Ethnicity & religion strand abstracts


Ethnicity & education: Monday 12 September 1:30pm


What's a degree worth? Early labour market outcomes of EU domiciled graduates in the UK
Renee Luthra, Institute for Social and Economic Research, University of Essex   

International students currently comprise one in eight students in UK higher education, and are concentrated in highly sought after STEM and business subjects. To maximize the societal benefit of this migration, it is critical to understand the labour market outcomes of international students, both those who remain as well as those who leave after graduation. Using rich, large-scale administrative data on graduate labour market outcomes, this paper asks “How do EU domiciled graduates fare in the UK labour market, as compared to British students as well as to graduates who returned to their home countries?” We use the 2004-2012 DLHE, which surveys all EU and UK domiciled students obtaining a higher qualification 6 months after graduation, following graduates even if they leave the UK. We compare the degree class, employment, and wages of British and EU domiciled graduates in the UK and abroad, adjusting for socioeconomic background, educational performance (for employment and wages), higher education institution characteristics and course of study. On average, we find that EU domiciled students are much more likely to obtain a first class degree, to be employed, and to earn higher wages than UK domiciled students, even after adjusting for institution quality and socioeconomic background. German domiciled students are the highest performers. Moreover, we find that EU domiciled students who remain in the UK are more likely to have first class degrees. Further analysis will explore the wage differences between EU domiciled students who remain in the UK to those who return home. 

Evolution of the Jewish faith schools in the UK: Numerical trends, policy debates and theoretical implications
L. Daniel Staetsky, Institute for Jewish Policy Research (JPR), London   

Jewish faith schools form a small proportion (0.7%) of the total number of state-funded faith schools. However, the Jewish school sector expanded dramatically over the years, with the number of pupils doubling since the mid-1990s. This upward trend is of interest to the Jewish communal organisations sponsoring the Jewish school sector, such as the Board of Deputies of British Jews and the Jewish Leadership Council. Just like their counterparts at the national level (the local authorities and the Department of Education), in planning for Jewish schools the Jewish organisations respond to the demographic developments in the British Jewish community. However, unlike their counterparts, they also need to take into account the changing ‘appetite’ (=preference) for Jewish schools among the Jews. The preference for Jewish education is rooted in identity, not demography. To assist the Jewish communal organisations JPR carried a project aiming to inform the educational policies in the British Jewish community. The project explores: 1. Whether the expansion of Jewish school sector reflects an increase in preference for Jewish education or just a matter of increase in fertility at an earlier point in time (= an increase in the number of Jewish children); 2. Whether the expansion is seen among strictly Orthodox and mainstream Jews alike; 3. What policy conclusions can be derived from the findings by the Jewish communal organisations. This paper will show that Jewish school sector presents an intriguing case where demography and identity, in combination, shape policy debates and policy solutions. 

Variation in gender disparities in school education across ethnic groups in India: An analysis of the inequality in educational process
Mousumi Dutta 1, Zakir Husain 2, 1 Economics Department, Presidency University, India, 2 Humanities & Social Sciences Department, Indian Institute of Technology, India   

This paper presents an alternative approach to studying gender discrimination in school education across ethnic groups (like Scheduled Castes, Scheduled Tribes and Muslims) in India, based on unit level data from the National Sample Survey Office Employment and Unemployment 68th round (2012-2013). Gender differentials in education are generally measured in terms of inequality in educational outcomes. Another approach focusses on the inequality in different stages of education, estimating transition probabilities at each level. Both approaches, however, have important limitations and should be integrated by decomposing the inequality in educational outcome as a weighted sum of the inequality in each transition stage (Buis, 2007). This enables us to simultaneously study inequalities in the highest level of outcome, examine how inequality varies over educational levels and assess the importance of each transition in the overall process of education. Results of a sequential logit show that gender disparity decreases as education level increases; surprisingly, it is lower among socially backward communities and also gets reversed in urban areas—with females having higher probabilities of completing school and high school levels. However, analysis of the importance of each transition in the overall process shows that inequality in enrolment remains a major factor in the persistence of gender disparities in educational outcomes. Further, analysis of principal status of children reveals that male drop-outs within backward ethnic groups occur mainly to enable the child to enter the informal labour force, while the majority of females who complete high school are confined to household activities. 

Ethnicity & space: Segregation & inequality dynamics. Monday 12 September 4:45pm 

Eastern European immigrants in a comparative cross-national context: A K-nearest neighbour perspective
Ian Shuttleworth 1, John Östh 2, 1 Queen’s University Belfast, 2 University of Uppsala 

Segregation studies have often sought to make cross-national or cross-city comparisons to answer questions as to whether one place is more or less segregated than somewhere else.  Despite the apparent simplicity of this question it is very hard to provide robust answers.  One set of problems is technical and concerns the methodology of comparison.  These issues include constructing a consistent geographical base and having data categories, for example ethnic groups, which can meaningfully be compared.  A second set of problems relate to the social, economic, political and demographic contexts for segregation where the groups which are the analytical focus differ in important ways which make comparison a more tenuous activity.  How far, for instance, can South Asian segregation in Northern Ireland be compared with community segregation in Northern Ireland?  The presentation explores these issues by describing and analysing patterns of Eastern European segregation in the UK, Ireland and Sweden using a K-nearest neighbour approach implemented using EQUIPOP.  This approach creates similarly-sized neighbourhoods thus helping to make better comparisons.  Beyond this it also considers how the settlement geography of immigrant groups which arrived at approximately the same time, and which can be defined in official statistics in closely comparable ways, differs between countries.  It then suggests reasons for differences and similarities. 

Race, ethnicity, xenophobia or social ties?: explaining ethno-specific mobility in England and Wales
Eric Kaufmann, Birkbeck University of London 

This paper shows that White Britons consistently move to whiter areas than minority Britons who originate in similar areas and share similar socioeconomic and demographic characteristics. Why? The work considers a descending order of social attachment, beginning with the macro - "race" - then ethnicity, and finally the micro level of attraction to friends/family. This paper uses longitudinal data - Understanding Society and the ONS LS - augmented by the Citizenship Survey - to ask which level of identity is most salient for movers. It also considers ethnic and religious commitment, and opposition to diversity (among whites), as potential factors. Results thus far suggest that ethnicity rather than race is most important for own-group attraction. Attitudes (religiosity, ethnic attachment, xenophobia) - while associated with movement to own-group areas - do not account for ethno-specific mobility. 

Local Ethnic Inequalities and Ethnic Minority Concentration in England and Wales, 2001-11
Kitty Lymperopoulou 1, Nissa Finney 2, Gemma Catney 3, 1 University of Manchester, 2 University of St Andrews, 3 University of Liverpool 

There has been a longstanding debate in the UK about the residential patterns of ethnic minority groups and the consequences of these patterns for social cohesion. However, there has been less focus on the role of residential segregation in perpetuating the socio-economic disadvantage of ethnic minority groups, and as yet there is little evidence on how residential segregation patterns relate to spatial patterns of ethnic inequality. This chapter draws on data from the 2001 and 2011 Censuses in the UK to explore how ethnic group concentrations within local authorities in England and Wales are related to patterns of ethnic inequality in employment, education, housing, and health. The Index of Dissimilarity is used to measure residential segregation for nine ethnic minority groups in 2011 and how it has changed since 2001, for neighbourhoods (output areas) within districts. Ethnic inequality is measured as the absolute difference in the proportion of White British people and people from ethnic minority groups within districts for employment, education, housing and health. Considerable variation by ethnic group and district is found in the association between segregation and socio-economic inequality, suggesting both positive and negative ‘effects’ of segregation. Key findings are that high residential segregation is associated with high ethnic inequalities in education within districts; high residential segregation is associated with low ethnic inequalities in employment, health and housing and this relationship strengthened over the 2000s; higher levels of ethnic inequality are found in districts which have had least reduction in segregation, or experienced an increase in segregation. 

Examining the roles of international and internal migration in the spatial distribution of the UK’s ethnic group populations
Nik Lomax, Pia Wohland, Philip Rees, School of Geography, University of Leeds 

Both internal and international migration play vital roles in explaining the changing spatial distribution of the UK’s ethnic groups. Previous attention has focussed on selected regions (primarily London) or cities. In this paper we present an analysis for the whole UK, at local authority and city region scale in England and at home country level outside England, for the decade between the 2001 and 2011 Censuses. The analysis draws on a detailed exercise in reconciling ethnic population change over that period, constrained to published migration tables as well as births and deaths. We present results which capture both the variety of experiences across ethnic groups and between different areas in the UK. Preliminary results show varied urban/rural patterns and migration propensity by age between ethnic groups. We then use the reconciliation to inform assumptions about both internal and international migration in an ethnic population projection model. By running scenarios which successively set international migration and then internal migration flows to zero, we can show how different the ethnic spatial distributions might be and hence measure their impact. Overall, ethnic diversity is projected to increase in a large number of UK local authority and city region areas, with substantial variation observed between areas. Results from this projection model will be important for policy decisions at both national and local level. 

Ethnicity, employment & social mobility. Tuesday 13 September 09:00am 

Re-thinking the concept of ethnicity for population studies and evidence-based policy Nissa Finney, University of St Andrews & ESRC Centre on Dynamics of Ethnicity 

Ethnicity is frequently employed in population studies, and particularly in studies on residential migration and neighbourbood change. It is employed in a number of functions: to describe individuals and neighbourhoods; to frame studies that explore social differences; to describe differential experiences; to explain social differences; and as a consequence of social processes. This paper argues that for the concept of ethnicity to be useful for understanding population patterns and processes it needs to be re-theorised in relation to all its functions. The paper begins this re-theorising by examining how ethnicity is employed in studies of residential patterns and processes, reviewing theories of ethnicity in population studies, and within ethnic, race and migration studies more broadly. The paper then draws on analysis of 2011 Census data together with migration history interviews with young adults to examine the nature and extent of the significance of ethnicity in residential patterns, decisions, experiences and aspirations. Thus, the paper makes two contributions: 1) a critique of the concept of ethnicity in population studies that illustrates how scale, life course, family relations, historical groundings and processes of racialisation need to be better accounted for; 2) in relation to the conference theme of demography and evidence-based policy, discussion of how ethnicity might be fruitfully captured in censuses, surveys and administrative data as we look to the 2021 Census and beyond.

Employment dynamics among ethnic minorities in the United Kingdom
Nicole Martin 1, Lucinda Platt 2, 1 Institute for Social and Economic Research, University of Essex, 2 The London School of Economics and Social Science 

It is well-established that ethnic minorities suffer higher rates of unemployment than the white British group in the labour market, even among highly qualified graduates. However, until now there has not been an opportunity to consider the duration of unemployment, and stability of employment. This is important because frequently changing jobs means that employees have to rebuild job-specific human capital, making them less productive in the short term (all other things being equal). This may also help account for observed pay gaps, if ethnic minority employees do not receive increments or promotions due to shorter time in post. Longitudinal data in Understanding Society, the UK’s longitudinal household study, allows us to use survival analysis to examine (i) how the duration of employment and unemployment spells vary by the ethnicity and religious group of employees, (ii) what factors protect against short spells of employment.  Does living in an area with more coethnic residents help ethnic minorities find a job, and are ethnic minority graduates able to access more stable employment than their counterparts without degrees? 

The evolution of unemployment inequality in Northern Ireland, 1991-2011
Neil Rowland, School of Geography, Archaeology and Paleoecology, Queen's University Belfast   

This paper assesses the relationship between religion and unemployment in Northern Ireland’s labour market. Since 1975, researchers have investigated why the Catholic rate of unemployment was significantly and consistently higher than the Protestant rate. By the mid-1990s, most concluded that a number of factors – including, controversially, religious discrimination – had combined to generate this inequality. However, no published research has used the latest available data to investigate whether Northern Ireland’s two main communities are still so sharply divided along these lines. This literature gap is addressed here using the Northern Ireland Longitudinal Study (NILS), which is a 28 per cent sample (approx. 500,000 individuals) of Northern Ireland’s census records linked to administrative data for the 1981-2011 censuses. Its large sample size and longitudinal structure provide a unique opportunity to analyse this research question at a level of detail unmatched by other individual-level datasets. To investigate, using the NILS sample for each of the 1991, 2001 and 2011 censuses, various regression models were constructed to estimate raw and conditional differentials between Catholic and Protestant unemployment rates. The results reveal that these differentials have narrowed through time, particularly between 1991 and 2001. However, despite past policymaking efforts, Catholics in the 2001 and 2011 samples face an approximate 3.0 percentage point higher probability of unemployment than Protestants after controlling for confounding factors. Although this indicates that full unemployment equality has yet to be achieved, it also indicates that intercommunal unemployment outcomes in Northern Ireland have never been as similar as they are at present. 

Religion and Educational Attainment as Interacting Factors in the Determination of Economic Opportunity and Social Mobility in the new Northern Ireland
John Moriarty 1, David Wright 1, Dermot O'Reilly 2, Allen Thurston 3, 1 Administrative Data Research Centre for Northern Ireland, Queen's University Belfast, 2 Centre for Public Health, Queen's University Belfast, 3 School of Education, Queen's University Belfast   

Research Question: This study examines whether these disparities between Protestants and Catholics have persisted through the post-conflict period, or whether differences in social mobility have diminished historical inequalities. Furthermore, we assess the role of education in bringing about these changes. Data are from a ten-year cohort of Northern Ireland Longitudinal Study (NILS) members aged 8-17 in 1991. Occupational class status is captured in 1991 and 2011, as well as religion of upbringing, educational attainment and key demographic characteristics. Methods: The relationship between religion of upbringing and occupational class ‘destination’ in 2011 was captured using ordinal logistic regression. Moderating effects of parents’ 1991 occupational class and educational attainment were tested through further iterations of the regression model. Results: We find evidence that, given class background, occupational class of Catholics is slightly greater, indicating an acceleration in social mobility. Catholics in this cohort were also less likely to leave education at an early stage, which is likely driving the mobility of this group. We also find shifts in occupational class structure over the period from 1991 to 2011, with both professional and routine occupations expanding and intermediate occupations contracting. Discussion and Applications: There is evidence suggesting that persistence in education among Catholics may be driving a levelling of class status between the dominant religious groupings in Northern Ireland, adding to the demand for effective interventions to keep young people in school for longer. 


Ethnicity, religion, attitudes & the life course. Wednesday 14 September 09:00am 

The effect of education on ethnic inequality
Malcolm Brynin, ISER University of Essex   

Ethnic minority men are relatively likely to be in poorly paying occupations, though ethnic minority women do far better relative to white British women. Yet both men and women from minority groups are more likely than the majority group to have a degree. Using the UK Labour Force Survey and Destination of Leavers from Higher Education survey the analysis seeks to reconcile this contradictory outcome. The results using the LFS suggest that the key problem is overqualification. Not only are graduates from minorities more likely to be overqualified for the specific work they do but they are also likely to be working in occupations where graduates are less in demand. OLS regression and decomposition analysis suggests that this is not all down to differential characteristics. However, the DLHE results indicate that ethnic minorities are more likely to go to a non-elite university. The paper seeks to assess the causes of this and its contribution to continuing, and in some cases, worsening pay gaps. 

Intergenerational transmission of attitudes toward migration as a potential antagonist of the demographic metabolism
Nadia Steiber, Erich Striessnig, Samir KC, Wittgenstein Centre for Demography and Global Human Capital (IIASA, OEAW/VID, WU)    

Due to their continuous aging, capital rich European societies increasingly rely on migration from societies with large youth bulges
to fill up their depleting human capital stocks. In this article we look not only at how public opinion toward migration has changed over time, but also distinguish the age and period dimensions of attitudinal change with regard to migration. Using different waves of the European Social Survey, in a first step we model age-period-cohort trends in attitudes toward migration controlling for important covariates. Subsequently, we project how attitudes toward migration can be expected to evolve in the future considering past changes along cohort lines and future age compositions. For this projection exercise we apply the notion of the “Demographic Metabolism”, introduced by Ryder (1965) and advanced into “A Predictive Theory of Socioeconomic Change” by Lutz (2013). Yet, whereas the demographic metabolism assumes that societies automatically become more tolerant as more tolerant younger cohorts replace less tolerant older cohorts, in this article we look at the possibility of intergenerational transmission of attitudes as a mechanism through which the workings of the metabolism may be slowed down (e.g., when the attitudes of upwardly mobile young individuals – with higher levels of education than their parents – are still strongly influenced by their parents’ attitudes that they have been socialized with before entering the schooling system). 

Prevalence, intensity and persistence of ethnic or racial harassment in Britain
Alita Nandi, Renee Luthra, Michaela Benzeval, ISER, University of Essex   

We use data from Understanding Society to examine the patterns of harassment among Britain’s ethnic minority groups today, thereby updating and extending the last study on this topic based on the 1994 FNSEM (Virdee 1998). We identify individual and area level characteristics associated with harassment defined as being physically or verbally attacked. Interestingly we find that white UK men are more likely to report harassment than ethnic minority men. But Muslim or Sikh men are more likely to report racial harassment, i.e., identify their ethnicity, nationality and religion as the reason. We also find that these risks are lower for those living with co-ethnics or having some educational qualification. Individual or area level deprivation does not matter. We also examine potential harassment defined as avoiding a place or finding it unsafe as the negative impact on health and wellbeing of this may be as high as that of actual harassment. We find that women are more likely than men to experience this kind of harassment. As a result, they are less likely to experience actual harassment. Although individual income does not change the likelihood of actual harassment it does reduce the likelihood of potential harassment. 16-19 year olds and 60+ year olds are less likely to be actually or potentially harassed than other age groups. Finally, the longitudinal design of the study enables us to examine repeated exposure to harassment and find similar associations as the single incident model. But we find that area level deprivation reduces the likelihood of multiple harassment incidents. 

Differences in expectations and outcomes with respect to childbearing between the religious ‘nones’ and the religiously active
Marion Burkimsher, University of Lausanne (Affil.)   

Across western Europe, childlessness is more common amongst the non-religious, and their family sizes are significantly smaller than the religiously active. The question is whether these outcomes reflect a lower desire for children amongst secular people and, if so, why this might be so. In Switzerland the Family and Generations Survey was carried out in 2013; the questions posed were similar to those in the Generations and Gender Survey. The FGS found that the non-affiliated childless were nearly three times as likely as Christian church-goers to state a preference for not wanting any children; however, a large majority still express a desire for at least one child. One module in the FGS asked respondents their expectations of the impact that having a first or subsequent child would have on their life. Using logistic regression, with control variables such as educational level, it was found that the religiously active were around twice as likely to expect greater life satisfaction from having a(nother) child. In contrast, those who professed no religion were twice as likely as Christian attendees to expect their financial situation to become ‘much worse’. The childless ‘nones’ were also twice as worried about the negative impact of a child on their sex life and their professional life. Further investigations are planned using the GGS data from other countries to investigate whether these attitudinal differentials by religiosity are common elsewhere. In the secularising environment of Europe, this could offer an underlying explanation of low fertility levels. 

Religious identity, practices and beliefs across the life course
Aradhna Kaushal, Dorina Cadar, Mai Stafford, Marcus Richards, MRC Unit for Life Long Health and Ageing at UCL    

Introduction: Findings from population surveys have shown a steady decline in religious affiliation in the UK. The aim of this paper was to investigate trends in religious identity, practices and beliefs across the life course and how these change throughout the life, using data from the National Survey of Health and Development (1946 British birth cohort). Method Participants were 1271 study members with complete data on religious upbringing, practices and beliefs from age 11 to 68. Study members were asked about religious upbringing at ages 11 and 36, strength of religious beliefs at age 26 and frequency of religious attendance at ages 36, 43 and 60-64. Questions on whether religion provides meaning or purpose in life and the frequency of prayer or meditation were asked at age 68. Results: Around 80% of study members reported a religious upbringing and attended Sunday school. By age 36, 66% reported a religious belief and only 24% attended a place of worship at least once a month. Regular religious attendance declined at age 43 and 60-64 to 16% and 15% respectively. Despite low levels of religious attendance, 35% of study members reported that religion provides some meaning or purpose in life and 24% report praying or meditating regularly at age 68. Conclusion: Findings show a decrease in religious practices and beliefs throughout the life course and a departure from the religion study members were brought up in. Future research should focus on determining if socio-economic factors are associated with religious identity, practices and beliefs.