FORUM

Volume 50, Number 2, 2008

www.wwwords.co.uk/FORUM

Big or Small: does the size

of a secondary school matter?
MARK NEWMAN

ABSTRACT The relationship between the size of a school and various education
outcomes continues to interest parents, campaigners and politicians. This article
summarises some of the arguments made in relation to the importance of school size
and explores the results of a systematic review of 31 research studies on the effects of
secondary school size. Overall, the review found that directions and patterns of effect
vary for different outcomes. The results of the review suggested that there was little
empirical evidence to justify policies aimed at changing or mandating particular school
sizes. However, given the continuing interest in the issue and indications that more
research is becoming available, the author suggests that continuing rigorous systematic
evaluation is needed to explore the association between school size and outcomes.

Introduction
An article in the Sunday Times in January 2008 reported that the optimal size of
secondary schools was about 800 children. The overall thrust of the article was
that secondary schools in England are becoming too big. The article included
the Sunday Times own data and the opinion of a ‘professor of education’ to
support the newspaper’s claim (Sian Griffiths, Sunday Times, 6 January, 2008).
Issues raised in the article reflect a long-standing public and media interest in
the issue of school size. One of the attractions of ‘size’ is that, in theory at least,
the issue is amenable to relatively simple controls. Decisions about the size of
any new school are made in consultation with local authorities, and size is often
cited as a justification for closing a particular school. In England, government
policy on school admissions has been gradually moving towards a model
whereby individual schools have more freedom to choose the size of their
intake. This could be viewed as a reduction in the country’s collective capacity
to control school size or, more benignly, as transferring the power over the
decision from politicians (local or national), to parents and schools themselves.

167
Mark Newman
Whatever view is taken, it is useful to know whether, and under what
circumstances, school size is in fact an important determinant of outcomes.

Arguments in Favour of and

against Bigger or Smaller Schools
Many different arguments have been offered as to why school size might affect
student achievement. Broadly speaking, it is argued that larger schools are able
to offer wider curricular and extra-curricular opportunities and a concomitant
increase in teacher specialisation. Against this it is argued that smaller schools
are able to offer a more personalised learning environment (Darling Hammond
et al, 2002) and that this facilitates greater interaction and participation by
students and teachers alike.
Another related group of arguments concerns the ways in which changes in
school size brought about through planning or through the operation of quasi-
market forces may affect the quality of educational provision. For example, it is
argued that ‘poorly performing’ schools will be unpopular and therefore decline
in size as parents choose to send their children to ‘successful’ schools. This, it is
argued, will lead the shrinking school to ‘improve its performance’. This type of
argument is linked to issues about the value of the wider functions of schools as
centres of community networks and the consequences for local communities of
school closure. This is where the school size debate intersects with general
issues about the public provision of school places, especially in rural areas, and
is a prominent feature of school size research in the USA.
The notion of ‘economies of scale’ provides the basic framework around
which arguments about costs of education in relation to school size are
constructed. Scale economies occur when the cost of enrolling an additional
student is lower than the average cost at that point, thereby resulting in the
average cost declining as enrolments increase. If it can be shown that the same
or better educational outcome can be achieved at a lower cost per pupil in larger
schools, there may be a case for setting a mandatory minimum school size, or
for limiting the size of schools if average costs are found to rise once a certain
level of enrolment is exceeded (McKenzie, 1995).

What Size are Secondary Schools?

For the 3,343 publicly-funded secondary schools in England (excluding schools
admitting only students with special needs) in January 2007, more than 50 per
cent had between 700 and 1,200 pupils. About 2 per cent of schools had more
than 1,800 pupils and about 1 per cent had fewer than 300 pupils (DCSF,
2007). Whilst the top end of the range in England may appear considerably
lower than its equivalent in the USA (which has schools of 5,000+ pupils), like
the USA there is a wide range in the size of secondary schools.
Knowing how many schools there are, and of what size, does not,
however, tell us very much. We need to know whether there is any relationship

168
 DOES THE SIZE OF A SECONDARY SCHOOL MATTER?

between the size of a school and any outcomes that we may think indicate
something about the quality of educational provision in that school. We might
do this by referring to theoretical arguments about the relationship between size
and outcomes, as in the examples above. But given that different, and sometimes
opposing, arguments are made we would also want to explore the relationship
empirically i.e. through research.

Research on School Size

The relationship between school size and outcomes could be empirically
investigated in a number of different ways. Quantitative approaches investigate
the relationship between school size as an ‘independent variable’ and a range of
‘dependent’ outcome variables. Data are generally cross-sectional, derived from
national school census returns and/or large scale school surveys. In some
instances data may be longitudinal, considering the effects as a school changes
size either through school consolidation or normal variations in enrolment.
Based on relevant theories (and on pragmatic considerations such as availability
of data), researchers control for particular variables in their models (either by
matching or by inclusion as an independent variable), and select an appropriate
method of statistical analysis.
Qualitative approaches may, through case studies of particular schools or
local authorities, provide a very detailed description of the relationships
between school size and other factors. These approaches may be particularly
helpful in identifying contextual factors affecting school size, the preoccupations
of key people (such as head teachers, teachers and pupils) in schools of different
sizes and, through observational studies, qualitative differences between schools
of different size.

Systematic Review
Before conducting any new primary research to answer a question it is useful to
consider the evidence provided by existing studies. There are numerous
‘reviews’ of research on school size (for example, Cotton, 1996; McGuire,
1989; Muir, 2000; Spielhofer et al, 2002; Tasker, 2003). These reviews all
have limitations of one kind or another. One thing they have in common is that
they do not provide a clear explanation of how the author identified, selected,
or synthesised the evidence they use; that is they were not reported in a way
that made it possible to assess their relative strengths and weaknesses as reviews.
Not surprisingly, the reviews come to different conclusions, but it is not clear to
the reader why they did so.
In 2003 the EPPI-Centre was commissioned by the (then) Department for
Education and Skills (DfES) to conduct a systematic review of the relationship
between school size (expressed as the number of students) and ‘outcomes’.[1] A
systematic review differs from the traditional literature review in that it is

169
Mark Newman
undertaken as a piece of research and thus uses a transparent and explicit
process or method. Specifically:
• it answers a research question;
• it is protocol driven;
• an explicit search strategy is used to identify all possible research evidence;
• an explicit set of criteria are used to select studies for inclusion in the review;
• a coding tool is used to obtain the same information about all the studies
included in the review;
• this information is used to make a judgement about the quality and relevance
of a particular study to the review question;
• the results of the individual studies are synthesised in a systematic and
transparent manner to generate a ‘new’ and/or summary answer to the
review question.[2]
The remainder of this article provides a summary of the findings of this review.

Results of the Systematic Review

The systematic review addressed the following question: ‘In OECD countries
since 1990 what is the relationship between secondary schools of different sizes
and various ‘outcome’ variables?’.[3] A total of 3,874 citations were identified
through systematic searches of six databases. After screening all of these, 31
studies were included in the review. The majority of studies in the review were
carried out in North America, with 21 from the USA and two from Canada. The
remaining studies came from England (six), Finland (one) and Australia (one).
The studies were grouped according to outcome measure used: student
attainment (19 studies), student drop-out and absence (five studies), student
violence (six studies), student and teacher perceptions of school climate (five
studies), economic outcomes (five studies), and school organisational practices
(two studies). Within each of these categories studies were sub-divided
according to their quality. The type of analysis used in the studies produced a
‘slope’ showing the direction of ‘effect’ as either increasing as school size
increases (positive, +), decreasing as school size increases (negative, -),
increasing and then decreasing as school size increases (quadratic, ∩), or some
combination of these. Figure 1 shows a summary of the slope or direction of
effect for each of these outcome groups. The summary direction of effect reflects
the review group’s judgment about what the evidence found in the studies
included in the review tells us about the relationship between secondary school
size and that particular outcome.
Overall, the results suggest a quadratic association between school size
and student attainment in exams. That is, attainment increases as school size
increases up to a certain point but then appears to decline. The findings from a
number of studies suggest that, within the secondary range, the association
between attainment and size may be different across age groups and for students
with different levels of socio-economic status (SES).

170
 DOES THE SIZE OF A SECONDARY SCHOOL MATTER?

Figure 1. Summary of directions of main effect for different outcome categories.

Note: this figure is for illustrative purposes only and is not based on literal
interpretations of quantitative coefficients reported.

Overall the results suggest a quadratic association between school size and
student absence. That is, absence decreases as school size increases up to a
certain point but then appears to increase. None of the studies examined
variation on this outcome (or any of the subsequent outcomes reported below)
by age or SES.
The summary of directions of effect for school size and student behaviour
in Figure 1 illustrates the inconsistency of results in the included studies. In
some studies there was increased ‘ violence’ in larger schools; for other studies
the association was in the opposite direction, i.e. there was more ‘violence’ in
smaller schools. Different studies use different definitions of violent or anti-
social behaviour and different methods of data collection, either of which may
explain the different directions of effect reported.

171
Mark Newman
Studies measuring what students and teachers thought about their school
found a negative association between school size and this outcome, despite
different conceptualisations, outcome measures and instruments used. Such a
pattern of results suggests that we can be fairly confident that teachers and
students in smaller schools have more positive perceptions of the school climate
than their counterparts in larger schools.
In all studies measuring the association between school size and costs, the
‘costs’ measured were limited to public expenditure on schools. The pattern of
results here suggests that as school size increases the school’s cost per pupil
decreases. The size of the association between average secondary school size
and costs differs slightly between studies. An increase in school size of 10 per
cent is estimated to reduce costs per student by between one per cent and four
per cent depending on the definition of cost used.

Discussion
When considering the implications of the results provided by the systematic
review it is important to take into account its limitations. The remit of this
review extended only to a consideration of studies that investigated empirically
the association between an outcome variable and school size. It is also important
to note that studies from the USA about the division of larger schools into much
smaller schools (so called ‘schools-within-schools’) were excluded from the
review because they did not meet the criteria for inclusion.
The review process itself had a number of limitations. Comparatively
limited searches were undertaken and we were not able to obtain all of the
papers identified. It is difficult to estimate the impact of this. A simple
preliminary analysis of the sensitivity of the bibliographic database search found
that between 60 and 70 per cent of papers cited in the included studies had
been identified. The design of studies included in the in-depth review allowed
for sophisticated analysis of the degree of association between the dependent
variable (for example, attainment) and school size. However, the studies did not
assign students or teachers at random to different size schools (i.e. they did not
use an experimental design), and therefore may be affected by selection bias.
Conclusions about causality must therefore be considered tentative.
Different studies use different analytical models, different methods of
analysis and different methods of constructing both the dependent and
independent variables. Making comparisons across studies is therefore difficult,
even when they use the same dependent variable. There are also differences in
the socio-economic and cultural contexts of schooling in different countries.
Taken together, these differences may limit the generalisability of conclusions
about the ‘effects’ of school size on school outcomes.
Another important limitation of the findings is that the individual studies
in the review only measure a limited range of outcomes. Attainment, for
example, is measured only in terms of performance in tests and examinations,
which is only one aspect of attainment. Similarly, the costs of schooling are

172
 DOES THE SIZE OF A SECONDARY SCHOOL MATTER?

measured only in terms of the financial costs to government. School size impacts
on a much wider range of costs and benefits than those included in existing
studies.

Conclusions
The systematic review was concerned with the overall relationship between
secondary school size and outcomes. In terms of the overall relationship, i.e.
across all outcomes, the association with school size would appear to be
inconsistent. In practical terms this suggests that for some types of outcomes
larger schools do better than smaller schools and for others smaller schools do
better.
Three key issues require further consideration. First, to be of practical
value we would need to know at what school size attainment is likely to be
maximised and absence (or student drop-out) is minimised. Being confident that
the relationship is quadratic does not tell us very much. For example, a
quadratic relationship with an optimal size range of about 400 students may
have quite different implications to one where the optimal range is about 2,000
students. As we have seen earlier, an optimum size range of 2,000 would
include just about all secondary schools in England and so would not be
particularly useful. The studies included in this review do not provide a clear
answer to this question. Second, we need to know whether the directions of
‘effect’ based on the ‘average’ secondary school apply to all types of secondary
school and/or subgroups of particular students. Third, we need to know the
‘effect’ on outcomes of changing the size of individual schools. Such an analysis
would need to include not only the effects on the school that had changed size
but also effects on neighbouring schools. Only two studies in the systematic
review attempted this. Bradley & Taylor (1998, 2004) provide empirical
evidence that schools which perform well (according to GCSE point scores)
relative to neighbouring schools grow more quickly in size than schools that
perform badly. What these studies do not tell us, however, is how changes in
school size within localities affected the overall performance of students in those
localities, nor was it clear in these studies whether higher GCSE performance
led to increased enrolment or higher enrolment led to increased GCSE
performance.
The findings of the systematic review would seem to refute some of the
more prevalent myths regarding the advantages and disadvantages of smaller
and larger schools. For example, the view that student attainment is universally
higher in smaller schools and that student behaviour is universally worse in
larger schools is inconsistent with the evidence in the systematic review. These
relationships appear to be much more complex than such simple arguments
suggest.
The EPPI-Centre review on secondary school size does not provide strong
evidence to support policy initiatives that aim to either increase or decrease the
size of schools and/or to close or change the structure of schools below or

173
Mark Newman
above a certain size. The review findings suggest that for some outcomes, there
may be advantages for smaller schools whilst for other outcomes larger schools
may be more advantageous. There are also a number of qualifications that need
to be taken into account when considering the practical application of these
results. Firstly, the ‘effects’ of changing school size within a single school and
the upheaval associated with this are not identified or investigated in this
review. Secondly, many of the associations (even if they are statistically
significant) are comparatively weak, and other factors are shown to be as
important or more important in predicting the outcome variable. In particular,
whilst many studies found a statistically significant association between school
size and attainment, this was found to be considerably weaker than the
association between SES and attainment across all the studies in the review.
The issue of school size continues to excite interest amongst policy
makers, campaigners, researchers and parents. This is not surprising given
continuous changes in demographic profiles, continuing concern over
inequalities in educational attainment and continued emphasis on economic
efficiency. Further work is need in this area, both quantitative and qualitative.
Since the systematic review was published, economists at OECD have analysed
data from the PISA study and suggested that larger schools are more ‘efficient’
(Sutherland & Price, 2007). Economists at the Centre for the Economics of
Education have analysed a new combination of large-scale, individual-level
administrative datasets linked by the Department for Children, Schools and
Families and the Department for Innovation, Universities and Skills including
data from the National Pupil Database and Pupil Level Annual School Census.
They argue the results of their analysis suggest that larger schools are more
effective at getting their pupils to University (Chowdry et al, 2008).
However ‘efficiency’ at obtaining exam grades, or getting to university,
are not the only things we might take note of when considering how big or
small we want our secondary schools to be. We might also be interested in
wider aspects of children’s education and want to know about non-
governmental ‘costs’ such as costs to families and communities of changes in
school size. Consolidation of any kind may lead to more travel for students
and/or loss of school buildings for community use for example. It is therefore
important that any claims made about the association between school size and
outcomes continue to be critically and systematically investigated using a range
of different perspectives.

Notes
[1] The Department for Education and Skills (DfES) and HM Treasury (the UK
government department responsible for much economic and financial policy) in
England funded the review, set the initial scope, and were involved in refining
the research questions in collaboration with the research team.
[2] Interested readers can find out more about how to conduct systematic reviews
on the methods pages of the EPPI-Centre website

174
 DOES THE SIZE OF A SECONDARY SCHOOL MATTER?

(http://eppi.ioe.ac.uk/cms/Default.aspx?tabid=89). Further details of the

methods used in the systematic review on school size can be found in the
detailed reports and papers published on the review (Garrett et al, 2004;
Newman et al, 2005).
[3] OECD Countries: Australia, Austria, Belgium, Canada, Czech Republic,
Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy,
Japan, Korea, Luxembourg, Mexico, Netherlands, New Zealand, Norway,
Poland, Portugal, Slovak Republic, Spain, Sweden, Switzerland, Turkey, United
Kingdom, United States.

Acknowledgements
The systematic review referred to in this article was completed by the author
with colleagues: Zoe Garrett, Diana Elbourne (EPPI-Centre), Ann West, Philip
Noden (London School of Economics), Steve Bradley & Jim Taylor (University
of Lancaster). The review was funded by the DfES (now DCSF) and HM
Treasury.

References
Bradley, S. & Taylor, J. (1998) The Effect of School Size on Exam Performance in
Secondary Schools, Oxford Bulletin of Economics and Statistics, 60(3), 291-324.
http://dx.doi.org/10.1111/1468-0084.00102
Bradley, S. & Taylor, J. (2004) The Economics of Secondary Schooling, in G. Johnes &
J. Johnes (Eds) International Handbook of Education Economics. Cheltenham: Edward
Elgar.
Chowdry, H., Crawford, C. & Vignoles, A. (2008) Widening Participation in Higher
Education: a quantitative analysis, Building Research Capacity, 13, 8-11.
Cotton, K. (1996) School Size, School Climate and Student Performance. North West Regional
Educational Laboratory. http://www.nwrel.org/scpd/sirs/10/c020.html (accessed
10 March 2008).
Darling-Hammond, L., Ancess, J. & Ort, S. W. (2002) Reinventing High School:
outcomes of the coalition campus schools project, American Educational Research
Journal, 39(3), 639-673. http://dx.doi.org/10.3102/00028312039003639
Department for Children, Schools and Families (DCSF) (2007) Schools and Pupils in
England, January (Final)
http://www.dcsf.gov.uk/rsgateway/DB/SFR/s000744/index.shtml (accessed 14
April 2008).
Garrett, Z., Newman, M., Elbourne, D., Bradley, S., Noden, P., Taylor, J. & West, A.
(2004) Secondary School Size: a systematic review.
http://eppi.ioe.ac.uk/cms/Default.aspx?tabid=320 (accessed 13 May 2008).
McGuire, K. (1989) School Size: the continuing controversy, Education and Urban Society,
21(2), 164-174. http://dx.doi.org/10.1177/0013124589021002005
McKenzie, P. (1995) Scale Economies in Australian Secondary Schools, Australian
Journal of Education, 39(2), 109-125.

175
Mark Newman
Muir, E. (2000) Smaller Schools: how much more than a fad?, American Educator,
Winter, 40-46.
Newman, M., Garrett, Z., Elbourne, D., Bradley, S., Noden, P., Taylor, J. & West, A.
(2006) Does Secondary School Size Make a Difference? A Systematic Review,
Educational Research Review, 1(1), 41-61.
http://dx.doi.org/10.1016/j.edurev.2006.03.001
Spielhofer, T., O’Donnell, L., Benton, T., Schagen, S., & Schagen, I. (2002) The Impact
of School Size and Single-Sex Education on Performance. Slough: National Foundation for
Educational Research.
Sutherland, D. & Price, R. (2007) Linkages between Performance and Institutions in the
Primary and Secondary Sector. Economics Department working papers No. 558
ECO/WKP(2007)18.
http://www.olis.oecd.org/olis/2007doc.nsf/ENGDIRCORPLOOK/NT00002C
0A/$FILE/JT03228886.PDF (accessed 10 March 2008).
Tasker, M. (2003) Smaller Structures in Secondary Education: a research digest. Bristol: Human
Scale Education.

MARK NEWMAN is a Senior Research Officer and Associate Director of the

EPPI-Centre, University of London. He manages the EPPI-Centre’s systematic
review programmes in Education and Social Policy. Correspondence: Dr Mark
Newman, EPPI-Centre, Social Science Research Unit, Institute of Education,
University of London, 18 Woburn Square, London WC1H 0NR, United
Kingdom ([email protected]).

176