and discussing the opportunities and challenges these methodologies present for population researchers.

Panel members:

Chair: Professor Elspeth Graham (ESRC Centre for Population Change, University of St. Andrews).

• Darren Smith (Department of Geography, Loughborough University).
• Stephanie Condon (Institut National d’Etudes Demographiques).
• Jo Sage (ESRC Centre for Population Change, University of Southampton).
• Paulina Trevena (TBC) (ESRC Centre for Population Change, University of Southampton).

Examples of potential topics for discussion include:

• Understanding reliability and validity in qualitative research;
• Epistemological challenges for mixed-method research;
• Integrating qualitative and quantitative approaches in mixed methods research design;
• Longitudinal qualitative research and mixed method research design;
• The use and development of innovative qualitative research methods in population studies;
• Sampling and selection issues in qualitative research;
• The use of (new) technologies in qualitative research.

Email: Dr. Jo Sage: j.sage@soton.ac.uk|