Professional Documents
Culture Documents
5 Structural and Motivational Mechanisms of Academic Achievement - A Mediation Model of Social Background Effects On Academic Achievement
5 Structural and Motivational Mechanisms of Academic Achievement - A Mediation Model of Social Background Effects On Academic Achievement
Abstract
This paper takes up ongoing discussions on the inequality of educational oppor-
tunities and formulates a conceptual model to link separate lines of research. Our
particular focus is on combining motivational and structural approaches into a
mediation model that explains differences in academic achievement. In the litera-
ture, four main mechanisms of social reproduction are discussed. Two main path-
ways refer to (1) parents’ expectations regarding their children’s academic success
and (2) replicating cultural capital through intra‐familial cultural practices. (3)
Parents’ perception of children’s abilities depends on social position and is influ-
ential for expectations of success. (4) For all three pathways, we expect effects on
students’ motivational characteristics, which in turn influence academic achieve-
ment. We test our conceptual model by structural equation modelling using lon-
gitudinal data from primary school students in Germany. Empirical evidence is
in line with the assumptions in the model. Cultural reproduction and expecta-
tions of success can be seen as the key components of the model. However, both
chains of reproduction are related to each other by parents’ perception of child’s
ability, and their effects are mediated by child’s motivational characteristics.
Keywords: Achievement; social background; path model; mediation; cultural
resources
Introduction
Hartmut Ditton (Ludwig-Maximilians-University Munich) Michael Bayer (Leibniz Institute for Educational Trajectories) and
Florian Wohlkinger (Ludwig-Maximilians-University Munich) (Corresponding author email: florian.wohlkinger@edu.lmu.de)
© London School of Economics and Political Science 2018 ISSN 0007-1315 print/1468-4446 online.
Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 101 Station Landing, Suite 300,
Medford, MA 02155, USA on behalf of the LSE. DOI: 10.1111/1468-4446.12506
2 Hartmut Ditton, Michael Bayer and Florian Wohlkinger
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Structural and motivational mechanisms of academic achievement 3
processes in the educational system: rational action theory does not make any
assertions regarding the development of individuals’ beliefs or attitudes.
The second sociological tradition follows the work of Pierre Bourdieu and
focuses on cultural reproduction processes. This theoretical branch does not draw
on the distinction between primary and secondary effects of social background,
but rather merges these effects into the figure of class‐specific habitus. In fact,
Bourdieu argues against assumptions made by rational action theory and espe-
cially against the assumption ‘that the agent acts freely, consciously, and […] “with
full understanding”, the action being the product of a calculation of chances and
profits’ (Bourdieu 2000: 138). The concept of habitus can be understood as a set
of ‘dispositions’ (Bourdieu 1998: 6) and is, despite its principal vagueness, a theo-
retical anchor for introducing motivational factors into the analyses of reproduc-
tion processes. Habitus is Bourdieu’s central concept of the cultural reproduction
of inequality, but it is rarely utilized in quantitative empirical research. Instead, his
concept of cultural capital plays a considerably more prominent role in studying
reproduction processes (cf. Van De Werfhorst, De Graaf and Kraaykamp 2001).
Since DiMaggio’s comparison of Bourdieu’s cultural reproduction thesis with
his own cultural mobility approach (cf. DiMaggio 1982), this branch of research
has especially focused on the effects of cultural capital on educational success (cf.
De Graaf, De Graaf and Kraaykamp 2000). Some studies have tried to opera-
tionalize the concept of habitus for quantitative research (cf. Gaddis 2013), but,
as Edgerton and Roberts note, ‘the bulk of research on habitus has been qual-
itative in nature’ (Edgerton and Roberts 2014: 194). One particular reason for
this shortage is the difficulties in operationalizing the habitus concept within the
framework of quantitative research approaches – a problem that undoubtedly still
exists (cf. Nash 1999). Goldthorpe (2007) criticizes Bourdieu’s concept of cultural
capital for its vagueness and stated that there is no (more) empirical evidence for
Bourdieu’s assumptions of structural stability, but Lareau and Weininger (2003)
showed that these kind of interpretations of Bourdieu’s conceptualization of cul-
tural capital can be traced back to the influential article of DiMaggio (1982). In
our analysis later on, we will refer to ‘cultural resources’, which we consider as
equivalent to Bourdieu’s (1986) notion of objectified cultural capital.
In empirical educational research, plenty of models and analyses have been
developed in the meantime. However, at the same time, fruitful insights from
the Wisconsin tradition or from research on motivation are not very prominent
in the current discussions. With regard to motivational processes as mediators
between socio‐structural resources and educational success, some references can
be found at the level of conceptual models. However, when it comes to empir-
ical analyses, not much of this remains. For example, Erikson and Rudolphi
illustrate that in Sweden, ‘secondary effects account for around a third of class
differences in the transition from compulsory education to upper secondary’
(Erikson and Rudolphi 2010: 301). They assume that processes of unobserved
early choice are at play here. These early choices reflect class‐specific educational
British Journal of Sociology 0(0) © London School of Economics and Political Science 2018
4 Hartmut Ditton, Michael Bayer and Florian Wohlkinger
Research on motivation
In the 1980s, Jacquelynne Eccles and colleagues (1982) began to analyse the
impact of motivational and social aspects on educational success and educational
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Structural and motivational mechanisms of academic achievement 5
decisions. They especially reflect the mediating role of parental perceptions of the
child and of child‐oriented parental behaviour. Therefore, in a differentiated way,
the Eccles model (cf. Eccles and Wigfield 2002: 119) elaborates those aspects of
family socialization that are only briefly sketched in the models discussed before.
At the same time, the child as the subject of education reappears in a systematic
way. Thus, reflecting parental perceptions of the child, parental practices in edu-
cation and children’s own motivation could be a good starting point for devel-
oping a more elaborated analytical model for social reproduction in education.
The assumptions in the Eccles model (cf. Eccles and Wigfield 2002: 119)
encompass parents’ attitudes towards parenting and their behaviour, but child
characteristics were also taken into account, for example: the child’s perceptions
of socializer’s beliefs, expectations, and attitudes on the one hand, and the child’s
general self‐schemata as well as the child’s short‐ and long‐term goals on the other
hand. Altogether, this model is characterized by a high degree of differentiation
with respect to the child and to family processes. Nevertheless, it is rather crude
concerning social‐structural parameters and the family’s living conditions. It is
therefore mainly the inclusion of schemata and concepts of the self that links
the model to the tradition of Bourdieu, and it is the expectation‐value compo-
nent that links to rational action theory and Boudon. In our view, the incorpo-
ration of motivational concepts – particularly the child’s self‐concept of ability
– may provide an opportunity to operationalize Bourdieu’s concept of habitus.
Numerous studies (e.g., Gottfried 1990; Marsh 1990) corroborate the self‐con-
cept of ability (cf. Gecas 1982) as an important factor with respect to academic
achievement. Relations between social background, family processes, and the
child’s and parents’ aspirations and expectations still have to be studied in more
detail and should be considered together in comprehensive analytical models.
Despite the progress in analysing educational decisions in recent years, the pro-
cesses of transmission still resemble a ‘black box’ (Irwin 2009). Concerning the
impact of family background on children’s educational success, a statement formu-
lated by Lareau 30 years ago is still valid: ‘Research on this issue focused primarily
on educational outcomes; very little attention was given to the processes through
which these educational patterns are created and reproduced’ (Lareau 1987: 73).
Processes of mediation are at play whenever transitions to further education
occur. Clarifying these processes would greatly contribute to a deeper under-
standing of educational decisions and social inequality and represent a step
towards opening up the ‘black box’ of reproduction mechanisms. The integra-
tion of motivational processes and the analysis of aspects of self‐concept can
be seen as a step towards more differentiated models and offer the possibility of
integrating findings from different research traditions.
British Journal of Sociology 0(0) © London School of Economics and Political Science 2018
6 Hartmut Ditton, Michael Bayer and Florian Wohlkinger
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Structural and motivational mechanisms of academic achievement 7
British Journal of Sociology 0(0) © London School of Economics and Political Science 2018
8 Hartmut Ditton, Michael Bayer and Florian Wohlkinger
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Structural and motivational mechanisms of academic achievement 9
The model does not suggest an unambiguous causal ordering of the fac-
tors involved. There may be reciprocal dependence between factors, especially
between academic achievement, motivational characteristics and expectations.
In a longitudinal perspective, it is highly plausible that aspirations and expecta-
tions of success are influenced by achievement as well. In addition, motivational
characteristics in the long run will depend on experienced academic achieve-
ment. Nevertheless, considering academic achievement as the dependent vari-
able, aspirations and motivational characteristics can be included as predictors
in the model. This does not preclude that they themselves depend on achieve-
ment in the next loop of development. The sequence presented in the model is
concentrated on influences on academic achievement and it corresponds with
the chronological order of the variables used in the structural equation model.
To analyse the patterns postulated within the model, we use data from a
longitudinal study in German primary schools. Analytically, we resort to the
structural equation modelling technique (SEM) and expect a reasonably good
fit for a model that includes the pathways postulated in our framework. We do
not argue that a reasonable fit for the SEM will indicate causality but at least
it would imply that the mechanisms assumed are supported by empirical data.
The data used in this analysis stem from a longitudinal study covering second
to fourth grade in primary schools in Germany (cf. Ditton 2007; Ditton and
Krüsken 2009). As part of this longitudinal study, a total of 1,453 primary‐
school students from two federal states in Germany (Bavaria and Saxony) as
well as their parents and teachers participated in yearly intervals between 2005
and 2007. Apart from gathering information from all three survey groups on
familial, regional and school‐related conditions regarding competence acqui-
sition as well as attitudes regarding academic achievement, additional compe-
tence tests of curricular validity were administered to students in the subjects
of German and mathematics. The sample was drawn using a stratified ran-
dom sampling method according to the attributes of urbanization, school size,
school type and region. From each participating school, one randomly selected
class was included in the sample. Overall, 77 school classes took part in all three
surveys. Data collection was carried out at the end of each school year in sec-
ond, third and fourth grade. Due to missing values and in some cases ambiguous
information on school class, the final sample in our model consists of N = 1,425
students from N = 75 school classes.
We tested a structural equation model with full information maximum likeli-
hood estimation (FIML; maximum likelihood with missing values in Stata 13).
For the estimation procedure, we chose robust and clustered standard errors
(StataCorp LP 2013: 96ff). The clustered option reflects the nesting of students
within school‐classes. The robust option reflects that two of the variables used in
British Journal of Sociology 0(0) © London School of Economics and Political Science 2018
10 Hartmut Ditton, Michael Bayer and Florian Wohlkinger
Dependent variable
We use the grade average in the core subjects (i.e., German, mathematics, and
local history/general science) in grade four as measure for ACADEMIC
ACHIEVEMENT. The information on half‐term reports in fourth grade was
issued by teachers, so it can be considered reliable. The mean grade average for
the sample under investigation is x = 2.59, the standard deviation is σ = 0.76,
and it ranges from 1.00 to 4.67.2 For our analysis, we recoded the scale so that
larger values indicate higher levels of achievement.
Independent variable
As the only purely exogenous variable and indicator of SOCIAL
BACKGROUND, we include the highest educational qualification completed in
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Structural and motivational mechanisms of academic achievement 11
the household in the model. The information was conducted in the parent ques-
tionnaires (we used the highest qualification across all three points of measure-
ment) and is integrated into the model as a grouped categorical variable. The
lowest status group features cases without any qualifications as well as quali-
fications up until and including basic secondary‐school level (in general: nine
years of schooling; 18.0 per cent of cases). The second status group encom-
passes qualifications of intermediate secondary‐school level (46.0 per cent of
cases). The third and highest status group corresponds to high secondary‐school
level (including university entrance‐level qualifications; 36.0 per cent of cases).
British Journal of Sociology 0(0) © London School of Economics and Political Science 2018
12 Hartmut Ditton, Michael Bayer and Florian Wohlkinger
the child’s joy of reading, which comprises five items and stems from the IGLU/
PIRLS study (cf. Bos et al. 2005). Its mean is x = 3.19, its standard deviation is
σ = 0.68, and Cronbach’s alpha is α = 0.74.
The perceived difference between parental home and school introduces a scale
comparing the situation at home and in school into the model (FAMILY‐
SCHOOL RELATION). The scale was developed following the work of
Hauser‐Cram, Sirin and Stipek (2003) and consists of five items that collect
information from children attending second grade. A four‐point scale is used
to measure agreement to comparisons drawn between the school and family
setting (for example: ‘In school, I always need to behave completely differently
than at home’). The mean of the scale is x = 2.60, the standard deviation is
σ = 0.76, and Cronbach’s alpha is α = 0.60. FAMILY‐SCHOOL RELATION
should depend on social background, cultural resources and reading behaviour
and have an influence on motivational characteristics.
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Structural and motivational mechanisms of academic achievement 13
Results
The setup of the structural equation model to test the assumptions about medi-
ating effects relates straightforwardly to the mechanisms and specific pathways
listed above. The dependent variable is students’ academic achievement at the end
of fourth grade. Families’ social background is the only exogenous variable. The
first pathway goes from social background to parents’ aspirations (second grade
two) and expectation of success (third grade). The second pathway is constituted
from social background to cultural resources within the family, child’s reading
behaviour and the family‐school relation. Perceived child’s abilities (pathway 3)
take an intermediate position between expectations and cultural resources and
cultural practices. Child’s motivational characteristics (pathway 4) forms the
final link in the mediation chain. We assume that motivational characteristics
are influenced by both the ‘cultural‐reproduction’ and the ‘ability‐expectation’
chain. As already mentioned, we do not neglect backward effects of the child’s
self‐concept on parental expectations or from achievement to aspirations and
expectations in a longitudinal perspective. Rather, our aim is to highlight and
test the importance of factors concerning expectations, cultural resources and
perceived ability for academic achievement at the end of primary school. This
should shed light on the complexity of mechanisms involved in the development
of academic achievement.
All variables in the model are observed (manifest) variables, with three excep-
tions: Cultural resources are handled as a latent variable which is measured by
book supply (loading: 0.78, s.e.: 0.02) and other cultural possessions at home
(loading: 0.74, s.e.: 0.02). The child’s reading behaviour is also handled as a latent
variable and was measured twice (at the end of second and third grade; loadings
(s.e.): 0.72 (0.04), 0.76 (0.04)). Finally, we combine academic self‐concept (load-
ing: 0.58, s.e.: 0.03) and school‐related helplessness (loading: −0.68, s.e.: 0.03)
into the latent variable motivational characteristics.
Figure II illustrates the results of our analysis with the structural equation
module in Stata 13. As already mentioned, robust standard errors clustered
within the school classes were estimated. Since Stata does not report detailed
measures of goodness of fit when robust standard errors are estimated, a model
with regular standard errors was estimated for comparison. The results of this
model are identical to the ones presented in Figure II (except for the standard
errors) and confirm that the fit of the model is quite reasonable (χ2 = 97.7;
d.f. = 37; RMSEA = 0.034 with 90% CI: 0.026 – 0.042).3 An inspection of the
modification indices reveals that only minor improvements of fit would occur
if any of the restrictions in the model would be released. The chart in Figure II
displays standardized path coefficients with p < 0.05.
Regarding the findings, we first want to point out that there is no remaining
direct effect of social background on academic achievement.4 Rather, the quite
British Journal of Sociology 0(0) © London School of Economics and Political Science 2018
14 Hartmut Ditton, Michael Bayer and Florian Wohlkinger
high association between SES and academic achievement (r = 0.37) can be seen
as entirely induced by the mediating variables included in the model. The vari-
ables integrated in the model explain 52 per cent of the variance in academic
achievement. Secondly, the results clearly support our expectation that substan-
tial mediating effects exist through cultural resources and expectation of success
(cf. Table 2 in the Appendix).5 Moreover, the two (main) pathways do not coexist
independently of each other, but there are effects of cultural resources on expec-
tation of success and aspirations. Thirdly, child’s perceived ability as well as
motivational characteristics are of importance in the mediation chain.
Cultural resources within the family substantially depend on SES (0.57) and
have a direct effect on child’s reading behaviour (0.30). Reading behaviour takes
a central role in the model, influencing parents’ perceived ability of the child
(0.23), probability of success (0.18) and child’s motivational characteristics
(0.34). The effects of SES on family‐school relation foremost are mediated by
cultural resources and reading behaviour, but a direct effect remains (0.09). With
respect to cultural resources, the effects on academic achievement are mostly
mediated through expectation of success and perceived ability, but beyond that,
there is also a direct effect (0.17).
The pathway through aspirations and expectation of success is also of high
importance. Expectation of success depends on SES in a direct way (0.24) and
is mediated by aspirations (0.28 * 0.32). Whereas aspirations have no further
direct effect on academic achievement, the direct effect of expectation of success
is substantial (0.31). Furthermore, the effects of aspirations and expectation of
success are partly mediated by motivational characteristics (0.11 * 0.39 and 0.28
* 0.39).
The child’s perceived ability has no direct effect on academic achievement, but
there are remarkable indirect effects through aspirations (0.13), expectation of
success (0.20) and motivational characteristics (0.18). Concerning motivational
characteristics, it turns out that there are effects from reading behaviour (0.34)
and aspirations (0.11) as well as expectation of success (0.28). Furthermore, per-
ceived ability (0.18) and family‐school relation (0.12) have an influence on moti-
vational characteristics. Altogether, motivational characteristics are fairly well
predicted in the model (variance explained: 0.49) and have substantial effects on
academic achievement (0.39).
Our model for the explanation of social‐background effects on academic
achievement clearly reveals that mediations via cultural resources and expec-
tation of success are of great importance. Moreover, both pathways are related
to perceived ability and therefore linked together. The most meaningful single
factors in the model are cultural resources, reading behaviour and expectation
of success. Motivational characteristics are also of high importance. In turn, as
is already well known, cultural resources and expectation of success are closely
related to family’s SES. Furthermore, reading behaviour is important in this
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Structural and motivational mechanisms of academic achievement 15
Discussion
British Journal of Sociology 0(0) © London School of Economics and Political Science 2018
16 Hartmut Ditton, Michael Bayer and Florian Wohlkinger
Furthermore, child’s motivation can be seen as a key factor for academic suc-
cess. Taking account of the mediation chains, no direct social‐background effect
on achievement remains, and it seems that it may be beneficial to take the model
proposed here as a valuable starting point for further research.
Limitations
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Appendix
[5]Omitted:
[2] Added: [3] Added: [4] Omitted: SES ->
[1] As shown SES -> SES -> all SES -> cultural expectancy
Model in Figure II achievement other variables paths paths
AIC 27296.42 27294.20 27296.75 27671.45 27439.85
BIC 27575.30 27578.35 27596.68 27939.81 27708.21
RMSEA 0.034 0.033 0.035 0.089 0.061
lower 0.026 0.025 0.026 0.082 0.054
upper 0.042 0.042 0.044 0.096 0.068
pclose 0.999 1.00 0.998 0.000 0.007
CFI 0.98 0.98 0.98 0.88 0.94
Notes
Bibliography
Bos, W., Hornberg, S., Arnold, K.‐H., Faust, 2006. The States of the Federal Republic
G., Fried, L., Lankes, E.‐M., Schwippert, K. and of Germany in National and International
Valtin, R. (eds) 2008 IGLU‐E 2006. Die Länder Comparison], Münster: Waxmann.
der Bundesrepublik Deutschland im nationalen Bos, W., Lankes, E.‐M., Prenzel,
und internationalen Vergleich [IGLU‐E/PIRLS M., Schwippert, K., Valtin, R., Voss,
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Structural and motivational mechanisms of academic achievement 19
British Journal of Sociology 0(0) © London School of Economics and Political Science 2018
20 Hartmut Ditton, Michael Bayer and Florian Wohlkinger
Goldthorpe, J.H. 2007 ‘“Cultural Capital”: Kloosterman, R., Ruiter, S., De Graaf, P.
Some Critical Observations’, Sociologica and Kraaykamp, G. 2009 ‘Parental
1(2): 1–23. Education, Children’s Performance and the
Gottfried, A.E. 1990 ‘Academic Intrinsic Transition to Higher Secondary Education:
Motivation in Young Elementary School Trends in Primary and Secondary Effects
Children’, Journal of Educational Psychology over Five Dutch School Cohorts (1965–
82(3): 525–38. 99)’, British Journal of Sociology 60(2):
Haller, A.O. and Butterworth, C.E. 1960 377–98.
‘Peer Influences on Levels of Occupational Kohn, M.L. 1969 Class and Conformity: A
and Educational Aspiration’, Social Forces Study in Values, Homewood, IL: Dorsey
38(4): 289–95. Press.
Hauser, R.M., Sewell, W.H. and Alwin, Kohn, M.L. and Schooler, C. 1969 ‘Class,
D.F. 1976 ‘High School Effects on Occupation, and Orientation’, American
Achievement’ in W.H. Sewell, R.M. Hauser Sociological Review 34(5): 659–78.
and D.L. Featherman (eds) Schooling and Kramer, R.‐T. and Helsper, W. 2011 ‘Kulturelle
Achievement in American Society, New Passung und Bildungsungleichheit ‐ Potenziale
York: Academic Press. einer an Bourdieu orientierten Analyse der
Hauser‐Cram, P., Sirin, S.R. and Stipek, D. Bildungungleichheit’ [‘Cultural Fit and
2003 ‘When Teachers’ and Parents. Values Education Inqeuality – Potentials of an Analysis
Differ: Teachers’ Ratings of Academic of Education Inequality Oriented in Bourdieu’]
Competence in Children from Low‐ in H.‐H. Krüger, U. Rabe‐Kleberg, R.‐T.
Income Families’, Journal of Educational Kramer and J. Budde (eds) Bildungsungleichheit
Psychology 95(4): 813–20. revisited. Bildung und soziale Ungleichheit vom
Hedström, P. and Swedberg, R. 1998 ‘Social Kindergarten bis zur Hochschule [Education
Mechanisms: An Introductory Essay’ in Inequality Revisited. Education and Social
P. Hedström and R. Swedberg (eds) Social Inequality from Kindergarten to University],
Mechanisms: An Analytical Approach to Wiesbaden: VS Verlag.
Social Theory, Cambridge: Cambridge Kunter, M., Schümer, G., Artelt, C.,
University Press. Baumert, J., Klieme, E., Neubrand, M.,
Hedström, P. and Ylikoski, P. 2010 ‘Causal Prenzel, M., Schiefele, U., Schneider,
Mechanisms in the Social Sciences’, Annual W., Stanat, P., Tillmann, K.‐J. and Weiß,
Review of Sociology 36(1): 49–67. M. 2002 PISA2000. Dokumentation der
Hurrelmann, B., Hammer, M. and Niess, Erhebungsinstrumente [PISA 2000. Documen-
F. 1995 Leseklima in der Familie [Reading tation of Survey Instruments], Berlin: Max‐
Climate in the Family], Gütersloh: Planck‐Institut für Bildungsforschung.
Bertelsmann‐Stiftung. Lareau, A. 1987 ‘Social Class Differences
Irwin, S. 2009 ‘Locating Where the Action in Family‐School Relationships: The
Is: Quantitative and Qualitative Lenses Importance of Cultural Capital’, Sociology
on Families, Schooling and Structures of of Education 60(2): 73–85.
Social Inequality’, Sociology 43(6): 1123–40. Lareau, A. 2000 Home Advantage:
Jackson, M. (ed.) 2013 Determined to Social Class and Parental Intervention in
Succeed? Performance versus Choice in Elementary Education (2nd edn), Lanham,
Educational Attainment, Stanford, CA: MD: Rowman & Littlefield.
Stanford University Press. Lareau, A. and Weininger, E.B. 2003
Kalter, F. and Kroneberg, C. 2014 ‘Between ‘Cultural Capital in Educational Research:
Mechanism Talk and Mechanism Cult. A Critical Assessment’, Theory and Society
New Emphases in Explanatory Sociology 32(5): 567–606.
and Empirical Research’, Kölner Zeitschrift Looker, E.D. and Pineo, P.C. 1983
für Soziologie und Sozialpsychologie 66(S1): ‘Social Psychological Variables and their
91–115. Relevance to the Status Attainment of
© London School of Economics and Political Science 2018 British Journal of Sociology 0(0)
Structural and motivational mechanisms of academic achievement 21
British Journal of Sociology 0(0) © London School of Economics and Political Science 2018