Dimensions of quality

by Graham Gibbs
The higher Education Academy

The views expressed in this publication are those of

the authors and not necessarily those of the Higher
Education Academy.

ISBN 978-1-907207-24-2

© The Higher Education Academy

September 2010

The Higher Education Academy

Innovation Way
York Science Park
Heslington
York YO10 5BR

www.heacademy.ac.uk

Tel: +44 (0)1904 717500

Fax: +44 (0)1904 717505

All rights reserved. Apart from any fair dealing

for the purposes of research or private study,
criticism or review, no part of this publication may
be reproduced, stored in a retrieval system, or
transmitted, in any other form or by any other
means, graphic, electronic, mechanical, photocopying,
recording, taping or otherwise, without the prior
permission in writing of the publishers.

To request copies of this report in large print or in a

different format, please contact the Academy.

Designed by Daniel Gray

Printed by The Charlesworth Group

2
Dimensions of quality

by Graham Gibbs

2 Foreword by Craig Mahoney

4 1. Executive summary

8 2. Introduction

11 3. The nature of dimensions of quality

14 4. Presage dimensions of quality

19 5. Process dimensions

38 6. Product dimensions of educational quality

43 7. Summary and conclusions

50 8. About the author

51 9. Acknowledgements

52 10. References
The higher Education Academy

Foreword

The perennial debate about what constitutes quality in undergraduate education has
been reignited recently, not least by a range of published research, Select Committee
activity, tightening of resource, and the large-scale review by Lord Browne.
As the organisation dedicated to enhancing the quality of students’ learning
experiences, the Higher Education Academy is pleased, through this piece of work, to
contribute further to this important debate.
Our starting-point is twofold: first, that higher education should be a
transformative process that supports the development of graduates who can make
a meaningful contribution to wider society, local communities and to the economy.
Second, that any discussion around quality needs to be evidence-informed. As a
result, we identified a need to synthesise and make sense of the scattered research
in the field of higher education quality. We wanted to find out what the research
evidence tells us and what further work we can do to apply the relevant findings in
our quest to improve the quality of student learning in UK higher education.
Graham Gibbs states that the most important conclusion of this report is that
what best predicts educational gain is measures of educational process: in other
words, what institutions do with their resources to make the most of the students
they have. Examining the evidence, he draws conclusions about some key topics that
have been the subject of much debate around quality. For example, he concludes
that the number of class contact hours has very little to do with educational quality,
independently of what happens in those hours, what the pedagogical model is, and
what the consequences are for the quantity and quality of independent study hours.
He also reiterates research (Nasr et al., 1996) that shows that teachers who have
teaching qualifications (normally a Postgraduate Certificate in Higher Education, or
something similar) have been found to be rated more highly by their students than
teachers who have no such qualification. I think this is a crucial point. At the Academy
we believe that high quality teaching should be delivered by academic staff who are
appropriately qualified and committed to their continuing professional development.
To this end we will continue to provide and develop an adaptable framework for
accredited teaching qualifications in HE, incorporating the UK Professional Standards
Framework and other relevant teaching qualifications. We will also continue to work
with HEIs to develop and manage CPD frameworks for learning and teaching.
The report also concludes that some dimensions of quality are difficult to
quantify, and it is therefore difficult to see what effect they might have. Aspects of

2
 Dimensions of qualit y

departmental culture are one such area: whether teaching is valued and rewarded,
whether innovation in teaching is systematically supported and funded, etc. The
Academy has already conducted research into the reward and recognition of teaching
which showed that over 90% of academic staff thought that teaching should be
important in promotions. We will continue to focus on this work.
Some of the findings of this report may confirm aspects of institutional policy
on enhancing quality, some of them will prompt new and different approaches to
focused investment of funding and expertise in order to maximise educational gain,
particularly at a time of diminishing resource. Some of them will call into question
the efficacy and appropriateness of practices and policies, and cause us to look not at
how much is spent per capita, but on how it is spent; less on how many contact hours
are provided but with whom and with what consequences for independent learning;
on the extent to which we truly support and adopt the kinds of pedagogic practices
that engender students’ intrinsic engagement in their learning.
Graham argues for a better focus on evidence in order to understand quality
properly, to ensure that our quality process are informed to a greater extent by what
we know about what constitutes effective practice and about the extent to which
these practices are employed, to make better and more coordinated use of the full
range of available data, and to understand the relationship between them.
This paper is primarily for an audience of senior managers of HEIs – the
colleagues who develop and implement the kinds of institutional policies that have
the propensity to improve student learning and who conceptualise the frameworks
to support that vital process. We hope that this report will meaningfully inform both
policy and practice and look forward to following up this work in the coming months
by engaging with you in debates and discussions about the dimensions of quality.

Professor Craig Mahoney

Chief Executive
The Higher Education Academy.

3
The higher Education Academy

1. Executive summary

“A … serious problem with national magazine rankings is that from
a research point of view, they are largely invalid. That is, they are
based on institutional resources and reputational dimensions which
have only minimal relevance to what we know about the impact
of college on students … Within college experiences tend to count
substantially more than between college characteristics.”
—Pascarella, 2001

1.1 The focus of the report

This report has been written to contribute to the current debates about educational
quality in undergraduate education in the UK, and about the need to justify increases
in resources on the basis of indicators of educational quality. This report will identify a
range of dimensions of quality and examine the extent to which each could be considered
a valid indicator, with reference to the available research evidence. It attempts to identify
which kinds of data we should take seriously and which we should be cautious of placing
weight on. Some of these dimensions we might be wise to pay attention to currently lack
a solid evidence base, especially in relation to research carried out in the UK context, and
so the report also identifies priorities for research and for data collection and analysis.

1.2 T he approach taken to considering dimensions of quality

The report identifies which dimensions of educational quality can reasonably be

used to compare educational settings. It adapts Biggs’s ‘3P’ model (Biggs, 1993) of
‘presage’, ‘process’ and ‘product’ to categorise the variables under consideration
(see Section 3.2). The report examines a wide range of potential indicators. Presage
variables define the context before students start learning, process variables describe
what goes on as students learn, and product variables relate to the outcomes of
that learning. For presage and process variables the available evidence is examined
concerning the validity of the variable: the extent to which it predicts student
learning outcomes and educational gains. Each product variable is examined for its
ability to indicate comparative quality.

4
 Dimensions of qualit y

1.3 T he limitations of presage and product variables

Presage variables such as funding, research performance and, the reputation that
enables an institution to have highly selective student entry, do not explain much
of the variation between institutions in relation to educational gains. Measures of
educational product such as grades do reflect these presage variables, but largely
because the best students compete to enter the best-funded and most prestigious
institutions and the quality of students is a good predictor of products. Measures
of product such as retention and employability are strongly influenced by a raft of
presage variables that go well beyond those used by HEFCE in setting performance
benchmarks. The lack of comparability of degree standards proves an obstacle to
interpretation of student performance data in the UK. This makes interpreting and
comparing institutional performance extremely difficult.

1.4 The importance of process variables

What best predicts educational gain is measures of educational process: what

institutions do with their resources to make the most of whatever students they
have. The process variables that best predict gains are not to do with the facilities
themselves, or to do with student satisfaction with these facilities, but concern a
small range of fairly well-understood pedagogical practices that engender student
engagement. In the UK we have few data about the prevalence of these educational
practices because they are not systematically documented through quality assurance
systems, nor are they (in the main) the focus of the National Student Survey.
Class size, the level of student effort and engagement, who undertakes the
teaching, and the quantity and quality of feedback to students on their work are all
valid process indicators. There is sufficient evidence to be concerned about all four of
these indicators in the UK.

1.5 The importance of multivariate analysis

Few relationships between a single dimension of quality and a single measure of either
educational performance or educational gain can be interpreted with confidence
because dimensions interact in complex ways with each other. To understand what
is going on and draw valid conclusions it is necessary to have measures of a range of
dimensions of quality at the same time and to undertake multivariate analysis. Large-
scale multivariate analyses have been repeatedly undertaken in the US, and have
successfully identified those educational processes that affect educational gains, and
those that do not or that are confounded by other variables. In contrast there has
been little equivalent analysis in the UK. This is partly because data in the UK that

5
The higher Education Academy

could form the basis of multivariate analysis for that purpose are currently collected
by different agencies and have never been fully collated.

1.6 The importance of educational gain

Because educational performance is predicted by the entry standards of students,

to compare institutional performance in a valid way it is necessary to measure
educational gain: the difference between performance on a particular measure before
and after the student’s experience of higher education. While the most influential
US studies measure educational gain in a variety of ways, there is very little evidence
available in the UK about educational gain.

1.7 Dimensions of quality in different kinds of institutions

Institutions have different missions, and comparing them using product dimensions of
quality that are the goals of only a subset of the institutions leads to conclusions of
doubtful value. Process dimensions give a fairer comparative picture of quality than
do presage or product dimensions. However, different pedagogic phenomena, and
hence different process variables, are likely to be salient in different institutions. For
example, only some of the very different ways in which The Open University or the
University of Oxford achieve such high National Student Survey ratings are relevant
to other kinds of university.

1.8 Dimensions of quality in different departments

Indicators of dimensions of quality often vary widely between departments within

the same institution, for a variety of reasons. Prospective students need quality
information about the specific degree programme they wish to study at an institution
rather than about institutional averages or about clusters of degree programmes
aggregated into ‘subjects’ as at present. Providing such information at a sufficient level
of granularity may be impractical.

1.9 Dimensions of quality that are difficult to quantify

Studies of the characteristics of both institutions and departments that have been found
to be outstanding in terms of valid dimensions of educational quality have identified
process variables that would be extremely difficult to quantify or measure in a safe way,
such as the extent to which teaching is valued, talked about and developed.

6
 Dimensions of qualit y

1.10 Evidence of the products of learning

One of the most telling indicators of the quality of educational outcomes is the work
students submit for assessment, such as their final-year project or dissertation. These
samples of student work are often archived, but rarely studied. There is considerable
potential for using such products as more direct indicators of educational quality than
proxies such as NSS scores.

1.11  he potential for improved quality, and the evaluation of

T
improvements in quality

There is clear evidence that educational performance and educational gains can be
enhanced by adopting certain educational practices. In the US the National Survey
of Student Engagement (NSSE) has been used successfully by many institutions
to identify where there are weaknesses in current educational processes and to
demonstrate the positive impact of the introduction of certain educational practices.
Pooling data across such innovations then provides a valid basis to guide other
institutions in the adoption of practices that are likely to be effective. The NSS
cannot be used in the UK in the same way, despite its reliability. There is a valuable
role to be fulfilled by national agencies in supporting the use of valid measures of the
impact of changed educational practices, and in pooling evidence across institutions.

1.12 T he potential for informing potential students about quality

It seems unlikely that comparative indicators of quality currently available in the UK

could provide prospective students with a valid basis to distinguish between individual
courses with regard to their educational quality. The collation of currently available
data into league tables is invalid and misleading. Even in the US where a range of
more valid indicators are more widely available, those responsible for collecting and
interpreting the data counsel strongly against their collation into a single league table.

7
The higher Education Academy

2. Introduction

The extent to which indicators of quality have shaped both the politics of higher
education and institutional priorities is not a new phenomenon (Patrick and Stanley,
1998). However, there is currently increased emphasis on the overall quality of
undergraduate education in the UK. Data from a number of recent surveys and
studies have raised challenging issues about:

—— differences in quality between institutions within the UK that in the past

have, rightly or wrongly, been assumed to be broadly comparable;
—— differences in quality between national higher education systems, to whom in the
past the UK has been assumed, rightly or wrongly, to be superior, in the context
of an increasingly competitive international higher education market place;
—— the adequacy of national quality regimes that have emphasised scrutiny of
an institution’s quality assurance to a greater extent than of its educational
processes or outcomes of the kind emphasised in some of the recent high
profile surveys and studies.

A Parliamentary Select Committee (House of Commons, 2009) has taken

evidence from a wide range of sources and reached challenging conclusions both about
the quality of UK higher education and how that quality can be assured in the future.
Among all the debate there has sometimes been uncritical acceptance of
some sources of evidence that cannot bear the weight of interpretation, and also
rejection of evidence that deserves to be taken more seriously. Even in public
reports argument has sometimes made no use of available evidence. To give one
example the Quality Assurance Agency (2009) has responded to data that suggest
both that UK students might study significantly less hard than their European
counterparts, and that there are wide differences between institutions and subjects
within the UK in relation to how many hours are studied (HEPI, 2006, 2007;
Brennan et al., 2009). From the perspective of the current report the key questions
in this case are:

—— does it matter that some students receive less class contact than others?
Are class contact hours an indicator of quality?
—— does it matter that some students put in less total effort than others? Are
total student learning hours an indicator of quality?

8
 Dimensions of qualit y

In Section 5.2 below, evidence is reviewed that might inform the QAA’s current
position on this issue.
Similarly the findings of a study of student experience by the National Union of
Students (NUS, 2008) might be interpreted differently if they were informed by the
available empirical evidence on the issues it addresses, such as the effects of paid
work on students’ study hours.
The literature on the validity of indicators of quality is vast, widely dispersed
and mostly American. It tends to be focused on specific purposes, such as critiquing
a particular university league table, critiquing a particular government-defined
performance indicator, establishing the characteristics of a particular student feedback
questionnaire, or examining the characteristics of a particular indicator (such as
research performance). Much of this literature is technical in nature and written for
a specialist audience of educational researchers. The current report attempts to
bring much of this diverse literature together encompassing many (though not all)
dimensions of quality. It is not intended to be an exhaustive account, which would be a
very considerable undertaking, and it is written for a general audience. It will not delve
into statistical and methodological minutiae, although sometimes an appreciation of
statistical issues is important to understanding the significance of findings.
This report is intended to inform debate by policy formers of four main kinds:
those concerned about the overall quality of UK higher education; those concerned
with institutional and subject comparisons; those concerned with funding on the basis
of educational performance and those within institutions concerned to interpret
their own performance data appropriately. It may also be useful to those directing
resources at attempts to improve quality as it identifies some of the educational
practices that are known to have the greatest impact on educational gains.
It is important here to be clear what this report will not do. It will not review
alternative quality assurance regimes or make a case for any particular regime.
In identifying dimensions of quality that are valid it will, by implication, suggest
elements that should be included in any quality assurance regime, and those that
should not be included.
The report will not be making overall comparisons between the UK and other
HE systems, between institutions within the UK, between subjects nationally or
between subjects or departments within institutions. Rather the purpose is to
identify the variables that could validly be used in making such comparisons.
The report is not making a case for performance-based funding. Reviews of
the issues facing such funding mechanisms can be found elsewhere (Jongbloed and
Vossensteyn, 2001). However, valid indicators of quality will be identified that any
performance-based funding system might wish to include, and invalid indicators will
be identified that any performance-based system should eschew.
Finally, the report is not making a case for the use of ‘league tables’ based on
combinations of quality indicators, nor does it consider the issues involved in the
compilation and use of existing or future league tables. Trenchant and well-founded

9
The higher Education Academy

critiques of current league tables, and of their use in general, already exist (Bowden,
2000; Brown, 2006; Clarke, 2002; Eccles, 2002; Graham and Thompson, 2001; Kehm
and Stensaker, 2009; Thompson, 2000; Yorke, 1997). Some of these critiques cover
similar ground to parts of this report in that they identify measures commonly used
within league tables that are not valid indicators of educational quality.
Throughout the report there is a deliberate avoidance of using individual
institutions in the UK as exemplars of educational practices, effective or ineffective,
with the exception of a number of illustrations based on The Open University and
the University of Oxford. Despite being far apart in relation to funding, they are
close together at the top of rankings based on the NSS. They have achieved this
using completely different educational practices, but these practices embody some
important educational principles. They are so different from other institutions that
there can be little sense in which they can be compared, or copied, except at the
level of principles. It is these principles that the report seeks to highlight, because
they illuminate important dimensions of quality.

10
 Dimensions of qualit y

3. The nature of dimensions of quality

3.1 Conceptions of quality

‘Quality’ is such a widely used term that it will be helpful first to clarify the focus
of this report. There have been a number of attempts to define quality in higher
education, or even multiple models of quality (e.g. Cheng and Tam, 1997). The most
commonly cited discussion of the nature of quality in higher education in the UK is
that by Harvey and Green (1993), and their helpful nomenclature will be employed
here. First, quality is seen here as a relative concept – what matters is whether one
educational context has more or less quality than another, not whether it meets
an absolute threshold standard so that it can be seen to be of adequate quality,
nor whether it is reaches a high threshold and can be viewed as outstanding and
of exceptional quality, nor whether a context is perfect, with no defects. What is
discussed here is the dimensions that are helpful in distinguishing contexts from each
other in terms of educational quality.
Quality may also be seen to be relative to purposes, whether to the purposes
and views of customers or relative to institutional missions. This report does
not take customer-defined or institutionally defined conceptions of quality as its
starting point. Rather an effort will be made to focus on what is known about what
dimensions of quality have been found to be associated with educational effectiveness
in general, independently of possible variations in either missions or customers’
perspectives. The report will then return to the issue of institutional differences and
will comment in passing on differences between students in the meaning that can be
attached to quality indicators such as ‘drop-out’.
A further conception of quality made by Harvey and Green is that of quality
as transformation, involving enhancing the student in some way. This conception
comes into play when examining evidence of the educational gains of students (in
contrast to their educational performance). This transformation conception of
quality is also relevant when examining the validity of student judgements of the
quality of teaching, where what they may want teachers to do may be known from
research evidence to be unlikely to result in educational gains. What is focused on
here is not necessarily what students like or want, but what is known to work in
terms of educational effectiveness.
It is usual to distinguish between quality and standards. This distinction is most
relevant in Section 6.1 on student performance, where the proportion of ‘good

11
The higher Education Academy

degrees’ can be seen to be in part a consequence of the qualities of what students

have learnt and in part a consequence of the standards applied in marking the
products of student learning. This report will not focus on standards that have been
the subject of much recent debate; for example, concerning the operation of the
external examiner system.

3.2 Categorising dimensions of quality: presage, process and product

Education is a complex business with many interacting dimensions of quality in many

varied contexts. To understand what is going on it is necessary to have a way of
conceiving of the variables involved and of organising and interpreting studies of the
relationships between these variables. This report will adopt the commonly used ‘3P’
model (Biggs, 1993), which approaches education as a complex system with ‘Presage’,
‘Process’ and ‘Product’ variables interacting with each other. The ‘3P’ model is
essentially the same as that used by large-scale studies in the US (e.g. Astin, 1977,
1993): the ‘Input-Environment-Output’ model. Presage variables are those that exist
within a university context before a student starts learning and being taught, and
include resources, the degree of student selectivity, the quality of the students, the
quality of the academic staff and the nature of the research enterprise. None of these
presage variables determine directly how the educational process may be conducted,
although they often frame, enable or constrain the form education takes.
Process variables are those that characterise what is going on in teaching
and learning and include class size, the amount of class contact and the extent of
feedback to students. Process variables also include the consequences of variables
such as class size for the way students go about their learning, e.g. how those
variables impact on the quantity and quality of their study effort and their overall
level of engagement.
Product variables concern the outcomes of the educational processes and
include student performance, retention and employability. Products can also include
psychometric measures of generic outcomes of higher education, such as students’
ability to solve problems. In some studies the key product measure is not student
performance, but educational gain: the difference between performance on a
particular measure before and after the student’s experience of higher education. The
difference between performance and gain will be crucial in understanding dimensions
of quality, as we shall see.
The categorisation of variables as presage, process or product is not always
straightforward. For example, some process variables such as the level of student
engagement may be related to other process variables, such class size, which may
in turn be related to funding levels. Which are the presage variables and which
the products? Class size is not seen as a presage variable in the 3P model as it is
in part a consequence of policy decisions about how to use resources and in part

12
 Dimensions of qualit y

a consequence of educational decisions about teaching methods. The presage

variable of resources does not necessarily predict either. Nor is student engagement
conceived of in the 3P model as a product. Both class size and student engagement
are conceived of as part of the processes that may influence education outcomes,
which are categorised as products.
In examining the usefulness of potential performance indicators involved
in ‘league tables’, presage, process and product variables have sometimes been
subdivided into more categories within a more complex model (Finnie and Usher,
2005; Usher and Savino, 2006), but for the purposes of this report the simple 3P
model will suffice.
This report examines a wide range of presage, process and product variables in
turn, and in doing so identifies relationships that are known to exist between them.

13
The higher Education Academy

4. Presage dimensions of quality

This section considers four presage dimensions of quality:
funding, staff:student ratios, the quality of teaching staff and the
quality of students

4.1 Funding

Institutional funding predicts student performance to some extent. It predicts

cohort size (Bound and Turner, 2005), and class size predicts student performance
(see Section 5.1). Funding also affects the kind of teachers the institution can afford
to undertake the teaching and this affects student performance (see Section 4.3).
How much funding per student is allocated to the provision of learning resources
also predicts student study effort, which in turn predicts student performance
(see Section 5.2). However, funding predicts performance largely because the best
students go to the best-resourced institutions and the quality of the students predicts
their performance (see Section 4.4). A series of large-scale US studies have found
little or no relationship between institutional funding and measures of educational
gain (Pascarella and Terenzini, 2005).
Even the ability of institutional levels of funding to predict student performance
is somewhat limited. A study in the US has compared groups of colleges with near
identical funding per student and found graduate completion rates varying between
35% and 70% (Ewell, 2008), so the differences in what colleges do with their funding
must be very wide. In addition, institutions with similar levels of performance display
widely varying levels of funding with some receiving only 60% of the revenues
per student that others receive, but achieving near identical performance on a
whole range of outcome measures. Twenty institutions that had been identified
as unusually educationally effective, in relation to student retention and levels of
student engagement, have been matched, in terms of a whole raft of other potentially
influential presage variables such as student selectivity, with a group of markedly
less effective institutions. There was found to be no difference in the level of funding
allocated to teaching between the effective institutions and the carefully matched less
effective group (NCHEMS, 2003).
What distinguished these effective institutions was that the funding was used
differently; for example, on faculty development (see Section 5.3.1), teaching and
learning centres and academic support staff such as tutors and counsellors (see
Section 5.7.3). These uses of funding were exactly what previous and subsequent
studies have found to be characteristic of “a campus ethos devoted to student
success” (Gansemer-Topf et al., 2004). This issue of ethos and values is returned to in
the conclusion.

14
 Dimensions of qualit y

4.2 Student:staff ratios

While at the level of the institution student:staff ratios (SSRs) may seem to be
an inevitable consequence of funding levels, institutions in practice spend funds
on buildings, on administration, on ‘central services’, on marketing, on teachers
undertaking research, and so on, to very varying extents, rather than spending it all
on teaching time. The doubling of tuition fees in the US in recent decades has not
been accompanied by any overall improvement in SSRs, but has largely been used
for administration and meeting accreditation requirements. Institutions spend very
different proportions of their available funding on teachers. So SSRs might be seen to
be a more direct indicator of educational quality than funding.
Low SSRs offer the potential to arrange educational practices that are known to
improve educational outcomes. First, close contact with teachers is a good predictor
of educational outcomes (Pascarella and Terenzini, 2005) and close contact is more
easily possible when there are not too many students for each teacher to make close
contact with. Low SSRs do not guarantee close contact, as Harvard’s recent self-
criticism has demonstrated, but they do make it possible.
Second, the volume, quality and timeliness of teachers’ feedback on students’
assignments are also good predictors of educational outcomes (see Section 5.6), and
again this requires that teachers do not have so many assignments to mark that they
cannot provide enough, high-quality feedback, promptly. Again, low SSRs do not
guarantee good feedback or feedback from experienced teachers. In the UK turn-
round times for feedback may be a matter of local policy rather than driven by SSRs
and turnaround times vary enormously between institutions (Gibbs and Dunbar-
Goddet, 2009).
Third, while low SSRs do not guarantee small classes, they certainly make them
possible, and class size predicts student performance (see Section 5.1 below).
However, once student entry characteristics are taken into account, educational
gains have been found to be largely unrelated to SSRs (Terenzini and Pascarella,
1994). This suggests either that institutions with low SSRs are not exploiting their
potential advantages through the use of effective educational practices or that SSR
figures hide other variations, or both.
SSRs reported at institutional level do not necessarily give a good indication
of the SSRs students actually experience. Patterns of work vary; for example,
academics do a greater proportion of administration, with fewer support staff, in
some institutions, effectively reducing their availability to students. They undertake
more research in some institutions while the proportion of their research time
funded by research income varies. The difference between students’ years of study
can be marked, with much greater funding per student characteristically being
allocated to third-year courses than to first-year courses, leading to better SSRs and
smaller classes in the third year (and the National Student Survey is administered
in the third year). Furthermore institutions do not allocate funding to departments

15
The higher Education Academy

in strict proportion to student numbers, but take advantage of well-recruiting and

income-generating courses, and so increase their SSRs. The overall consequence is
that effective SSRs, as students experience them, within a year of their study, within a
department, may bear little resemblance to SSRs reported at institutional level.
There are also commonly expressed concerns about the accuracy and meaning of
SSR data as reported to institutions, and as made public by HESA, due to variations
in how they are calculated, complicated by variations in the nature of employment of
teachers and how they are deployed (see Section 4.3 below).

4.3 Quality of teaching staff

Bald SSR data are unhelpful in that they disguise the realities of who the staff are
with whom students have contact. For example, undergraduates at Yale often do
not receive feedback from tenured faculty until their third year. In US research
universities the teaching undertaken by graduate teaching assistants is a constant
quality concern and is regularly cited in student exit surveys as their number one
complaint about the quality of teaching.
An hour of a graduate teaching assistant may cost a fraction of an hour of a
tenured academic, and most institutions are quick to exploit this. Recent surveys
(HEPI, 2006, 2007) reveal wide variations between institutions in the proportion of
teaching that students experience that has been undertaken by research students
as opposed to tenured academics. The majority of small group teaching was found
to be undertaken by teachers other than academics at Russell Group and pre-
1992 universities. At the University of Oxford the extent to which students take a
‘surface approach’ to their study, emphasising only memorisation (see Section 5.5.2
below), is linked to the proportion of their tutorials taken by teachers other than
College Fellows (Trigwell and Ashwin, 2004). A much lower proportion of teaching is
undertaken by research students at Oxford than at other Russell Group universities.

I am currently an undergraduate at a Russell Group University... I have three

hour-long lectures (taught by world-class professors) and three hour-long group
seminars (taught by unpaid postgraduate students) a week. That is it. If we are
going to compromise social mobility and academic freedom in order to fund our
universities better, we damn well ought to receive a standard of education that’s
worth the sacrifice.
—Letter to Education Guardian, 25 May, 2010

In teaching-oriented institutions with few research students who might do the

teaching, a significant proportion of teaching may be undertaken by what the US
terms ‘adjunct faculty’ who may have portfolio teaching careers spanning a number
of institutions, with an office in none of them. In vocational and creative arts courses

16
 Dimensions of qualit y

students may experience a significant proportion of their teaching undertaken by

professionals who may spend only a few hours a week on campus. Adjunct faculty
and professionals may bring special characteristics and talents to their work, but they
may never have the opportunity to develop a course over several years so that it
provides a rich and effective learning environment for students. They may not be paid
to meet students out of class or to provide detailed comments on their assignments.
They may never attend departmental meetings or fully understand how the degree
programme works to which their teaching is supposed to contribute. Students may
not know who the full-time tenured academics are and who are not – they will simply
experience teaching of varying quality and commitment. While the HEPI data uncover
variation in the proportion of teaching undertaken by research students – who
students can see are mainly young – it is difficult for student surveys to identify the
proportion of teaching undertaken by adjunct faculty and HEPI data do not illuminate
this issue.
In the US the proportion of part-time and non-tenure-track ‘adjunct faculty’ has
increased enormously in public institutions. Institutions with higher proportions of
adjunct faculty have lower graduation rates when other factors are held constant. An
increase of 10% in part-time faculty is associated with a reduction of 3% in graduation
rates. Within institutions, first-year students taught to a greater extent by adjunct
faculty are less likely to persist into the second year (Ehrenberg, 2006).

4.4 Quality of students

In the US, by far the best predictor of students’ educational outcomes whether the
measure is grades, a psychometric test of principled reasoning, or career success, is
their school SAT score when they enter college, with correlations in the range 0.85
to 0.95. In other words up to 90% of all variation in student performance at university
can sometimes be explained by how they performed before they entered university.
In the UK the link is less strong, but there has for decades been clear evidence of
the extensive impact of schooling on student performance in higher education, both
in terms of school leaving grades and type of school (Smith and Naylor, 2005). In the
UK students from independent schools perform less well than do students from state
schools with equivalent entry grades (Hoskins et al., 1997; Smith and Naylor, 2005).
The question that then arises is whether any of this enhanced performance is due
to qualities of the institution other than their ability to be highly selective on entry.
Large-scale longitudinal studies of a diverse range of institutions have tested students
about their academic behaviour and experience (including their engagement, see
Section 5.5.3 below) from a total of nearly 300 colleges and involving data from nearly
80,000 students (for a summary of this work see Kuh and Pascarella, 2004). These
studies have found very little relationship between educational selectivity (i.e. quality
of student intake) and the prevalence of what are known to be educationally effective

17
The higher Education Academy

practices. Selectivity was found to be negatively associated with some practices, such
as the amount of teacher feedback to students, and even where there were found to
be small positive relationships (for example with higher expectations on students),
selectivity only accounted for 2% of the variance in educational practices.
It might be argued that selective institutions do not need special educational
practices because their students are able enough to engage themselves. However, the
degree of selectivity does not predict the degree of student engagement – students
are just as engaged (or unengaged) in non-selective institutions (Pascarella et al.,
2006). So while league tables in the UK invariably include A-level point scores as an
indicator of educational quality, if the US evidence is anything to go by they tell us
almost nothing about the quality of the educational process within institutions or the
degree of student engagement with their studies.
It might be argued that there are educational benefits to a student of being
surrounded by other able students. This could raise students’ expectations of
themselves (one the of ‘Seven Principles’)1, and it is known that in group work it
is the previous educational attainment of the best student in the group that best
predicts the group grade, not the average level of prior attainment or the level of
the weakest student (Gibbs, 2010). We would then need to look at the extent to
which the educational process maximises how students could gain from each other,
for example through collaborative learning. The extent of collaborative learning is a
good predictor of educational gains (the ‘Seven Principles’ again). However, it will not
help a student much if the other students are highly able if they then engage largely
in solitary competitive learning. The US data cited above make it clear that students
are not more likely to be involved in collaborative learning, or to be engaged by it, in
institutions with more selective entry, in which the students are more able.
Students bring more to higher education than their A-level scores. It is likely that
their cultural capital, their aspirations, self-confidence and motivations all influence
their performance and interact with teaching and course design variables.

1 The ‘Seven Principles of Good Practice in Undergraduate Education’ (Chickering and Gamson
1987a, 1987b, 1991) are based on a very wide review of empirical evidence, and have been
used widely in the US and elsewhere as guides to the improvement of university teaching. The
principles are that good practice: encourages student-faculty contact, encourages cooperation
among students; encourages active learning; provides prompt feedback; emphasizes time on task;
communicates high expectations; and respects diverse talents and ways of learning.

18
 Dimensions of qualit y

5. Process dimensions

This section considers the effects on educational
effectiveness of class size, class contact hours, independent study
hours and total hours, the quality of teaching, the effects of the
research environment, the level of intellectual challenge and student
engagement, formative assessment and feedback, reputation, peer
quality ratings and quality enhancement processes

5.1 Class size

Meta-analysis of large numbers of studies of class-size effects has shown that the
more students there are in a class, the lower the level of student achievement (Glass
and Smith, 1978, 1979). Other important variables are also negatively affected by class
size, such as the quality of the educational process in class (what teachers do), the
quality of the physical learning environment, the extent to which student attitudes
are positive and the extent of them exhibiting behaviour conducive to learning (Smith
and Glass, 1979). These negative class-size effects are greatest for younger students
and smallest for students 18 or over (ibid.), but the effects are still quite substantial in
higher education. Lindsay and Paton-Saltzberg (1987) found in an English polytechnic
that “the probability of gaining an ‘A’ grade is less than half in a module enrolling
50-60 than it is in a module enrolling less than 20” (p218). All subsequent UK studies
have reported sizable negative correlations between class size (as measured by
the number of students registered on a course) and average student performance,
in most but not all subjects, and in most but not all contexts (Gibbs et al., 1996;
Fearnley, 1995). Large classes have negative effects not only on performance but also
on the quality of student engagement: students are more likely to adopt a surface
approach in a large class (Lucas et al., 1996) and so to only try to memorise rather
than attempt to understand (see Section 5.5.2 on depth of approach to learning).
At a micro-level there is evidence that the educational process is compromised as
class size increases. In higher education discussion groups, for example, a whole range
of things go wrong as class size increases. There is a much lower level of participation
by all but a minority of students and the contributions that students do make tend to
concern clarification of facts rather than exploration of ideas (Bales et al., 1951).

19
The higher Education Academy

US research shows that higher education students give lower overall ratings
to teachers of large classes (Wood et al., 1974; Feldman, 1984). However, there are
reasons to question the meaning of this finding. The same teachers are given higher
ratings when they teach smaller classes. As such ratings of teachers are relatively
reliable and stable this suggests that students’ ratings of teachers in large classes
are reflecting something other than the teachers themselves. A qualitative study of
students’ experience of large classes (Gibbs and Jenkins, 1992) has thrown light on
variables other than the teaching. There may be intense student competition for
limited library and other resources in large classes, and teachers may have to rely
on a few textbooks if students are to read anything. The amount and promptness of
feedback on assignments is likely to decline, as teacher time is squeezed. The nature
of assessments may change from engaging open-ended projects to quick tests, as
marking otherwise takes too long. Close contact with teachers outside of class and
access to remedial tutoring and advice may be more limited. Large classes may be
associated with weak social cohesion, alienation and a breakdown in social behaviour,
leading to cheating, hiding library books, and so on. All this is more to do with
what happens outside of class on courses with large enrolments, rather than what
happens in class, but it is classroom activity that is the focus of most school-based
research and US higher education research. Where out-of-class studying is the major
component of student learning the crucial variable may be course enrolment rather
than class size. US data show that cohort size is strongly negatively correlated with
student performance (Bound and Turner, 2005).
Another difference between school and higher education in relation to class-
size effects is that in higher education the range of class sizes being studied is very
much wider: perhaps 20 to 1,000 instead of 10 to 40 in schools. Different variables
inevitably become prominent in such very large classes. In school, students may
experience all their classes as much the same size. In higher education what may
matter most is not the size of the largest lecture that is attended on any particular
course but the size of the smallest seminar group or problem class that they attend
within the same course. Open University students may attend a course with an
enrolment of over 10,000, but they usually only experience a tutor group of 24,
and each tutor usually has only one tutor group so they can get to know students
individually. At the Open University it would probably make a difference if this tutor
group was 12 or 48 but not if total enrolment was 500 or 20,000.
Classrooms used for specialist purposes, such as laboratories and studios, usually
limit the number of students it is possible to teach at once, regardless of how many
students have enrolled, and although laboratories have become much larger, there
are limits to class-size effects while within the lab. However, increased enrolments
with fixed specialist spaces have the inevitable consequence of reducing the amount
of time students have access to these specialist facilities. This has transformed art and
design education. Instead of students ‘owning’ a permanent space they can take time
to become creative and competent in, when enrolment increases they visit a shared

20
 Dimensions of qualit y

space occasionally. The number of students in the studio at any one time may not have
changed much but art students’ experience has been changed out of all recognition.
Gibbs et al. (1996) found that in Art, Design and the Performing Arts, each additional 12
students enrolled on a course gave rise to a decline of 1% in average marks.
Negative class-size effects are not inevitable and a certain amount is known about
how to support good quality learning despite large classes (Gibbs and Jenkins, 1992).
The Teaching More Students initiative in the early 1990s trained 9,500 polytechnic and
college lecturers on the assumption that such improvements were possible despite
larger classes (Gibbs, 1995). The National Centre for Academic Transformation in the
US has helped scores of institutions to redesign large-enrolment, first-year courses.
They have shown that it is possible to improve student outcomes while reducing
teaching contact time and reducing funding. The Open University has retained the
number and nature of assignments per course, the amount, quality and turnaround
time of feedback from tutors, and the small size of tutor groups, through strict course
approval rules, with course enrolments that are seldom below 500.
The conundrum, of course, is that in the UK overall student performance
has increased at the same time that overall class size has increased. This issue is
addressed in Section 6.1.

5.2 Class contact hours, independent study hours and total hours

The number of class contact hours has very little to do with educational quality,
independently of what happens in those hours, what the pedagogical model is, and
what the consequences are for the quantity and quality of independent study hours.
Independent study hours, to a large extent, reflect class contact hours: if there is
less teaching then students study more and if there is more teaching students study
less, making up total hours to similar totals regardless of the ratio of teaching to
study hours (Vos, 1991). However, some pedagogic systems use class contact in ways
that are very much more effective than others at generating effective independent
study hours. A review of data from a number of studies by Gardiner (1997) found
an average of only 0.7 hours of out-of-class studying for each hour in class, in US
colleges. In contrast each hour of the University of Oxford’s tutorials generate on
average 11 hours of independent study (Trigwell and Ashwin, 2004) and Oxford’s
students have been found to put in the greatest overall weekly effort in the UK
despite having comparatively fewer class contact hours (HEPI, 2006, 2007). What
seems to matter is the nature of the class contact. ‘Close contact’ that involves
at least some interaction between teachers and students on a personal basis is
associated with greater educational gains (Pascarella, 1980) independently of the total
number of class contact hours (Pascarella and Terenzini, 2005); the provision of close
contact is one of the ‘Seven principles of good practice in undergraduate education’
(Chickering and Gamson,1987a, 1987b, 1991).

21
The higher Education Academy

Seven principles of good practice in undergraduate education

Good Practice Encourages Student-Faculty Contact

Good Practice Encourages Cooperation among Students
Good Practice Encourages Active Learning
Good Practice Gives Prompt Feedback
Good Practice Emphasizes Time on Task
Good Practice Communicates High Expectations
Good Practice Respects Diverse Talents and Ways of Learning

—Chickering and Gamson, 1987a

As an illustration of the lack of relationship between class contact hours and

outcomes, medical education worldwide has migrated from traditional didactic
pedagogies, characterised by large numbers of large class lectures, towards
problem-based pedagogies, characterised by a much smaller number of small
interactive classes, usually within the same resource envelope. This change has been
accompanied by a substantial increase in independent learning hours and evidence of
greater pedagogical effectiveness measured in a variety of ways (Dochy et al., 2003).
Similarly the Open University has managed to top the National Student Survey league
tables while having very much the lowest class contact hours in the UK.
This is not the same as arguing that you can cut class contact hours from an
existing unchanged pedagogy without making any difference to student learning, or
that increasing hours will make no difference. If students read primarily in order to
discuss what they have read in a seminar, and the seminar is taken away, then they
will probably read a good deal less and learn less as a consequence. Very little class
contact may result in a lack of clarity about what students should be studying, a lack
of a conceptual framework within which subsequent study can be framed, a lack of
engagement with the subject, a lack of oral feedback on their understanding, and
so on. It depends what role the class contact is performing. What matters is the
quantity and quality of engagement generated by the particular uses to which class
contact is put.
What seems to be more important than class contact hours is the total number
of hours that students put in, both in and out of class (the quality of that effort is
considered in Section 5.5 below). The number of hours that students study out of class
is usually measured in one of two ways: by asking students to estimate, retrospectively,
how much they think they have been studying, often ‘on average’, or by asking students
to keep a log, as they study, of what they do with their time, over a relatively short
period of time (such as a week). While logs are likely to be more accurate than
retrospective estimates, they also run the risk of changing student study behaviour
as students become more aware of their pattern and level of effort. Indeed keeping

22
 Dimensions of qualit y

a log is a common learning activity on study skills courses. When asking students
to estimate their study hours retrospectively, the form of the question used varies
between different surveys and the timing of the surveys varies in relation to how long
ago students are attempting to remember or how wide a spread of courses they are
being asked to make average estimates across. Students who attend less and study less
may be missed by surveys while conscientious students who attend more and study
more may be more likely to return surveys. The impact of such potential biases is not
well researched and the reliability of study-hours data is not known.
The question: ‘Are higher study hours associated with better student learning
and performance?’, can be posed in two rather different ways. First: ‘Are the students
who study longer hours the ones that perform best?’ The answer to this question
is not straightforward (Stinebrickner and Stinebrickner, 2008), because very able
students may be able to meet assessment requirements without having to study
very hard, while less able students may put in many hours unproductively (Ashby et
al., 2005). There is also evidence that students who, inappropriately, take a ‘surface’
approach to their studies (see Section 5.5.2 below) find this so unproductive that they
gradually reduce their effort after initially working hard and end up studying fewer
hours than students who take a ‘deep’ approach (Svensson, 1977).
If, however, the question is framed differently as: ‘If a student were to study
more hours, would they perform better?’ or even ‘If average study hours on a degree
programme were higher, would average performance be higher?’, the answer is much
more clearly ‘Yes’. ‘Time on task’ is one of the evidence-based ‘Seven Principles of
Good Practice in Undergraduate Education (Chickering and Gamson, 1987). The
reasonable assumption here is that if you don’t spend enough time on something then
you won’t learn it, and that increasing the number of hours students spend studying
is one of the most effective ways of improving their performance. North American
research and development work on ‘student engagement’ (see Section 5.5.3 below)
uses student effort as an important indicator of engagement.
The Bologna process has used total student effort (class contact hours plus
independent study hours) as its metric for defining the demands of a Bachelors
degree programme, set at 1,500 to 1,800 hours a year: 4,500 to 5,200 hours over
three years. A series of studies have found that UK students’ total weekly effort in
hours is lower than in comparison with either the particular European countries
studied or in comparison with overall European norms (Brennan et al., 2009;
Hochschul-Informations-System, 2005; Sastry and Bekhradnia, 2007; Schomburg
and Teichler, 2006). These findings deserve to be taken seriously because they are
relatively consistent across different studies and methodologies, carried out in
different countries, or across repetitions of the same study in different years.
It should be possible to iron out gross differences between institutions and
subject areas, as the number of study hours per credit, and hence the number
of hours required for a Bachelors programme, are clearly defined in course
documentation. However, the range in weekly study effort between English

23
The higher Education Academy

institutions, within subjects, found in the HEPI studies is wide, for example from 14
hours a week to nearly 40 hours per week within Philosophy (Sastry and Bekhradnia,
2007). Differences between subjects are also wide. Broad differences in total study
hours between science and technology programmes (which tend to have both high
class contact hours and weekly demands for work such as problem sheets and
laboratory reports) and the humanities (which tend to have both lower class contact
hours and less regular assignments such as essays) are well known and have been
reported frequently over the years (e.g. Vos,1991). However, the differences between
subjects identified by the HEPI surveys are substantial, with some subjects having
national average weekly study efforts of only around 20 hours per week. Twenty
hours per week within the comparatively short UK semesters equates to around
500 hours a year: one third of the minimum specified under the Bologna Agreement.
To achieve the Bologna specification of a minimum of 4,500 hours for a Bachelors
programme, students in these subjects in the UK would have to study for nine years.
Differences on this scale cannot easily be argued away by claiming that UK students
are somehow inherently superior or that UK educational practices are somehow
inherently more efficient, in the absence of any evidence to back up such claims.
A survey of international students who have experienced both a UK higher
education institution and another EU higher education institution (Brennan et al.,
2009) found that such students are more likely to rate UK Bachelors programmes as
‘less demanding’ and less likely to rate them as ‘more demanding’, a finding that does
not justify the lower number of hours involved. UK students have been reported
to have done more work than was required of them to a greater extent than in any
other European country (ibid.). Yet the total number of hours studied in the UK is still
below European norms, which suggests that the UK requirements must be lower.
If it were the case that less able students needed to study more, then one
would find the larger study hours figures in institutions that have students with
weaker educational backgrounds. Instead the reverse is the case, with higher weekly
study hours reported in institutions with students with the strongest educational
backgrounds (Sastry and Bekhradnia, 2007). The most likely explanation therefore is
that the demands made on students are different in different institutions, and that even
weaker students are able to meet these demands while studying, in some institutions,
and in some subjects, a third of the hours the Bologna Agreement specifies.
There are a number of possible explanations of why such students might study
few hours:

—— Some degree programmes rely heavily on coursework assessment, with

over 90% of marks on some degree programmes coming from coursework
rather than on examinations. While this may have come about for sound
educational reasons, it may allow students to study only one or two topics
in depth per course, because that is all there are the teaching resources
to mark, reducing student effort in weeks when there are no formal

24
 Dimensions of qualit y

assessment demands. Students have been found to work more regularly and
cover the syllabus to a greater extent when there is a higher proportion of
marks from examinations (Gibbs and Lucas, 1997).

—— High levels of detail in course specifications, of learning outcomes and

assessment criteria, in response in part to QAA codes of practice, allow
students to identify what they ought to pay attention to, but also what they
can safely ignore. A recent study has found that in such courses students may
narrow their focus to attention to the specified assessed components at the
expense of everything else (Gibbs and Dunbar-Goddet, 2007). Students have
become highly strategic in their use of time and a diary study has found students
to progressively abandon studying anything that is not assessed as they work
their way through three years of their degree (Innis and Shaw, 1997).

—— Studies of the effects of paid work undertaken by full-time students have

reported a substantial reduction in study hours in relation to the extent of
their paid work, although effects vary somewhat between contexts and degree
programmes (Carney et al., 2005; Curtis and Williams, 2002; Ford et al., 1995;
Hunt et al., 2004; Paton-Salzberg and Lindsay, 1993). A typical study is that of
Svanum and Bigatti (2006), who found that paid work reduced course effort
and reduced grades. In the US, students respond to the need to ‘work their
way through college’ by taking fewer credits at a time and gaining their degree
over more years (and across several institutions) as a consequence. In the UK
full-time students seem to assume that they can undertake paid work while
meeting the demands of a full-time course and still graduate in three years.

—— Students who live at home, rather than on a residential campus, are likely to
experience competing demands on their time, and less social and academic
integration (Tinto, 1975). The institutions in the UK with the lowest average
study hours include universities in urban conurbations with a substantial
proportion of students living at home.

—— The universities with low average study hours are often also institutions with
low annual investment per student in libraries and other learning resources. This
would make it more difficult for students to gain access to the resources they
need for their study: the book will be out and the study space with a computer
will be occupied. Data from HESA, HEPI and the National Student Survey have
been analysed for the purpose of the current report on this issue. They showed
that institutional funds allocated to learning resources, per student, predict total
student learning hours (with correlations of +0.45 for the social sciences and
humanities subjects analysed). Funding for learning resources also predicts average
students’ responses to the NSS question on the quality of learning resources,

25
The higher Education Academy

although less well. The institution with the highest weekly average study hours also
has the greatest annual investment in learning resources and the highest National
Student Survey ratings for ‘learning resources’: the University of Oxford.

While the focus of this report is on undergraduate programmes, there has recently
been a good deal of attention paid to the relative quality of UK Masters level courses,
given that they are usually planned to be considerably shorter in duration than their
mainland European counterparts. For example, the Chair of the UK Council for Graduate
Education has argued that these gross differences do not matter because UK Masters
courses are ‘more intensive’, and claimed that the overall amount of learning time is
roughly equal between the UK and mainland Europe (Olcott, 2010). This unsubstantiated
claim could be checked by repeating the Sastry and Bekhradnia undergraduate study of
study hours, cited above, in Masters courses and adding questions to measure the extent
to which students take a deep approach to their studies (see Section 5.5.2 below).

5.3 Quality of teaching

5.3.1. Quality of teaching: experience and training

Teachers who have teaching qualifications (normally a Postgraduate Certificate in

Higher Education, or something similar) have been found to be rated more highly by
their students than teachers who have no such qualification (Nasr et al., 1996). This
finding was in a context where obtaining such a qualification was largely voluntary,
and those who have the qualification might be considered to be different in some way
from those who have not, and this could be argued to invalidate the comparison. The
difference might concern the extent of professionalism or commitment to teaching,
but nevertheless there was no control group in the study. A longitudinal study that
overcomes this objection has examined the impact over time on students’ ratings
of their teachers, and on teachers’ thinking about teaching, of (mainly) compulsory
initial training during their first year of university teaching, in eight countries. It found
improvements on every scale of the ‘Student Evaluation of Educational Quality’, a
questionnaire developed in the US (Marsh, 1982) and tested for use in the UK (Coffey
and Gibbs, 2000), and improvements in the sophistication of teachers’ thinking (as
measured by the ‘Approaches to Teaching Inventory’, a measure of teaching that
predicts the quality of student learning, Trigwell et al., 2004). This improvement
in measures of teaching quality could not be attributed to mere maturation or
experience as teachers in a control group in institutions without any initial training
were found to get worse over their first year, on the same measures (Gibbs and
Coffey, 2004). The quality of school teaching in the UK is in part underpinned by
a belief that initial or in-service training is essential and the, admittedly limited,
evidence suggests that the same belief would be justified in UK higher education.

26
 Dimensions of qualit y

5.3.2. Quality of teaching: research record

There is little or no relationship between measures of the quality or quantity of

teachers’ research and measures of the quality of their teaching (for a review of 58
studies of the evidence, see Hattie and Marsh, 1996).

…the common belief that teaching and research were inextricably intertwined is
an enduring myth. At best teaching and research are very loosely coupled.
—Hattie and Marsh, 1996, p529

Some excellent researchers make excellent teachers and some do not. Despite
critiques of the measures of research and teaching that are normally used, none of
the critics have managed to develop or use alternative measures that demonstrate a
relationship between research and teaching. A minority of undergraduate students
have been reported to value their teachers being active researchers provided this
does not interfere with their studies (for example, through their teacher being absent
while undertaking research) (Lindsay et al., 2002), but there is no evidence that this
improves their learning.

5.3.3. Quality of teaching: judged by students

Despite the common disparagement of student ratings of teachers, they can be

highly reliable (in the sense that students agree with each other about who the
best teachers are, agree with teachers’ peers, and make the same judgements on
different occasions) and relatively immune from biases of various kinds. Students
can readily tell the difference between teachers they like and teachers who they
think are good teachers, and the common criticism that student feedback is simply
a popularity parade is largely unfounded (see Marsh, 1987, for a review of the vast
literature on the reliability and validity of student feedback questionnaires). These
observations are based on the use of thoroughly developed and tested feedback
questionnaires. However, in the UK almost all such questionnaires are ‘home-
grown’ and are likely to be of doubtful reliability and open to all kinds of biases.
Different questionnaires are used in different institutions and there is therefore no
basis for comparison of the quality of teachers between institutions or subjects.
If student ratings of teachers were considered a useful indicator of comparative
quality then there would need to be an agreement for everyone to use the same
questionnaire, such as the Student Evaluation of Educational Quality (SEEQ)
(Marsh, 1982; Coffey and Gibbs, 2000).
Several well-developed, reliable, student feedback questionnaires have reasonable
levels of validity in that scores on scales within the questionnaires (involving adding
ratings from clusters of similar questions) predict aspects of student learning process

27
The higher Education Academy

(such as the extent of studying following teaching), learning outcomes (such as grades)
and other worthwhile consequences (such as the likelihood of students choosing to
study further courses with the same teacher). The proportion of variance in such
measures of the products of good teaching, that is explained by student ratings, varies
across different questionnaire scales and different measures of products, but it is
usually high enough to take measures of teaching based on student ratings seriously
(Abrami et al., 1990).
There is an important distinction to be made here between student ratings of
the extent to which teachers engage in activities that are known to improve learning
(such as providing enough prompt feedback on assignments), which tend to be reliable
and valid, and global judgements of whether teaching is ‘good’, which are open to all
kinds of subjective variation in the interpretation of what ‘good’ means. Students also
change over time in their sophistication as learners, for example in their conception
of learning (Säljö, 1979) and in their conception of knowledge (Perry, 1970). As
they change, so their conceptions of what ‘good teaching’ consists of evolve (Van
Rossum et al., 1985). What an unsophisticated student might consider to be good
might consist of the teacher providing all the content in lectures and then testing for
memory of that content, while a more sophisticated student might see good teaching
as involving supporting independent learning and the development of a personal
stance towards knowledge. What unsophisticated students want their teachers to do
is often bad for their learning and responding to their global ratings uncritically is not
the way to improve quality. When a survey reports a single global rating of the extent
to which student think all the teaching over three years is simply ‘good’, these very
different student conceptions of good teaching are muddled together and the average
rating is then very difficult to interpret. In contrast so-called ‘low inference’ questions
that refer to specific teacher behaviours, such as the promptness of their feedback,
are much easier to interpret.

5.4 Research environment

As we have seen above (in Section 5.3.2) there is no relationship between measures
of an individual academic’s research and measures of their teaching. However, it could
be argued that it is not individual researchers’ teaching that matters here, but the
research environment generated by the majority of teachers in a department being
research active. This might be considered a presage variable, but as we shall see, what
matters is the educational process, not prior research performance.
At the level of departments within an institution the situation is the same as
it is at the level of individual teachers. The best research departments may or may
not be the best teaching departments: there is no correlation between measures
of a department’s research and measures of its teaching (Ramsden and Moses,
1992). There are suggestions that there may be relationships between the extent of

28
 Dimensions of qualit y

research intensiveness of a department and students’ sense of benefit, in a way that

is not mediated by the teaching, but this is in any case only for some students (e.g.
Dunbar-Goddet and Trigwell, 2006), and this relationship appears to be a function
of the focus of attention of the students rather than a consequence of different or
better teaching: if students are interested in scholarship they will be more likely to
say that they benefit from a scholarly environment.
At the level of whole institutions, those where there is a strong orientation
towards research often reveal a weak emphasis on teaching, and vice versa – there
is a strong negative relationship in relation to institutional priorities and this has
measurable effects on educational gains:

… a college whose faculty is research-orientated increases student dissatisfaction

and impacts negatively on most measures of cognitive and affective development.
—Astin, 1993, p363

There is evidence that the research environment can impact positively on

undergraduate students, in a way that can be measured, where maximising the
benefits to undergraduates of research strengths has been made a deliberate
priority. For example, MIT enables about 80% of its undergraduates to engage in a
real research project as an intern or junior research student placed with a research
group, and they have good evidence concerning how students benefit (Bergren et al.,
2007). The National Science Foundation in the US has similar evidence across many
institutions that such initiatives:

… have a major impact on most participants’ confidence and their understanding

of research-related issues, increase their interest in careers in research and science
and engineering, and lead them to raise their degree expectations.
—Ramaley, 2004, cited in Jenkins, 2004

(See also Bauer and Bennett, 2003; Hathaway et al., 2002.) The key point here
is that such benefits have to be deliberately engineered – they do not accrue by
magic simply because research is going on as well as teaching. The institutional
indicator of quality in these studies is the existence of an undergraduate research
opportunities scheme, not the strength of the institution’s research. Similarly
the positive relationship found at the University of Oxford between students’
experience of research-active staff and the extent to which they take a deep
approach to learning (Trigwell, 2005) is a consequence of the collegial system
fostering active inclusion in a community of (research) practice, not simply of the
existence of a research enterprise.
For these reasons departmental RAE scores or other measures of research
activity or performance in the environment students study within are not, on their
own, valid indictors of educational quality.

29
The higher Education Academy

5.5 Level of intellectual challenge

It can be argued that an institution, or even an entire higher education system, is

of higher quality because it offers a greater level of intellectual challenge. In this
section the notion of intellectual challenge is operationalised in three ways: challenge
as determined by the level of the curriculum students are studying, challenge as
indicated by the depth of approach students take to their studies, and challenge as
reflected in the level of student engagement with their studies.

5.5.1. Level of the curriculum

In some subject areas with relatively hierarchical knowledge, formal course

documentation can be very informative about the level of the curriculum. It is
then possible to compare the level of intellectual challenge by examining course
descriptions to see how far through the subject students are expected to progress.
It is in principle possible to study curriculum documentation within well-defined
disciplines, from different periods, in order to ascertain whether there has been a
watering down of intellectual demands over time. There are several limitations to this
kind of study:

—— documentation changes over time in what is described and in what detail,

and also differs between institutions;
—— fashions change over time within subjects about what should be included,
which makes changes of level of demand difficult to judge;
—— even in supposedly hierarchical subjects it is sometimes difficult to agree what
is necessary to study first, or what counts as more advanced. In Medicine,
for example, problem-based learning may nowadays confront first-year
students with problems of a complexity that in past curricula, that had clear
divisions between pre-clinical and clinical stages, only fourth-year students
encountered. However, the first-year students now tackle these complex
problems at a lower level, with a lesser knowledge base. Calibrating the level
of the problem and the level at which they are tackled is not straightforward.

Nevertheless there is probably more scope for determining quality by examining

curriculum documentation than is currently exploited, provided it is undertaken by
those within disciplinary communities that understand the complexities involved.
There are several typologies for educational objectives or learning outcomes
(for example that of Bloom et al., 1956) that could be used to compare the level
of difficulty of what students are supposed to do with the content (for example,
remember it, apply it or critique it). It is possible to distinguish levels of intellectual
demand between different levels of BTEC courses in the same subject by comparing

30
 Dimensions of qualit y

the proportion of educational objectives specified at each of a number of levels (such

as ‘knowledge’ and ‘synthesis’). Higher levels of courses are supposed to have a higher
proportion of their curricula objectives at a higher level in the hierarchy. Proposals
for new courses can be turned down if too many objectives are set at too low a level.
Where a curriculum authority has the power to impose a common set of
definitions and a common format of documentation, as BTEC does, this is just
about possible to implement. Where each institution and each degree programme
has the freedom to specify learning outcomes independently of any common
conceptual framework or set of definitions, such comparisons become impossible,
except in an impressionistic way. For learning outcomes that concern generic skills,
such as ‘communication’ and ‘group work’, consistent judgement of levels against
standards is even more difficult, independently of exemplars that go beyond curricula
documentation. Alverno College in the US has demonstrated that even for such
generic outcomes or capabilities, it is possible to define standards in unambiguous
ways and to train teachers to use these standards in consistent ways, but unless
Alverno’s approach, or something like it, were to be adopted by all, comparison
across contexts would still not be possible.
Nevertheless there is plenty of scope for subject associations to make more use
of curriculum documentation in judging the intellectual demands of programmes,
at least in principle. To judge them in practice it is also necessary to examine the
products of student learning (see Section 7.7).

5.5.2. Depth of approach to studying

In the 1970s Ferenc Marton and his colleagues in Goteborg distinguished between a
‘surface approach’ to learning in which students intend to reproduce material, and
a ‘deep approach’ in which students intend to make sense of material: a distinction
between a focus of attention on the sign or what is signified. To illustrate the
consequences for student learning outcomes, a student who takes a surface approach
to reading an article with a principle-example structure (such as a case study) may
remember the example, while the student who takes a deep approach is more likely
to understand the principle (Marton and Wenestam, 1978). A surface approach has
been demonstrated in a wide variety of studies to have depressingly limited and
short-lasting consequences even for memory of facts. A deep approach is essential
for long-term and meaningful outcomes from higher education (see Gibbs et al. (1982)
and Marton et al. (1984) for overviews of this literature).
Students are not ‘surface students’ or ‘deep students’ – approach to learning
is in the main a context-dependent response by the student to perceived demands
of the learning context (Ramsden, 1979). The relevance to dimensions of quality is
that it is possible to identify those features of courses that foster a surface or a deep
approach. Students tend to adopt a surface approach to a greater extent when there

31
The higher Education Academy

is, for example, an assessment system that rewards memorisation, such as superficial
multiple-choice-question tests. In contrast students tend to adopt a deep approach,
for example, when they experience good feedback on assignments, and when they
have a clear sense of the goals of the course and the standards that are intended to
be achieved. These influential characteristics of courses are the focus of the Course
Experience Questionnaire (CEQ) (Ramsden, 1999), originally developed in studies
at Lancaster University in the 1970s, through which students indicate the extent to
which these course features are experienced. Reasonably close relationships have
been found between scores on scales of the CEQ and the extent to which students
take a deep and surface approach to their studies, and so CEQ scale scores that
focus on certain course features can act as a rough proxy for educational outcomes,
because approach predicts outcomes to some extent. The CEQ became the basis of
the questionnaire used annually throughout Australian higher education to measure
comparative quality of degree programmes, published in annual reports aimed at
students. It has been used for some years within some institutions as a performance
indicator for allocating a proportion of funding for teaching to departments, as at
the University of Sydney. It has now been adopted nationally in Australia as one
component of performance indicators for allocating over A$100 million of teaching
funding (in 2008) to universities each year. It has become the driving force behind
evidence-based institutional efforts to improve teaching that focus on course design
rather than on individual teacher’s skills (Barrie and Ginns, 2007). A modified version
of the CEQ (the OSCEQ) has been used annually at the University of Oxford.
It is often assumed that the validity of the National Student Survey (NSS) is
based on the same research and evidence. Up to a point this is true. However, the
characteristic of students’ intellectual engagement with their studying that best predicts
their learning outcomes, the extent to which they take a deep approach, is not included
as a scale in the NSS (and nor is it in the CEQ). Some characteristics of what have been
found to be effective courses, such as concerning feedback, are included in the NSS.
However, most of the scales of the original version of the CEQ that relate somewhat
to the extent to which students take a deep approach, such as ‘Clear Goals and
Standards’ or ‘Appropriate Workload’ are not included in the NSS (and neither are they
in the most recent versions of the CEQ). In fact both questionnaires lack most of the
scales that would strengthen their validity. The missing scales are currently included as
options in both questionnaires, but this means that comparable data are not published
or available for comparison between institutions or courses.
Even some of these missing scales have a somewhat tenuous claim to validity
today. For example in the 1970s it was found that if students were grossly
overburdened then they might abandon a deep approach and adopt a surface
approach to their studies. However, 30 years later excessive workload seems a
distant memory (see Section 5.2), so the ‘Appropriate Workload’ scale no longer
seems likely to predict, to a worthwhile extent, which students will adopt a surface
approach, and hence their learning outcomes.

32
 Dimensions of qualit y

There have been no recent studies to confirm the original findings concerning
relationships between features of courses, student responses and learning outcomes
in current contexts. There have been no direct studies of the validity of the NSS in
relation to its ability to predict educational gains. There have been no studies that
demonstrate that if evidence-based practices are adopted, and NSS scores improve,
this will be associated with improved educational gains. For that kind of evidence we
have to look to measures of student engagement.

5.5.3. Student engagement

There is currently a high level of interest in the US in measuring student engagement

as a crucial indicator of educational quality, embodied in the use of a questionnaire:
the National Survey of Student Engagement (NSSE). The research underlying the
NSSE, and its wide-scale use and perceived value, has involved three main stages.
First, very large-scale studies examined a huge range of process variables on
campus and also assessed a huge range of measures of educational gain, across a very
wide range of courses and institutions, to identify which process variables relate to
any of the measures of gain. No research encompassing so many variables at once,
or on a similar scale, has been conducted outside of the US. These vast studies have
been repeated over three decades and very similar findings have identified the same
few process variables that are most closely linked to educational gains (Pascarella
and Terenzini, 2005). Essentially the crucial variable is ‘student engagement’ and it
has proved possible to identify the process variables involved in engaging students,
such as the level of academic challenge, the extent of active and collaborative learning
and the extent and quality of student-faculty interaction. These process variables
are prominent in the influential evidence-based ‘Seven principles of good practice in
undergraduate education’ elaborated above (Chickering and Gamson, 1987, 1991).
These principles have been used by many institutions in the US as indicators
of quality in reviewing and improving their educational practices, and there has
been much pooling of institutional experience in using these principles. Pascarella
et al. (2008) list a plethora of studies that demonstrate that if you act on the ‘Seven
principles of good practice’ then this improves student outcomes.
Second, a questionnaire has been developed through which students can
indicate the extent to which these crucial process variables have been experienced
and the ways in which they are engaged in their studying: the National Survey of
Student Engagement (NSSE, 2007). This questionnaire has been used widely – by
774 universities and colleges in 2008 alone. It has also been widely used to monitor
student engagement as changes have been made to educational practices in attempts
to improve student learning. Sufficient studies of appropriate rigour have been
published to make it possible to bring the data together and make generalisable
conclusions. Meta-analysis of studies of innovation and change support the original

33
The higher Education Academy

identification by the earlier research of a predictable range of educational processes

that, when adopted or enhanced, make a measurable difference to students
engagement (Carinin et al., 2006).
The third stage of this research has involved validity studies – examining if there
is a relationship between engagement and educational outcomes: between scores
on the NSSE and both student performance and psychometric measures such as of
effective reasoning and problem solving, ‘inclination to inquire’ and lifelong learning.
A number of studies have shown clear links between NSSE scores and outcomes
such as first-year success (e.g. LaNasa et al., 2007). Pascarella et al. (2010) have
gone further and in a much larger study, involving 19 colleges of a wide variety of
types, demonstrated strong relationships between NSSE scale scores and a range of
educational gains, involving before and after measurements. The authors concluded:

NSSE results regarding educational practices and student experiences are good
proxy measures for growth in important educational outcomes.

In other words if you want to know the ‘value added’ by students’ higher
education experience then the NSSE will provide a good indication without needing
to use before and after measures of what has been learnt.
It is interesting to note, with reference to the self-imposed limitations of the
NSS and CEQ, that the scale on the NSSE that has the closest relationship with
educational gains concerns ‘deep learning’ (Pascarella et al., 2008).

5.6 Formative assessment and feedback

The educational intervention in schools that has more impact on student learning
than any other involves improving formative assessment and especially the provision
of more, better and faster feedback on student work (Black and Wiliam, 1998;
Hattie and Timperley, 2007). ‘Good practice provides prompt feedback’ is one of
the evidence-based ‘Seven principles of good practice in undergraduate education’
(see above). On degree programmes where the volume of formative assessment is
greater, students take a deep approach to their studies to a greater extent (Gibbs and
Dunbar-Goddet, 2007) and deep approach is a good predictor of learning outcomes
(see Section 5.5.2 above). Enhanced feedback can also improve student retention
(Yorke, 2001).
The number of occasions during a three-year Bachelors programme in the UK
on which students are required to undertake an assignment purely for the purpose
of learning, with feedback but without marks, varies widely between institutions.
One study has found a range from twice in three years at one English university
to over 130 times at another (Gibbs and Dunbar-Goddet, 2009). In another UK
study using the same assessment audit methodology (TESTA, 2010), the volume of

34
 Dimensions of qualit y

written feedback on assignments over three years varied from below 3,000 words
per student to above 15,000 words, and for oral feedback varied from 12 minutes
per year per student to over ten hours per year (Jessop et al., 2010). These are much
wider variations between institutions than exist in their funding per student, their
SSRs, their class contact hours or their independent study hours. The issue addressed
by the NSS that reveals the greatest area of student disquiet is feedback.
As resources per student have declined there have been economies of scale in
teaching that are difficult to achieve in assessment: assessment costs go up pretty much
in proportion to the number of students. This places enormous time pressures on
teachers. Quality assurance systems in most institutions have not prevented the volume
of formative assessment from declining substantially, despite the QAA Code of practice.
An exception is The Open University where the number of assignments per module,
and the volume and quality of tutor feedback on all assignments have been maintained
over 30 years. This has been achieved across all courses by formal requirements of
their course approval process and by several quality assurance processes. The Open
University has exceptionally high NSS scores for assessment and feedback.

5.7 Other process dimensions of quality

5.7.1. Reputation

Seeking the views of research peers is a common method used to judge a department
or university’s research quality and the same methodology could in principle be
used to judge educational quality. The highly influential university ranking system
in the US provided by the US News and World Report, ‘America’s Best Colleges’,
invests heavily in surveys of Deans and Presidents in establishing college reputations.
However, the reputational ranking that derives from these surveys correlates closely
with the size of institution’s federal research grants (Graham and Thompson, 2001)
and can also be predicted by undergraduate selectivity, per student expenditure
and number of doctoral awarding departments (Astin, 1985), none of which predict
educational gains. Reputational data have a very poor reputation as a valid indicator
of educational quality.

5.7.2. P eer ratings (where these focus on process dimensions)

Many quality assurance systems make use of expert peer judgement of the quality
of educational provision in a degree programme, at the time of a periodic review of
some kind, based on a wide range of evidence and documentation and sometimes
including observation of teaching. The relationship between these ratings and the
evidence on which they are based is not easy to establish as they are inherently

35
The higher Education Academy

subjective and global, and based on different combinations of evidence, with different
weightings, in different contexts, by different groups of peers. However, there may be
potential for the application of professional expertise in such subjective judgements
to reach more valid conclusions than could be achieved merely on the basis of
individual quantitative measures. This is what Teaching Quality Assessment (TQA)
ratings attempted to provide in a quantitative way. In league tables in England the
six four-point rating scales involved in TQA have usually been combined into a single
score out of 24, and institutional averages out of 24 have been used as indicators
of educational quality. Subsequent analysis of average TQA scores for institutions
has revealed that they are very largely predictable on the basis of student entry
standards (A-level points scores) and research performance (RAE scores), together
or separately, without reference to any measures of educational process (Drennan
and Beck, 2001; Yorke, 1997, 1998). In other words, TQA scores largely reflect
reputational factors. This would not be a terminal problem if research performance
and quality of students were valid indicators of educational quality but, as we have
seen above, they are not. The inability of reputational factors to provide a valid
indicator of educational quality is highlighted above in Section 5.7.1. The inability of
peer judgements to be immune from reputational factors undermines their credibility.
TQA scores were also subject to other confounding variables, such as institutional
size, which have not been taken into account either in moderating overall scores, or
in league tables based on TQA scores (Cook et al., 2006).

5.7.3. Student support

The QAA have highlighted, in their reviews of what has been learnt from institutional
audits (QAA, 2003), the important role played by adequate student support services
of various kinds: study skills development, counselling, English language support,
support for students with special needs, and so on. There a number of reasons
why it is difficult to estimate the extent to which student services play a role in
educational effectiveness or gain. Support services are configured in many different
ways, for example subsumed within academic roles or centralised in generic service
units. They are described using different terminology: for example, there are few US
equivalents of the UK’s traditional personal tutor role, and few UK equivalents of the
role of ‘student advising’ in the US. This makes collating evidence across contexts, or
comparing like with like, somewhat challenging. Data concerning the positive impact
of student support from large US studies are difficult to relate to the nature of UK
provision. The impact of such services also relates closely to the nature of student
intake. Slender provision at one institution might be perfectly adequate because it
only has a zephyr of demand to deal with, while at another institution even extensive
and professionally run support services may face a gale of demand and expectations
and so may fall short despite extensive institutional commitment. There is clear

36
 Dimensions of qualit y

evidence of the role of various kinds of student support, for example concerning the
impact on student performance of the development of students’ study skills (Hattie et
al., 1996). However, what support services are appropriate, and how they might best
be delivered, can be highly context- and discipline-specific. For this reason no general
empirical conclusions will be drawn here.

5.7.4. Quality enhancement processes

Much of the past focus of attention of the Council for National Academic Awards,
and today the Quality Assurance Agency, has been on quality processes, such as
the operation of the external examiner system and the use of student evaluation
of teaching, that are intended to assure quality. The assumption is that if such
processes are securely in place, then an adequate level of quality can be more or
less guaranteed. There is some evidence to support this kind of assumption. As was
discussed in Section 4.1 above, in institutions where student engagement is found to
be high and educational gains are high, one finds a higher than average investment of
resources in quality enhancement processes such as faculty development and teaching
and learning centres (Gansemer-Topf et al., 2004). There is also evidence that some
of the prescribed quality enhancement processes have a positive measurable impact,
but only under certain circumstances. For example, collecting student feedback on
teaching has little or no impact on improving teaching (Weimer and Lenze, 1997)
unless it is accompanied by other process such as the teacher consulting with an
educational expert, especially when preceded by the expert observing teaching and
meeting students (Piccinin et al., 1999).
The extent of institutional adoption of quality enhancement processes through
teaching and learning strategies has been documented for English institutions,
(HEFCE, 2001; Gibbs et al., 2000), but there is currently no evidence that the extent
of adoption of these processes relates to any other measures of process or product.

37
The higher Education Academy

6. Product dimensions of educational quality

6.1 Student performance and degree classifications

In the UK the measure most commonly used to indicate the quality of the outcome
of higher education is the proportion of students gaining upper second class or first
class degrees. The proportion of students who gain ‘good degrees’ has increased
very markedly over time, although unevenly across institutions and subjects (Yorke,
2009). At the same time presage and process indicators of quality (such as funding
per student, the quality of student intake, class size, SSRs, amount of close contact
with teachers and amount of feedback on assignments) have declined. Yorke (2009)
suggests a whole list of reasons why this counter-intuitive phenomenon has occurred.
For example, the proportion of assessment marks derived from coursework has
increased and coursework usually produces higher marks than examinations (Gibbs
and Lucas, 1997). Most of the possible explanations currently lack data through which
they could be tested.
The key problem appears to be that there has been little to stop grade inflation.
The external examiner system has not proved capable of maintaining the standards
that are applied by markers to whatever quality of student work is being assessed.
As a consequence degree classifications cannot be trusted as indicators of the
quality of outcomes. A whole raft of unjustifiable variations exists in the way student
degree classifications are generated. For example, a Maths students is more than
three times as likely to gain a first class degree than a History student (Yorke et
al., 2002; Bridges et al., 2002) and there are idiosyncratic institutional algorithms
for adding marks from different courses (Yorke et al., 2008) that can make as much
as a degree classification difference to individual students (Armstrong et al., 1998).
The best predictor of the pattern of degree classifications of an institution is that
they have produced the same pattern in the past (Johnes, 1992), and institutions’
historical patterns are not easily explicable.
It has been argued that there is no longer any meaningful sense in which degree
standards are comparable (Brown, 2010). There has been persistent criticism of the
meaning and interpretability of degree classifications as indicators of educational
outcomes (e.g. House of Commons, 2009) and these arguments have been largely
accepted, e.g. by the QAA (2006), and so the arguments will not be rehearsed here.
What is clear is that degree classifications do not currently provide a sound basis for
indicating the quality of educational outcomes of a UK institution.

38
 Dimensions of qualit y

6.2 Student retention and persistence

The Open University and the University of Oxford have comparable NSS student
ratings for the perceived quality of their educational provision, but are at opposite
ends of rankings in terms of student retention, with about 98% of entering
undergraduates completing in three years at Oxford, almost double the proportion
of new students completing a ten-month course at The Open University. Student
retention (in relation to persisting from one year to the next and completion rates
within normal time frames) vary very considerably from one institution to another
even when educational provision is judged to be similarly excellent or similarly poor.
Institutional comparisons are made difficult by the varied nature of student cohorts.
Broadly, national retention rates vary in inverse relation to age participation rates
(OECD, 2000): the broader the range of student ability is entering higher education,
the lower is the overall retention rate. In addition, different institutions take their
students from different subsets of the overall ability range.
Students vary not just in terms of their record of past educational success, but
in other variables known to affect retention such as whether they live on campus
(Chickering, 1974) and whether they are undertaking paid work to support their
studies (Paton-Saltzberg and Lindsay, 1993).
In the US it is no longer the case that the majority of students gain the credits
they need for a qualification from a single institution. So ‘drop-out’ is not only the
norm but is, for many students, expected and even planned for as they accumulate
credits wherever and whenever is convenient. This is not yet the norm in the UK,
but ‘drop-out’ does not have the same meaning or significance for an increasing
proportion of students as it does for policy makers (Woodley, 2004). It is not simply
that part-time students complete at different rates than do full-time students, but
that ‘retention’ has a different significance for them.
A variable known to influence retention is whether students are socially
and academically well integrated (Tinto, 1975). Social and academic integration is
affected by living off campus, living at home, and taking time out to earn enough
to continue studying. The prevalence of these variables is very varied across
institutions, and it is difficult to take all such variables fully into account in judging
institutional retention performance.
Student variables also affect retention within institutions and are so influential that
in the US commercial companies (such as the Noel-Levitz organisation) offer services
to institutions to collect management information and other student data concerning
their educational qualifications, preparedness and attitudes, in order to predict which
students are most likely to drop out so that scarce additional support can be directed
at the students most likely to benefit. A mathematical, data-driven approach of this kind
at The Open University has identified very wide differences between entering students
in relation to the probability of them completing a single course. This prediction has
been used to decide which students to contact and support, with measurable positive

39
The higher Education Academy

consequences for overall retention (Simpson, 2003; Gibbs et al., 2006). The types of
student variables that predict drop-out go well beyond the kind of data that HEFCE
have available to calculate institutional benchmarks for retention. So even the extent
to which institutions exceed or fall short of their retention benchmarks can only be a
crude and incomplete measure of their educational quality.
Not all of the institutional variation in retention is due to student variables.
Efforts to improve retention have been evaluated for 30 years in the US, and while
overall retention rates have remained largely static, this hides substantial progress
in improving retention in some institutions. A good deal is now known about what
kinds of institutional efforts are likely to improve retention and persistence in the
US (Barefoot, 2004) and, with a much lesser evidence base, in the UK (Yorke,
1999). Making good use of thorough information about students so as to target
timely individualised support and intervention is one of the most effective practices.
Other effective practices closely resemble those identified as improving student
performance and educational outcomes in general (LaNasa et al., 2007), discussed
in Section 5.5.3 above. In particular, collaborative and interactive learning and close
contact with teachers increases social and academic integration. As pointed out
above, such interventions have a greater impact on less able students.
If variations between students, and especially psychological variables such as
motivation and commitment, and social variables, such as where students live and
how much time they have available to study, could be fully taken into account, then
retention performance could be used as an indicator of educational quality. However,
with the data currently available this is not yet practicable.

6.3 Employability and graduate destinations

The extent to which graduating students are able to obtain employment reasonably
quickly, in graduate jobs, in fields relevant to their degree subject, and with a salary
that justifies their investment of time and money in their higher education, is a
commonly used dimension of quality. The difficulty with employability data, as with
retention data, is their interpretation. Different methods of collecting data, and in
particular the timing of the data collection, makes a considerable difference, and
the process usually relies on surveys involving student self-reporting. However
employability data are collected, interpreting differences between institutions is
problematic for a wide variety of reasons (Smith et al., 2000):

—— It is affected by degree classification, and student performance is closely

linked to students’ prior educational qualifications (which also affects
employability), which varies greatly between institutions (see Section 4.4
above). In addition, degree classifications vary in their meanings across
institutions (see Section 6.1 above).

40
 Dimensions of qualit y

—— It is greatly affected by institutional reputation, which is a very poor

indicator of educational quality (see Section 5.7.1 above). There is little
evidence that employers have an accurate and up-to-date picture of the
educational quality of the institutions they employ graduates from, especially
if they derive this picture from currently invalid league tables.
—— It is affected by region and locality, due to variations in local employment
markets and the proportion of students who live at home and who are less
likely to move away either to study or to find employment.
—— It changes over time, due to changes in the employment market, and
probably differentially between institutions, without any changes in the
effectiveness of higher education institutions.
—— It is affected by students’ social class, and the mix of social class varies
between institutions.
—— It is affected by students’ age, and age profiles vary between institutions.
—— It is affected by student affluence, with some students needing to take any
employment quickly while others can afford to wait for graduate-level and
subject-relevant employment, or to continue their education.
—— It is affected by subject mix, with higher levels of graduate unemployment,
non-graduate-level employment, and employment in fields unrelated to the
degree subject, more often in some subjects than in others.

Interpreting the differences that exist between employability indicators in the

UK and mainland Europe is also difficult because the higher education systems
are different in crucial ways. For example, about half of UK graduates from non-
vocational subjects felt that their first jobs were not appropriate to a degree-level
education compared with about a quarter of such graduates from other European
countries. For vocational science graduates the proportion was only 17% and
10% respectively, but still higher for the UK than for the rest of Europe (Brennan
and Tang, 2008a, 2008b). The overall difference, across all subjects, is probably
because European graduates tend to be older, much more likely to have Masters-
level qualifications by the time they enter the job market, and more likely to have
experienced work-based education relevant to their subject during their more
extended education. There is also a different subject mix between the UK and
Europe with more non-vocational Arts graduates in the UK (with unemployment
rates twice that of some other subjects). This employability difference between the
UK and Europe disappears about five years after graduation, at which point 96% of
UK graduates are in employment (ibid.), which supports the interpretation that there
are influential differences between the UK and mainland Europe regarding graduates’
age and vocational experience at the time of graduation.
There are US examples of research that validate institutional employability
missions by identifying long-term consequences for the way graduates demonstrate,
in their professional lives, abilities learnt at college (e.g. Mentkowski and Doherty,

41
The higher Education Academy

1984). There are only very modest studies of this kind in the UK (e.g. Jenkins et al.,
2001), and certainly not enough to make institutional comparisons or even to validate
institutional claims about the efficacy of their employability missions.
The Higher Education Statistics Agency is able to take into account some
variables (subject of study, qualifications on entry and age on entry) in setting
institutional performance benchmarks for employability, but not others. Smith et
al. (2000) have made a more mathematically sophisticated attempt to take more
variables into account, but still leave out crucial variables about which data are not
easy to obtain. Interpreting an institution’s graduate employment performance in
relation to HEFCE benchmarks is fraught with problems.
Finally, the loose fit that characterises the UK’s higher education and its jobs
market has been interpreted by some commentators not as a problem, but as
providing flexibility for graduates to cope with a fluid employment market that
is constantly changing in relation to the capabilities that are required. This issue
concerns the difference between expertise for efficiency, which is what employers
recruiting graduates normally demand, and adaptable expertise, that enables an
individual to operate effectively in unpredictable new situations (Schwartz et al.,
2005). It takes very different kinds of educational process to develop these two
forms of expertise. There is a lack of evidence about the long-term consequences for
graduate employment of either narrowly focused vocational education or education
that emphasises efficiency in generic ‘employability skills’, rather than emphasising
the higher order intellectual capabilities involved in adaptable expertise. This makes
relying on HESA’s very short-term employment data a risky thing to do.

42
 Dimensions of qualit y

7. Summary and conclusions

7.1 The importance of process variables

Much of this report demonstrates what commentators in the US have been arguing
for many years. Presage variables such as funding, research performance and the
reputation that enables an institution to have highly selective entry, do not explain
much of the variation between institutions in relation to educational gains. Measures
of educational product such as grades and career earnings reflect these presage
variables, because the best students compete to enter the best funded and most
prestigious institutions and the quality of students is the best predictor of products.
Measures of product such as retention and employability are strongly influenced by a
raft of variables that make interpreting an institution’s performance extremely difficult.
The most important conclusion of this report is that what best predicts
educational gain is measures of educational process: what institutions do with their
resources to make the most of whatever students they have. The process variables
that best predict gains are not to do with the facilities themselves, or to do with
student satisfaction with these facilities, but concern a small range of fairly well-
understood pedagogical practices that engender student engagement.
In the UK we have few data about the prevalence of these educational practices
because they are not systematically documented through quality assurance
systems and nor are they (in the main) the focus of the NSS. The best measure of
engagement, the NSSE, is used only to a very limited extent in the UK.

7.2 The importance of multivariate analysis

Much of the UK data about relationships between presage and process variables, or
between either presage or process variables and product variables, looks at one pair
of variables at a time – for example, the relationship between a measure of research
performance (e.g. the RAE) and a measure of teaching quality (e.g. TQA scores). Such
relationships are invariably confounded with related variables, for example with the
quality of students attracted to the high-status institutions that have high research
performance. As a consequence few relationships between two variables can be
interpreted with confidence. The few UK studies that have examined a number of
variables at a time using some form of multivariate analysis (e.g. Drennan and Beck,

43
The higher Education Academy

2001; Yorke, 1998) have confirmed that apparently strong relationships between pairs of
variables (e.g. between a measure of research and a measure of teaching) are confounded
by other variables that could equally be responsible for apparent relationships (e.g. a
measure of quality of student intake). In the US there have been far more, larger and
more complex, multivariate analyses that take into account a whole raft of variables
at the same time and which, as a consequence, are able to tease out those variables
that are confounded with others and those that are not. We are therefore largely
dependent on US data and analyses for our understanding of the complex relationships
between dimensions of quality. Some of the necessary data that would allow a more
comprehensive multivariate analysis in the UK have already been collected and collated
(for example by HEFCE, HESA, the NSS and by HEPI), but it currently resides in
different data-bases. It would be helpful to combine these databases so as to allow
multivariate analysis, and to align data collection methods to make this easier to do.

7.3 The importance of educational gain

While some UK data include measures of educational product, there are very few UK
studies that have included measures of educational gain. This matters because the best
predictor of product is the quality of students entering the institution, and the quality
of students varies greatly between institutions, so that if you only have a measure of
product, such as degree classifications, rather than of gains, then you cannot easily
interpret differences between institutions. When UK studies do attempt to measure
gain they involve different measures on entry than on leaving higher education (for
example A-level point scores and degree classifications, respectively). Furthermore
the most common measure of product, degree classification, varies in its meaning and
standard across subjects and across institutions (Yorke, 2009). It is therefore difficult to
interpret even these comparative measures of gain. Studies in the US in contrast are far
more likely to use psychometric measures of generic educational outcomes (such as a
test of critical thinking) with the same measure, and with the same standards being used
across different subjects and institutions, and also using the same measure both before
and after experiencing three or four years of college. In this way a reliable measure of
educational gain, and comparison between institutions in relation to educational gain,
is possible. Again we are heavily dependent on US studies for evidence of which quality
dimensions predict educational gain, and especially on the vast studies, and reviews of
evidence, undertaken by Astin (1977, 1993) and Pascarella and Terenzini (1991, 2005).

7.4 Dimensions of quality in different kinds of institutions

Relying on US data might not matter if institutions and educational processes were
essentially the same on either side of the Atlantic. However, it seems likely that

44
 Dimensions of qualit y

the dimensions that define quality in a valid way are different in different kinds of
institutions. For example, even within the US the quality indicators that appear valid
for large, national, research universities (in the sense that they predict educational
performance tolerably well) do not work as well, or at all, in regional schools and
non-selective colleges (Schmitz, 1993). Similarly the normal lack of a relationship
between an emphasis on research and an emphasis on teaching does not seem
to apply to a small group of well-endowed liberal arts colleges that emphasise
close contact between teachers and students (Astin, 1993). Different pedagogical
phenomena are likely to be salient in different contexts, with somewhat different
patterns of relationships between process and product, dependent on context.
It is not just that different educational processes might have more influence on
educational gains in some types of institution than in others. Measures of educational
gain themselves might also need to be different between institutions if they are to
have meaning. Institutional missions vary, particularly with regard to the relative
importance of employability and subject knowledge. It would be surprising if the
same measures of educational gain were equally appropriate in all UK institutions.
For example, The Open University’s mission, emphasising openness, means that it
would not seek to increase student retention and performance through increasing
selectivity because that would reduce its openness. Its own indicators of quality
are distinctive, and are different even from those used by HEFCE in determining its
funding. The problem here is that funding mechanisms are driven by indicators of
quality that cut across institutions’ missions.
US research has done well to identify any consistent patterns at all across varied
contexts. However, the limits of what is possible to conclude, on average, have been
highlighted by those conducting the research (Pascarella, 2001). The same caution
should accompany extrapolation of findings about key indicators of quality from
varied US contexts to varied UK contexts.

7. 5 D i m e n s i o n s o f q u a l i t y i n d i f f e r e n t d e p a r t m e n t s

Much of the literature cited above, and most of the debate, has focused on
institutional differences in quality. However, it is clear that departments can differ
hugely within the same institution. Regarding NSS scores, there are institutions
that have the highest-rated department in England in one subject and the lowest
rated in another subject, despite sharing the same institutional quality indicators.
Educational leadership of departments makes a difference, creating cultures that value
teaching, that engage in a constant process of improving teaching, and that create
rich and engaging learning environments, to some extent whatever the institutional
environment and presage variables (Ramsden, 1998; Gibbs et al., 2008b).
Interestingly the two institutions frequently referred to in this report, and
that appear at the top of the NSS ranking, the University of Oxford and The

45
The higher Education Academy

Open University, have ‘institutional pedagogies’: common patterns of teaching and

assessment across all departments. All their departments score similarly highly on
the NSS. This may be in part due to deliberate policies to protect the institutional
pedagogy (which is known to work well) from outside pressures, with the effect of
limiting departmental variation in teaching quality. Princeton University similarly has
an approach to quality assurance through its course approval process that is intended
to maintain ‘the Princeton model’ rather than encourage diverse innovation.
Much of the emphasis in the UK on recognising and rewarding quality in teaching
has focused on individual teachers (e.g. through the National Teaching Fellowship
Scheme) or institutions (through league tables of one kind or another). There are
examples elsewhere of national and institutional schemes to identify educational quality
(e.g. in Finland and at the University of Oslo, respectively) that focus on departments,
or ‘learning environments’ at the level of a degree programme (Gibbs, 2008).

7.6 Dimensions of quality in different subjects

The Carnegie Foundation for the Advancement of Teaching established a large-scale

initiative on the assumption that the pedagogies of disciplines are different: that, for
example, educational quality is achieved through different strategies and practices in
the creative arts than in the sciences, and differently in English than in Law (Huber
and Morreale, 2002). At a sufficient level of abstraction there are similar underpinning
educational principles across all disciplines (Gibbs, 1999), but they are embodied in
such varied educational practices, and are salient to such different extents in different
disciplines, that disciplines in effect achieve educational quality in different ways. If
you then attempt to measure quality across disciplines, for example by using the
Course Experience Questionnaire, you find that some disciplines emerge consistently
better than others, across different studies and different institutions. Either one
has to accept that certain subjects are always taught less well than others, which
seems highly unlikely, or that different measures of quality are better aligned with the
consequences of some (disciplinary) pedagogic practices than with others. Product
measures are particularly susceptible to disciplinary variation. The distribution of
degree classifications is markedly different between subjects in the UK (Yorke et al.,
2002; Bridges et al., 2002). Employability means different things, and comes about in
different ways, following the study of different disciplines. Comparing quality between
disciplines is fraught with difficulties.
It seems likely that many students choose subjects first and institutions second,
not the other way round. Current institutional quality data and rankings are the
wrong place to start if informing potential students is the priority. Furthermore
the definition of a ‘subject’ in the information available to students often does not
correspond very closely with the programme in which students may be interested.
Brown et al. (2009) give an example of a ‘subject category’ used for reporting

46
 Dimensions of qualit y

NSS ratings that in one institution is an average drawn from 11 different degree
programmes. Students need good data about programmes more than they do about
institutions or even about broad ‘subjects’, and the NSS currently does not provide
that, for technical reasons that will be difficult to overcome. Political demands for
‘better information for customers’ cannot be met with current data gathering and
analysis methods partly because they aggregate data in too coarse a way. Once
data are aggregated in a fine enough way to be useful, there are then bound to be
problems with sample sizes. This problem may be intractable and is one of a number
of similar problems that make it difficult to provide information about quality in
accessible and usable forms even when it has been collated (Brown, 2007).
The quality of individual courses or modules also varies within degree
programmes, and the extent of this variation may be related to degree coherence.
This report has focused on institutions and degree programmes rather than on
variables that primarily affect individual courses.

7.7 Dimensions of quality that are difficult to quantify

Most of this report has focused on dimensions of quality that are fairly readily
operationalisable in a way that enables them to be measured quantitatively, so that
statistical relationships can be established with other dimensions that are similarly
easy to measure. There are other dimensions of quality that are important, at least
in some contexts, but that are difficult or impossible to quantify. For example,
throughout literature involving case studies of excellent teaching at department level
there are references to aspects of departmental culture: whether teaching is valued
and rewarded, whether teachers regularly talk to each other about teaching and its
improvement, whether innovation in teaching is systematically supported and funded,
whether educational effectiveness is the subject of serious scholarly evaluation, and
so on (Hannan and Silver, 2000). Qualities of departmental leadership of teaching
make a considerable difference (Ramsden, 1998; Gibbs et al., 2008a), and some efforts
have been made to measure teachers’ perceptions both of departmental leadership of
teaching and of the teaching environment that frames the kind of teaching and learning
that is likely to take place (e.g. Prosser and Trigwell, 1997; Martin et al., 2003).
Sometimes highly effective educational systems are driven almost entirely by
values, such as ‘liking young people’, virtually independently of the pedagogic practices
employed or the resources available. In an international study of departments that were
identified by their institution as of exceptionally high quality in relation to teaching,
students in one of the departments said that their teachers were not especially
good but that it didn’t matter because they felt included in an exciting community of
scholars (Gibbs et al., 2008a). Studies at Oxford Brookes University concerning why
some subjects regularly produced better student performance than others found no
differences in any quantitative measure of presage variables. However, a qualitative

47
The higher Education Academy

follow-up study found that the high performing subjects were characterised by healthy
‘communities of practice’ involving much discussion of how to solve teaching problems
so as to make the entire programme work well for students. In contrast, subjects
with consistently low average marks were characterised by a corresponding lack of
talking about teaching, and a fragmented focus on individual courses (Havnes, 2008).
It may be difficult or impossible to measure such influential variables in ways that
allow safe comparison between contexts, although it may be possible to measure
their consequences, for example in relation to student engagement.

7.8 Evidence of the products of learning

Among the most telling of all indicators of the quality of educational outcomes must
be students’ final-year dissertations and project reports. It is a distinctive feature
of UK higher education (and in the past a requirement of the CNAA for honours
degree classification) that students undertake a very substantial piece of independent
study in their final year. Even at US Ivy League institutions undergraduate students
would usually need to take a fourth, Honours, year to tackle such a challenging piece
of work. It is often a culmination and integration of all they have learnt, especially
in applied and creative fields of study. There is an almost total lack of evidence
concerning the relative quality of such products across institutions, within subjects.
An attempt, for this report, to obtain such evidence from subject centres elicited
not a single example, and the few published studies illustrate the embryonic nature of
efforts (e.g. Woolf et al., 1999). Dissertations and project reports are often archived
and are available for study – although currently not comprehensively across all
institutions. Such products would be amenable to systematic peer review within each
subject’s academic community, in a way that the external examiner system signally
fails to do (Warren-Piper, 1994). Such products would also be amenable to review
by educational researchers using a generic framework for categorising the quality of
learning outcomes such as the SOLO (structure of the observed learning outcome)
taxonomy (Biggs and Collis, 1982), which is capable of distinguishing levels of quality
across different forms of assessment product within subjects, and even across subjects.

7.9  he potential for improved quality, and the evaluation of

T
improved quality

The lack of a relationship between research performance, funding, SSRs and student
selectivity, on the one hand, and student engagement and educational gains on
the other, that makes these presage variables such poor indicators of quality, is
not inevitable – it is not like an invariant physical law, i.e. it does not apply to all
circumstances, for all time. It is in part a consequence of comparatively well-funded,

48
 Dimensions of qualit y

selective, research-oriented institutions not exploiting their potential advantages

to the full. If they were to change their practices in line with what is known about
educational effectiveness, in other words emphasise process variables rather than
relying on presage variables to do the job on their own, then this relationship could
change, provided that their potential advantages enabled them to implement the
processes more thoroughly than others.
There is a rapidly growing evidence base in the US about the impact on
educational gains of changes in evidence-based educational processes (cf. Carinin et
al., 2006). It is becoming ever clearer, as a consequence, what educational processes
it is sensible to introduce with a reasonable likelihood of improving educational gains.
There is no such evidence base in the UK, or even a methodology for building one.
While in England there are some deliberate institutional efforts to improve NSS
scores they do not always appear to be based on any valid conceptual framework
concerning what processes are likely to produce what outcomes and the NSS does
not currently have validity as an indicator of quality in the sense that it predicts
outcomes, let alone gains. An example of an alternative approach in the UK is the
TESTA project based at the University of Winchester (TESTA, 2010), which employs
established methodologies for auditing assessment practices and for measuring
student learning responses to those practices, both before and after evidence-
based changes in assessment, involving whole degree programmes, across a number
of institutions. Progress has been made in the US partly because a large number
of institutions have been prepared to adopt the same evidence-based conceptual
framework and use the same valid evaluation tools. This allows meaningful evaluation
data to be pooled, and so enabling conclusions to be drawn that stand out from the
noise of contextual variation. The Higher Education Academy is starting to take a
lead in identifying and disseminating suitable evaluation tools and methodologies, and
creating an evidence infrastructure within which data from locally conducted studies
could be collated.

49
The higher Education Academy

8. About the author

Graham Gibbs has spent 35 years in research and development work to improve the
quality of teaching, student learning and assessment in higher education.
He has been centrally involved in a series of national teaching development
initiatives, including the Teaching More Students Project and HEFCE’s Institutional
Learning and Teaching Strategy initiative, and in the co-ordination of the Fund for
the Development of Teaching and Learning. He is the founder of the Improving
Student Learning Symposium and of the International Consortium for Educational
Development in Higher Education. He has been awarded an Honorary Doctorate by
Sheffield Hallam University for his leadership of the development of teaching in the
UK, and by the University of Utrecht for his international leadership of efforts to
improve university teaching.
He retired from his position as Professor and Director of the Oxford Learning
Institute, at the University of Oxford, in 2007.

50
 Dimensions of qualit y

9. Acknowledgements

In undertaking background research for this report I received support from Bahram
Bekhradnia, John Brennan, David Watson and Mantz Yorke, and also from Roger
Brown who in addition provided invaluable guidance in revising drafts. They helped
me to locate its focus in relation to existing literature and in relation to the nature
of current debates about quality and standards in higher education. Their wisdom is
greatly appreciated.
I would also like to thank Higher Education Academy staff for their support, and
in particular Dr Rachel Segal for her editorial guidance

51
The higher Education Academy

10. References

Abrami, P.C., d’Appolonia, S. and Cohen, to improvements in teaching and learning in

P.A. (1990) Validity of student ratings of classrooms. Quality in Higher Education. 13 (3),
instruction: what we know and what we do pp275–286.
not. Journal of Educational Psychology. 82 (2), Bauer, K.W. and Bennett, J.S. (2003) Alumni
pp219–231. Perceptions Used to Assess Undergraduate
‘adjwilson’. (2010) If the cap fits [online]. London: Research Experience. Journal of Higher
Education Guardian. Available from: Education. 74 (2), pp210–230.
www.guardian.co.uk/education/2010/may/25/ Bergren, M., Snover, L. and Breslow, L. (2007)
education-letters-gibb-adhd-tuition-fees Undergraduate research opportunities at
[12 August 2010]. MIT. Illuminatio. Spring 2007, pp6–8.
Armstrong, M., Clarkson, P. and Noble, M. Biggs, J.B. (1993) From theory to practice: a
(1998) Modularity and credit frameworks: the cognitive systems approach. Higher Education
NUCCAT survey and 1998 conference report. Research and Development. 12 (1), pp73–85.
Newcastle-upon-Tyne: Northern Universities Biggs, J. and Collis, K.F. (1982) Evaluating the quality
Consortiuum for Credit Accumulation and of learning: the SOLO Taxonomy (Structure of
Transfer. the observed learning outcome). New York:
Ashby, E., Plant, K., Ericsson, A., Hill, L. and Academic Press.
Asberg, K. (2005) Why study time does not Black, P. and Wiliam, D. (1998) Assessment and
predict grade point average across college classroom learning. Assessment in Education.
students: Implications of deliberate practice 5 (1), pp7–74.
for academic performance. Contemporary Bloom, B.S., Englehart, M.D., Furst, E.J., Hill, W.H.
Educational Psychology. 30 (1), pp96–116. and Krathwohl, D.R. (1956) Taxonomy of
Astin, A. (1977) Four critical years. San Francisco: educational objectives. Handbook 1: Cognitive
Jossey-Bass. domain. New York: David McKay.
Astin, A.W. (1985) Achieving educational excellence: Bound, J. and Turner, S. (2005) Cohort crowding:
a critical assessment of priorities and practices how resources affect collegiate attainment.
in higher education. San Francisco and London: Journal of Public Economics. 91 (5–6), pp877–899.
Jossey-Bass. Bowden, R. (2000) Fantasy Higher Education:
Astin, A. (1993) What matters in college. San university and college league tables. Quality in
Francisco: Jossey-Bass. Higher Education. 6 (1), pp41–60.
Bales, R.F., Strodtbeck, F.L., Mills, T.M. and Brennan, J. and Tang, W. (2008a) The employment of
Roseborough, M. (1951) Channels of UK graduates: comparisons with Europe. REFLEX
communication in small groups. American Report to HEFCE No 1. Bristol: HEFCE.
Sociological Review. 16 (4), pp461–468. Brennan, J. and Tang, W. (2008b) Subject differences
Barefoot B.O. (2004) Higher education’s revolving in graduate employment across Europe. REFLEX
door: confronting the problem of drop-out Report to HEFCE No 2. Bristol: HEFCE.
at US colleges and universities. Open Learning, Brennan, J., Patel, K. and Tang, W. (2009) Diversity
19, (1), pp65–77. in the student learning experience and time
Barrie, S. and Ginns, P. (2007) The linking of devoted to study: a comparative analysis of the
national teaching performance indicators UK and European evidence. Bristol: HEFCE.

52
 Dimensions of qualit y

Bridges, P., Cooper, A., Evanson, P., Haines, C., Coffey, M. and Gibbs, G. (2000) The evaluation of
Jenkins, D., Scurry, D., Woolf, H. and Yorke, M. the Student Evaluation of Educational Quality
(2002) Coursework marks high, examination questionnaire in UK higher education.
marks low: discuss. Assessment and Evaluation in Assessment and Evaluation in Higher Education.
Higher Education. 27 (1), pp35–48. 26 (1), pp89–93.
Brown, R. (2006) League Tables – do we have to Cook, R., Butcher, I. and Raeside, R. (2006)
live with them? Perspectives: Policy and Practice Recounting the scores: an analysis of the
in Higher Education. 10 (2), pp33–38. QAA subject review grades 1995–2001.
Brown, R. (2007) The information fallacy. Oxford: Quality in Higher Education. 12 (2), pp135–144.
Higher Education Policy Institute. Available Curtis, S. and Shani, N. (2002) The effect of
from: www.hepi.ac.uk/484-1291/The- taking paid employment during term-time on
Information-Fallacy.html [June 2010]. students’ academic studies. Journal of Further
Brown, R. (2010) Comparability of degree standards? and Higher Education. 26 (2), pp129–138.
Oxford: Higher Education Policy Institute. Curtis, S. and Williams, J. (2002) The reluctant
Available from: www.hepi.ac.uk/455-1838/ workforce: undergraduates’ part-time
Comparability-of-degree-standards.html employment. Education and Training. 44 (1),
[June 2010]. pp5–10.
Brown, R., Carpenter, C., Collins, R. and Winkwist- Dochy, F., Segers, M., Van den Bossche, P. and
Noble, L. (2009) Recent developments in Gijbels, D. (2003) Effects of problem-based
information about programme quality. Quality learning: a meta-analysis. Learning and
in Higher Education. 13 (2), pp173–186. Instruction. 13 (5), pp533–568.
Carinin, R., Kuh, G. and Klein, S. (2006) Student Drennan, L.T. and Beck, M. (2001) Teaching
engagement and student learning: testing Quality Performance Indicators – key
the linkages. Research in Higher Education. 47 influences on the UK universities’ scores.
(1), pp1–32. Quality Assurance. 9 (2), pp92–102.
Carney, C. McNeish, S. and McColl, J. (2005) The Dunbar-Goddet, H. and Trigwell, K. (2006)
impact of part-time employment on students’ A study of the relations between student
health and academic performance: a Scottish learning and research-active teachers.
perspective. Journal of Further and Higher Paper presented at the 14th International
Education. 29 (4), pp307–319. Improving Student Learning Symposium,
Cheng, Y.C. and Tam, W.M. (1997) Multi-models Bath, 4–6 September.
of quality in education. Quality Assurance in Eccles, C. (2002) The use of university rankings
Education. 5 (1), pp22–31. in the United Kingdom. Higher Education in
Chickering, A.W. (1974) Commuting versus resident Europe. 27 (4), pp423–432.
students: Overcoming the educational inequities of Ehrenberg, R.G. (2006) What’s Happening in Public
living off campus. San Francisco: Jossey-Bass. Higher Education? Westport, CT: Praeger.
Chickering, A.W. and Gamson, Z.F. (1987a) Seven Ewell, P. (2008) No correlation: musings on some
Principles for Good Practice in Undergraduate myths about quality. Change. November-
Education. Racine, WI: The Johnson December 2008, 40 (6), pp8–13.
Foundation Inc. Fearnley, S. (1995) Class size: the erosive effect
Chickering, A.W. and Gamson, Z.F. (1987b) Seven of recruitment numbers on performance.
principles for good practice in undergraduate Quality in Higher Education. 1 (1), pp59–65.
education. AAHE Bulletin. 39 (7), pp3–7. Feldman, K. (1984) Class size and college students’
Chickering, A.W. and Gamson, Z.F. (1991) evaluations of teachers and courses: a closer
Applying the seven principles for good practice look. Research in Higher Education. 21 (1),
in undergraduate education. San Francisco: pp45–116.
Jossey-Bass. Finnie, R. and Usher, A. (2005) Measuring the
Clarke, M. (2002) Some guidelines for academic Quality of Post-secondary Education: Concepts,
quality rankings. Higher Education in Europe. Current Practices and a Strategic Plan. Kingston,
27 (4), pp443–459. ON: Canadian Policy Research Networks.

53
The higher Education Academy

Ford, J., Bosworth, D. and Wilson, R. (1995) Part- Gibbs, G., Habeshaw, T. and Yorke, M. (2000)
time work and full-time higher education. Institutional learning and teaching strategies
Studies in Higher Education. 20 (2), pp187–202. in English higher education. Higher Education.
Gansemer-Topf, A., Saunders, K., Schuh, J. 40 (3), pp351–372.
and Shelley, M. (2004) A study of resource Gibbs, G., Knapper, C. and Picinnin, S. (2008a)
expenditure and allocation at DEEP colleges. Departmental leadership for quality teaching:
Ames, IA: Educational Leadership and Policy an international comparative study of
Studies, Iowa State University. effective practice. London: Leadership
Gardiner, L.F. (1997) Redesigning higher education: Foundation. Available from: www.lfhe.
producing dramatic gains in student learning. ac.uk/research/projects/gibbsoxford.html
ASHE-ERIC Higher Education Report 7. [May 2010].
Washington DC: Association for the Study of Gibbs, G., Knapper, C. and Picinnin, S. (2008b)
Higher Education. Disciplinary and contextually appropriate
Gibbs, G. (1995) National scale faculty approaches to leadership of teaching in
development for teaching large classes. research-intensive academic departments in
In: Wright, A. (ed.) Teaching Improvement higher education. Higher Education Quarterly.
Practices. New York: Anker. 62 (4), pp 416–436.
Gibbs, G. (1999) Are the pedagogies of the Gibbs, G., Lucas, L. and Simonite, V. (1996) Class
disciplines really different? In: Rust, C. size and student performance: 1984–94.
(ed.) Improving Student Learning Through the Studies in Higher Education. 21 (3), pp261–273.
Disciplines. Oxford: Oxford Centre for Staff Gibbs, G., Morgan, A. and Taylor, E. (1982)
and Learning Development. A review of the research of Ference
Gibbs, G. (2008) Designing teaching award schemes. Marton and the Goteborg Group: a
York: Higher Education Academy. phenomenological research perspective
Gibbs, G. (2010) The assessment of group on learning. Higher Education. 11 (2),
work: lessons from the literature. Oxford: pp123–145.
Assessment Standards Knowledge Exchange. Gibbs, G., Regan, P. and Simpson, O. (2006)
Available from: www.brookes.ac.uk/aske/ Improving student retention through
documents/Brookes%20groupwork%20 evidence based proactive systems at the
Gibbs%20Dec%2009.pdf [June 2010]. Open University (UK). College Student
Gibbs, G. and Coffey, M. (2004) The impact of Retention. 8 (3), pp359–376.
training of university teachers on their teaching Glass, G.V. and Smith, M. L. (1978) Meta-Analysis
skills, their approach to teaching and the of Research on the Relationship of Class Size
approach to learning of their students. Active and Achievement. San Francisco: Far West
Learning in Higher Education. 5 (1), pp87–100. Laboratory for Educational Research and
Gibbs, G. and Dunbar-Goddet, H. (2007) The effects Development.
of programme assessment environments on student Glass, G.V. and Smith, M. L. (1979) Meta-analysis
learning. York: Higher Education Academy. of Research on the Realationship of Class-
Gibbs, G. and Dunbar-Goddet, H. (2009) Size and Achievement. Evaluation and Policy
Characterising programme-level assessment Analysis. 1, pp2–16.
environments that support learning. Graham, A. and Thompson, N. (2001) Broken
Assessment and Evaluation in Higher Education. ranks: US News’ college rankings measure
34 (4), pp481–489. everything but what matters. And most
Gibbs, G. and Jenkins, A. (eds.) (1992) Teaching universities don’t seem to mind. Washington
Large Classes: maintaining quality with reduced Monthly. 33 (4), pp9–14.
resources. London: Kogan Page. Grunig, S.G. (1997) Research, reputation and
Gibbs, G. and Lucas, L. (1997) Coursework resources: the effect of research activity on
assessment, class size and student perceptions of undergraduate education and
performance: 1984–94. Journal of Further and institutional resource acquisition. Journal of
Higher Education. 21 (2), pp183–192. Higher Education. 33 (9), pp9–14.

54
 Dimensions of qualit y

Hannan, A. and Silver, H. (2000) Innovating in of Session 2008-09. Volume 1. London: The
Higher Education: teaching, learning and Stationery Office.
institutional cultures. Buckingham: The Society Huber, M.T. and Morreale, S. (2002) Disciplinary
for Research into Higher Education/Open styles in the scholarship of teaching and learning:
University Press. exploring common ground. Washington DC:
Harvey, L. and Green, D. (1993) Defining quality. American Association for Higher Education
Assessment and Evaluation in Higher Education. and the Carnegie Foundation for the
18 (1), pp9–34. Advancement of Teaching.
Hathaway, R.S., Nagda, B.A. and Gregerman, S.R. Hunt, A., Lincoln, I. and Walker, A. (2004) Term-
(2002) The relationship of undergraduate time employment and academic attainment:
research participation to graduate and evidence from a large-scale survey of
professional education pursuit: an empirical undergraduates at Northumbria University.
study. Journal of College Student Development. Journal of Further and Higher Education. 28 (1),
43 (5), pp614–631. pp3–18.
Hattie, J. and Marsh, H.W. (1996) The relationship Innis, K. and Shaw, M. (1997) How do students
between research and teaching: a meta- spend their time? Quality Assurance in
analysis. Review of Educational Research. 66 (4), Education. 5 (2), pp85–89.
pp 507–542. Jenkins, A. (2004) A guide to the research on
Hattie, J. and Timperley, H. (2007) The power of teaching-research relations. York: Higher
feedback. Review of Educational Research. 77 Education Academy.
(1), pp 81–112. Jenkins, A., Jones, L. and Ward, A. (2001) The
Hattie, J., Bibbs, J. and Purdie, N. (1996) Effects Long-term Effect of a Degree on Graduate
of Learning Skills Interventions on Student Lives. Studies in Higher Education. 26 (2),
Learning: A Meta-analysis. Review of pp147–161.
Educational Research. 66 (2), pp99–136. Jessop, T. & El-Hakim, Y. (2010) Evaluating and
Havnes, A. (2008), There is a bigger story behind. improving the learning environments created
An analysis of mark average variation across by assessment at programme level: theory
Programmes. European Asssociation for Research and methodology. European Association for
into Learning and Instruction Assessment Research into Learning and Instruction
Conference. University of Northumbria. Assessment Conference, University of
HEFCE (2001) Analysis of Strategies for Learning Northumbria, June 2010.
and Teaching. Report 01/37a. Bristol: Higher Johnes, G. (1992) Performance indicators in higher
Education Funding Council for England. education: a survey of recent work. Oxford
HEPI (2006) The Academic Experience of Students Review of Economic Policy. 8 (2), pp19–33.
in English Universities (2006 Report). Oxford: Jongbloed, B.W.A. and J.J. Vossensteyn (2001)
Higher Education Policy Institute. Keeping up Performances: an international
HEPI (2007) The Academic Experience of Students survey of performance based funding in
in English Universities (2007 Report). Oxford: higher education. Journal of Higher Education
Higher Education Policy Institute. Policy and Management. 23 (2), pp127–145.
Hochschul-Informations-System (2005) Eurostudent Kehm, B.M. and Stensaker, B. (2009) University
2005: Social and Economic conditions of student rankings, diversity and the landscape of higher
life in Europe 2005. Hannover: HIS. education. Rotterdam: Sense Publishers.
Hoskins, S., Newstead, S.E. and Dennis, I. (1997) Kuh, G.D. and Pascarella, E.T. (2004) What
Degree Performance as a Function of Age, does institutional selectivity tell us about
Gender, Prior Qualifications and Discipline educational quality? Change. September-
Studied. Assessment & Evaluation in Higher October 2004, 36 (5), pp52–58.
Education. 22 (3), pp317–328. LaNasa, S., Olson, E. and Alleman, N. (2007) The
House of Commons Innovations, Universities, impact of on-campus student growth on
Science and Skills Committee (2009) first-year engagement and success. Research
Students and Universities. Eleventh Report in Higher Education. 48 (8), pp941–966.

55
The higher Education Academy

Lindsay, R. and Paton-Saltzberg, R. (1987) NCHEMS (2003) Do DEEP institutions spend more
Resource changes and academic performance or differently than their peers? Boulder, CO:
at an English Polytechnic. Studies in Higher National Centre for Higher Education
Education. 12 (2), pp213–27. Management Systems.
Lindsay, R., Breen, R. and Jenkins, A. (2002) NUS (2008) NUS Student Experience Report.
Academic research and teaching quality: the London: National Union of Students.
views of undergraduate and postgraduate OECD (2000) Education at a Glance 2000. Paris:
students. Studies in Higher Education. 27 (3), Organisation for Economic Cooperation and
pp309–327. Development.
Lucas, L., Jones, O., Gibbs, G., Hughes, S. and Olcott, D. (2010) Par for the course. Times Higher
Wisker, G. (1996) The effects of course Education. 8 April, p32.
design features on student learning in large Pascarella, E.T. (1980) Student-faculty informal
classes at three institutions: a comparative contact and college outcomes. Review of
study. Paper presented at the 4th Educational Research. 50 (4), pp545–595.
International Improving Student Learning Pascarella, E.T. (2001) Identifying excellence in
Symposium, Bath. undergraduate education: are we even close?
Marsh, H.W. (1982) SEEQ: a reliable, valid and Change. 33 (3), pp19–23.
useful instrument for collecting students’ Pascarella, E.T. and Terenzini, P. (1991) How college
evaluations of university teaching. British affects students. San Francisco: Jossey-Bass.
Journal of Educational Psychology. 52, pp77–95. Pascarella, E.T. and Terenzini, P. (2005) How college
Marsh, H.W. (1987) Students’ evaluations of affects students: a third decade of research,
university teaching: research findings, Volume 2. San Francisco: Jossey-Bass.
methodological issues, and directions for Pascarella, E.T., Cruce, T., Umbach, P., Wolniak,
future research. International Journal of G., Kuh, G., Carini, R., Hayek, J., Gonyea, R.
Educational Research. 1 (3), (entire issue). and Zhao, C. (2006) Institutional selectivity
Martin, E., Trigwell, K., Prosser, M. and Ramsden, and good practices in undergraduate
P. (2003) Variation in the experience of education: How strong is the link? Journal of
leadership of teaching in higher education. Higher Education. 77 (2), pp251–285.
Studies in Higher Education. 28 (3), pp247–259. Pascarella, E.T., Seifert, T.A. and Blaich, C. (2008)
Marton, F. and Wenestam, C. (1978) Qualitative Validation of the NSSE benchmarks and deep
differences in the understanding and approaches to learning against liberal arts
retention of the main point in some texts outcomes. Paper presented at the annual
based on the principle-example structure. meeting of the Association for the Study of
In: Gruneberg, M.M., Morris, P.E. and Sykes, Higher Education, Jacksonville, FL. Available
R.N. (eds.) Practical aspects of memory. from: www.education.uiowa.edu/crue/
London: Academic Press. publications/index.htm [March 2010].
Marton, F., Hounsell, D. and Entwistle, N. (1984) Pascarella, E.T., Seifert, T.A. and Blaich, C. (2010)
The experience of learning. Edinburgh: Scottish How effective are the NSSE benchmarks in
Academic Press. predicting important educational outcomes?
Mentkowski, M. and Doherty, A. (1984) Careering Change. January-February 2010, 42(1),
After College: Establishing the Validity of Abilities pp16–22. Available from: www.changemag.
Learned in College for Later Careering and org/index.html [March 2010].
Professional Performance. Final Report to the Paton-Saltzberg, R. and Lindsay, R. (1993) The
National Institute of Education. Milwaukee, WI: Effects of Paid Employment on the Academic
Alverno College. Performance of Full-time Students in Higher
Nasr, A., Gillett, M. and Booth, E. (1996) Education. Oxford: Oxford Polytechnic.
Lecturers’ teaching qualifications and Patrick, J.P. and Stanley, E.C. (1998) Teaching and
their teaching performance. Research research quality indicators and the shaping
and Development in Higher Education. 18, of higher education. Research in Higher
pp576–581. Education. 39 (1), pp19–41.

56
 Dimensions of qualit y

Perry, W.G. (1970) Forms of intellectual and ethical Simpson, O. (2003) Student retention in open and
development in the college years: a scheme. distance learning. London: Routledge Falmer.
New York: Holt, Rhinehart and Winston. Smith, M.L. and Glass, G.V. (1979) Relationship
Piccinin, S., Cristi, C. and McCoy, M. (1999) The of class-size to classroom processes, teacher
impact of individual consultation on student satisfaction and pupil affect: a meta-analysis.
ratings of teaching. International Journal of San Francisco, CA: Far West Laboratory for
Academic Development. 4 (2), pp75–88. Educational Research and Development.
Prosser, M. and Trigwell, K. (1997) Perceptions of Smith, J. and Naylor, R. (2005) Schooling effects on
the teaching environment and its relationship subsequent university performance. Economics
to approaches to teaching. British Journal of of Education Review. 24 (5), pp549–562.
Educational Psychology. 67, pp25–35. Smith, J., McKnight, A. and Naylor, R. (2000)
QAA (2003) Learning from Subject Review 1993– Graduate employability: policy and
2001. Gloucester: Quality Assurance Agency performance in higher education in the UK.
for Higher Education. The Economic Journal. 110 (464), pp382–411.
QAA (2006) Background Briefing Note: The classification Stinebrickner, R. and Stinebrickner, T. (2008)
of degree awards. Gloucester: Quality Assurance The Causal effect of Studying on Academic
Agency for Higher Education. Performance. The B.E. Journal of Economic
QAA (2009) Thematic enquiries into concerns Analysis and Policy. 8 (1), Article 14.
about academic quality and standards in higher Svanum, S. and Bigatti, S.M. (2006) The Influences
education in England. Gloucester: Quality of Course Effort and Outside Activities on
Assurance Agency for Higher Education. Grades in a College Course. Journal of College
Ramsden, P. (1979) Student learning and Student Development. 47 (5), pp564–576.
perceptions of the academic environment. Svensson, L. (1977) On qualitative differences
Higher Education. 8 (4), pp411–427. in learning: III – Study skill and learning.
Ramsden, P. (1998) Learning to lead in higher British Journal of Educational Psychology. 47,
education. London: Routledge. pp223–243.
Ramsden, P. (1999) A performance indicator of Terenzini, P. T, and Pascarella, E. T. (1994) Living
teaching quality in higher education: the myths: Undergraduate education in America.
Course Experience Questionnaire. Studies in Change. 26 (1), pp28–32.
Higher Education. 16 (2), pp129–150. TESTA (2010) Transforming the Experience of
Ramsden, P. and Moses, I. (1992) Associations between Students Through Assessment. Available
research and teaching in Australian higher from: www.winchester.ac.uk/studyhere/
education. Higher Education. 23 (3), pp273–295. ExcellenceinLearningandTeaching/research/
Säljö, R. (1979) Learning about learning. Higher Pages/TESTA.aspx [12 August 2010].
Education. 8 (4), pp443–451. Tinto, V. (1975) Dropout from higher education:
Sastry, T. and Bekhradnia, B. (2007) The academic a theoretical synthesis of recent research.
experience of students in English universities. Review of Educational Research. 45 (1),
London: Higher Education Policy Institute. pp89–125.
Schmitz, C. (1993) Assessing the validity of Higher Thompson, N. (2000) Playing with numbers: how
Education Indicators. Journal of Higher US News mis-measures higher education and
Education. 64 (5), pp503-521. what we can do about it. Washington Monthly.
Schomburg, H. and Teichler, U. (2006) Higher 32 (9), pp16–23.
education and graduate employment in Europe: Trigwell, K. (2005) Teaching–research relations,
Results from graduate surveys from twelve cross-disciplinary collegiality and student
countries. Dordrecht: Springer. learning. Higher Education. 49 (3), pp235–254.
Schwartz, D.L., Bransford, J.D. and Sears, D. (2005) Trigwell, K. and Ashwin, P. (2004) Undergraduate
Efficiency and innovation in transfer. Stanford, students’ experience at the University of
CA: Stanford University. Available from: www. Oxford. Oxford: Oxford Learning Institute.
stanford.edu/~danls/Efficiency%20and%20 Available from: www.learning.ox.ac.uk/oli.
Innovation%204_2004.pdf [April 2010]. php?page=365 [April 2010].

57
The higher Education Academy

Trigwell, K. & Prosser, M. (2004) The development classifications, 1994–95 to 2006–07, for
and use of the Approaches to Teaching England, Wales and Northern Ireland. York:
Inventory. Educational Psychology Review, 16, 4, Higher Education Academy.
pp409–424. Yorke, M., Bridges, P. and Woolf, H. (2000) Mark
Usher, A. and Savino, M. (2006) A world of difference: distributions and marking practices in UK
a global survey of university league tables. Toronto, higher education; some challenging issues.
ON: Educational Policy Institute. Available Active Learning in Higher Education. 1 (1),
from: www.educationalpolicy.org/pdf/world-of- pp7–27.
difference-200602162.pdf [12 August 2010]. Yorke, M., Barnett,G., Bridges, P., Evanson,
Van Rossum, E.J., Deijkers, R. and Hamer, R. P.,Haines, C., Jenkins,D., Knight, P., Scurry,
(1985) Students’ learning conceptions and D., Stowell, M. and Woolf, H. (2002) Does
their interpretation of significant educational grading method influence honours degree
concepts. Higher Education. 14 (6), pp617–641. classification? Assessment and Evaluation in
Vos, P. (1991) Curriculum Control of Learning Higher Education. 27 (3), pp269–279.
Processes in Higher Education. 13th Yorke, M., Woolf, H., Stowell, M., Allen, R.,
International Forum on Higher Education of Haines, C., Redding, M., Scurry, D.,Taylor-
the European Association for Institutional Russell, G., Turnbull, W. and Walker, W.
Research. Edinburgh. (2008) Enigmatic Variations: Honours Degree
Warren-Piper, D. (1994) Are Professors Professional? Assessment Regulations in the UK. Higher
The Organisation of University Examinations. Education Quarterly. 63 (3), pp157–180.
London: Jessica Kingsley.
Weimer, M. and Lenze, L.F. (1997) Instructional
interventions: a review of literature on
efforts to improve instruction. In: Perry, R.P.
and Smart, J.C. (eds.) Effective Teaching in
Higher Education: Research and Practice. New
York: Agathon Press.
Wood, K., Limsky, A.S., and Straus, M. A. (1974).
Class Size and student evaluations of faculty.
Journal of Higher Education. 43, pp524–34.
Woodley, A. (2004) Conceptualising student
drop-out in part time distance education:
pathologising the normal? Open Learning. 19
(1), pp47–63.
Woolf, H., Cooper, A., Bourdillon, B., Bridges, P.,
Collymore, D., Haines, C., Turner, D. and
Yorke, M. (1999) Benchmarking academic
standards in History: an empirical exercise.
Quality in Higher Education. 5 (2), pp145–154.
Yorke, M. (1997) A good league table guide? Quality
Assurance in Higher Education. 5 (2), pp61–72.
Yorke, M. (1998) The Times ‘League Table’ of
universities, 1997: a statistical appraisal.
Quality in Education. 6 (1), pp58–60.
Yorke, M. (1999) Leaving early: undergraduate non-
completion in higher education. London: Falmer.
Yorke, M. (2001) Formative assessment and its
relevance to retention. Higher Education
Research and Development. 20 (2), pp115–126.
Yorke, M. (2009) Trends in honours degree

58
Dimensions of qualit y

59
The higher Education Academy

60
Dimensions of qualit y

61
The Higher Education Academy supports the sector in
providing the best possible learning experience for all
students. It does this by:

—— p roviding national leadership in developing and

disseminating evidence-informed practice about
enhancing the student learning experience
—— operating as an independent broker, enabling
expertise to be shared across institutions and
subject areas
—— working at multiple levels, with individual
academics, subject communities,
departments, faculties and institutions
—— working across all parts of the UK, recognising
the distinctive policy contexts and priorities of
the devolved administrations but also provising
opportunities to share expertise among them.

The Academy is an independent organisation funded

by grants from the four UK higher education funding
bodies, subscriptions from higher education institutions,
and grant and contract income for specific initiatives.

www.heacademy.ac.uk