Predicting students' academic performance based on school and socio-demographic characteristics

Abstract

Students' trajectories into university are often uniquely dependent on school qualifications though these alone are limited as predictors of academic potential. This study endorses this, examining associations between school grades, school type, school performance, socio-economic deprivation, neighbourhood participation, sex and academic achievement at a British university. Consistent with past research, large entry-level differences between students are generally narrowed by final year at university. Students from the most deprived areas performed less well than more affluent students. Asian and black students performed less well than white students. Female students performed better than their male counterparts. Contrasting with past research, though school performance was positively associated with entry grades, students from low-performing schools were more likely to achieve the highest degree classifications. Additionally, independent school students performed less well than comprehensive school students at final year despite entering with higher grades. These variations exemplify how patterns observed nationally may differ between universities.

Full text

Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=cshe20
Download by: [82.41.43.9] Date: 06 December 2016, At: 08:42
Studies in Higher Education
ISSN: 0307-5079 (Print) 1470-174X (Online) Journal homepage: http://www.tandfonline.com/loi/cshe20
Predicting students' academic performance based
on school and socio-demographic characteristics
Tamara Thiele, Alexander Singleton, Daniel Pope & Debbi Stanistreet
To cite this article: Tamara Thiele, Alexander Singleton, Daniel Pope & Debbi Stanistreet
(2016) Predicting students' academic performance based on school and socio-
demographic characteristics, Studies in Higher Education, 41:8, 1424-1446, DOI:
10.1080/03075079.2014.974528
To link to this article: http://dx.doi.org/10.1080/03075079.2014.974528
© 2014 The Author(s). Published by Taylor &
Francis.
Published online: 27 Nov 2014.
Submit your article to this journal Article views: 2453
View related articles View Crossmark data
Citing articles: 2 View citing articles

Predicting students’academic performance based on school and
socio-demographic characteristics
Tamara Thiele
a
*, Alexander Singleton
b
, Daniel Pope
c
and Debbi Stanistreet
c
a
Department of Psychological Science, University of Liverpool, Eleanor Rathbone Building,
Bedford Street South, Liverpool L69 7ZA, UK;
b
Department of Geography and Planning,
University of Liverpool, Jane Herdman Building, Liverpool L69 3GP, UK;
c
Department of
Public Health and Policy, Institute of Psychology, Health and Society, University of
Liverpool, Whelan Building, Liverpool L69 3GB, UK
Students’trajectories into university are often uniquely dependent on school
qualifications though these alone are limited as predictors of academic potential.
This study endorses this, examining associations between school grades, school
type, school performance, socio-economic deprivation, neighbourhood
participation, sex and academic achievement at a British university. Consistent
with past research, large entry-level differences between students are generally
narrowed by final year at university. Students from the most deprived areas
performed less well than more affluent students. Asian and black students
performed less well than white students. Female students performed better than
their male counterparts. Contrasting with past research, though school
performance was positively associated with entry grades, students from low-
performing schools were more likely to achieve the highest degree classifications.
Additionally, independent school students performed less well than comprehensive
school students at final year despite entering with higher grades. These variations
exemplify how patterns observed nationally may differ between universities.
Keywords: education; attainment; contextual background; inequality
Despite a dramatic increase in higher education (HE) participation in England over the
last half century, the under-representation of students from socio-economically disad-
vantaged backgrounds remains a glaring reality (Blanden and Machin 2004; Breen and
Jonsson 2005; Croxford and Raffe 2013; Haveman and Smeeding 2006; Singleton
2010a). These students are known as Widening Participation (WP) students, who
along with students with disabilities and some ethnic minority groups are currently
under-represented in HE (Gorard 2008; Mason and Sparkes 2002). Differences in
HE participation are largely attributed to the poorer school-level academic qualifica-
tions obtained by a large proportion of students within low socio-economic status
(SES) classifications and are associated with educational disadvantage (Chowdry
et al. 2013; Steele, Vignoles, and Jenkins 2007; Sutton Trust 2005). Further, research
comparing the academic performance of students from different school types and
© 2014 The Author(s). Published by Taylor & Francis.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommon s.
org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
*Corresponding author. Email: t.thiele@liverpool.ac.uk
Studies in Higher Education, 2016
Vol. 41, No. 8, 1424–1446, http://dx.doi.org/10.1080/03075079.2014.974528

backgrounds in HE suggests that school qualifications do not necessarily represent ‘true
academic potential’(Hoare and Johnston 2010; Ogg, Zimdars, and Heath 2009; Peers
and Johnston 1994; Zimdars 2007). In particular, the finding that students from inde-
pendent schools tend to enter university with higher grades than students from (non-
fee paying) state schools but perform less well once at university when entry grades
are held constant is regarded as rationale for utilising contextual data alongside
school grades in the university admissions process (HEFCE 2003,2014; Hoare and
Johnston 2010; Naylor and Smith 2005, Sutton Trust 2010a).
Contextual data places academic attainment into the context of the circumstances in
which the results were obtained, including comparative school and socio-economic
data, with the principal aim of promoting fair access to HE (Bridger, Shaw, and
Moore 2012). The implementation of this alongside school grades is often rec-
ommended on the grounds that currently university admissions systems in the UK
focus almost entirely on students’past academic results, often without giving any con-
sideration to the context in which these were achieved (Chowdry et al. 2013; Gorard
2008; HEFCE 2014; Mullen 2011; Sutton Trust 2010b).
Though the usage of contextual data in admission has been historically controver-
sial, associated with positive discrimination and social engineering (Henry 2013), it has
progressively become more acceptable and is now being actively encouraged provided
that institutional policy is transparent and based on evidence (Department for Business,
Innovation & Skills 2011). A survey looking at the use of contextual data in admissions
carried out by SPA in 2012 found that out of 17 higher education institutions (HEIs)
approximately one-third (37%) were using contextual data in admissions and 57%
were planning to use it (Moore, Mountford-Zimdars, and Wiggans 2013). Though
the type of contextual information used varies widely between institutions, this gener-
ally includes information relating to students’personal details, school and college data,
and area-level data (Bridger, Shaw, and Moore 2012; Moore, Mountford-Zimdars, and
Wiggans 2013). However, research examining this and relationships between contex-
tual background characteristics and achievement in general is highly limited (Gorard
2008; Moore, Mountford-Zimdars, and Wiggans 2013; Zimdars 2007). Addressing
these issues is critical as it could help mitigate inequalities persistent in UK HE partici-
pation (Sutton Trust 2005). The extent of such inequalities is reflected empirically with
more than two-fifths of students studying at Oxbridge Universities being privately edu-
cated, despite the fact that just 7% of schools in the UK are independent (Sutton Trust
2008).
Greater degrees of socio-economic inequality and social stratification have been
associated with pervasive negative educational, health and crime-related outcomes (Fein-
stein 2003; Cabinet Office 2011; Wilkinson and Pickett 2009). Promoting fair access to
HE is considered imperative as a means to reducing these inequalities and associated det-
rimental repercussions (Haveman and Smeeding 2006; Kelly 2012). Concomitantly,
increasing equality of opportunity is important for raising skill levels, contributing to
national productivity and social mobility (Dorling 2010). These are considered priority
issues in the UK where currently the usage of contextual data as part of the university
admissions process is promoted as part of a broader widening participation policy
agenda (Cable and Willets 2011, Croxford and Raffe 2013; Milburn 2009). This is
specifically promoted within government reports proposing ‘HEIs should continue to
use, and where possible, expand the range of all the information available to them to
identify the best students with the greatest potential to reach the highest academic
achievement’(Department for Children, Schools and Families 2008,12).
Studies in Higher Education 1425

The ever-increasing pressure to widen access to prospective students from socio-
economically deprived groups has been greatly augmented by a substantial increase
in tuition fees from £3600 to a maximum of £9000 per annum (Cable and Willets
2011; Harrison 2011). Consequentially, identifying and targeting people from socio-
economically disadvantaged areas with academic potential has become of even
greater financial importance to HEIs, as potential to charge the full uncapped amount
is only permissible if the Office for Fair Access (OFFA) considers that programmes
are being made available to everyone with academic potential (Browne 2010;
Clayton 2012). Moreover, the usage of contextual data in the admissions process
could help widen participation and identify students who may require academic
support (Henry 2013). However, justifying the implementation of contextual data in
university admissions necessitates a robust evidence base, which can adequately
demonstrate the impact of students’background characteristics on academic perform-
ance (Bridger, Shaw, and Moore 2012). This paper seeks to expand this evidence base
by examining relationships between a selected range of school and socio-demographic
factors identified as predictors of educational disadvantage and academic performance
at a British university.
Contextual background characteristics
The literature identifies a range of background characteristics that influence educational
disadvantage and differentiated performance including school effects, socio-economic
background and personal attributes.
In comparison to students from more affluent backgrounds, a disproportionate
number of students from socio-economically disadvantaged backgrounds attend
poor-performing schools and come from neighbourhoods with low participation in
HE (HEFCE, 2010), and this can impact on their chances of entering HE (Forsyth
and Furlong 2003; Gorard 2012; Leathwood 2004; Voigt 2007). Although there is
an overlap between school type and school performance, where fee-paying schools
are predominantly higher performing, the associations between school type and
school performance with academic performance at university has been found to
differ between studies (HEFCE 2003,2005,2014; Smith and Naylor 2001). Indeed,
the average performance of students at a school does not appear to have a consistent
effect on academic attainment in HE (Ogg, Zimdars, and Heath 2009). There is dis-
agreement regarding the direction of the effect of school performance on academic
attainment (HEFCE 2003; Smith and Naylor 2001) including whether or not school
performance has a significant effect at all (HEFCE 2014; Hoare and Johnson 2010).
Furthermore, even though fee-paying schools tend to have better overall perform-
ance, a ‘school type effect’has been documented whereby for a given set of A-level
results, the degree performance of students who attended state schools has been
found to be higher, compared to those who attend private schools, when all other
factors are held equal (HEFCE 2003,2005,2014; Hoare and Johnston 2010; Naylor
and Smith 2005;Sutton Trust 2010b). This ‘school type effect’has been evidenced
in numerous studies, where it is considered to make a ‘strong case’for making lower
offers to individuals from disadvantaged backgrounds as on average their performance
at HE would at least match that of an independent school student (HEFCE 2003,2005;
Henry 2013; Kirkup et al. 2010; Naylor and Smith 2005; Smith and Naylor 2001). The
justification is based largely on the assumption that independent school pupils are at an
advantage over students from state schools with a similar level of ability, who may be in
1426 T. Thiele et al.

an environment that prevents them from achieving grades reflective of their true aca-
demic potential (McNabb, Pal, and Sloane 2002). School type differences in HE
achievement appear to be less marked between students with the highest A-level
achievement and HEIs with highest entry requirements. This has led researchers to
question whether the ‘school type effect’exists at these institutions in the past
(HEFCE 2003,2014; Parkes 2011).
A further variable that is associated with disadvantage and individual performance
is socio-economic background. This attribute also often interacts with school effects
given patterns of social selection associated with school admissions policy (Singleton
et al. 2011). Various studies have found that students from the least affluent socio-econ-
omic groups tend to perform less well than their more affluent peers (HEFCE 2014;
Hoare and Johnston 2010; Smith and Naylor 2001). However, much of the research
examining these socio-economic differences in HE attainment has used the National
Statistics Socio-Economic Classification (NS-SEC), the method currently used to ident-
ify SES during the university admissions process (Harrison 2011; Harrison and Hatt
2009; Hoare and Johnston 2010; Singleton 2010b). However, a number of flaws
have been identified with the use of NS-SEC as a contextual background characteristic,
particularly as around 25% of students do not provide this self-identified non-manda-
tory information on application to HE, and those who omit this, often fit into target
WP populations (Harrison and Hatt 2009,2010; Singleton 2010b).
An alternative approach to NS-SEC utilises postcodes, linking individuals to a dom-
icile location by geo-coding home postcode. However, in presenting such analysis, this
is accompanied by an important caveat that the measure relates to the context of an area
in which a student lived, rather than an attribute they personally possess (Gorard 2012;
Osborne and Shuttleworth 2004). That said, for the majority of undergraduate admis-
sions, NS-SEC is also not an individual measure, as this relates to parental occupation,
although geographic context could perhaps be considered applicable at a household
scale.
By attaching locations to the domicile postcodes of students, these can be linked to a
range of indicators of locational context, each of which pertains to a spatial unit of a
given zonal size. Such indicators include the Index of Multiple Deprivation (IMD),
which is a well-recognised measure of deprivation, comprising data pertaining to
seven different dimensions (Income, Employment, Health and Disability, Education,
Skills and Training, Barriers to Housing and Services, Living Environment and
Crime) (Flouri, Mavroveli, and Midouhas 2013). IMD scores are derived at the scale
of Lower Layer Super Output Areas (LSOAs), which are areas containing between
400 and 1200 households. The IMD has, however, received surprisingly little attention
in educational research compared to other fields, despite being recommended by the
HEFCE (2007) as a means of identifying people from NS-SEC groups 4–7 (Broecke
and Nicholls 2007; Feinstein 2003; Harrison 2011; Lupton 2004).
Afurther measure that has received relatively little attention in educational research,
despite it being devised by HEFCE to identify those from backgrounds with lower
levels of participation in HE, is the Participation Of Local Areas classification
(POLAR 3) (Corver 2010). POLAR 3 was created by HEFCE, by ranking 2001
Census Area Statistic (CAS) Wards by their young participation rates for the combined
2005–2009 cohorts. There are a total of 8850 CAS wards in England and Wales with an
average population of just under 6000 (Finney and Jivraj 2013). The POLAR 3 classi-
fication reports the rates of participation for those wards and is typically divided into
quintiles. There are also limited examples of research using the POLAR classification,
Studies in Higher Education 1427

which is particularly surprising considering this is used by HEFCE for calculating
widening participation funding, and by the Higher Education Statistics Agency
(HESA) to measure institutional performance (HEFCE 2012,2014). However, recently
HEFCE (2014) used both POLAR 3 and the Indices of Deprivation Affecting Children
as postcode-based measures of disadvantage and found that students from neighbour-
hoods with lower levels of participation in HE and students from less affluent areas,
respectively, were consistently less likely to achieve a 2:1 or a first-class degree at
university.
Finally, personal characteristics such as sex and ethnicity are also known to influ-
ence academic performance (Ackerman, Kanfer, and Beier 2013). Research suggests
that on average females generally achieve higher grades than males throughout edu-
cation, with some studies reporting that males may be more likely to achieve a first-
class degree (Dayioğlu and Türüt-Aşik 2007; Gneezy, Niederle, and Rustichini
2003; Hu and Wolniak 2013; McCrum 1994,1996; McNabb, Pal, and Sloane 2002;
Mellanby, Martin, and O’Doherty 2000; Pomerantz, Altermatt, and Saxon 2002;
Sheard 2009). Though the present study does not focus on ethnicity, significant differ-
ences in performance and participation have been documented between ethnic groups.
In the UK, white students as an overall category have been found to perform slightly
better than students who were not self-identified as white (Broecke and Nicholls
2007; HEFCE 2014; Jacobs 2008; Richardson 2008).
Previous studies have examined associations between students’background charac-
teristics and academic performance nationally and at individual universities (HEFCE
2003,2005,2014;Henry2013; Smith and Naylor 2001). However, no previous case
studies have been found which use both postcode-based measures of disadvantage
along with school background information to identify educational disadvantage
despite the limitations associated with measures such as NS-SEC and known differences
in student composition existent between HEIs (Gibbons and Vignoles 2012; Reay,
Crozier, and Clayton 2010;Reayetal.2001; Singleton 2010a,2010b). This is critical
from an admissions perspective as it is the responsibility of individual HEIs to ensure
that their fair admissions policies are grounded in empirical evidence and it is in their
interests to target those students with the academic potential to perform well in their
studies. The present study at a British university endorses this by investigating the
extent to which students’contextual background characteristics influence academic/
degree performance.
Method
Measuring and modelling contextual background and achievement
This study examines data from a British university, one of the six original ‘red brick’
civic universities and a founding member of the Russell Group. Traditionally, such elite
universities in the UK have tended to have an over-representation of students from
more affluent backgrounds and are more selective, with higher entry requirements
(Sutton Trust 2010b). However, the fact that the university campus is based in a city
with some of the most socio-economically deprived areas in the country means that tra-
ditionally the university has attracted a relatively high proportion of applicants from
low SES backgrounds.
Table A1 shows how the University of Liverpool (UoL) compares in the recruit-
ment of under-represented groups both nationally and in relation to other Russell
1428 T. Thiele et al.

Group universities. Compared to other Russell Group universities, the UoL has a higher
intake of students from low participation neighbourhoods (LPN) (POLAR 3), students
from lower socio-economic groups (NS-SEC 4–7) and students from low-income
households. Despite this, compared to national averages, the intake of students from
disadvantaged backgrounds at the UoL is generally lower. However, the proportion
of students from state schools at the UoL is similar to the national average, in this
way in particular the UoL differs from other Russell Group universities.
Data for the study were obtained from the university central student database, which
includes all necessary student background information and tracks performance from the
point of application through to graduation. For the purposes of this study, only students
registered on full-time three-year classified degrees entering the university between
2004/2005 and 2009/2010, and then graduating three years after their entry were
included. This was the last entry year that allowed analysis of both entry and exit
points. There were no significant changes to the university’s admission policies or
grading criteria during this time period, so data were stratified by year of entry but
also treated as a single data set. The data set contains socio-demographic (sex, age,
ethnicity and domicile), school attended, prior attainment (based on Universities and
Colleges Admissions Service [UCAS] Tariff Points), and HE performance
information for 5369 students.
Where data were missing for key variables, students were excluded from the analy-
sis. This was primarily socio-economic information as not all postcodes could be
matched to IMD and POLAR 3 scores, though a small proportion of academic infor-
mation was also missing. Analyses were designed to explore research questions
centred on the relationships between school type, school performance, socio-economic
background and academic performance at university. The full list of variables included
in the analysis is described in Table A2 (see Appendix).
In order to make comparisons between degree programmes and students as fair as
possible, students registered on four- and five-year programmes including Veterinary
Science, Medicine and Dentistry were excluded from the data set. Secondly, only
students with a postcode within England were included in the analyses as the IMD is
produced separately in each of the four UK administrations. Students from outside
the UK were also excluded. Finally, only data for students who completed three-year
degrees programmes successfully were included in this study.
Univariable logistic regression was carried out summarising the association
between contextual background characteristics and academic performance; this was
defined as good (2:1, first classification) versus other. Multivariable logistic regression
was carried out to identify which factors were independently associated with academic
performance. All analyses were undertaken using SPSS (version 21).
Results
There was no evidence of collinearity between the explanatory factors used in the
analysis (p> .05).
Students in the data set were predominantly self-classified as white (91.5%) and
aged below 21 (92.4%). The percentage of males and females in the study was rela-
tively uniform (58.4% females) in aggregate, though differences in the proportion of
males and females across university faculties varied.
Table A3 presents a descriptive summary of the association between each of the
contextual background characteristics and academic performance. Significant
Studies in Higher Education 1429

differences were observed in the Universities and Colleges Admissions Service
(UCAS) tariff points of students from different school backgrounds, quintiles of
socio-economic deprivation, neighbourhoods with different levels of participation in
HE, different ethnicities and between males and females.
The majority of students came from comprehensive schools and sixth form colleges
(3431, 75.2%). Students who attended grammar schools and sixth form colleges came
into university with the highest UCAS tariff points (Table A2). However, similar findings
were not reflected in university attainment, as students from comprehensive schools
achieved the highest average final year grades, and generally performed better than stu-
dents from all other school types. Conversely, students from independent schools andstu-
dents from the category of schools ‘state other’achieved the lowest average grades at
university compared to students from other school types. Moreover, these students
were also significantly more likely to achieve degree classifications below a 2:1.
Students who attended schools that were considered high performing in terms of
A-level performance/equivalent entered university with higher A-level (or equivalent)
grades compared to those who went to lower performing schools (p< .0005). However,
by final year at university, differences in overall mark averages were no longer
statistically significant. A similar pattern was observed when comparing the academic
performance of students with different levels of neighbourhood participation in HE
(POLAR 3). Here, students from high participation neighbourhoods (HPN) entered
university with significantly higher UCAS tariff points than students from LPN;
however, by the final year at university, differences between students from LPN and
HPN were no longer statistically significant.
With respect to multiple deprivation (IMD), the number of students within each
quintile increased as deprivation decreased, so there were 2.17 times more students
in quintile 5 (least deprived) than quintile 1 (most deprived). There was also a positive
relationship between IMD quintile and UCAS tariff points such that students from the
least deprived areas entered university with the highest UCAS tariff points and conver-
sely, students from the most deprived areas entered with the lowest UCAS tariff points.
By contrast, material deprivation predicted only slight differences in academic achieve-
ment once students were at university. Indeed, only students from the most deprived
socio-economic quintile achieved slightly less well on average and were more likely
to achieve lower second class or a lower degree classification, but this was not statisti-
cally significant.
With regard to ethnicity, there were significant group differences in students’UCAS
tariff points, which decreased but were largely consistent at university. Black students
entered university with the lowest number of UCAS tariff points and once at university
achieved the second lowest average attainment after Asian students. Students from both
of these ethnic groups were also more likely to get a degree of 2:2 or lower.
Finally, a consistent statistically significant association was observed for sex in
relation to academic attainment in both school and university attainment. Males entered
university with significantly lower grades than females, achieved lower average marks
at university and were also less likely to get a ‘good degree’(2:1 or above).
Table A4 summarises the results for contextual background factors in relation to final
degree classification. Compared to students from the most deprived quintile (IMD), stu-
dents from all of the other IMD quintiles were slightly more likely to obtain a good
degree; however, this association was only statistically significant for IMD quintiles 4
and 5. Secondly, compared to students who had attended comprehensive schools, stu-
dents from the four other types of school were less likely to obtain a good degree, but
1430 T. Thiele et al.

this association was only statistically significant for students from independent schools.
Thirdly, ethnicity was significantly associated with degree performance, where com-
pared to white students, Asian and black students were more than 50% less likely to
achieve a good degree. Finally, sex and UCAS tariff points were both found to predict
significant differences in the probability of getting a good degree.
There were no significant differences in the likelihood of achieving a good degree at
university between groups of students who came from neighbourhoods with low/high
participation and between those students who attended schools with low/high levels of
performance (Table A4).
Multivariable logistic regression was carried out to estimate how students’background
characteristics including neighbourhood participation (POLAR 3), deprivation, edu-
cational background and personal characteristics influenced their odds of getting a good
degree. Table A5 presents these results incorporating the seven background characteristics
simultaneously and degree performance as a binary outcome (1st and 2:1 versus all others).
Whilst the majority of associations between socio-economic deprivation and edu-
cational performance were initially found to be statistically significant in the univari-
able analysis, in multivariable analysis socio-economic deprivation was observed to
exert less of an influence on the chances of getting a good degree after allowing for
the effects of the other variables (Table A4). Compared to students from the most
deprived socio-economic quintile (quintile 1), students from quintile group 4 were
most likely to achieve a good degree (odds ratio (OR) = 1.34; 95% CI = 0.99–1.82);
comparisons with other quintiles did not achieve statistical significance.
Compared to comprehensive school students, multivariable analyses revealed that
students from all other types of school had significantly lower odds of achieving a
good degree (with the exception of the category ‘state other’where the association
was not statistically significant) (OR = 0.58; 95% CI = 0.27–1.24). The difference
was greatest between students from comprehensive schools and students from indepen-
dent schools who were found to be 40% less likely to achieve a good degree (OR =
0.61; 95% CI = 0.48–0.77).
Though performance of school did not significantly predict differences in edu-
cational performance univariately, there was a significant association in the multivari-
able analysis. Here it was found that students from schools that were high performing
were significantly less likely to achieve a good degree than those from low-performing
schools (OR = 0.78; 95% CI = 0.62–0.98). Associations between neighbourhood par-
ticipation (POLAR 3) and degree classification remained non-significant.
Ethnicity remained a significant predictor of degree performance in the multivari-
able analysis. Compared to white students, Asian students were 48% less likely to
achieve a ‘good degree’(OR = 0.52, CI = 0.33–0.82) and similarly black students
were 53% less likely to achieve a good degree (OR = 0.47, CI = 0.24–0.89)
Students’sex also remained a significant predictor in multivariable analysis. Com-
pared to males, females were more than 50% more likely to achieve a good degree (OR
= 1.52; 95% CI = 1.30–1.79). Finally, students’UCAS tariff points (entry-level per-
formance) were also significantly associated with university performance in the multi-
variable analysis (OR = 1.01; 95% CI = 1.01–1.01).
Discussion
The principal aim of this research was to explore the relationship between students’con-
textual background characteristics and academic performance at university in order to
Studies in Higher Education 1431

identify which characteristics were associated with students’chances of achieving a
‘good degree’(upper second- or first-class degree). No other case studies have been
found where this is explored using both postcode-based measures of disadvantage and
school background information. Hence, a critical part of this research involved investi-
gating whether patterns identified in previous studies were also evidenced at this British
university and exploring potential variations which could exist as a consequence of the
differences in student intake and performance which are known to exist even between
elite universities (Hoare and Johnston 2010; Singleton 2010b).
Principal findings from results
A crucial part of the analysis involved addressing the extent to which school grades are
representative of ‘true academic’potential by comparing group differences in attain-
ment at school compared to university. Statistically significant associations were
observed between all of the contextual background characteristics: IMD, school type,
school performance, neighbourhood participation, sex and ethnicity and students’
school grades (UCAS tariff points). With the exception of IMD and POLAR 3, all of
these variables were also significantly associated with university attainment, though
compared to differences observed in entry grades, these associations differed substan-
tially in terms of size/direction. Additionally, consistent with other studies, school
grades were also found to be a strong and significant predictor of academic performance
at university (HEFCE 2012,2014; Kirkup et al. 2010; McKenzie and Schweitzer 2001).
Socio-economic differences persisted in final year performance at university, but
only approached statistical significance between students from the most deprived
areas and those from the second least deprived group ceteris paribus. Additionally, stu-
dents from the most deprived areas were found to be more likely to achieve degree
classifications of 2:2 or below. Unlike the IMD, POLAR 3, hence coming from neigh-
bourhoods with low or high levels of participation in HE, did not predict significant
differences in final year performance.
Conversely, the type of school that students attended had a significant and differen-
tial impact on school achievement compared to university achievement. Overall, stu-
dents from comprehensive schools were more likely to achieve a good degree than
students from all other school types. Students from independent schools were found
to be less likely to achieve a ‘good degree’compared to students from comprehensive
schools despite being more likely to enter university with higher grades. To an extent,
this is similar to the relationship between school performance and academic achieve-
ment as students from low-performing schools achieved slightly higher final averages
than their counterparts from high-performing schools.
With regard to ethnicity, there were significant differences in students’UCAS tariff
points and similar differences were observed in university attainment where Asian stu-
dents and black students were significantly less likely to achieve good degrees compared
to white students. Finally, one variable that did not follow any of the patterns evidenced
hitherto was sex. Males entered university with lower grades than females, and once at
university were also less likely to achieve either a first or an overall good degree.
How do these findings relate to the current evidence base/other studies?
Differences in academic achievement by SES have been reported in numerous studies
(Delaney, Harmon, and Redmond 2011; Hoare and Johnston 2010). These are typically
1432 T. Thiele et al.

evidenced from a young age and span across a number of subjects (Aikens and Barbarin
2008; Chowdry et al. 2013; Coley 2002; Evans 2008). Consequently, the differences in
academic performance observed in this study are not surprising, particularly as differ-
ences that persisted by the third year at university were primarily between students who
came from the most deprived areas and students from some of the most affluent areas.
Similar findings have been reported in past research and attributed to differences in a
range of factors including family support, family history in HE (Allardice and
Blicharski 2000; HEFCE 2014; Richardson, Abraham, and Bond 2012), term time
working (Moreau and Leathwood 2006; Salamonson et al. 2012), and differences in
expectations and aspirations (Pampaka, Williams, and Hutcheson 2012; Thomas 2001).
In this research, the gap in academic attainment between students from neighbour-
hoods with low or high levels of participation in HE (POLAR 3) was only significant at
entry level. Though slight differences prevailed even by final year at university, these
were not significant. HEFCE (2014) found little variation in academic performance
when POLAR quintiles were examined together and entry grades were taken into
account. However, they also found that in particular students from the areas with
lowest participation rates (POLAR quintile 1) performed significantly less well and
achieved the lowest proportions of high degree classifications.
Ethnic differences in achievement are widespread and have been reported in numer-
ous studies (Broecke and Nicholls 2007; Hoare and Johnston 2010; Woolf, Potts, and
McManus 2011). Though variations exist with regard to the particular ethnic groups
that perform less well, overall, white students have generally been found to perform
better than students from other ethnic minorities (HEFCE 2014; Jacobs 2008; Richard-
son 2008). The latter was not evidence in this study at entry level or in university attain-
ment. However, compared to white students, black and Asian students were particularly
less likely to achieve a good degree, and the proportion of black and Asian students
achieving degree classifications of 2:2 or below was alarmingly high. Ethnic differ-
ences in achievement such as these are not atypical; these have been identified in
other studies and require further exploration.
Though previous studies have reported males as being up to 50% more likely to
achieve first-class degrees than females (McCrum 1996; Mellanby, Martin, and O’Doh-
erty 2000), in more recent studies females have been found to outperform males, con-
sistent with the findings from this study (Dayioğlu and Türüt-Aşik 2007; Sheard 2009).
The gap in academic performance between males and females is alarming and though
this does not relate to WP per se requires further exploration, particularly, as differences
in subject choice between males and females are acknowledged and studies have found
that trends may vary by age, between subjects and have differential effects on employ-
ability outcomes (Ackerman, Kanfer, and Beier 2013; Hu and Wolniak 2013; Richard-
son and Woodley 2003). Consequently, future research should explore interactions
between sex, subject choice, SES and outcomes.
The findings pertaining to the relationship between school performance and aca-
demic achievement are difficult to reconcile unreservedly with past research, as it is
not only highly limited, but findings have been mixed and largely inconsistent
(HEFCE 2003; Hoare and Johnston 2010; Smith and Naylor 2001). According to
HEFCE (2003), findings have been particularly mixed because the effect of school per-
formance varies largely, depending on factors such as A-level points, students’sex and
subject. The findings of this research are consistent with those of Smith and Naylor
(2001) who also found evidence of a positive association between attendance at
lower performing schools and degree performance. They argue that when comparing
Studies in Higher Education 1433

two students with the same A levels, the student who is less advantaged, coming from a
state school with lower overall performance, is more likely to have greater underlying
ability. This suggests that the school qualifications achieved by students from low-per-
forming schools may not represent their true/academic potential. However, these results
contrast with the findings of HEFCE (2003) regarding the direction of the association
between school performance and degree attainment and other studies where no signifi-
cant association was found (HEFCE 2014; Hoare and Johnston 2010). This highlights
the need for further research exploring variations in school performance.
Students from independent schools did not enter the university with the highest
grades, as was the case in other studies (HEFCE, 2003,2005,2014; Hoare and Johnston
2010; Naylor and Smith 2005; Smith and Naylor 2001). However, consistent with past
research, once at university, students from independent schools achieved lower results
than students from all but one of the other school types, including comprehensive
school students (Hoare and Johnston 2010). This effect is said to occur largely
because independent school students are at an advantage over students from state
schools and this advantage is reflected in their qualifications and progression to HE
(Dorling 2010). This advantage is associated with factors including the quality of edu-
cation students receive, types of subjects offered, a greater focus on preparing students
for university and indeed altogether better resourcing as educational spending (23%) on
privately educated children in Britain is more than almost any other rich nation in
Europe (Hoare and Johnston 2010). According to Ogg, Zimdars, and Heath (2009),
teaching effects at independent schools inflate the qualifications obtained by their stu-
dents. Either or both of these arguments could explain why comprehensive school stu-
dents enter university with lower results than independent school students but all other
factors held equal, finalise with higher results.
Implications of these findings
Results from the present study support the notion that variables such as school grades
are not the only causal factors behind patterns in academic attainment in HE and should
be accompanied by information that puts this attainment into context. This is evidenced
by the fact that even though all of the variables, namely IMD, school type, school per-
formance, neighbourhood participation, sex and ethnicity, were significantly associated
with entry grades (UCAS tariff points), overall, differences between groups largely
decreased at university. Additionally, associations between school performance and
school type differed at university compared to entry level. Specifically, students from
low-performing schools and independent schools were less likely to achieve a ‘good
degree’compared to students from comprehensive and high-performing schools,
despite entering university with higher grades. Hence, together, these results support
the implementation of contextual information in university admissions. This is
widely advocated in educational policy by the government and different HEIs where
similar findings have been documented and considered to make a ‘strong case’for
making reduced offers to students from particularly disadvantaged backgrounds
(HEFCE 2014; Henry 2013; Hoare and Johnston 2010; Kirkup et al. 2010; Naylor
and Smith 2005; Smith and Naylor 2001;Sutton Trust 2010b).
Despite the increased interest in ‘contextualised admissions’, there is little publicly
available research detailing why and how particular background characteristics are used
at individual HEIs (Moore, Mountford-Zimdars, and Wiggans 2013). Additionally,
critics argue that making reduced offers to students from socio-economically/
1434 T. Thiele et al.

educationally disadvantaged backgrounds discriminates against students from affluent
backgrounds/independent schools and may reduce academic excellence at HEIs (Sin-
gleton, 2010b). The findings of the present study represent a powerful riposte to
such arguments, providing additional support for the ‘school type effect’and the
notion that school grades may not reflect true academic potential.
Limitations and directions for future research
The present research has various limitations that must be taken into consideration when
interpreting these findings. Firstly, it is not possible to control for all factors that affect
university attainment. Some prominent factors that were not controlled for include
working during term time (Moreau and Leathwood 2006), living at home (Holdsworth
2006), student engagement (Hu and Wolniak 2013; Johnson and Reynolds 2013),
family history in HE (Allardice and Blicharski 2000; Delaney, Harmon, and
Redmond 2011) and individual characteristics including intelligence (Farsides and
Woodfield, 2003; Haworth et al. 2011; Mega, Ronconi, and De Beni 2014). Further-
more, a main limitation of this study is that it did not include students that entered uni-
versity via non-standard routes, even though research suggests that these students may
be more likely to have suffered from educational disadvantage (Broecke and Nicholls
2007; Gorard 2012). Secondly, this study included only those students who success-
fully completed their degrees; and not those who failed or dropped out; future research
should examine trends in academic achievement in these groups of students. Hence, it is
important to take into account that these findings are not representative of all university
applicants. Another limitation of this research is that the IMD relates only to LSOAs
and not postcodes or smaller geographical units (Gorard and See 2009; Hoare and John-
ston 2010; Smith and Naylor 2001). Indeed, Gorard (2012) highlights that some of the
most deprived families actually live in heavily polarised areas, such as inner London
boroughs. Despite this, the IMD was found to be a useful tool for identifying significant
differences in performance. Similarly, POLAR 3 is also restricted in this sense as it is
also based on aggregate data. Thus, it must be considered that trends relating to both
IMD and POLAR 3 do not necessarily relate to individuals themselves but rather to
the areas in which they are based. A final and common limitation relevant to the
present study lies in the high proportion of missing data as this could significantly
bias analyses and results, and is something that must be taken into account (Gorard
2008,2012).
The need for further research exploring educational disadvantage and variations in
academic performance is indisputable, as a number of questions remain unanswered.
This is partly due to the strict exclusion criteria that were used to make comparisons
between students as fair as possible. Future studies should also focus on those students
who entered university via non-standard routes and compare performance of students
with different types of school qualifications as trajectories throughout HE may be influ-
enced by these factors. Secondly, it is critical for future research to explore why differ-
ences in achievement at university occur in order to support students and identify those
at risk of dropping out, failing and/or not achieving a ‘good degree’. Thirdly, associ-
ations between contextual background and academic performance in programmes
extending beyond three years including medicine, dentistry and veterinary science
require further exploration. Addressing the paucity of research on this is essential as
these programmes are highly oversubscribed, selective and competitive, having
higher entry requirements than most other programmes. Finally, it is important to
Studies in Higher Education 1435

note that even within elite universities, there are major differences in student compo-
sition and student performance. Consequently, though the present case study illustrates
important differences between different groups of students at a British university, ana-
lyses must be expanded to include other universities. In particular, analysis should
focus on the most competitive and selective universities, known as elite universities
as these are often criticised for having comparatively less WP students to other univer-
sities (Singleton 2010b). The lack of research on this is problematic and must be
addressed as convincing evidence is necessary for guiding the decision-making
process in the implementation of contextual data alongside school qualifications.
Concluding comments
Understanding factors which are associated with differences in HE participation and
performance is crucial, particularly given the expansion that the British HE system
has undergone in the last decade, changing financial regimes, and the inequalities
which persist (Breen and Jonsson 2005; Sutton Trust 2010b). Though there is
general awareness that prior opportunities and social background impact on academic
performance and subsequently access to HE, utilising contextual data for admissions in
an evidence-based manner is less well understood (Bridger, Shaw, and Moore 2012;
Zimdars 2007). This may represent one of the underlying reasons why current contex-
tual considerations are so limited (Zimdars 2007).
However, the usage of contextual information in admissions can be regarded as a
mechanism for ameliorating the current admissions systems by addressing limitations
related to the usage of examination marks as these alone are not considered an appro-
priate proxy of an applicant’s true academic potential (Ogg, Zimdars, and Heath 2009).
The present study provides insight into the associations between different background
characteristics and academic outcomes, contributing to the evidence base that advocates
the implementation of contextual data alongside school grades during the admissions
process. This illustrates how contextual data can be utilised to identify students with
school qualifications that may not reflect the extent of their academic potential, but
also to help identify those at risk of underperforming once they are in HE (Bridger,
Shaw, and Moore 2012; Lupton 2004; Ogg, Zimdars, and Heath 2009). Thus, going
beyond purely theoretical analysis, the practical repercussions of the present research
could help raise academic attainment to higher levels, and more generally, improve
the student experience (Bridger, Shaw, and Moore 2012). Unfortunately, research on
this remains highly limited despite the fact it is critical to HEIs, the UK government
and, most importantly, to students themselves as changes in admissions may impact
on their life chances and subsequent career opportunities (Jacobs 2008; Mullen
2011). Further research is necessary to ensure that university policies are based on
firm evidence to safeguard fair access to HE.
References
Ackerman, P. L., R. Kanfer, and M. E. Beier. 2013. “Trait Complex, Cognitive Ability, and
Domain Knowledge Predictors of Baccalaureate Success, STEM Persistence, and Gender
Differences.”Journal of Educational Psychology 105 (3): 911–27. doi:10.1037/a0032338.
Aikens, N. L., and O. Barbarin. 2008. “Socioeconomic Differences in Reading Trajectories: The
Contribution of Family, Neighbourhood, and School Contexts.”Journal of Educational
Psychology 100 (2): 235–51.
1436 T. Thiele et al.

Allardice, M., and J. Blicharski. 2000. “‘Tracking Students’Progression: Learning their
Lessons.”Journal of Widening Participation and Lifelong Learning 2 (3): 32–37. doi:10.
1037/0022-0663.100.2.235.
Blanden, J., and S. Machin. 2004. “Educational Inequality and the Expansion of UK Higher
Education.”Scottish Journal of Political Economy 51 (2): 230–49. doi:10.1111/j.0036-
9292.2004.00304.x.
Breen, R., and J. O. Jonsson. 2005. “Inequality of Opportunity in Comparative Perspective:
Recent Research on Educational Attainment and Social Mobility.”Annual Review of
Sociology 31: 223–43.
Bridger, K., J. Shaw, and J. Moore. 2012. Fair Admissions to Higher Education: Research to
Describe the Use of Contextual Data in Admissions at a Sample of Universities and
Colleges in the UK. Cheltenham: Supporting Professionalism in Admissions. http://www.
spa.ac.uk/documents/Full_SPA_Contextual_data_Research_Report.Feb2012.pdf
Broecke, S., and T. Nicholls. 2007. Ethnicity and Degree Attainment. Department for Education
and Skills: Research Report RW92.
Browne, J. 2010. Securing a Sustainable Future for Higher Education –The Independent
Review of Higher Education Funding and Student Finance. London: Department for
Business, Innovation and Skills. http://www.bis.gov.uk/assets/biscore/corporate/docs/s/10-
1208-securing-sustainable-higher-education-browne-report.pdf
Cabinet Office. 2011. Opening Doors, Breaking Barriers: A Strategy for Social Mobility –
Update on Progress since April 2011. London: Cabinet Office. http://dera.ioe.ac.uk/id/
eprint/14537
Cable, V., and D. Willets. 2011. Guidance to the Director of Fair Access. London: Department
for Business, Innovation & Skills. https://www.gov.uk/government/publications/guidance-
to-the-director-of-fairaccess
Chowdry, H., C. Crawford, L. Dearden, A. Goodman, and A. Vignoles. 2013. “Widening
Participation in Higher Education: Analysis Using Linked Administrative Data.”Journal
of the Royal Statistical Society: Series A (Statistics in Society) 176 (2): 431–57. doi:10.
1111/j.1467-985X.2012.01043.x.
Clayton, M. 2012. “On Widening Participation in Higher Education Through Positive
Discrimination.”Journal of Philosophy of Education 46 (3): 414–31. doi:10.1111/j.1467-
9752.2012.00858.x.
Coley, R. J. 2002. An Uneven Start: Indicators of Inequality in School Readiness. Princeton, NJ:
Educational Testing Service, Policy Information Center. https://www.ets.org/Media/
Research/pdf/PICUNEVENSTART.pdf
Corver, M. 2010. Trends in Young Participation in Higher Education: Core Results for
England. Bristol: HEFCE. http://www.hefce.ac.uk/pubs/hefce/2010/10_03//
Croxford, L., and D. Raffe. 2013. “Differentiation and Social Segregation of UK Higher
Education, 1996–2010.”Oxford Review of Education 39 (2): 172–92. doi:10.1080/
03054985.2013.784193.
Dayioğlu, M., and S. Türüt-Aşik. 2007. “Gender Differences in Academic Performance in a
Large Public University in Turkey.”Higher Education 53: 255–77. doi:10.2307/29735052.
Delaney, L., C. Harmon, and C. Redmond. 2011. “Parental Education, Grade Attainment an
Earnings Expectations Among University Students.”Economics of Education Review 30
(6): 1136–52.
Department for Business, Innovation & Skills. 2011. Higher Education: Students at the Heart of
the System. London: Department for Business, Innovation & Skills. https://www.gov.uk/
government/uploads/system/uploads/attachment_data/file/32409/11-944-higher-education-
students-at-heart-of-system.pdf
Department for Children, Schools and Families. 2008. National Council for Educational
Excellence Recommendations. Nottingham: Department for Children, Schools and
Families. http://webarchive.nationalarchives.gov.uk/20130401151715/https://www.eduat
ion.gov.uk/publications/eOrderingDownload/NCEE-Recommendations.pdf
Dorling, D. 2010. Injustice: Why Social Inequality Persists. Bristol: Policy Press.
Farsides, T., and R. Woodfield. 2003. “Individual Differences and Undergraduate Academic
Success: The Roles of Personality, Intelligence, and Application.”Personality and
Individual Differences 34 (7): 1225–43. doi:10.1016/S0191-8869(02)00111-3.
Studies in Higher Education 1437

Feinstein, L. 2003. “Inequality in the Early Cognitive Development of British Children in the
1970 Cohort.”Economica 70: 73–97. doi:10.1111/1468-0335.t01-1-00272.
Finney, N., and S. Jivraj. 2013. “Ethnic Group Population Change and Neighbourhood
Belonging.”Urban Studies 50 (16): 3323–3341.
Flouri, E., S. Mavroveli, and E. Midouhas. 2013. “Residential Mobility, Neighbourhood
Deprivation and Children’s Behaviour in the UK.”Health & Place 20 (March): 25–31.
Forsyth, A., and A. Furlong. 2003. Losing Out? Socioeconomic Disadvantage and Experience
in Further and Higher Education. Bristol: Policy Press.
Gibbons, S., and A. Vignoles. 2012. “Geography, Choice and Participation in Higher Education
in England.”Regional Science and Urban Economics 42 (1–2): 98–113.
Gneezy, U., M. Niederle, and A. Rustichini. 2003. “Performance in Competitive Environments:
Gender Differences.”The Quarterly Journal of Economics 118 (3): 1049–74.
Gorard, S. 2008. “Who Is Missing from Higher Education?”Cambridge Journal of Education
38 (3): 421–37. doi:10.1080/03057640802286863.
Gorard, S. 2012. “Who Is Eligible for Free School Meals? Characterising Free School Meals as a
Measure of Disadvantage in England.”British Educational Research Journal 38 (3): 1003–
17.
Gorard, S., and B. H. See. 2009. “The Impact of Socio-economic Status on Participation and
Attainment in Science.”Studies in Science Education 45 (1): 93–129.
Harrison, N. 2011. “The Mismeasure of Participation: How Choosing the ‘Wrong’Statistic
Helped Seal the Fate of Aimhigher.”Higher Education Review 45 (1): 30–61.
Harrison, N., and S. Hatt. 2009. “Knowing the ‘Unknowns’:Investigating the Students Whose
Social Class Is Not Known at Entry to Higher Education.”Journal of Further and Higher
Education 33 (4): 347–57.
Harrison, N., and S. Hatt. 2010. “‘Disadvantaged Learners’: Who Are We Targeting?
Understanding the Targeting of Widening Participation Activity in the United Kingdom
Using Geo-demographic Data from Southwest England.”Higher Education Quarterly 64
(1): 65–88. doi:10.1111/j.1468-2273.2009.00431.x.
Haveman, R. H., and T. M. Smeeding. 2006. “The Role of Higher Education in Social
Mobility.”The Future of Children 16 (2): 125–50. doi:10.2307/3844794.
Haworth, C. M. A., K. Asbury, P. S. Dale, and R. Plomin. 2011. “Added Value Measures in
Education Show Genetic as Well as Environmental Influence.”PLoS ONE 6 (2), art. no.
e16006. doi:10.1371/journal.pone.0016006.
HEFCE (Higher Education Funding Council for England). 2003. Schooling Effects on Higher
Education Achievement (HEFCE Report 2003/32). Bristol: HEFCE. http://webarchive.
nationalarchives.gov.uk/20100202100434/http://www.hefce.ac.uk/pubs/hefce/2003/0332.
html
HEFCE (Higher Education Funding Council for England). 2005. Schooling Effects on Higher
Education Achievement: Further Analysis –Entry at 19 (HEFCE Report 2005/09). Bristol:
HEFCE. http://heer.qaa.ac.uk/SearchForSummaries/Summaries/Pages/SCEE157.aspx
HEFCE (Higher Education Funding Council for England). 2007. Higher Education Outreach:
Targeting Disadvantaged Learners (HEFCE Report 2007/12). Bristol: HEFCE. http://www.
hefce.ac.uk/pubs/hefce/2007/07_12/
HEFCE (Higher Education Funding Council for England). 2010. Trends in Young Participation
in Higher Education: Core Results for England (HEFCE Report 2010/03). Bristol: HEFCE.
https://www.hefce.ac.uk/pubs/year/2010/201003/
HEFCE (Higher Education Funding Council for England). 2012. POLAR 3: Young
Participation Rates in Higher Education (HEFCE Report 2012/26). Bristol: HEFCE.
http://www.hefce.ac.uk/pubs/year/2012/201226/
HEFCE (Higher Education Funding Council for England). 2014. Differences in Degree
Outcome: Key Findings (HEFCE Report 2014/03). Bristol: HEFCE. http://www.hefce.ac.
uk/pubs/year/2014/201403/#d.en.86821
Henry, J. 2013. “Comprehensive School Pupils Do Better at University, Two New Studies
Confirm.”The Guardian, June 15. Accessed http://www.guardian.co.uk/education/2013/
jun/16/accesstouniversity-private-schools/
Hoare, A., and R. Johnston. 2010. “Widening Participation Through Admissions Policy –A
British Case Study of School and University Performance.”Studies in Higher Education
36 (1): 21–41. doi:10.1080/03075070903414297.
1438 T. Thiele et al.

Holdsworth, C. 2006. “‘Don’t You Think You’re Missing Out Living at Home?’Student
Experiences and Residential Transitions.”The Sociological Review 54 (3): 495–519.
doi:10.1111/j.1467-954X.2006.
Hu, S., and G. C. Wolniak. 2013. “College Student Engagement and Early Career Earnings:
Differences by Gender, Race/Ethnicity, and Academic Preparation.”The Review of
Higher Education 36 (2): 211–33. doi:10.1353/rhe.2013.0002.
Jacobs, S. 2008. Ethnicity, Gender and Degree Attainment Project: Final Report. Manchester:
Higher Education Academy. http://www.heacademy.ac.uk/assets/documents/research/
Ethnicity_Gender_Degree_Attainment_Jan08.pdf
Johnson, M. K., and J. R. Reynolds. 2013. “Educational Expectation Trajectories and
Attainment in the Transition to Adulthood.”Social Science Research 42 (3): 818–35.
doi:10.1016/j.ssresearch.2012.12.003.
Kelly, A. 2012. “Measuring ‘Equity’and ‘Equitability’in School Effectiveness Research.”
British Educational Research Journal 38 (6): 977–1002. doi:10.1080/01411926.2011.
605874.
Kirkup, C., R. Wheater, J. Morrison, and B. Durbin. 2010. Use of an Aptitude Test in University
Entrance –A Validity Study: Updated Analyses of Higher Education Destinations,
Including 2007 Entrants. Slough: National Foundation for Educational Research.
Leathwood, C. 2004. “A Critique of Institutional Inequalities in Higher Education (or an
Alternative to Hypocrisy for Higher Educational Policy).”Theory and Research in
Education 2(1):31–48. doi:10.1177/1477878504040576.
Lupton, R. 2004. Schools in Disadvantaged Areas: Recognising Context and Raising Quality
(Case Paper, 76). London: Centre for Analysis of Social Exclusion, London School of
Economics and Political Science. http://eprints.lse.ac.uk/6321/
Mason, C., and V. J. Sparkes. 2002. “Widening Participation in Physiotherapy Education: Part
1: Introduction.”Physiotherapy 88 (5): 273–75. doi:10.1016/S0031-9406(05)61419-8.
McCrum, N. G. 1994. “The Academic Gender Deficit at Oxford and Cambridge.”Oxford
Review of Education 20 (1): 3–26.
McCrum, N. G. 1996. “Gender and Social Inequality at Oxford and Cambridge Universities.”
Oxford Review of Education 22 (4): 369–97.
McKenzie, K., and R. Schweitzer. 2001. “Who Succeeds at University? Factors Predicting
Academic Performance in First Year Australian University Students.”Higher Education
Research & Development 20: 21–33. doi:10.1080/07924360120043621.
McNabb, R., S. Pal, and P. Sloane. 2002. “Gender Differences in Educational Attainment: The
Case of University Students in England and Wales.”Economica 69 (275): 481–503.
Mega, C., L. Ronconi, and R. De Beni. 2014. “What Makes a Good Student? How Emotions,
Self-regulated Learning, and Motivation Contribute to Academic Achievement.”Journal of
Educational Psychology 106 (1): 121–31.
Mellanby, J., M. Martin, and J. O’Doherty. 2000. “The ‘Gender Gap’in Final Examination
Results at Oxford University.”British Journal of Psychology 91 (3): 377–90. doi:10.
1348/000712600161880.
Milburn, A. 2009. Unleashing Aspirations: The Final Report of the Panel on Fair Access to the
Professions. London: Cabinet Office. http://hdl.voced.edu.au/10707/19259
Moore J., A. Mountford-Zimdars, and J. Wiggans. 2013. Contextualised Admissions:
Examining the Evidence. Cheltenham: Supporting Professionalism in Admissions
Programme.
Moreau, M. P., and C. Leathwood. 2006. “Balancing Paid Work and Studies: Working-Class
Students in Higher Education.”Studies in Higher Education 31 (1): 23–42.
Mullen, F. 2011. Participation in Higher Education: Statistics (Briefing 10/08). Scottish
Parliament Information Centre. http://www.scottish.parliament.uk/SPICeResources/
Research%20briefings%20and%20fact%20sheets/SB10-08.pdf
Naylor, R., and J. Smith. 2005. “Schooling Effects on Subsequent University Performance:
Evidence for the UK University Population.”Economics of Education Review 24 (5):
549–62.
Ogg, T., A. Zimdars, and A. Heath. 2009. “Schooling Effects on Degree Performance: A
Comparison of the Predictive Validity of Aptitude Testing and Secondary School Grades
at Oxford University.”British Educational Research Journal 35 (5): 781–807.
Studies in Higher Education 1439

Osborne, B., and I. Shuttleworth. 2004. “Widening Access to Higher Education in the UK:
Questioning the Geographic Approach.”Higher Education Management and Policy 16
(1): 101–18.
Parkes, G. 2011. Academic Performance of Undergraduate Students at Cambridge by School/
College Background. Cambridge: University of Cambridge. http://www.admin.cam.ac.uk/
offices/admissions/research/docs/school_performance.pdf
Peers, I. S., and M. Johnston. 1994. “Influence of Learning Context on the Relationship Between
A-Level Attainment and Final Degree Performance: A Meta-analytic Review.”British
Journal of Educational Psychology 64 (1): 1–18.
Pomerantz, E. M., E. R. Altermatt, and J. L. Saxon. 2002. “Making the Grade but Feeling
Distressed: Gender Differences in Academic Performance and Internal Distress.”Journal
of Educational Psychology 94 (2): 396–404.
Reay, D., G. Crozier, and J. Clayton. 2010. “‘Fitting In’or ‘Standing Out’: Working-Class
Students in UK Higher Education.”British Educational Research Journal 36 (1): 107–
24. doi:10.1080/01411920902878925.
Reay, D., J. Davies, M. David, and S. J. Ball. 2001. “Choices of Degree or Degrees of Choice?
Class, ‘Race’and the Higher Education Choice Process.”Sociology 35 (4): 855–74. doi:10.
1177/0038038501035004004.
Richardson, J. T. E. 2008. Degree Attainment, Ethnicity and Gender: A Literature Review.
York: Higher Education Academy.
Richardson, M., C. Abraham, and R. Bond. 2012. “Psychological Correlates of University
Students’Academic Performance: A Systematic Review and Meta-analysis.”
Psychological Bulletin 138 (2): 353–87. doi:10.1037/a0026838.
Richardson, J. T. E., and A. Woodley. 2003. “Another Look at the Role of Age, Gender and
Subject as Predictors of Academic Attainment in Higher Education.”Studies in Higher
Education 28 (4): 475–93.
Salamonson, Y., B. Everett, J. Koch, S. Andrew, and P. M. Davidson. 2012. “The Impact of
Term-Time Paid Work on Academic Performance in Nursing Students: A Longitudinal
Study.”International Journal of Nursing Studies 49 (5): 579–85.
Sheard, M. 2009. “Hardiness Commitment, Gender, and Age Differentiate University Academic
Performance.”The British Journal of Educational Psychology 79 (1): 189–204.
Singleton, A. D. 2010a. “The Geodemographics of Educational Progression and Their
Implications for Widening Participation in Higher Education.”Environment and Planning
A42 (11): 2560–80.
Singleton, A. D. 2010b. Educational Opportunity: The Geography of Access to Higher
Education. Farnham: Ashgate.
Singleton, A. D., P. A. Longley, R. Allen, and O. O’Brien. 2011. “Estimating Secondary School
Catchment Areas and the Spatial Equity of Access.”Computers, Environment and Urban
Systems 35 (3): 241–49.
Smith, J., and R. Naylor. 2001. “Determinants of Degree Performance in UK Universities: A
Statistical Analysis of the 1993 Student Cohort.”Oxford Bulletin of Economics and
Statistics 63 (1): 29–60.
Steele, F., A. Vignoles, and A. Jenkins. 2007. “The Effect of School Resources on Pupil
Attainment: A Multilevel Simultaneous Equation Modelling Approach.”Journal of the
Royal Statistical Society: Series A (Statistics in Society) 170 (3): 801–24.
Sutton Trust. 2005. Applications, Offers and Admissions to Research Led Policies. London:
Sutton Trust. http://heer.qaa.ac.uk/SearchForSummaries/Summaries/Pages/AWP26
Sutton Trust. 2008. University Admissions by Individual Schools. London: Sutton Trust. http://
www.suttontrust.com/research/universityadmissionsbyindividualschools
Sutton Trust. 2010a. Comprehensive Pupils Outperform Independent and Grammar Pupils in
University Degrees. London: Sutton Trust. http://www.suttontrust.com/news/news/
comprehensive-pupils-outperform/
Sutton Trust. 2010b. Responding to the New Landscape for University Access. London: Sutton
Trust. http://socialwelfare.bl.uk/subjectareas/serviceactivity/education-skills/suttontrust/
responding10.aspx
Thomas, L. 2001. “Power, Assumptions and Prescriptions: A Critique of Widening Participation
Policy-making.”Higher Education Policy 14 (4): 361–76. doi:10.1016/S0952-8733(01)
00026-5.
1440 T. Thiele et al.

Voigt, K. 2007. “Individual Choice and Unequal Participation in Higher Education.”Theory and
Research in Education 5 (1): 87–112.
Wilkinson, R., and K. Pickett. 2009. The Spirit Level: Why More Equal Societies Almost Always
Do Better. London: Allen Lane.
Woolf, K., H. W. W. Potts, and I. C. McManus. 2011. “Ethnicity and Academic Performance in
UK Trained Doctors and Medical Students: Systematic Review and Meta-analysis.”BMJ
342: d901. doi:10.1136/bmj.d901.
Zimdars, A. 2007. “Challenges to Meritocracy? A Study of the Social Mechanisms in Student
Selection and Attainment at the University of Oxford.”PhD thesis, Open Access
Publications, University of Oxford, Oxford.
Studies in Higher Education 1441

Appendix
Table A1. UoL widening access performance.
1.
LPN
2. Lower socio-economic
groups (NS-SEC 4–7)
3. State schools/
colleges
4. Low-income
households
UoL 8.5% 22.0% 87.6% 27.4%
English Russell
Group
5.5% 18.3% 72.5% 21.4%
England 10.2% 30.9% 88.5% 31.6%
Note: 1–4 from the UoL 2014 Access Agreement.
Table A2. Description of outcome (educational performance) and predictor (contextual
factors) variables.
Variables Description
Outcome variables
Average performance Students’university academic performance was represented as a
percentage indicating the average mark achieved for each year of
their degree. Most analyses focused on final year performance as this
has a 70% weighting on the overall degree
Final degree
classification
Degrees were classified according to the UK undergraduate degree
classification system; first class typically being awarded to those
who achieved 70% and above, 2:1 to those who achieved between
60% and 69%, 2:2 awarded to those achieving 50% and 59%, and
third-class degrees awarded to students achieving between 40% and
49%. For most analyses, a binary classification (1st and 2:1 versus
others) was used
Predictor variables
UCAS tariff points UCAS tariff points are a system used for allocating points to post-16
qualifications in the UK (e.g. for A levels, A = 120, B = 100, C = 80,
etc.). This was calculated from students’three highest qualifications
and used as a measure of prior achievement for entry to HE
School type The type of school students’attended for their A levels were organised
into five categories including independent schools, state grammar
schools, state comprehensives, sixth form colleges and the category
labelled state other (includes voluntary aided schools, voluntary
controlled schools, technical colleges and adults colleges)
School performance School performance data were used to contextualise prior attainment,
represented by the overall percentage of students gaining 5A*–Eor
more at A levels or equivalent. Based on this, a binary classification
was created where ‘high’-performing schools, represented those
schools where 82.5% of students and above achieved 5A*–Eor
more at A level or their equivalent. ‘Low’-performing schools were
those where less than 82.5% of students achieved 5A*–E or more at
A level or their equivalent based on averages reported in Department
for Education performance tables
(Continued.)
1442 T. Thiele et al.

Table A2. (Continued .)
Variables Description
Neighbourhood
participation
POLAR 3 data were matched to the CAS wards to illustrate the typical
HE participation profile within which students were domiciled.
POLAR 3 data were reported as five quintiles: ordered from ‘1’
(lowest participation) to ‘5’(highest participation). A binary
classification was created to compare performance of students
residing in areas of lowest participation (1 and 2) to others (3, 4 and 5).
Quintiles 1 and 2 are those areas, which attract additional widening
participation funding for each student domiciled within them
Multiple deprivation The IMD was used to identify the multiple facets of total deprivation.
Students’postcodes were matched to LSOAs, which contain an
average of 1500 households. These were then used to append IMD
scores provided that students had a valid English postcode. There are
32,482 LSOAs in England. IMD ranks LSOA with 1 as most
deprived and 32,482 as least deprived. For the analyses IMD scores
were divided into quintiles, where quintile 1 includes the most
deprived areas and quintile 5 includes the least deprived
Sex/ethnicity Sex and ethnicity were self-reported by students during the university
application process. Students’ethnicities were categorised as one of
the following: white, Asian, black, Chinese, and mixed and other
Table A3. Descriptive breakdown of characteristics of study sample for students in all three-year
degree programmes.
Indicator of student performance
UCAS tariff
points
Final year
average
Degree –
first class
Degree –
class 2:1
Degree –
class
2:2/3rd
Variable No. MSD MSD No. % No. % No. %
School type:
Independent 564 359.40 74.92 61.59 6.48 53 9.40 345 61.17 166 29.43
Grammar 511 389.73 77.74 62.52 6.20 60 11.45 336 64.12 128 24.43
Comprehensive 2350 348.99 87.21 62.73 6.61 327 13.84 1506 63.73 530 22.43
Sixth form 1081 389.54 89.09 62.20 6.65 136 12.58 673 62.26 272 25.16
State (other) 55 335.64 82.07 61.79 6.53 4 7.27 34 61.82 17 30.91
p< .0005 p= .01 p= .01 p= .01 p= .01
School performance:
High 3526 375.03 83.62 62.58 7.03 439 11.84 2394 64.58 874 23.58
Low 1822 333.26 96.61 62.36 6.44 136 15.61 520 59.70 215 24.68
p< .0005 p= .42 p< .01 p< .01 p< .01
Deprivation:
a
1 655 301.67 153.58 61.48 7.75 89 13.61 381 58.26 184 28.13
2 687 336.84 122.84 62.50 6.54 90 13.12 437 63.70 159 23.18
3 917 340.25 122.14 62.51 6.51 125 13.66 561 61.31 229 25.03
4 1153 350.81 113.98 62.83 6.21 150 13.01 753 65.31 250 21.68
5 1423 361.74 108.14 62.50 6.42 189 13.28 909 63.88 325 22.84
p< .0005 p< .01 p< .01 p< .01 p< .01
(Continued.)
Studies in Higher Education 1443

Table A3. (Continued.)
Indicator of student performance
UCAS tariff
points
Final year
average
Degree –
first class
Degree –
class 2:1
Degree –
class
2:2/3rd
Variable No. MSD MSD No. % No. % No. %
POLAR 3:
b
High 4010 364.70 87.83 62.46 6.37 510 12.72 2539 63.33 960 23.95
Low 1222 356.78 92.31 62.13 7.22 175 14.37 739 60.67 304 24.96
p= 0.01 p= .02 p= .18 p= .18 p= .18
Sex:
Males 1985 351.80 88.94 61.77 7.15 281 12.7 1290 58.1 649 29.23
Females 2949 370.10 88.73 62.93 6.11 423 13.4 2082 66.2 641 20.38
p< .0005 p< .0005 p< .0005 p< .0005 p< .0005
Ethnicity
White 4913 297.38 40.32 62.63 6.14 644 13.11 3125 63.62 1143 23.27
Asian 127 299.00 45.94 60.03 6.55 12 9.60 61 48.80 52 41.60
Black 65 266.84 52.66 60.76 5.87 7 10.77 32 49.23 26 40.00
Chinese 48 280.00 44.49 62.60 9.07 12 25.00 25 52.08 11 22.92
Mixed 111 361.74 108.14 62.67 7.13 17 15.32 68 61.26 26 23.42
Other 105 293.22 45.43 63.21 6.53 12 11.43 61 58.10 32 30.48
p< .0005 p< .0005 p< .0005 p< .0005 p< .0005
a
Defined by quintiles of IMD (1 = most deprived …5 = least deprived).
b
Neighbourhood HE participation.
Table A4. Unconditional bivariate logistic regression models for student characteristics with
final degree performance (2:1 and 1st versus lower classification).
Indicator variable OR ‘good degree’
Variable No. % OR 95% CI p-Value
State comprehensive (reference) 2334 51.5 1
Sixth form college 1068 23.6 0.85 0.72–1.01 .07
State other 54 1.2 0.65 0.36–1.17 .15
State grammar 521 11.5 0.87 0.69–1.08 .21
Independent school 556 12.3 0.69 0.56–0.85 < .0005
School performance:
High (reference) 3663 81.0 1
Low 857 19.0 0.96 0.80–1.14 .62
Deprivation:
a
1 (reference) 642 13.5 1
2 678 14.2 1.28 0.99–1.64 .06
3 907 19.0 1.134 0.90–1.43 .28
4 1145 24.0 1.37 1.09–1.71 .01
5 1401 29.4 1.33 1.07–1.65 .01
POLAR 3:
b
High (reference) 3964 76.8
Low 1198 23.2 0.97 .83–1.13 .65
(Continued.)
1444 T. Thiele et al.

Table A4. (Continued.)
Indicator variable OR ‘good degree’
Variable No. % OR 95% CI p-Value
Sex:
Males (reference) 2179 41.1 1
Females 3119 58.9 1.58 1.39–1.80 <.0005
Ethnicity
White (reference) 4913 91.5 1
Asian 127 2.4 0.43 0.30–0.62 <.0005
Black 65 2 0.45 0.27–0.75 <.01
Chinese 48 0.9 1.07 0.53–2.15 .86
Mixed 111 2.1 1.02 0.65–1.61 .93
Other 105 2 0.75 0.48–1.17 .20
UCAS points (continuous) 4952 92.2 1.01 1.01–1.01 <.0005
a
Defined by quintiles of IMD (1 = most deprived …5 = least deprived).
b
Neighbourhood HE participation.
Table A5. Multiple logistic regression including all student characteristics (deprivation (IMD),
school grades, school type, school performance, neighbourhood participation and sex) and final
year performance (2:1 and 1st versus lower categories).
Indicator variable OR ‘good degree (1st or 2:1)’
Variable No. % OR 95% CI p-Value
State comprehensive (reference) 1829 49.0 1
Sixth form college 968 26.0 0.67 0.55–0.82 <.0005
State other 35 0.9 0.58 0.27–1.24 .16
State grammar 416 11.6 0.71 0.54–0.94 .016
Independent school 482 12.9 0.61 0.48–0.77 <.0005
School performance:
Low (reference) 644 17.3 1
High 3086 82.7 0.78 0.62–0.98 .03
Deprivation:
a
1 (reference) 452 12.1 1
2 520 13.9 1.25 0.92–1.70 .16
3 725 19.4 1.03 0.76–1.39 .85
4 918 24.6 1.34 0.99–1.82 .06
5 1115 30.0 1.17 0.87–1.59 .31
POLAR 3:
b
Low (reference) 820 22.0 1
High 2910 78.0 1.08 0.86–1.34 .52
Sex:
Males (reference) 1520 40.8 1
Females 2210 59.3 1.52 1.30–1.79 <.0005
Ethnicity
White (reference) 4913 91.5 1
Asian 127 2.4 0.52 0.33–0.82 <.0005
(Continued.)
Studies in Higher Education 1445

Table A5. (Continued.)
Indicator variable OR ‘good degree (1st or 2:1)’
Variable No. % OR 95% CI p-Value
Black 65 2 0.47 0.24–0.89 .002
Chinese 48 0.9 1.33 0.52–3.38 .55
Mixed 111 2.1 1.33 0.73–2.40 .35
Other 105 2 0.84 0.47–1.53 .57
UCAS points (continuous) 4952 92.2 1.01 1.01–1.01 <.0005
a
Defined by quintiles of IMD (1 = most deprived …5 = least deprived).
b
Neighbourhood HE participation.
1446 T. Thiele et al.