Jump to ContentJump to Main Navigation
Unequal AttainmentsEthnic educational inequalities in ten Western countries$

Anthony Heath and Yaël Brinbaum

Print publication date: 2014

Print ISBN-13: 9780197265741

Published to University Press Scholarship Online: May 2015

DOI: 10.5871/bacad/9780197265741.001.0001

Show Summary Details
Page of
date: 27 May 2020

Ethnic Differences in Completion of Upper Secondary Education

Ethnic Differences in Completion of Upper Secondary Education

Chapter:
(p.149) 6 Ethnic Differences in Completion of Upper Secondary Education
Source:
Unequal Attainments
Author(s):

Yaël Brinbaum

Anthony Heath

Publisher:
British Academy
DOI:10.5871/bacad/9780197265741.003.0006

Abstract and Keywords

In this chapter we explore the completion of upper secondary education. The central question here is whether the high rates of ethnic minority continuation into upper secondary education lead to a closing of the attainment gaps with the majority group. The story which emerges is that there are considerable continuities over the secondary school career. Ethnic disadvantages have not cumulated, but neither have they been eliminated. Nevertheless, the picture at the end of upper secondary education is not simply a repeat of the picture at the end of lower secondary. Overall, we find that in Canada, France and the USA the ethnic penalties found in test scores at the end of lower secondary have been mitigated by the end of upper secondary, while in Sweden the movements are more likely to be in the opposite direction.

Keywords:   ethnic inequality, second generation, educational career, integration context, upper secondary education

Introduction

SUCCESSFUL COMPLETION OF UPPER SECONDARY EDUCATION has been identified as a key stage in educational progression in most Western countries. Qualifications such as the Abitur in Germany and the ‘Bac’ in France qualify young people for access to tertiary education and in turn for entry into the most prestigious (and well-rewarded) professional and managerial jobs of the salariat. Moreover, completed upper secondary education and the corresponding credentials have value in their own right for direct access to the labour market, and generally give access to more skilled occupations than do lower-level qualifications or incomplete upper secondary education.

While these propositions hold generally true in Western societies (see, for example, the data reported in Heath & Cheung 2007), it is important to recognise that there are important institutional differences between our countries. Indeed, in some countries it is not entirely clear what ‘successful completion’ actually means. There are also major differences in the proportions of young people who complete, and the kinds of qualification that they obtain, and thus implications for how much of an advantage completion gives young people in the labour market.

In this chapter we take completion to constitute the award of a relevant national certificate or qualification at level 3A or 3B in the International Standard Classification of Education (ISCED) scheme (Schneider & Kogan 2008).1 We include acquisition of the appropriate vocational qualifications (3B) as well as of the high-prestige academic qualifications like the Abitur and Bac. Our measure thus includes quite a diverse array of qualifications. There are also major variations not only between the nature of the qualifications but also of the grades obtained in a particular qualification. These can (p.150) be very consequential for the prestige of the job and even more so for the prestige of the university that one can subsequently attend. For example, top grades at the British academic A-level certificate are necessary in order to gain access to an elite university while lower grades give access to lower-status institutions. There are therefore some important qualitative distinctions within our measure of ‘completed upper secondary education’. Unfortunately, information on the grades obtained is rarely available in national datasets and we are not able to pursue these issues in this chapter. Nevertheless, we can be confident that possession of one of these upper secondary certificates will improve the chances of the holder in the labour market in comparison with young people who have left school without such a qualification.

As in previous chapters, we would expect to find both ‘origin’ and ‘destination’ effects on the chances of successful completion. Considering origin effects first, ethnic minorities who have been found in previous chapters to be particularly successful, such as the Chinese or Indians, might be expected to continue their success in upper secondary education. Similarly, groups such as students of Turkish background who were behind with respect to test scores (Chapter 3) and/or who had low transition rates to upper secondary (Chapter 4) could be expected to have lower rates of successful completion of upper secondary. We expect to see considerable continuity with the results of previous chapters, but it will also be important to check whether these ethnic penalties and premia have diverged further over the course of the upper secondary career, or whether there has been some mitigation. Have disadvantaged groups like those of Turkish background fallen further behind, or have they caught up? Have the highly successful groups like the Chinese pulled further ahead or have they slipped back?

There is also a question mark about some groups who, as we saw in Chapters 4 and 5, had positive ‘secondary effects’ despite negative ‘primary effects’. Students such as Maghrebians in France or Caribbeans in England and Wales may have higher continuation rates into upper secondary than would have been expected given their test scores at age 15, but perhaps they will find it difficult to keep up with their academically higher-scoring peers. In other words, there is a question as to whether high continuation rates can be converted into high completion rates. More generally, do the high aspirations and ambitions that we believe lie behind the positive secondary effects of ethnicity observed in Chapters 4 and 5 get translated into high completion rates at the end of upper secondary education?

We might also expect to find some destination effects, just as in Chapter 5. Essentially, some systems such as the French, Belgian and Dutch (and to a lesser extent the British) maintain high academic standards for the school-completion exams, with independently marked formal exams which a relatively small proportion of the age group typically achieves. Other systems (p.151) (such as the US, Canadian and Swedish) emphasise excellence and standards to a lesser extent, with grading largely being undertaken by teachers and with a much higher proportion of the age group completing upper secondary education. What we might expect is that in the systems with formal examinations it will be harder for minorities to convert high ambitions into successful outcomes, whereas in the more comprehensive ‘open-to-all’ systems the high aspirations will lead to high completion rates. Waters et al. (2013) have suggested the hypothesis that minorities do well when they are free to choose, but do less well when it is institutions that do the choosing and grading.

In the sections which follow we begin by briefly reviewing the definition and measurement of completion in the different countries. We then compare the proportion of young people with completed upper secondary education by ethnic origin and country, focusing on the differences or similarities between the groups and the countries. After a comparative and static analysis of completion we adopt a longitudinal perspective, when the data allow this approach, comparing the ethnic penalties and premia at the different stages of the secondary school career. We compare the following countries: Belgium and the Netherlands (early selection countries), England and Wales, France and Sweden (late-selection countries) and Canada and the USA (non-tracked systems). We do not have the requisite data for Finland, Germany (except for looking at gross differences) or Switzerland.

A typology of upper secondary outcomes

Before turning to the results of our analysis, we need to present a more detailed account of what ‘completion’ means in the different educational systems. This is shown in Table 6.1. Our definition of completion is in essence that students have obtained whatever the upper secondary leaving certificates are in each country. We include both academic qualifications and the corresponding certificates or diplomas awarded at the end of upper secondary vocational courses. While this is not ideal, as it lumps together qualifications of very different characteristics and qualities, it does have the advantage that it can be applied to the North American as well as to the European systems, unlike a measure which focuses simply on the acquisition of high-prestige academic certificates.

In Table 6.1 we can see that completion rates vary considerably between countries. They are the highest in the two North American countries with their comprehensive, non-tracked, upper secondary systems where 90 per cent or more of the age-cohort graduate from upper secondary education. It should also be noted that these systems are the least ‘credentialised’ without the formal high stakes examinations like the German Abitur or French Bac. (p.152)

Table 6.1. Upper secondary school-completion certificates

Country

Name of uppersecondaryqualification

Provides access to university

Percentage achieving the qualification(s)

Value in the labour market*

Belgium

Upper secondary diploma (from either vocational or academic tracks)

Yes though rarely in practice if a vocational diploma

81

1.38

Canada

High school diploma

Yes

87

1.01

England and Wales

A levels or vocational level 3 qualifications such as GNVQs

Depends on grades and number of passes, and specific course requirements

42

0.94

France

Baccalauréat (including general, technological and vocational baccalaureats)

Yes but chances of success depend on type of baccalaureat

65

1.79

Germany

Abitur or Fachhochschulreife

Abitur to university, Fachhochschulreife to universities of applied sciences

81

1.50

Netherlands

Upper secondary school-leaving certificate (VWO)

Yes

54

1.43

Sweden

Upper secondary certificate

Yes, except for around 15% with insufficient pass grades; vocational tracks give more limited access to degrees

67

1.80

USA

High school diplomaor GED

Yes

94

0.96

(*) These figures are estimates of the ‘ethnic penalties’ given in Heath & Cheung (2007). They are the parameter estimates for men, comparing the effects of upper secondary with lower secondary qualifications on access to the salariata

High school diplomas are awarded at the end of upper secondary education, but these are awarded by schools rather than by centralised examination boards as in some European countries. In Canada and the USA there is also the General Educational Development test (GED). This is a secondary school ‘equivalency’ test which is open to school-leavers and certifies high school level skills. The test must be taken in person at a designated GED testing centre, and once the test is successfully completed the young person earns a High School Equivalency Certificate.

At the other extreme we have England and Wales, which as we saw in Chapter 4 had one of the highest rates of early school-leaving. This probably reflects the fact that England and Wales have high stakes examinations (the (p.153) General Certificate of Secondary Education—GCSE) taken at age 15–16 which gives young people qualifications that have considerable value in the labour market. In France and the Netherlands there are also vocational qualifications (at ISCED level 3C) which are typically taken before the age of 18 but which have value in the labour market. In France, for example, the vocational qualifications of CAP (certificat d'aptitude professionnelles) and BEP (brevet d'études professionnelles) give their holders improved chances of finding work (see Kieffer 2008: Table 1). These are completed by many children with working-class backgrounds and this is also true for some ethnic minorities such as the Portuguese.

Completion of upper secondary in different countries

Figures 6.1 presents the proportion of the age-cohort who completed upper secondary education by ethnic origin and country of destination. We divide Figure 6.1 into three panels, distinguishing countries (as in Chapter 5) by time of selection. As in previous chapters, Figure 6.1 gives the gross rates before taking any account of statistical controls for social background. It is also important to note that these are percentages of the age-cohort as a whole, not simply proportions of those who stayed on in upper secondary education.

Ethnic Differences in Completion of Upper Secondary Education

Figure 6.1(a). Proportion of the age-cohort who complete upper secondary education: countries with early selection.

(p.154) We begin with Figure 6.1(a), by looking at Belgium, Germany and the Netherlands, the three countries in our study where there is early tracking. The story told by Figure 6.1(a) is a familiar one from previous chapters. In Belgium the three main minorities, Italians, Moroccans and Turks, all have lower rates of completion than the majority group, the most disadvantaged group being those of Turkish background (with only 60% completing compared with the majority's 83%). In Germany it is a similar picture with only the minority from the former Soviet Union (with 68%) matching the majority group (67%). In the Netherlands, on the other hand, the gaps between majority and minorities are small, although the rate is systematically the lowest for the Turkish second generation in all three countries. The Moroccan second generation, however, seems to be more successful in the Netherlands than in Belgium.

Turning to Figure 6.1(b), we again see some familiar patterns. In France most ethnic groups have lower completion rates than the majority group, while in Sweden it is primarily the Nordic groups who have lower completion rates. In England and Wales, on the other hand, all groups other than the Bangladeshis have rates of completion as good as or better than the majority group.

Lastly we come to the two North American countries with their comprehensive, rather than tracked, upper secondary schooling systems where completion rates are very high. All groups alike have high completion rates and there are relatively small inter-group differences, much less, for example, than in England and Wales with its low rate of completion. In Canada, almost every minority group has a higher rate of completion than the majority group, while in the USA there is more variation, with Mexicans in particular being a less successful group.

If we then control for socioeconomic background, as in previous chapters, we find that the gross differences are substantially reduced, eliminated, or even inverted for some groups in some countries. (The details are given in Table A6.1 in the appendix.)

For example, the coefficients become positive and significant for the Caribbean second generation in Canada, for the Bangladeshi second generation in England and Wales, for the Maghrebians in France and for the East Asians in Sweden. More generally, many children of Asian groups have school success (or high completion), confirming previous research from the USA (e.g. Portes & Rumbaut 2001; Kao & Thompson 2003). This is also consistent (for the Southeast Asian second generation) in France, based on a recent survey (Brinbaum et al. 2012).

At the opposite end of the scale, some groups are less likely to complete upper secondary schooling, particularly the Turkish second-generation group in the different European countries. They still encounter ethnic penalties— although net coefficients are reduced—in Belgium, the Netherlands and Sweden (a pattern also confirmed by recent new evidence in France). (p.155)

Ethnic Differences in Completion of Upper Secondary Education

Figure 6.1(b). Proportion of the age-cohort who complete upper secondary education: countries with later selection.

(p.156)
Ethnic Differences in Completion of Upper Secondary Education

Figure 6.1(c). Proportion of the age-cohort who complete upper secondary education: countries with comprehensive systems.

(p.157) The pattern is more nuanced for some groups, such as the North Africans. This group is, however, heterogeneous in the different countries. Ethnic penalties, albeit reduced when compared with the gross difference, still exist for them in Belgium and in Sweden, while the gross differences are eliminated in the Netherlands and are even inversed in France. However, this success is nuanced insofar as they are less likely to achieve the general baccalauréat which gives greater chances of success in higher education (Brinbaum & Kieffer 2009). Besides, the advantage concerns mainly the girls.

Educational careers: cumulative disadvantages or progress?

Our key concern in this chapter is whether minorities have closed the gaps with their respective majority groups over the course of upper secondary education, and in particular whether the positive ‘secondary effects’ of ethnicity on continuation and track choice which we saw in Chapters 4 and 5 have been converted into positive secondary effects at completion. We therefore show in Table 6.2 the sequence of net coefficients (i.e. the measures of ethnic penalties and premia) at the four stages of the secondary school career; namely achievement at the end of lower secondary (from Chapter 3 and shown in column 1), early school-leaving (from Chapter 4 and shown in column 2), choice of academic or vocational track (Chapter 5, shown in column 3) and completion of upper secondary (Chapter 6, shown in column 4). These net coefficients are the ones estimated in models which control for parental socioeconomic circumstances, family composition and gender; they do not include controls for test performance. (We have to drop Germany at this stage since data were available only for gross differences at completion, not for net differences after controls for social background.)

When inspecting the patterns shown in Table 6.2, we need to be cautious not to over-interpret changes in the magnitude of the coefficients. In a sense we are comparing apples with oranges: the net coefficients in column 1 (achievement at the end of lower secondary) are based on standardised scores, whereas those in column 4 are based on a binary outcome (completion or not) with a great deal of heterogeneity within the category. In the case of column 3 (track choice), early school-leavers are excluded and so we have a selective sample, not the whole of the age-cohort. We will not therefore devote as much attention to the coefficients for track choice as we do to those for continuation. Track choice has particular implications for Chapter 7, on tertiary education, where it will be of great interest to see whether ambitious choices of academic tracks in secondary education have been followed by successful university careers.

(p.158)

Table 6.2. Net upper secondary completion coefficients compared with net coefficients at earlier stages.

Test score/grades

Continuation

Academic/vocational

Completion

France

Column 1

Column 2

Column 3

Column 4

Dom-Tom

−0.08

0.01

0.09

−0.01

Portuguese

−0.03

0.22

0.27

0.14

Maghrebian

−0.12

0.38

0.38

0.27

Sub-Saharan African

−0.27

0.03

0.22

0.12

Canada

West European

0.03

−0.07

South European

−0.17

0.46

Other European

0.08

0.27

Caribbean

−0.25

0.54

South Asian

0.13

1.47

Chinese

0.23

0.02

Filipino

−0.19

0.66

East and Southeast Asian

0.24

0.71

Other

−0.14

0.71

Mixed

0.07

0.07

USA

Other white

−0.20

−0.35

Black

−0.43

0.17

Mexican

−0.49

−0.08

Other Latino

−0.41

−0.22

East Asian

0.39

0.58

Other Asian

−0.07

0.41

Belgium

Italian

−0.17

North African

−0.54

0.05

−0.28

Turkish

−0.77

−0.23

−0.36

Other

−0.52

0.29

−0.13

Mixed

0.25

0.06

−0.18

Netherlands

Moroccan

−0.35

−0.26

0.00

−0.16

Surinamese & Antillean

−0.16

−0.33

0.27

−0.07

Turkish

−0.20

−0.17

0.06

−0.44

England and Wales

Black Caribbean

−0.22

0.20

0.10

−0.07

Black African

−0.01

0.78

0.33

−0.14

Indian

0.29

0.69

0.59

0.36

Pakistani

0.11

0.33

0.32

−0.11

Bangladeshi

0.39

0.70

0.35

0.50

Chinese

0.65

0.79

0.86

0.50

Sweden

Finnish

−0.09

−0.11

0.06

−0.15

Danish

−0.15

−0.19

−0.05

−0.21

Norwegian/Icelandic

−0.05

0.00

0.01

−0.19

South European

0.08

−0.05

0.52

−0.14

Sweden

Column 1

Column 2

Column 3

Column 4

Polish

0.20

0.09

0.53

0.07

East European

0.07

−0.03

0.46

−0.04

(Ex-)Yugoslav

0.04

0.01

0.47

−0.03

Iranian

0.28

0.32

0.96

0.04

Iraqi

−0.03

−0.01

0.72

−0.18

Turkish

−0.05

−0.24

0.73

−0.27

North African

0.05

−0.10

0.82

−0.31

From Africa's Horn

0.19

0.23

1.09

0.04

Other African

0.01

−0.04

0.59

−0.29

Chilean

−0.19

−0.15

0.37

−0.33

East Asian

0.58

0.38

0.96

0.37

Southeast Asian

0.47

0.40

0.69

0.15

Other Asian

0.35

0.24

0.96

−0.02

Note. Figures in bold are significant at the 0.01 level.

(p.159) In exploring the changing coefficients across the upper secondary career our major interest will be in discovering whether significant negative coefficients in columns 1 and 2, indicating ethnic penalties, have turned into significant positive coefficients in column 4, indicative of ethnic premia (and vice versa). It would certainly be reasonable to regard this as progress, although of course there would still be supplementary issues (which we are unable to address with the currently available data) about qualitative differences in types of school-completion certificate.2

A second issue which we need to bear in mind is the nature of the datasets used when estimating the coefficients. For the first three columns we typically use panel studies, which track the same individuals over time. These panels invariably suffer from attrition, as some of the original panel members cannot be traced in later waves. Attrition is not necessarily a problem if it is random with respect to the variables of interest. But it is probable that low-achieving students are relatively likely to disappear from the panel, and the results may therefore be biased. For column 4 we can sometimes use fresh, independent samples which will therefore avoid the attrition problem. However, other problems can emerge when comparing independent samples if there are methodological differences in coverage, sampling or fieldwork procedures. We can probably have greatest confidence in Swedish results, since these are based on (p.160) the Registers and cover the whole population (although even with the Swedish data there is a potential issue since a smaller set of years is pooled for the analysis of completion than for grades).

A suitably cautious interpretation of Table 6.2 suggests that there are clear signs of progress in some countries (Canada, France and the USA), a more ambiguous story with other countries, and some evidence of minorities perhaps slipping back in Sweden.

Countries with evidence of minority progress

France shows a rather clear picture of progress over the secondary school career. The black group (Sub-Saharan Africans) who had a significant negative coefficient with respect to test scores at the end of lower secondary education is not significantly different from the majority group at completion. And the two other groups (of Portuguese and Maghrebian background) have moved from non-significant or negative coefficients to significant positive coefficients, indicating that they have higher completion rates than their peers from the majority group in similar socioeconomic circumstances. Furthermore, there is no compelling evidence that the positive (or neutral) secondary effects of ethnicity observed in continuation into upper secondary education (column 2) have evaporated at completion: the pattern at completion in France looks very similar to the pattern for early school-leaving. Minorities in France have not, as far as we can tell, slipped back in the course of their upper secondary school career and have sustained the gains made at the transition into upper secondary. However, we should remember that there is considerable differentiation between the different French Bacs, and there is evidence that minorities may be channelled into the lower-status technological Bacs (Brinbaum & Kieffer 2009).

Canada shows a very similar story to France, although in the case of Canada we have results only for test scores (column 1) and completion (column 4). As with France, we have several cases where significant negative coefficients in column 1 have become significant positive ones in column 4 (South Europeans) or non-significant coefficients in column 1 have become significantly positive (South Asians, Filipinos and ‘other’). There is just one minority moving in the opposite direction with the significant positive coefficient for Chinese with respect to test scores becoming non-significant (and effectively zero) at completion.3

(p.161) The USA also shows some evidence of minority progress. For several groups there is little or no change in the net coefficients between columns 1 and 4, but for Mexicans and other Latinos we see significant negative coefficients for test scores at the end of lower secondary replaced by nonsignificant coefficients for completion. There is also evidence that the second-generation black group (not to be confused with the long-established African American group) has moved from a significant ethnic penalty to a non-significant ethnic premium.4

Hints of progress

Belgium also offers some indications of progress although, as we can see, the coefficients for Moroccans and Turkish-origin groups are significantly negative in both columns 1 and 4. The evidence for progress is based solely on the magnitude of the coefficients, which is less convincing given the very different metrics involved at these two stages of the educational career.

The Netherlands also show considerable similarities with Belgium. While there are hints of a reduction in ethnic penalties between columns 1 and 4, with negative coefficients for the Moroccan and Caribbean groups in column 1 being followed by non-significant ones in column 4, we are not inclined to place too much reliance on this as there are negative signs for both the non-significant coefficients.5 For both Belgium and the Netherlands we are inclined to an ‘open’ verdict.

Little change overall

The results for England and Wales in contrast suggest that the positive ethnic premia observed with respect to early school-leaving (and choice of academic courses) are not sustained at completion. For completion, the net coefficients look very similar to those in column 1 for GCSE scores while the coefficients in columns 2 and 3 tend to be larger and more positive. Thus for black Africans we start in column 1 with a non-significant coefficient; this is followed by significant positive ones in columns 2 and 3, before returning to a non-significant one in column 4.

To be sure, the significant negative coefficient for black Caribbeans in column 1 ends up as a non-significant one in column 4, but this is matched by (p.162) a similar move in the opposite direction for the group of Pakistani origin. And there is no clear pattern of change for the Indian, Chinese or Bangladeshi groups.6 Overall it does look as though, in England and Wales, the very positive story told in Chapter 4 about high minority continuation rates is not sustained at the end of the school career.

Some minorities slipping back

Sweden was our most successful environment for minorities with respect to grades at the end of lower secondary education and it also has many examples of positive secondary effects of ethnicity (ethnic premia) with respect to choice of the academic track. However, there is evidence from Table 6.2 of a few minorities slipping back when it comes to completion. To be sure, for most groups, the net coefficients for grades, early school-leaving and completion are pretty similar to each other but there are some examples of moves towards greater disadvantage. Thus, for the groups with backgrounds in Iran or the Horn of Africa, significant positive coefficients at the first two stages have turned into non-significant ones at the second stage and for the groups from Iraq and other African backgrounds non-significant coefficients at the first two stages have turned into significant negative ones at completion. There is also movement in the ‘wrong’ direction for South and East Europeans. Moreover, there is not a single example of a move in the opposite direction.

It is also quite striking that the array of positive coefficients for track choice in column 3 is not mirrored by corresponding positive coefficients in column 4. A more detailed analysis reveals that the non-completion of some ethnic minority groups occurs within both the academic and vocational tracks in Sweden (Jonsson & Rudolphi 2010). Thus, the high minority ambitions and aspirations which we suspect lie behind the positive coefficients for track choice have not been able to manifest themselves in completion rates.7

(p.163) Conclusions

We need to be suitably cautious in our interpretation of the evidence, given the various methodological caveats which we have discussed earlier. It is certainly safe to say that for many groups there is considerable continuity between the patterns found with respect to test scores, early school-leaving and final completion of upper secondary courses. Minorities such as the Turkish and North African groups who experienced ethnic penalties at the earlier stages are still the ones most likely to experience them at completion, while the Asian groups who were most likely to have ethnic premia at the earlier stages are still the most likely to have premia with respect to completion. There is no compelling general evidence either of the cumulation of disadvantage across the upper second career, or of its mitigation. We do not see much in the way of the positive secondary effects of ethnicity at completion which were so evident with respect to choice of track.

On the other hand, there are some intriguing suggestions that in France, Canada and the USA (and perhaps in Belgium and the Netherlands) there are movements towards the reduction of ethnic penalties, while in Sweden the movements are, if anything, in the opposite direction. While there may well be methodological reasons for this, it is striking that the two countries with clearest evidence of mitigation of ethnic penalties are Canada and the USA, where there is a comprehensive system of upper secondary education, high overall rates of completion and the absence of high stakes examinations.

This is consistent with the arguments of Waters et al. (2013) about the roles of choice and selection. Where minorities are free to choose, as in the American and Canadian upper secondary schools, they appear to choose to persevere with their education. So it appears that in these systems we do see evidence of positive ‘secondary effects’ of ethnicity, potentially a result of the high aspirations accompanying positive selection as suggested in Chapter 5. This also seems to apply in the French system, where there are multiple options and routes to an upper secondary certificate.

However, the French system reminds us that there can be considerable differentiation within the broad category of ‘completion’ with its distinctions between academic, technological and vocational Bacs. While minorities may be successful in securing an upper secondary school-leaving certificate, there may well be qualitative differences between the certificates obtained by minorities and the majority. These in turn may have implications for the next stage of the educational career or for entry into the labour market. As Lucas (2001; 2009) argued in his thesis of ‘effectively maintained inequality’, the apparent closing of quantitative gaps between young people from different social backgrounds may mask emerging inequalities in access to qualitative distinctions between educational strata of differing prestige.

(p.164) References

Bibliography references:

Brinbaum, Y. & Kieffer, A. (2009), ‘Trajectories of Immigrants' Children in Secondary Education in France: Differentiation and Polarization’, Population, 64(3): 507–54.

Brinbaum, Y., Moguerou, L. & Primon, J.L. (2012), ‘Les enfants d'immigrés ont des parcours scolaires différenciés selon leur origine migratoire’, in Immigrés et descendants d'immigrés en France (Paris, INSEE Références—Édition 2012), 43–59.

Heath, A.F. & Cheung, S.-Y. (2007), Unequal Chances: Ethnic Minorities in Western Labour Markets, Proceedings of the British Academy 137 (Oxford, Oxford University Press).

Jonsson, J.O. & Rudolphi, F. (2010), ‘Structural Integration in Multi-ethnic Society: Educational Careers of Children to Immigrants in Sweden’, Manuscript (Swedish Institute for Social Research).

Kao, G. & Thompson, J. (2003) Racial and Ethnic Stratification in Educational Achievement and Attainment. Annual Review of Sociology, 29: 417–42.

Kieffer, A. (2008), ‘Applying the ISCED-97 to France: Some Issues and Propositions’, in S.L. Schneider (ed.), The International Standard Classification of Education (ISCED 97): An Evaluation of Content and Criterion Validity for 15 European Countries, (Mannheim, MZES), 103–24.

Lucas, S.R. (2001), ‘Effectively Maintained Inequality: Education Transitions, Track Mobility, and Social Background Effects’, American Journal of Sociology, 106: 1642–90.

Lucas, S.R. (2009), ‘Stratification Theory, Socioeconomic Background, and Educational Attainment: A Formal Analysis’, Rationality and Society, 21: 459–511.

Portes, A. & Rumbaut, R. (2001), Legacies: The Story of the Immigrant Second Generation (Berkeley, CA, University of California Press).

Schneider, S.L. & Kogan, I. (2008), ‘The International Standard Classification of Education 97: Challenges in the Application to National Data and the Implementation in Cross-national Surveys’, in S.L. Schneider (ed.), The International Standard Classification of Education (ISCED 97). An Evaluation of Content and Criterion Validity for 15 European Countries (Mannheim, MZES), 13–46.

Waters, M.C., Heath, A., van Tran, C. & Boliver, V. (2013), ‘Second-generation Attainment and Inequality: Primary and Secondary Effects on Educational Outcomes in Britain and the U.S.’ in R. Alba & J. Holdaway (eds), The Children of Immigrants in School (New York, New York University Press), 120–59.

(p.166)

Table A6.1. Gross and net effects of ethnic origins from probit analyses of completing upper secondary education by country and origin.

Belgium

Canada

England and Wales

France

Netherlands

Sweden

USA

Gross

Net

Gross

Net

Gross

Net

Gross

Net

Gross

Net

Gross

Net

Gross

Net

Majority group (reference)

0

0

0

0

0

0

0

0

0

0

0

0

0

0

W European

−0.05

−0.07

−0.21

−0.01

−0.34

−0.35

Finnish

−0.39

−0.15

Danish

−0.47

−0.21

Norwegian

−0.43

−0.19

S European

−0.45

−0.17

0.21

0.46

−0.27

0.14

−0.39

−0.14

E European

−0.22

−0.04

Polish

−0.13

0.07

Ex-Yugoslav

−0.27

−0.03

E Asian

0.15

0.37

0.57

0.58

Chinese

0.07

0.02

0.37

0.50

SE Asian

0.64

0.71

−0.21

0.15

0.14

0.41

S Asian

1.41

1.47

−0.19

−0.02

Indian

0.29

0.36

Pakistani

−0.23

−0.11

Bangladeshi

0.06

0.50

W Asian

−0.51

−0.24

Iranian

−0.19

0.04

Iraqi

−0.41

−0.18

Turkish

−0.68

−0.36

−0.63

−0.44

−0.49

−0.27

Caribbean

0.42

0.54

−0.38

−0.07

0.09

−0.01

−0.01

−0.07

0.01

0.17

S-S African

−0.23

−0.14

−0.34

0.12

−0.55

−0.29

N African

−0.61

−0.28

−0.26

0.27

−0.32

−0.16

−0.58

−0.31

From Africa's Horn

−0.38

0.04

S American

−0.48

−0.21

Chilean

−0.66

−0.33

Mexican

−0.57

−0.08

Filipino

0.75

0.66

Other Latino

−0.58

−0.22

Source

Census 1991–2001

YITS 2004

YCS cohort 10

Panel95Youth 2002

SIM 2006

Register STAR

ELS2002

Notes. Figures in bold are significant at the 0.05 level. In the USA, the ‘white’ group has been allocated to the West European row, and the ‘black’ group to the Caribbean row of the table.

Notes:

Proceedings of the British Academy, 196, 149–166. © The British Academy 2014

(1) However, we exclude qualifications at ISCED level 3C, such as the BEP and CAP in France, which are typically completed at an earlier age.

(2) We should be more cautious when comparing coefficients from columns 3 and 4 since early school-leavers are dropped from the sample when estimating the models from which column 3 coefficients are derived. It is possible that the column 4 coefficients are driven entirely by early leaving and that the qualitative differences in track choice observed in column 3 are maintained, although becoming invisible in column 4 because they are pooled together in a single category.

(3) The Canadian results are based on the YITS panel study, which had a modest degree of attrition and thus some potential for bias.

(4) As with Canada, however, these results are based on a panel study which inevitably suffered from attrition.

(5) The coefficients in columns 1–3 for the Netherlands are taken from a panel study, while that in column 4 is from an independent representative survey.

(6) The England and Wales coefficients shown in Table 6.2 are taken from the YCS panel study, which is subject to considerable attrition. However, we can use an alternative source, the Longitudinal Study (LS), for estimating completion and this yields very similar net coefficients, the one major difference being that in the LS there is a significant negative coefficient for black Caribbeans.

(7) We should note that the number of individuals in the dataset for the analysis of completion is only half that for grades. This will inevitably affect levels of significance. We should not, therefore, place any weight on a finding that a significant positive coefficient for grades has turned into a non-significant one for completion. However, moves from non-significant to significant negative coefficients can be accorded considerable weight.