International Journal of Educational Research Open 2 (2021) 100087

Contents lists available at ScienceDirect

International Journal of Educational Research Open

journal homepage: www.elsevier.com/locate/ijedro

Mind the gap: Examining migrant-native disparities in reading

performance and subjective wellbeing among 15-year-old students in
different education systems
Zi Wang1,∗
French Research Institute on East Asia, National Institute for Oriental Languages and Civilisations (INALCO), 2 Rue de Lille, 75007, Paris, France

a r t i c l e i n f o a b s t r a c t

Keywords: Contemporary European and North American societies have experienced changing demographics due to migra-
Happiness tion. School classrooms reflect this dynamic. To-date, few studies have examined migration effects on both read-
life satisfaction ing skills and subjective wellbeing from an international comparative perspective. Using the 2018 Programme
migration
for International Student Assessment (PISA) dataset, this paper addresses the issue by examining migrant-native
PISA 2018
gaps in reading, happiness, and life satisfaction among 15-year-old students in four education models. Regression
reading performance
subjective wellbeing results illustrate that migrant students in general tend to experience a double inequality: they have lower read-
ing performance and report lower life satisfaction when compared to native students (although in some cases
they report higher happiness levels). However, such effects are also context-dependent. While a migration back-
ground has significant negative effects on students’ reading in Germanic and Nordic schools, its negative effects
in Anglo-Saxon systems are above all manifested in students’ level of life satisfaction. In addition, compared to
first-generation migrants, being of second generation is positively associated with reading performance in all ed-
ucation models, with the strongest effect in the Nordic countries. Such positive generation effects are statistically
insignificant, on the other hand, on students’ wellbeing. Other individual factors such as socio-economic status
and gender present mixed correlations with reading and wellbeing. Such findings invite scholars and practition-
ers to reflect, through the prism of migration, on how reading proficiency and wellbeing can differ in diverse
educational approaches, and correspondingly seek context-specific intervention measures.

1. Introduction require special attention to their emotional and educational wellbeing

needs (Wang, 2020)
Contemporary societies are becoming increasingly diverse in the age To-date, few studies have examined migration effects on both read-
of global migration (Castles et al 2014; De Haas et al 2019). Schools are ing skills and subjective wellbeing from an international comparative
no exception, as urban classrooms in Europe and North America reflect perspective. This paper fills the gap by examining the education phe-
the changing demographics of the society at large. Yet, education sys- nomenon of migrant-native disparity in reading and subjective well-
tems are often blamed for their slow response to new social realities and being. The present study situates this issue in a cross-country com-
hence referred to as a major weakness in integration policies (Solano and parative global context. Migrants who are better integrated education-
Huddleston 2020). Research on youths with migration background has ally (in terms of performance and feelings of wellbeing) also have
shown that they are more likely to be disadvantaged vis-à-vis their na- higher chances of later-stage economic assimilation in host societies
tive peers in the domain of education (Heckmann 2008). Such a gap can (Schneeweis 2011). Against this backdrop, this paper contends that
be manifested in diverse forms: academic achievement, enrolment in migrant-native gaps in reading skills (and by extension, literacy) and
type of schools, duration of schooling, dropout rates, as well as the type wellbeing are major issues in a society’s attempts to formulate sustain-
of final qualification attained (Entorf and Minoiu 2005; Schnepf 2007). able migration and integration policies.
In addition, it has also been established that migrant students experi- Analysing data from the 2018 Programme for International Stu-
ence a wellbeing deficit compared to native students: they either tend dent Assessment (PISA), this study has the following objectives:
to report lower levels of affective happiness (Rodriguez et al 2020) or (i) to recognise patterns and indicators of inequality in reading

∗
Corresponding author.
E-mail addresses: [email protected], [email protected]
1
ORCID: https://orcid.org/0000-0001-7008-9627

https://doi.org/10.1016/j.ijedro.2021.100087
Available online 18 November 2021
2666-3740/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

performance and subjective wellbeing between migrants and natives, secondary phases strongly associate with employment, income, and
and (ii) identify which type of educational approach exacerbates or health in later stages of life (Cutler and Lleras-Muney 2006; Lange and
mitigates the migrant-native gaps in reading and subjective wellbeing. Topel 2006). Importantly, for countries with sizeable migrant popula-
Reading skills and international migration experience are inextricably tions and/or youths with migration backgrounds, education is key to
linked, and reading was the key assessment focus in PISA 2018. Happi- their integration and future success in social mobility (Machin et al
ness and life satisfaction are strong indicators of student wellbeing (see 2012). Understandably, migrant-native gaps in schools, whether in
section 3.2.1 for details). Against this backdrop, this research addresses terms of academic performance or wellbeing, deserve the attention of
the following questions: social scientists and policymakers alike. To what extent is there a perfor-
• To what extent is there a gap in reading performance and subjective mance gap today? This section presents visualisations of the PISA 2018
wellbeing between migrant and native students in major migrant data on mean reading scores and responses to the questions on happi-
destination countries in Europe and North America? ness and life satisfaction according to migrant and native populations.
• How is the migrant-native gap in reading and students’ subjective Such direct comparisons of reading and subjective wellbeing be-
wellbeing manifested in different education systems? tween migrants and natives at the outset enable the observation of over-
all patterns. Notably, it allows us to observe the extent to which gaps in
Encompassing all three aspects of migration, reading, and subjective reading and subjective wellbeing to migrants’ disadvantage exist among
wellbeing, results strongly suggest that, on the whole, migration has the PISA 2018 participating countries/regions. Representing students’ read-
biggest negative effect on students’ reading performance, followed by ing performance (Fig. 1) and the two dimensions of subjective wellbeing
life satisfaction, while it has a slightly positive effect on happiness. Such (Figs. 2 and 3) responses as they are, a migrant-native gap, to migrant
findings notwithstanding, this article posits that different education sys- students’ disadvantage, can be observed in a majority of countries in all
tems provide conducive contexts in different ways (reading or wellbe- three areas of reading, happiness, and life satisfaction:
ing) for migrant students. Based on the comparative analysis, I show The three Figures above represent the state of affairs concerning gaps
that the Anglo-Saxon approach (Ireland, UK, USA) proves conducive in reading scores and subjective wellbeing between migrant and native
in achieving parity in reading performance and level of happiness be- 15-year-old students in the entire PISA 2018 sample. The numerical val-
tween migrants and natives. Students in the Germanic-influenced (Aus- ues for each bar indicate the difference in mean reading score (Fig. 1),
tria, Germany, Netherlands, Switzerland) and Nordic (Finland and Swe- happiness (Fig. 2), and life satisfaction (Fig. 3) levels between natives
den) models, in turn, are less likely to experience a life satisfaction gap and migrants. A positive value indicates a higher mean for migrants, a
due to their migration status. Results also reveal that, across the board negative value indicates a lower mean for migrants. Generally speaking,
(but especially in the Nordic system), being of second-generation is pos- migrants fared worse than natives in reading, life satisfaction and hap-
itively associated with higher reading scores vis-à-vis first-generation piness in a majority of countries who participated in PISA 2018. This
migrants. This generation effect on subjective wellbeing is, on the other implies that migrant students in most countries are indeed subject to a
hand, insignificant. Furthermore, effects of individual factors, notably double inequality: in both reading and subjective wellbeing. Particu-
speaking a different language at home (often a proxy of international mi- larly noteworthy is the group of Western European, Nordic, and North
gration) and grade retention, on reading and wellbeing display strengths American countries. In these richest and most highly developed regions
and significance to varying degrees in different systems. in the world, the migrant-native gaps are persistent. This issue is all the
Findings of this study inform context-specific policy formulations more topical because these countries’ classrooms have sizeable youths
in mitigating inequality between migrants and natives in both reading of migration background. The next section reviews current scholarship
competence and subjective wellbeing. Education performance and well- on this issue.
being during adolescent years play a crucial role in later-stage achieve-
ments and quality of life (Eurostat 2019; Heinrich and Galan 2018).
For migrant youths, the stakes are even higher, as education integra- 2.1. Review of current scholarship
tion in schools impacts later-stage economic activity and assimilation in
the receiving country (Schneeweis 2011). This article’s conclusions con- A burgeoning trend of scholarship at the crossroads of migration and
tribute to scholarly debates and the formulation of more inclusive poli- education examined effects of migration on various aspects of youths’
cies aimed at alleviating the migrant-native gap in both reading literacy educational attainment. Mainly set in the context of rich European,
and wellbeing domains. More inclusive policies that take into account North American and Oceanian countries (among the largest migrant-
vulnerable groups (in this case migrants) are also more sustainable poli- receiving states), migration status in schools is broadly operationalised
cies in the promotion of education achievement, integration, and quality in the literature as students with an immigration background. This en-
of life in school in general. compasses both first-generation (foreign-born students with foreign-
This paper is organised as follows: Section 2 provides an overview born parents), and second-generation (students born locally but with
of the current reading and subjective wellbeing gaps between migrants both parents foreign-born) students1 (Frones et al 2020; OECD 2019).
and natives according to PISA 2018 data. I then review extant literature The broad consensus established by extant literature is that migra-
which thematised migration, academic performance, student wellbeing, tion status negatively correlates with academic performance (Dockery,
as well as comparative education. The section shows that empirical stud- Koshy, & Li, 2020; Heckmann, 2008; Jonsson & Rudolphi, 2011;
ies examining the nexus of migration, aspects of academic performance, Rodríguez, 2020; Volante, Klinger, & Bilgili, 2018). For instance,
and subjective wellbeing from an international comparative perspective Frones et al (2020) highlighted a general and consistent gap in read-
are still few and far between. In section 3, I present the sample, explain ing performance, to the disadvantage of immigrant minority students,
the variables adopted in the analyses, as well as empirical strategies in Denmark, Sweden and Norway throughout the three rounds of PISA
utilised in this study. Section 4 presents the results of empirical anal- in 2000, 2009, and 2018. In more specific terms, it has been found that
yses. Section 5 discusses research and policy implications of findings migrant students’ country of origin has an impact on their performance
beyond the confines of this work, limitations of the present study, as too. In Sweden, students of non-European migration origin have the low-
well as recommendations for future research. est school performance compared to natives and migrants of European
origin (Jonsson and Rudolphi, 2011). This effect need not be negative,
2. The migrant-native gaps in reading and subjective wellbeing: however, when it comes to aspirations: among the same group of non-
Evidence from PISA 2018 European migrant students who could enrol in upper secondary schools,

Education is crucial in the development of human capital. It is

1
acknowledged that educational attainments in secondary and post- This study adopts a similar approach to migration status. See section 3.

2
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

Fig. 1. Comparison of mean reading scores between natives and migrants among PISA 2018 countries/regions.

Fig. 2. Comparison of happiness between natives and migrants among PISA 2018 countries/regions.

a higher proportion of them chose the academic track, compared to na- grant students negatively impact academic achievement (Froehlich et al
tives and students of Nordic origin who displayed equal preference for 2018), so much so that even as the majority student population’s per-
academic and vocational programmes (Jonsson and Rudolphi, 2011). formance improved significantly since PISA 2000, performance inequal-
Along these lines, Moon et al (2009) emphasised different predictors ities to the disadvantage of migrant and socio-economic minorities still
of academic achievement for different migrant groups in the American persist (Davoli and Entorf 2018).
context. Whereas acculturation, parenting style, and parents’ income If the generally negative effect of migration persists in academic
are predictors of Korean migrant students’ academic achievement in performance, does it also impact students’ subjective wellbeing? On
American schools, parent education and length of stay in the country the level of education systems, Montt and Borgonovi (2018) argue
proved better predictors for the Mexican sample. Cultural influence in that most countries can only be successful in one area: either nur-
migrants’ country of origin is thematised by Drockery et al (2020), who turing high academic (mathematics) performance (the case of East
found that ‘children from migrant backgrounds affording lower school- Asian systems) or high levels of student wellbeing (Austria, Norway,
ing access to children of their own gender achieve lower scores’ in all Spain). Insofar as wellbeing as a predictor of achievement is concerned,
three PISA domains (Drockery et al 2020: 39). In Germany, research Samuel et al (2013) posit that subjective wellbeing (happiness and
has shown how educators’ negative stereotypes about non-European mi- life satisfaction) predict, to a large extent, intergenerational education

3
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

Fig. 3. Comparison of life satisfaction between natives and migrants among PISA 2018 countries/regions.

attainment in the case of Switzerland. However, in the afore-mentioned proach, the sample countries were categorised into four clusters of ed-
state-of-the-art, the migration perspective is often not considered. ucation typology2 with respect to diversity management:
Indeed, studies at the intersection of migration, academic perfor-
mance, and aspects of wellbeing/satisfaction/happiness are few and far • The early tracking system: Austria, Germany, Luxembourg, Nether-
between. Yet, to understand if migrant students have a higher tendency lands, Switzerland.
to experience a double inequality (as regards academic performance and • The common core-uniform integration system: France, Greece, Italy,
subjective wellbeing), research focusing on all three aspects are needed. Portugal, Spain
Among the few studies that did, Rodriguez et al 2020 used the PISA • The common core-flexible integration system: Ireland, United King-
2018 sample for Spain to show the existence of such a performance dom (UK), United States of America (USA)
inequality as regards mathematics and science, with migrants scoring • The individualised system: Finland, Sweden
lower. Moreover, natives experience higher positive affect (happiness).
Countries in the early tracking system (mostly Germanic-influenced)
The migrant-native gap in life satisfaction, however, was negligible. On
use tracking at the end of primary school as an early selection process
the one hand, such studies analysing single-country data have the advan-
to sort students into different pathways according to their academic
tage of detailing more nuanced effects of migration on both academic
results (Mons 2007). In the common core-uniform integration system
skills and student wellbeing, and to further observe inter-generational
(mostly Southern European), students follow a relatively standardised
differences among migrant students. These are insightful, as they are set
core education curriculum until 14-16 years old, with grade repetition
against the backdrop of national socio-economic, political and cultural
as the single most frequent measure to separate students with weaker
specificities.
academic abilities from the cohort. In contrast, schools in the common
On the other hand, single-country case studies do not allow for the
core-flexible integration system (mostly Anglo-Saxon) attend to diverse
observation of generalisable patterns across different education sys-
academic abilities by flexibly grouping students either within classes or
tems. For instance, do the afore-mentioned findings on the migrant-
within schools, without resort to hard and fast early tracking or grade
native gaps in academic performance and positive affect manifest dif-
repetition practices. Students in this system are also exposed to a com-
ferently when we enlarge the sample size of countries and educational
prehensive core curriculum until 15-16 years old. Lastly, schools in the
approaches? The current research landscape still lacks studies that ad-
individualised system (Nordic countries) rarely adopt tracking, flexible
dress such questions on the broader scale. This paper addresses the issue
grouping within schools, or grade retention to deal with heterogeneity.
by analysing the 2018 PISA dataset to answer the set of questions set
Instead, they rely more on individualised measures, such as individual
out in section 1 above. Adopting a multi-country comparative approach
or small-group tutoring, to ensure all students progress (Dupriez et al
has the advantage of allowing for the examination of large-range insti-
2008; Mons 2007).
tutional variations that is absent within single countries (Hanushek and
Along these lines of classification, studies using older PISA datasets
Woessmann 2011).
established that the migration disadvantage in reading and mathematics
is larger in continental European countries (early tracking and common
2.2. Comparative education systems
core-uniform integration systems) than in English-speaking countries
(common core-flexible system), while socio-economic status has the
Set against the backdrop of comparative analysis of different or-
largest impact on reading in English-speaking countries (plus Germany)
ganisations of school systems, this study focuses on a selection of 15
and least in Nordic societies (Entorf and Minoiu 2005; Schnepf 2007).
countries taken from the PISA 2018 dataset. These countries form clus-
Due to the lack of focus on subjective wellbeing in these studies, such
ters which display different heterogeneity management (managing stu-
results do not allow us to conclude if an education system that is
dents’ heterogenous academic ability) practices (Dupriez et al 2008;
good in alleviating migrant disadvantage in, say, reading performance
Mons 2007). This paper terms such a characteristic as diversity manage-
is also conducive in mitigating the wellbeing gap between migrants
ment. Diversity management in this context refers to students’ diverse
and natives. It is expected that education systems with such different
academic abilities. On this basis, the selected countries’ education sys-
tems are classified based on (i) how they provide a core curriculum (to
whom and until what age), and (ii) how they manage diversity in stu- 2
Note that the naming of categories is based on how education systems deal
dents’ academic ability (by early tracking, by standard measures like with their students’ diverse academic ability in the context of this research. This
grade repetition, or via individualised assistance). Following this ap- is by no means the only way to name such education systems.

4
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

institutional mechanisms in dealing with students’ diverse aptitudes will Using mean PVs in analyses might lead to (i) underestimation of stan-
have different impacts on migrant students’ reading and aspects of well- dard deviation, (ii) underestimation of within-school variance, and (iii)
being. overestimation of correlations between student performance and some
background variables (OECD 2009: 128). The first two issues, however,
are of little concern to this research. First, standard deviations and stan-
3. Data, variable descriptions, and methods
dard errors are averaged out due to the high number of observations in
the regression models of this study. Second, school-level factors are not
3.1. Sample: Programme for International Student Assessment (PISA)
part of the analysis, hence within-school variance is of little concern. The
2018
only potential issue is the third point. To address this concern, all 10 PVs
were separately tested in the regression model where reading score is
The Organization for Economic Cooperation and Development
the dependent variable. Results were then compared to the same model
(OECD) administers the large-scale PISA study in three-year intervals.
where the mean reading score (derived from the 10 PVs) was used. The
The goal is to assess key skills in reading, mathematics, and science,
comparison shows no change in both the significance and direction of
as well as various aspects of educational competencies of 15-year-olds
coefficients, and only slight and insignificant changes in the magnitudes
among participating countries and regions. The international and stan-
of the coefficients in a minority of cases. Crucially, these changes all lie
dardised nature of assessment also facilitates cross-country comparisons
within the 95% confidence interval. Hence, insofar as objectives of this
to improve educational policies and outcomes (Gomez-Fernandez and
study are concerned, using mean PVs does not lead to an overestima-
Mediavilla 2021). To-date, it is one of the largest samples compris-
tion of correlations between student performance and the background
ing students from diverse national, cultural, linguistic, and educational
variables used in this study.
backgrounds (Borgonovi and Han, 2021). Data used in this study come
Together with reading performance, subjective wellbeing is an im-
from the latest round of PISA 2018, and are all extracted from pub-
portant analytical outcome in this article. This study understands subjec-
licly available data files on OECD’s website. The subset of 15 countries
tive wellbeing as manifested by the following two indicators: life satis-
(mentioned in section 2.2) in this study adds to a total of approximately
faction and happiness. This is in line with current practices in the empiri-
118,000 students in the 15 countries for the analytic sample.
cal operationalisation of subjective wellbeing in social sciences research
(Happiness Research Institute 2020; Veenhoven 2012; 2015). In current
3.2. Variable description literature, happiness relates more to emotional moods (hence affective
dimension of wellbeing), whereas life satisfaction involves reflections on
3.2.1. Outcome variables one’s achievements and possessions and comparisons with other refer-
To test for correlations between migration on the one hand, and read- ence groups (hence cognitive dimension of wellbeing) (Adler and Selig-
ing and subjective wellbeing on the other hand, multiple linear regres- man 2016; Happiness Research Institute 2020). It is expected that these
sions were utilised. Hence, we do not only examine the relationship two dimensions yield different results as regards the migrant-native gap.
between two variables (e.g., migration and reading) in model 1, but Findings hence contribute to further theorising in extant literature on
crucially, how such a relationship holds when effects of other variables subjective wellbeing by showing how different dimensions of life con-
(e.g., socio-economic status, gender) were controlled in model 2. This tentment (whether affective of cognitive) correlate to different indi-
study has three outcome variables: reading performance, happiness, and cators. This study recognises that both happiness and life satisfaction
life satisfaction. Reading performance is used in this study for the follow- fluctuate in natural settings and the accuracy of individual respondents
ing two reasons. First, among the three competency areas tested by PISA, is hard to confirm. However, given such a representative sample size,
international migration experience of students and their families has reliability limits are expected to average out (Kahneman and Krueger
a more direct and significant impact on language-based reading skills 2006).
(Frones et al 2020). This is because linguistic knowledge (vocabulary In PISA 2018, students were asked the following question on hap-
and grammar) is more closely associated with reading comprehension piness: “Thinking about yourself and how you normally feel, how of-
(taught in language and humanities-based classes) than in mathematics ten do you feel happy?” Responses are computed on a 1-4 scale from 1
and science. It is expected that international migration experience, cou- “never”, 2 “rarely”, 3 “sometimes” to 4 “always”. As for life satisfaction,
pled with insufficient host society language knowledge (also the variable students were asked: “Overall, how satisfied are you with your life as a
“language at home”, see below), result in a reading performance gap. whole these days?” The original responses in the dataset ranged from 0
PISA 2018 defined reading literacy as understanding, using, evaluating, to 10, with increasing values corresponding to higher levels of life satis-
reflecting on and engaging with texts in order to achieve one’s goals, faction. In this study, life satisfaction was rescaled to match the 4-point
to develop one’s knowledge and potential, and to participate in society scale for happiness. Hence, for life satisfaction, answers from 0 to 4 were
(OECD 2019). Understood in this way, reading proficiency measures stu- grouped as 1 “Not satisfied”, 5 and 6 as 2 “Somewhat satisfied”, 7 and
dents’ ability to engage in a wide variety of activities both within and 8 as 3 “Moderately satisfied”, and 9 and 10 as 4 “Very satisfied”, as per
beyond the confines of their classroom. Potential reading gaps hence OECD’s practice (OECD 2019).
underscore future challenges in leading a socially active and enriching
life.
Second, reading was the main focus of academic competency in 3.2.2. Key independent and control variables
PISA 2018. Hence, compared to mathematics and science, a wider The key independent variable in this study is migration status. Fol-
range of tasks in reading than mathematics or sciences was assessed lowing OECD’s categorisation, a student is classified as immigrant if he
(Borgonovi and Han, 2021). What is more, all participating 15-year-olds or she was born outside the test country with two foreign-born parents
received at least some reading assessment, whereas not all were assessed (first-generation immigrant student) or if he or she was born in the test
in mathematics or science. In the interest of generating as rich a set of country but whose both parents were foreign-born (second-generation
analyses as possible, and simultaneously remaining relevant and com- immigrant student) (OECD 2019). As per OECD definition, those who
parable across the board, reading performance was therefore the most do not belong to either category was classified as non-immigrant native
suitable choice insofar as the aims of this study are concerned. Individ- student in this analysis. For the first set of regressions (Tables 2-5), a
ual mean reading scores for each respondent were computed using the dichotomous indicator was used to show if a student is a native (coded
average of 10 plausible values (PVs) for the reading assessment. The as 1) or migrant (coded as 2).
author is aware of the three potential issues for averaging the PVs and In addition, the following control variables and coding were utilised
using such mean values in analyses, as raised by the OECD (OECD 2009). in this study. Gender is defined by 1 for “female” and 2 for “male”

5
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

Table 1 selections and empirical strategy. The aim is to examine effects of and
Overview of descriptive statistics of main variables used in the study. relationships between migration, reading, and subjective wellbeing from
Variable N Mean Std. Dev. Min Max a comparative perspective of different education systems.
To find out, this study utilised a three-step process comprising mul-
Outcome variables
tilinear regression analyses. A first regression was conducted for the
Reading score 118579 494.848 93.320 70.284 816.441
Happiness 118579 3.37 0.630 1 4 whole sample to observe general patterns of the migrant-native in-
Life satisfaction 118579 2.85 1.022 1 4 equality. Next, separate regressions for each outcome variable from a
Key independent variable four-cluster comparative perspective were administered. Lastly, an addi-
Migration status 117256 1.15 0.361 1 2
tional regression analysis compared how second-generation local-born
Control variables
Gender 118579 1.50 0.500 1 2
students (of migrant parentage) fared against first-generation migrant
Language use at home 118460 1.17 0.373 1 2 students on the three outcome variables. The inter-group results were
Fear of failure 116919 -0.052 1.008 -1.894 1.891 then compared with each other, as well as against the backdrop of the
Grade repetition 118289 0.15 0.356 0 1 broader picture. In all regressions, direct correlations between the inde-
Family wealth 118483 0.132 0.866 -7.035 4.753
pendent variable (migration) and the three outcome variables (reading,
ESCS 117468 0.033 0.971 -5.476 3.717
Student’s expected ISEI 95854 65.004 19.835 11.01 88.96 happiness, life satisfaction) were tested in model 1, whereas model 2
checked for effects of key control variables. Findings derived from such
an analytical approach enhance our understanding of both the social re-
students3 . It is well established in extant scholarship that female stu- ality of gaps in reading performance and subjective wellbeing migrants
dents not only lack behind male counterparts in mathematics and sci- face, as well as the complexity of such inequalities manifested in differ-
ence performance, but also display higher levels of fear of failure, es- ent education systems.
pecially in high-achieving contexts (Borgonovi & Han, 2021; Gamazo &
Martínez-Abad, 2020; Karakolidis, Pitsia, & Emvalotis, 2016). However,
the performance gap is reversed to females’ advantage when it comes to 4. Results
reading (Gamazo and Martinez-Abad 2020). Yet, the clear-cut effects of
gender on both reading and subjective wellbeing are not yet well stud- 4.1. Migration status and inequality in reading competence and subjective
ied when migration background is considered. Individuals’ language use wellbeing: general patterns
at home ranges from 1 for “test language” and 2 for “other languages”.
This could act as another indicator for international migration experi- Table 2 below comprises results of the first regression with the total
ence. Grade repetition distinguishes between 0 for “no repetition” and sample of all country cases:
1 for students who have repeated a grade. The three variables of fear of Results in Table 2 answer the first research question: To what extent
failure, family wealth and economic, social and cultural status index4 is there a gap in reading performance and subjective wellbeing between
(ESCS) were created by PISA and standardised to take a mean of 0 and native and migrant students in major migrant destination countries in
standard deviation of 1. Hence, aligning students’ individual answers Europe and North America? The general picture emerging from the anal-
on this scale, positive scores represent higher fear of failure, more fam- ysis is that, all things considered, migrant students experience a double
ily wealth, and a higher score on the ESCS index in comparison to the gap in both reading performance and life satisfaction (model 2). Regard-
OECD mean. Lastly, students’ expected occupation, represented by the less of education systems, having a migration background is associated
International Socio-Economic Index of Occupational Status (ISEI index), with a 12.82-point drop in mean reading score, as well as a 0.077 drop
is included to consider effects of students’ aspirations on reading perfor- on the life satisfaction scale. Such negative correlations are statistically
mance and subjective wellbeing. In the PISA 2018 dataset, the students’ significant at the 99.9%. However, there is a caveat. Results in model
expected ISEI index reflects the status of occupation of either a respon- 2 also indicate that, for both reading and life satisfaction, the effect of
dent’s father or mother, whichever is higher. It is computed by the OECD migration status does not have the strongest magnitude once we con-
and ranges from a low of 11.01 (corresponding to “street cleaner”) to sider other factors. For reading, a one-unit change in grade repetition,
a high of 88.96 (corresponding to “judge”). These additional indicators the ESCS of a family, as well as the language spoken at home all result in
are pertinent because they imply socially disadvantaged positions in ad- a stronger magnitude of change in reading scores than migration status
dition to, or in combination with, migration. Thus, including them al- does (at the same 99.9% significance level). While ESCS has a positive
lows for more nuanced findings. It is instructive to highlight here that correlation with reading, grade repetition and a different language spo-
this paper focuses on effects of individual level factors and migration ken at home have negative correlations.
experience/background on the said outcome variables. Table 1 below For life satisfaction, a one-unit change in happiness, gender, and fear
provides an overview of all the outcome, key independent, and control of failure is associated with larger effects on life satisfaction (at the same
variables included in this study: 99.9% significance level) than migration status does. While happiness
Multicollinearity tests were performed on the set of independent and has a positive correlation with life satisfaction, gender (being a male)
control variables to check for correlations among them. With the vari- and fear of failure have negative correlations.
ance inflation factor (VIF) values ranging between 1.009 and 1.610, it The situation is different for happiness, since migration status is pos-
is concluded that this study’s regression models do not contain indepen- itively associated with happiness (model 2) when controlled for other
dent and control variables that are highly correlated to each other. factors. This positive relationship is, however, on a lower significance
level (95%) than in the case of reading and life satisfaction (both 99.9%).
3.3. Analytical approach
4.2. Migrant-native gaps in different education systems
Comparisons of the migrant-native gap in reading and subjective
wellbeing in section 2 served as the basis for the subsequent variable The second analytical step was a comparative analysis of four edu-
cational clusters based on the diversity management approach. Results
3
The other responses for gender: “valid skip”, “not applicable”, “invalid”, and
presented in Tables 3, 4, and 5 below answer the second research ques-
“no response” were excluded as they would not yield meaningful analyses. tion: How is the migrant-native gap in reading and students’ subjective
4
Aggregate indicator reflecting students’ socio-economic and cultural status. wellbeing manifested in different education systems?
This is compiled by PISA based on responses on parents’ educational attainment, Table 3 illustrates education system comparisons for reading perfor-
occupation, and the availability of a range of material resources at home. mance. Results in model 1 reveal that migrant students experience the

6
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

Table 2
Results of multilinear regression for the whole sample. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001

Reading Happiness Life satisfaction

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

∗∗∗ ∗∗∗ ∗∗∗ ∗ ∗∗∗
Migration status -37.941 -12.820 -0.039 0.012 -0.122 -0.077∗ ∗ ∗
Reading 0.00001 0.0001∗ ∗ ∗
Happiness 0.217 0.714∗ ∗ ∗
Life satisfaction -1.217∗ ∗ ∗ 0.287∗ ∗ ∗
Gender -12.322∗ ∗ ∗ -0.068∗ ∗ ∗ -0.181∗ ∗ ∗
Language use at home -20.314∗ ∗ ∗ -0.035∗ ∗ ∗ 0.023∗ ∗
Fear of failure 2.720∗ ∗ ∗ -0.038∗ ∗ ∗ -0.180∗ ∗ ∗
Grade repetition -61.693∗ ∗ ∗ 0.013∗ -0.060∗ ∗ ∗
Family wealth -9.569∗ ∗ ∗ 0.028∗ ∗ ∗ 0.034∗ ∗ ∗
ESCS 23.922∗ ∗ ∗ -0.013∗ ∗ ∗ 0.041∗ ∗ ∗
Student’s expected ISEI 1.178∗ ∗ ∗ 0.001∗ ∗ ∗ 0.0004∗ ∗
N 117255 94716 117255 94716 117255 94716
R-squared 0.022 0.280 0.001 0.234 0.002 0.276

Table 3
Results of multilinear regression for the four-group comparison on reading. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001

Early tracking Common core-Flexible integration Common core-Uniform integration Individualised system

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗
Migration status -51.924 -19.208 -7.571 1.661 -39.627 -15.379 -81.364 -51.448∗ ∗ ∗
Happiness -0.867 1.515 1.290∗ 5.998∗ ∗ ∗
Life satisfaction 0.414 -2.622∗ ∗ ∗ -1.773∗ ∗ ∗ 2.086
Gender -13.488∗ ∗ ∗ -7.647∗ ∗ ∗ -10.542∗ ∗ ∗ -23.933∗ ∗ ∗
Language use at home -31.993∗ ∗ ∗ -26.426∗ ∗ ∗ -11.851∗ ∗ ∗ -42.214∗ ∗ ∗
Fear of failure 2.507∗ ∗ ∗ 6.248∗ ∗ ∗ 1.707∗ ∗ ∗ 5.543∗ ∗ ∗
Grade repetition -48.818∗ ∗ ∗ -57.372∗ ∗ ∗ -65.874∗ ∗ ∗ -61.019∗ ∗ ∗
Family wealth -12.962∗ ∗ ∗ -13.105∗ ∗ ∗ -8.665∗ ∗ ∗ -15.340∗ ∗ ∗
ESCS 26.385∗ ∗ ∗ 30.989∗ ∗ ∗ 17.734∗ ∗ ∗ 25.815∗ ∗ ∗
Student’s expected ISEI 1.430∗ ∗ ∗ 1.302∗ ∗ ∗ 1.147∗ ∗ ∗ 1.258∗ ∗ ∗
N 24056 18621 22409 18030 60502 50224 10285 7724
R-squared 0.057 0.344 0.001 0.215 0.019 0.289 0.075 0.299

Table 4
Results of multilinear regression for the four-group comparison on happiness. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001

Early tracking Common core-Flexible integration Common core-Uniform integration Individualised system

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

∗∗∗ ∗∗ ∗∗∗ ∗
Migration status -0.037 0.015 0.013 0.044 -0.063 0.020 -0.030 0.033
Reading score 0.00004 0.00006 0.00007∗ 0.0003∗ ∗ ∗
Life satisfaction 0.303∗ ∗ ∗ 0.285∗ ∗ ∗ 0.278∗ ∗ ∗ 0.324∗ ∗ ∗
Gender -0.116∗ ∗ ∗ -0.095∗ ∗ ∗ -0.044∗ ∗ ∗ -0.083∗ ∗ ∗
Language use at home -0.034∗ ∗ -0.016 -0.031∗ ∗ ∗ -0.014
Fear of failure -0.042∗ ∗ ∗ -0.042∗ ∗ ∗ -0.040∗ ∗ ∗ -0.050∗ ∗ ∗
Grade repetition -0.034∗ ∗ -0.030 0.013 0.012
Family wealth 0.029∗ ∗ ∗ 0.006 0.045∗ ∗ ∗ 0.030∗ ∗ ∗
ESCS -0.008 0.018∗ ∗ -0.017∗ ∗ ∗ 0.010
Student’s expected ISEI 0.0003 -0.0001 0.0003∗ 0.001
N 24056 18621 22409 18030 60502 50224 10285 7724
R-squared 0.001 0.249 0.000 0.264 0.0001 0.218 0.000 0.285

Table 5
Results of multilinear regression for the four-group comparison on life satisfaction. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001

Early tracking Common core-Flexible integration Common core-Uniform integration Individualised system

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ ∗
Migration status -0.082 -0.010 -0.066 -0.075 -0.212 -0.112 -0.062 0.058
Reading score 0.00005 0.0003∗ ∗ ∗ 0.0002∗ ∗ ∗ 0.0002
Happiness 0.719∗ ∗ ∗ 0.803∗ ∗ ∗ 0.681∗ ∗ ∗ 0.731∗ ∗ ∗
Gender 0.240∗ ∗ ∗ 0.186∗ ∗ ∗ 0.156∗ ∗ ∗ 0.252∗ ∗ ∗
Language use at home 0.012 0.025 0.008 -0.104∗
Fear of failure -0.159∗ ∗ ∗ -0.192∗ ∗ ∗ -0.150∗ ∗ ∗ -0.207∗ ∗ ∗
Grade repetition -0.105∗ ∗ ∗ -0.012 -0.075∗ ∗ ∗ -0.062
Family wealth 0.029∗ ∗ ∗ 0.056∗ ∗ ∗ 0.064∗ ∗ ∗ 0.0005
ESCS 0.029∗ ∗ ∗ 0.038∗ ∗ ∗ 0.029∗ ∗ ∗ 0.051∗ ∗ ∗
Student’s expected ISEI 0.0004 0.001∗ ∗ 0.001∗ ∗ 0.001∗
N 24056 18621 22409 18030 60502 50224 10285 7724
R-squared 0.001 0.287 0.000 0.308 0.004 0.250 0.000 0.338

7
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

Table 6
Results of multilinear regression for the four-group comparison on reading, with the migration variable comparing second-generation to first-generation
migrants. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001

Early-tracking system Common core/Flexible integration system Common core/Uniform integration system Individualised system

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

∗∗ ∗∗∗ ∗ ∗∗∗ ∗∗∗ ∗∗∗
Second-generation 3.942 6.641 13.762 7.301 16.219 7.455 47.587 34.269∗ ∗ ∗
Happiness -1.657 1.127 0.697 4.955
Life satisfaction 0.382 -2.542 -1.442 2.526
Gender -12.764∗ ∗ ∗ -7.036∗ -10.453∗ ∗ ∗ -22.735∗ ∗ ∗
Language use at home -39.319∗ ∗ ∗ -20.602∗ ∗ ∗ -13.744∗ ∗ ∗ -62.539∗ ∗ ∗
Fear of failure 2.563∗ 6.128∗ ∗ ∗ 1.870 6.733∗
Grade repetition -48.606∗ ∗ ∗ -55.319∗ ∗ ∗ -65.237∗ ∗ ∗ -42.787∗ ∗
Family wealth -13.414∗ ∗ ∗ -14.479∗ ∗ ∗ -8.354∗ ∗ ∗ -18.664∗ ∗ ∗
ESCS 28.908∗ ∗ ∗ 32.022∗ ∗ ∗ 18.456∗ ∗ ∗ 29.501∗ ∗ ∗
Student’s expected ISEI 1.381∗ ∗ ∗ 1.263∗ ∗ ∗ 1.127∗ ∗ ∗ 1.147∗ ∗ ∗
N 6733 5445 3349 2761 6521 5378 1199 955
R-squared 0.000 0.337 0.005 0.211 0.008 0.284 0.061 0.302

largest performance gap in the individualised education system (Fin- Finally, concerning life satisfaction (Table 5), results in model 1 show
land, Sweden), followed by in the early tracking system (Austria, Ger- that migrant students report lower life satisfaction than natives in all
many, Luxembourg, Netherlands, Switzerland), in the common core- country groups, although this negative correlation is less significant in
uniform integration system (France, Greece, Italy, Portugal, Spain), and the individualised system. Comparing results in models 1 and 2, it tran-
the smallest gap in the common core-flexible integration system (Ire- spires that migration status contributes to the migrant-native gap in life
land, UK, USA). Comparing results in models 1 and 2, it is observable satisfaction only in the two common core education systems. In the early
that the migrant-native gap in reading, to migrants’ disadvantage, re- tracking and individualised systems, the effects of migration on life satis-
mains statistically significant and quantitatively important when con- faction lost significance. Happiness (the other aspect of subjective well-
trolling and not controlling for background variables in the individu- being in this paper) displays the strongest effect on and positive corre-
alised, early tracking, and the common core-uniform integration sys- lation with life satisfaction in all groups. Just like in happiness, gender
tems. The correlation between migration status and reading is, on the and fear of failure have particularly strong effects on 15-year olds’ life
other hand, positive in the common core-flexible integration countries, satisfaction. Whereas the gender effect is reversed (boys tend to report
albeit statistically insignificant. higher life satisfaction), students who report higher fears of failure have
Results in model 2 also exemplify effects on reading of factors other lower life satisfaction, as in the case of happiness. However, in contrast
than migration. In all cases, having repeated a grade is associated with to happiness, the ESCS has significant and positive correlation with stu-
the largest drop in mean reading score: from 65.874 points in com- dents’ life satisfaction in all education systems. In addition, in the two
mon core-uniform integration systems to 48.818 points in early tracking common core systems, higher reading scores are associated with higher
countries. Another noteworthy factor is the language spoken at home. levels of life satisfaction.
Speaking a home language that is different from the school and test lan-
guage is associated with a drop in reading scores across all systems: 4.3. Generational effects in migrants’ reading performance and subjective
from 42.214 points in the individualised system to 11.851 points in wellbeing
the common core-uniform integration system. Moreover, in all groups,
the ESCS index positively correlates with reading performance, with After having examined how migrant-native differences are mani-
the biggest effects observed in the common core-flexible integration fested in reading and subjective wellbeing across the four educational
(30.989 points) and early tracking (26.385 points) systems. Lastly, being approaches, the third, and final analytical step, asked: among migrants,
a male has a negative impact on reading in all cases, with the biggest im- do second-generation students perform better in reading and report
pact in the individualised system. This is in line with extant scholarship higher wellbeing levels? To do so, a set of similar regression analy-
(Cardoso 2020). ses as those in section 4.2 was conducted. In this step, only the two
Table 4 shows education system comparisons for happiness. After subgroups of migrant students – first and second-generations – were
having considered effects of control variables, the correlation between considered (see section 3.2.2 for definitions). Compared to the previous
migration status and happiness is either statistically insignificant (in regressions, the only change made was to the independent variable. In
early tracking and individualised systems) or is positive at 99 per- this case, the variable “second generation” distinguishes between “0”
cent (common core-flexible integration) or 95 percent (common core- for first-generation migrants and “1” for second-generation migrants.
uniform integration) significance. Therefore, migration alone does not Incorporating this change, regression results across the four educational
contribute to a significant migrant-native gap in happiness, regardless groups in the third analytical step were presented in Tables 6, 7, and 8.
of the education system concerned. If anything, being of migration her- The key finding stemming from this set of regression analyses is that
itage is positively correlated to happiness levels in two out of the four being a second-generation student of immigrant parentage has a positive
country groups. Life satisfaction displays strong and positive effects on association with reading performance, but barely with levels of happi-
happiness at the highest significance level in all groups. The positive ness or life satisfaction. Figures in Table 6 highlight that, in all four
association between reading score and happiness is highly significant country groups, being of second-generation positively impacts reading
in the individualised system, and somewhat significant in the common score vis-à-vis that of first-generation migrants. The positive effect was
core-uniform integration system. In all groups, being a male student and the strongest in the individualised system of Nordic countries, with more
reporting higher fear of failure correlate to lower levels of happiness. than 34 points change, as compared to 6-7 points change in the three
Unlike in reading, the ESCS does not manifest the same unequivocal other systems (at times at even lower significance levels). Table 7 illus-
effect on happiness: positive correlation in the common core-flexible in- trates that only in the early tracking group was there a significant pos-
tegration system, negative correlation in the common core-uniform inte- itive correlation between being a second-generation and happiness lev-
gration system, and statistically insignificant in the two other education els. Lastly, generation effects showed no significance on life satisfaction
systems. changes in any country group (Table 8). Hence, in terms of generation

8
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

Table 7
Results of multilinear regression for the four-group comparison on happiness, with the migration variable comparing second-generation to first-generation
migrants. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001

Early-tracking system Common core/Flexible integration system Common core/Uniform integration system Individualised system

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

∗∗∗ ∗∗
Second-generation 0.086 0.043 -0.030 -0.016 0.005 -0.001 0.058 0.017
Reading score -0.0001 0.0001 0.0001 0.0002
Life satisfaction 0.303∗ ∗ ∗ 0.286∗ ∗ ∗ 0.278∗ ∗ ∗ 0.323∗ ∗ ∗
Gender -0.113∗ ∗ ∗ -0.099∗ ∗ ∗ -0.044∗ ∗ -0.083∗
Language use at home -0.025 0.008 -0.026 0.013
Fear of failure -0.040∗ ∗ ∗ -0.043∗ ∗ ∗ -0.040∗ ∗ ∗ -0.050∗
Grade repetition -0.026 -0.025 0.013 0.025
Family wealth 0.025∗ 0.005 0.043∗ ∗ ∗ 0.029
ESCS -0.005 0.018 -0.017 0.010
Student’s expected ISEI 0.0004 0.0001 0.0003 0.001
N 6733 5445 3349 2761 6521 5378 1199 955
R-squared 0.004 0.252 0.000 0.264 0.000 0.218 0.001 0.280

Table 8
Results of multilinear regression for the four-group comparison on life satisfaction, with the migration variable comparing second-generation to first-
generation migrants. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001

Early-tracking system Common core/Flexible integration system Common core/Uniform integration system Individualised system

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

∗∗∗
Second-generation 0.105 0.025 -0.046 -0.015 0.012 -0.004 0.052 -0.045
Reading score 0.0001 0.0003 0.0002 0.0003
Happiness 0.725∗ ∗ ∗ 0.805∗ ∗ ∗ 0.685∗ ∗ ∗ 0.736∗ ∗ ∗
Gender 0.239∗ ∗ ∗ 0.190∗ ∗ ∗ 0.154∗ ∗ ∗ 0.248∗ ∗ ∗
Language use at home 0.007 -0.043 -0.016 -0.088
Fear of failure -0.158∗ ∗ ∗ -0.188∗ ∗ ∗ -0.151∗ ∗ ∗ -0.210∗ ∗ ∗
Grade repetition -0.110∗ ∗ ∗ -0.013 -0.083∗ -0.082
Family wealth 0.028 0.060∗ 0.070∗ ∗ ∗ 0.005
ESCS 0.029 0.034 0.032∗ 0.043
Student’s expected ISEI 0.0002 0.001 0.001 0.001
N 6733 5445 3349 2761 6521 5378 1199 955
R-squared 0.002 0.289 0.000 0.307 0.000 0.250 0.000 0.334

effects examined in this study, it can be said that such effects positively tegration in schools in Finland, Sweden, followed by in countries such
affect reading performance, but have little to no significant impact on as Canada, the UK, and the USA (Solano and Huddleston 2020). Yet,
the two wellbeing variables. Therefore, when considering effects on all this study’s findings show that in the individualised system (Finland and
three dependent variables – reading, happiness, and life satisfaction – Sweden), migration has the largest negative impact on reading. The neg-
the crucial point here is whether or not a student is a migrant (hence ative effect of speaking a different language at home5 than the school
more likely to experience gaps compared to natives), rather than which and test language is also the largest in these two countries. This could
generation a migrant student belongs to. partly be explained by the fact that native students in Finland and Swe-
den perform exceptionally well. Hence, although more support schemes
5. Discussion and conclusion for migrant youths may be in place in these countries, the performance
inequality is still considerable (also manifested in Fig. 1). Another pos-
The data analysis in this study provides evidence to suggest the as- sible explanation is that the individualised nature of support, includ-
sociation between migration on the one hand, and reading performance ing language and literacy support, could mean that standardised quality
and subjective wellbeing on the other. Results revealed that, compared control is lacking. This paper advances the knowledge in extant litera-
to natives, migrant students achieve lower reading performance, experi- ture by highlighting the importance to consider migrant students’ per-
ence lower life satisfaction, but report higher happiness (albeit at lower formance relative to that of native students (and not only focusing on
significance level and magnitude) across the board. Crucially, however, migrants’ performance alone) when comparing and assessing education
comparative analyses highlight the fact that effects of migration on read- systems’ ability to address migrant students’ academic needs.
ing and subjective wellbeing manifest differently for different outcome Results further highlight that several other factors affect reading, at
variables and in different education systems. While migration back- times more than migration status does. The most important of which
ground is not associated with a negative reading gap for migrants in the is grade repetition. That grade retention has the highest negative ef-
common core-flexible integration system, it is associated with a 51.448- fect on reading in common core-uniform integration systems, and the
point drop in mean reading score in the individualised system, and a least negative effect in early tracking systems is not surprising. In the
19.208-point drop in the early tracking system. However, whereas mi-
gration negatively affects students’ life satisfaction in the two common
core systems, its negative association with life satisfaction loses signifi- 5
It has to be acknowledged that certain officially multilingual countries in the
cance in the early tracking and individualised systems.
analytic sample (such as Spain and Switzerland) may have speakers of one na-
Prior comparative research affirmed that migrant students experi- tional language attending school in another language region, hence not related
ence the largest reading disadvantage in continental Europe (the early to international migration in the sense of this paper. From the PISA dataset,
tracking and common core-uniform systems in this study) (Entorf and however, it is only possible to know whether a student’s home language is the
Minoiu 2005; Schnepf 2007). It has also been shown that migrant stu- test language, hence a finer distinction could not be made. Care is therefore
dents have the most conducive environment for successful academic in- needed when interpreting effects of this variable.

9
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

former system, secondary school students by and large follow the same ternational migration, in which migrant students might more likely be
curriculum. The standard approach to manage diverse abilities is grade placed in a disadvantaged position in gaining reading literacy, due to
retention, hence respondents in these countries who repeated a grade (initial) language barriers. Lastly, as secondary school students are at a
tend to be particularly weak. In contrast, early tracking systems manage critical stage of growing up, what they experience, especially as regards
diverse aptitudes with a crucial streaming procedure at the end of pri- personal and emotional wellbeing, will have an impact on life achieve-
mary school. It can thus be expected that 15-year-olds in the same aca- ments in later adult years (Brule and Veenhoven 2014; Heinrich and
demic track in early tracking countries display smaller gaps in aptitude, Galan 2018; Lindorff 2020)
even for those who repeated a grade. Moreover, speaking a different lan- Moreover, results suggest that different combinations of indicators –
guage at home has a significant negative association with reading per- migration background plus grade repetition, home language use, and/or
formance across all systems. Such effects hence point to the importance socio-economic status – should be used to identify most vulnerable
of targeted language support measures in schools in all education sys- groups (in reading and subjective wellbeing) in different education sys-
tems, especially for first generation migrants. Moreover, the follow-up tems, so that more nuanced intervention measures can be devised. This
regression tests on generation effects revealed that, compared to first- has practical relevance in making potential changes to pedagogical ap-
generation migrant students, being of second generation is associated proaches. For instance, policymakers in systems with individualised ap-
with higher reading performance. Although this positive and significant proach might want to consider introducing more standardised support
correlation is manifested in all four country groups, the generation ef- programmes (e.g., extra language and reading remedial classes for those
fect appears to be the strongest in the individualised education system in disadvantaged positions, of which migration can be one indicator)
comprising the two Nordic countries. Results suggest that although the coupled with quality control/evaluations by senior educators. Practi-
general migrant-native gap in reading is the largest in the individualised tioners in schools adopting the common core approaches could pay
system, this system might also be more conducive to upward mobility particular attention to students from migration and lower ESCS back-
for second-generation migrants, in terms of acquiring reading compe- grounds as regards their state of wellbeing, and if need be, design special
tence. counselling sessions.
As regards subjective wellbeing, results revealed a high level of posi- This study has several limitations. First, it only considered reading
tive correlation between the two constituent aspects: happiness and life competence as an indicator of academic performance. As explained ear-
satisfaction, regardless of education system and after controlling for all lier, international migration is directly linked to issues of language be-
other variables. Insofar as migrants’ subjective wellbeing inequality is haviour, notably the acquisition of new languages in host societies. This
concerned, the negative impact of migration on life satisfaction (and in turn concerns language-related performance such as reading. The fact
the significant correlation) manifests in the two common core systems that reading was the focus of PISA 2018 also meant that it was more
only. As for happiness, migration experience has a slight positive im- reliable than mathematics and science scores (see section 3.2.1). Do
pact on the two common core systems, and loses statistical significance correlations between migration background and mathematics and sci-
in the early tracking and individualised systems. More than migration ence manifest in the same way and to the same significance level in
status, a series of other factors displayed consistent effects on the two di- each of the four clusters? This is a noteworthy issue for future research
mensions of subjective wellbeing across all education systems. Whereas to pick up. Second, it is worth reiterating that findings in this paper
female students reported higher levels of happiness but lower levels of pertain to the sample of 15 countries analysed in this study. They are
life satisfaction in all cases, students who had higher fear of failure re- therefore not generalisable to the entire PISA 2018 sample (approxi-
ported lower levels of both happiness and life satisfaction. While ex- mately 70 countries/regions). Further enquiry with mixed-method re-
tant research established the profile of students who are likely to report search in context-specific educational settings across regions would al-
higher fear of failure (Borgonovi and Han, 2021), results here contribute low more generalisable conclusions across the entire PISA sample of
to scholarship by highlighting the importance to address the issue from countries/regions. What is more, the cross-sectional data analysed here
a wellbeing perspective, for it negatively affects students’ happiness and do not permit the establishment of causality. Further examination of
life satisfaction across the board. causations of the correlations that emerged from the analysis is needed.
Results of this study advance current debates by casting the question A lead for future studies in examining possible causality would be to
of migrant-native inequality in a new light: that effects of migration on consider societal attitudes towards migration and diversity in the edu-
reading performance and subjective wellbeing manifest differently in cational contexts of the different countries (or groups) discussed here.
education systems with different diversity management practices. While This could be achieved with a mixed-method approach incorporating
the common core-flexible integration system proves good in alleviating qualitative research (in-depth interviews and classroom observations)
the migrant-native gap in reading, migrant students in the early tracking to explore if and how educators’ and peers’ attitudes towards migration
and individualised systems are less likely to experience a life satisfaction contribute to performance and wellbeing gaps observed in this study.
gap due to their migration status. Lastly, the two common core systems Also, in keeping with the current research focus on effects of individ-
appear to successfully mitigate negative effects of migration on happi- ual level migration on reading performance and subjective wellbeing,
ness. school-level factors were not included in the present analysis. This could
This paper contends that it is important for researchers, practition- be examined by future studies with corresponding foci on school-level
ers, and policymakers in education to consider the wellbeing and read- factors.
ing performance of secondary school students for several reasons. First, In addition, this study shows that even in similar educational sys-
recent studies have shown that adolescents’ wellbeing has positive asso- tems, the effect of migration and gender can be positive for one aspect
ciations with their academic performance, as well as other positive edu- of subjective wellbeing and negative for the other. This prompts the
cational outcomes such as engagement in social life and smoother transi- thought that although happiness and life satisfaction may overlap, they
tion from one school phase to another (Clarke 2020; Lindorff 2020). Sec- are not the same. Future studies could build on this empirical evidence
ond, implications of students’ reading performance go beyond the con- to further theorise and operationalise subjective wellbeing measures in
fines of secondary school classrooms. As I pointed out in section 3.2.1, the context of comparative education research with migration and gen-
assessments in PISA 2018 defined reading as understanding, using, eval- der perspectives. I also acknowledge that care needs to be taken when
uating, reflecting on and engaging with texts in order to achieve one’s interpreting cross-cultural comparisons on happiness and life satisfac-
goals, to develop one’s knowledge and potential, and to participate in tion. Partly for this reason, the scope of this study is limited to Euro-
society. In a broader sense, secondary students’ reading proficiency is pean and North American contexts in which happiness and life satisfac-
one useful tool to gauge their present and future ability to engage in tion are construed in broadly similar ways (Uchida and Ogihara 2012).
public social life. This is all the more pertinent in the context of in- To be sure, internal differences on understandings and expressions of

10
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

wellbeing exist, but extant scholarship has shown that large-scale com- indicators_-_education#Education_in_the_context_of_quality_of_life (Accessed: 24
parisons of average wellbeing (like this present study with PISA 2018) February 2021).
Froehlich, L., et al. (2018). Stereotype threat-effects for Turkish-origin migrants in Ger-
across countries, especially in Western contexts, are by and large reliable many: Taking stock of cumulative research evidence. European Educational Research
(Veenhoven, 2012b; Veenhoven, Land, & Sirgy, 2012a). Journal, Article 1474904118807539.
This article highlights the fact that inequality, to migrants’ disad- Frones, T. S., Rasmusson, M., Bremholm, J., Frønes, T. S., et al. (2020). Equity and Diver-
sity in Reading Comprehension—A Case Study of PISA 2000–2018. In Equity, equality
vantage, in reading performance and subjective wellbeing exists even in and diversity in the nordic model of education (pp. 305–335). Cham: Springer Interna-
some of the worlds’ richest and most developed countries. This comes at tional Publishing.
a critical juncture in which migrant students’ academic continuity and Gamazo, A., & Martínez-Abad, F. (2020). An Exploration of Factors Linked to Academic
Performance in PISA 2018 Through Data Mining Techniques. Frontiers in Psychology,
wellbeing needs tend to be more at stake than the general population, as
11.
a result of intermittent school closures due to the novel coronavirus dis- Geis-Thöne, W. (2020). Häusliches Umfeld in der Krise: Ein Teil der Kinder braucht
ease (Covid-19) (Wang, 2020). The Covid-19-induced school closures mehr Unterstützung (Family environment during times of crisis: One group of
children need more support). Cologne: Institut der deutschen wirtschaft. (German
not only threaten the learning rhythm of all students (Kuhlfeld and
Economic Institute) Available at Https://www.iwkoeln.de/fileadmin/user_upload/
Tarasawa 2020), but more importantly, exacerbate the existing struc- Studien/Report/PDF/2020/IW-Report_2020_Haeusliche_Lebenswelten_Kinder.pdf
tural inequality between native and migrant students (Geis-Thöne 2020; (Accessed: 6 August 2020).
European Commission 2020). Reading literacy has a fundamental role Gómez-Fernández, N., & Mediavilla, M. (2021). Exploring the relationship between Infor-
mation and Communication Technologies (ICT) and academic performance: A multi-
in promoting individual development both in academic attainment and level analysis for Spain. Socio-Economic Planning Sciences.
beyond the confines of classrooms (Vedechkina and Borgonovi 2021). Hanushek, E. A., & Woessmann, L. (2011). The Economics of International Differences
Seen in this light, results of this paper invite scholars and practitioners in Educational Achievement. In Handbook of the economics of education (pp. 89–200).
Elsevier.
in education to reflect, through the prism of migration, on how reading Happiness Research Institute (2020). ‘How we measure well-being’. Available at:
proficiency and wellbeing can differ in diverse educational approaches, Https://www.happinessresearchinstitute.com/happinessresearch (Accessed: 25 June
as well as the implications on longer-term personal developments in and 2020).
Heckmann, F. (2008). Education and the integration of migrants: Challenges for eu-
out of the classroom. ropean education systems arising from immigration and strategies for the suc-
cessful integration of migrant children in european schools and societies. Bam-
Declaration of interests berg: University of Bamberg. Available at Http://www.efms.uni-bamberg.de/pdf/
NESEducationIntegrationMigrants.pdf (Accessed: 13 June 2020).
Heinrich, P., & Galan, C. (2018). Being young in super-aging japan: Formative events and
The authors declare that they have no known competing financial cultural reactions (eds). London: Routledge.
interests or personal relationships that could have appeared to influence Jonsson, J. O., & Rudolphi, F. (2011). Weak Performance–Strong Determination: School
the work reported in this paper. Achievement and Educational Choice among Children of Immigrants in Sweden. Eu-
ropean Sociological Review, 27(4), 487–508.
Kahneman, D., & Krueger, A. B. (2006). Developments in the Measurement of Subjective
Acknowledgement of Funding Well-Being. Journal of Economic Perspectives, 20(1), 3–24.
Karakolidis, A., Pitsia, V., & Emvalotis, A. (2016). Examining students’ achievement in
mathematics: A multilevel analysis of the Programme for International Student As-
The research leading to these results received funding from the Ger- sessment (PISA) 2012 data for Greece. International Journal of Educational Research,
man Research Foundation (Deutsche Forschungsgemeinschaft, DFG) under 79, 106–115.
Research Grant Number CO 100/13-1.. The funder was not involved in Kuhfeld, M., & Tarasawa, B. (2020). The COVID-19 slide: What summer learning
loss can tell us about the potential impact of school closures on student academic
the research design, implementation, writing and dissemination in any
achievement. Portland: Northwest Evaluation Association (NWEA) Available at
way. Additional acknowledgement text. Https://www.nwea.org/content/uploads/2020/05/Collaborative-Brief_Covid19-
Slide-APR20.pdf (Accessed: 24 June 2020).
References Lange, F., Topel, R., Hanushek, E., & Welch, F. (2006). Chapter 8 The Social Value of Ed-
ucation and Human Capital. In Handbook of the economics of education (pp. 459–509).
Adler, A., & Seligman, M. E. P. (2016). Using wellbeing for public policy: Theory, mea- Elsevier.
surement, and recommendations. International Journal of Wellbeing, 6(1), 1–35. Lindorff, A. (2020). The impact of promoting student wellbeing on student academic
Borgonovi, F., & Han, S. W. (2021). Gender disparities in fear of failure among 15-year-old and non-academic outcomes: An analysis of the evidence. Oxford: Oxford University
students: The role of gender inequality, the organisation of schooling and economic Press Available at: Https://oxfordimpact.oup.com/wp-content/uploads/2020/10/
conditions. Journal of Adolescence, 86, 28–39. Wellbeing-Impact-Study-Report.pdf (Accessed: 15 October 2021).
Brulé, G., & Veenhoven, R. (2014). Participatory Teaching and Happiness in Developed Machin, S., Salvanes, K. G., & Pelkonen, P. (2012). Education and Mobility. Journal of the
Nations. Advances in Applied Sociology, 04(11), 235–245. European Economic Association, 10(2), 417–450.
Cardoso, M. E. (2020). Policy Evidence by Design: International Large-Scale Assessments Mons, N. (2007). Les nouvelles politiques éducatives: La france fait-elle les bons choix? (New
and Grade Repetition. Comparative Education Review, 64(4), 598–618. education policies : Is france making the right choice?. Paris: PUF.
Castles, S., Haas, H., & Miller, M. J. (2014). The age of migration: International population Montt, G., & Borgonovi, F. (2018). Combining Achievement and Well-Being in the Assess-
movements in the modern world: 5. Basingstoke: Palgrave Macmillan. ment of Education Systems. Social Indicators Research, 138(1), 271–296.
Clarke, T. (2020). Children’s wellbeing and their academic achievement: The danger- Moon, S. S., Kang, S.-. Y., & An, S. (2009). Predictors of Immigrant Children’s School
ous discourse of “trade-offs” in education. Theory and Research in Education, 18(3), Achievement: A Comparative Study. Journal of Research in Childhood Education, 23(3),
263–294. 278–289.
Cutler, D., & Lleras-Muney, A. (2006). Education and health: Evaluating theories and evidence Organization for Economic Development and Cooperation (OECD). (2009). PISA data anal-
(p. w12352). Cambridge: MA: National Bureau of Economic Research. w12352. ysis manual: Spss. Paris: OECD Publishing.
Davoli, M., & Entorf, H. (2018). The PISA Shock, Socioeconomic Inequality, and School Organization for Economic Development and Cooperation (OECD). (2019). PISA 2018
Reforms in Germany. Institute of Labor Economics Policy Paper, (140), 19. results: Https://www.oecd.org/pisa/publications/pisa-2018-results.htm. (Accessed:
De Haas, H., Castles, S., & Miller, M. J. (2019). The age of migration: International population 02.02.2021).
movements in the modern world. London: Red Globe Press. Rodríguez, S., et al. (2020). Performance and well-being of native and immigrant stu-
Dockery, A. M., Koshy, P., & Li, I. W. (2020). Culture, migration and educational per- dents. Comparative analysis based on PISA 2018. Journal of Adolescence, 85, 96–
formance: A focus on gender outcomes using Australian PISA tests. The Australian 105.
Educational Researcher, 47(1), 39–59. Samuel, R., Bergman, M. M., & Hupka-Brunner, S. (2013). The Interplay between Educa-
Dupriez, V., Dumay, X., & Vause, A. (2008). How Do School Systems Manage Pupils’ tional Achievement, Occupational Success, and Well-Being. Social Indicators Research,
Heterogeneity? Comparative Education Review, 52(2), 245–273. 111(1), 75–96.
Entorf, H., & Minoiu, N. (2005). What a Difference Immigration Policy Makes: A Com- Schneeweis, N. (2011). Educational institutions and the integration of migrants. Journal
parison of PISA Scores in Europe and Traditional Countries of Immigration. German of Population Economics, 24(4), 1281–1308.
Economic Review, 6(3), 355–376. Schnepf, S. V. (2007). Immigrants’ educational disadvantage: An examination across ten
Wang, Zi (2020). Addressing migrants’ well-being during COVID-19: An analysis of countries and three surveys. Journal of Population Economics, 20(3), 527–545.
Chinese communities’ heritage language schools in Germany. Migration Studies. Solano, G., & Huddleston, T. (2020). Migrant integration policy index 2020. Migration Policy
10.1093/migration/mnaa033. Group Available at: Https://www.mipex.eu/education Accessed: 26 February 2021.
European Commission (2020). COVID-19’s impact on migrant communities. Available Uchida, Y., & Ogihara, Y. (2012). Personal or interpersonal construal of happiness: A
at: Https://ec.europa.eu/migrant-integration/news/covid-19s-impact-on-migrant- cultural psychological perspective. International Journal of Wellbeing, 354–369.
communities (Accessed: 6 August 2020). Vedechkina, M., & Borgonovi, F. (2021). A Review of Evidence on the Role of Digital Tech-
Eurostat (2019). Quality of life indicators - education - Statistics Explained. Available nology in Shaping Attention and Cognitive Control in Children. Frontiers in Psychology,
at: Https://ec.europa.eu/eurostat/statistics-explained/index.php/Quality_of_life_ 12, 1–22.

11
Z. Wang International Journal of Educational Research Open 2 (2021) 100087

Veenhoven, R., Land, K. C., & Sirgy, M. J. (2012a). Happiness: Also Known as “Life Satisfac- Veenhoven, R., Glatzer, W., et al. (2015). The Overall Satisfaction with Life: Subjective
tion” and “Subjective Well-Being. In A. C. Michalos (Ed.), Handbook of social indicators Approaches (1). In Global handbook of quality of life: Exploration of well-being of nations
and quality of life research (pp. 63–77). Dordrecht: Springer Netherlands. and continents (pp. 207–238). Dordrecht: Springer.
Veenhoven, R. (2012b). Cross-national differences in happiness: Cultural measurement Volante, L., Klinger, D., & Bilgili, O. (2018). Immigrant student achievement and education
bias or effect of culture? International Journal of Wellbeing, 333–353. policy (eds). Cross-Cultural Approaches.

12