Appendix 1

SCHOOL SIZE AND SCHOOL EFFECTIVENESS: Summary of main findings of key research on schools'

September 2005

Index

Backgroundpages 3-7

  1. Overall performance and achievement
  2. National research findingspage 8
  3. Kent research on school size page 8
  4. Performance – primarypage 10
  5. Overall finding page 16
  6. Performance – secondarypage 16
  7. Overall findingpage 23
  1. Teaching
  2. Teaching – primarypage 24
  3. Teaching – secondarypage 27
  4. Overall finding page 28
  1. Delivery of the National Curriculum
  2. National research findingspage 29
  3. Curriculum offer in Hertfordshire page 29
  4. Overall finding page 30
  1. Cost
  2. Institutional costs page 31
  3. Support costs page 32
  4. Formula funding page 32
  5. Overall finding page 33
  1. Stability
  2. Budgetary impact of loss of pupil numberspage 34
  3. Pupil movement and resultspage 35
  4. Planning, Preparation and Assessment timepage 40
  5. Teacher absence page 42
  6. Recruitment page 42
  7. Overall finding page 44
  1. Leadership and Management
  2. Leadership and Management – primarypage 45
  3. Leadership and Management – secondarypage 46
  4. Overall finding page 48
  1. Student Behaviour
  2. Pupil behaviour – primarypage 48
  3. Pupil absence – primarypage 50
  4. Pupil behaviour – secondarypage 50
  5. Pupil absence – secondarypage 52
  6. Overall finding page 52
  1. Conclusionpages 53-56
  1. Appendices

Appendix A: Referencespage 56

Appendix B: DfES Data on Key Stage 2 results (English & maths)page 57

Appendix 1: Overall performance and value-added measurespage 58

Appendix 2: Teachingpage 61

Appendix 3: Leadership and Managementpage 63

Appendix 4: Student Behaviourpage 64

Appendix 5: Stabilitypage 65

Background

Introduction

  1. This report summarises the main findings of key research on the impact of school size on school performance. It covers both the primary and secondary phases. The research that has been considered in this report has been accessed through various sources including the Wheathampstead research library, the Internet (including the British Education Index) and University research departments.
  1. There is little published evidence available which relates directly to the impact of school size on performance. Many articles and research documents written or commissioned by organisations such as the National Small Schools Forum, Human Scale Education and the National Association of Small Schools, which promote education in small schools in the UK, were discounted for this paper. This was because they were biased towards the effectiveness of small schools and focused on emotive issues without significant statistical evidence to back up their claims.
  1. Although research in the US on school size is more developed than in the UK there is a much wider variation in the size of school and in the socio-economic and cultural contexts of schooling, which makes it difficult to relate conclusions to the UK context. The small schools movement in the United States is also very active and many articles that were found through Internet searches advocate the merits and perceived advantages of small schools.
  1. For these reasons this paper focuses mainly on research looking at the impact of size on school in England,rather than in the United States or other countries.
  1. Independent research referred to and used in this report mainly comes from organisations which have no preconceptions or agenda about school size; have a strong reputation within the research community; are unbiased and use broad evidence bases. The only major exception to this is Dr. Ian Craig's research Primary School Size & its Relationship to School Effectiveness: An exploration of optimal size for primary schools for Kent County Council. His research is referred to in this report because Kent is a statistical neighbour of Hertfordshire. The research was done in order to build evidence to support school planning and organisational proposals for primary provision within Kent County Council and so was particularly pertinent to Hertfordshire.
  1. Appendix A of this paper lists the research used.

Hertfordshire data

  1. Research on Hertfordshire schools' performance according to school size was put together using Ofsted inspection grades, Hertfordshire Framework for Monitoring, Challenge, Support and Intervention (MCSI) and DfES value-added measures. For the Ofsted grades, schools were categorised according to their school roll at the time of their inspection. For the value-added measures, schools were categorised according to their 2004 roll number. The MCSI framework is updated annually and the data used in the report is for 2004.

National research utilised

  1. Key national research referred and used in this report includes work carried out by the NfER, Ofsted and the Audit Commission. In addition, research from the EPPI-centre (which is part of the Social Science Research Unit, Institute of Education, University of London) is utilised for the secondary phase.
  1. There are a number of caveats that need to be taken into consideration when analysing the research and the data used in this report. These include:
  • Variations such as the quality of teaching, leadership and funding per pupil within a school will have an impact on school performance;
  • The current funding system with the minimum funding guarantee for every school, regardless of size, results in more funding per pupil in smaller schools than in larger schools;and
  • Ian Craig in his research, Primary School Size & its Relationship to School Effectiveness: An exploration of optimal size for primary schools identified, within the Small Church Study,[1] that it is undeniable that small schools have smaller classes than large schools.
  1. Wrigley and Fjeld Lofsnaes in Schools as social and learning communities also make the point that at the secondary phase in the UK there is a market system in operation, where more popular schools are encouraged to expand up to (and sometimes beyond) their physical capacity, leaving less popular schools with many empty places. They go on to say that additional pupils attracted to the growing schools tend to be children of more affluent or knowledgeable or determined parents and thus achievement leads to a change in size, rather than simply size affecting achievement.
  1. While it is essential to take note of this context it is possible to compare performance within these caveats. DfES value-added measures and Ofsted grades were used to consider whether school size has an impact on standards within JMI primary and secondary schools. The simple analysis of Hertfordshire JMI schools using Ofsted grades and also value-added measures shows that the largest schools did the best overall according to Ofsted grades and average value-added measures (although the smallest schools achieved a higher percentage of schools achieving 100+ in terms of value-added measures). A similar picture emerges for the secondary phase with the largest secondary schools in Hertfordshire performing the best.
  1. There is also some significant independent evidence to suggest that at secondary school phase there is an optimum size of school and that any larger or smaller than this optimum will have an adverse impact on standards.
  1. This report considers the research findings under a number of criteria which researchers have used as measurement of schools' performance. These are overall and value-added achievement, teaching, delivery of the National Curriculum, cost, stability, leadership and management and pupils' behaviour and attendance.

Definition of a small school

  1. The first issue to consider is the definition of a small school.There are a number of definitions of a small school, but no consensus. According to the LGA report Education in Rural Communities published in 2000, the Audit Commission uses 90 pupils as the threshold for a small primary. Primary schools with fewer than 100 pupils are small in OFSTED terms while those with fewer than 50 are very small.
  1. According to the Ofsted definition,the LGA report calculated in 2000 that there were 2700 small primary schools in England and Wales with fewer than 100 pupils on roll of which 700 were very small, having fewer than 50 on roll. The Church of England’sdefinition of a small school is one where the head is a ‘teaching head’, that is, when a substantial proportion of the head’s time is spent on teaching.
  1. The LGA report notes that the DfEE(as was) has provided extra resources for small schools and that a small school is defined as one with fewer than 200 pupils on roll (primary) or 600 (secondary) or 700 if they have a sixth form.
  1. According to this definition, there were 7,900 small primary schools in England (43 per cent of the total) and about 850 small secondary schools (24 per cent) in 2000. These figures include middle schools.
  1. Local authoritieshave their own definitions of small primary schools, usually for budgetary purposes. Some examples of thresholds for primary schools for small school support in 2000 are:
  • Cumbria - 100
  • Cornwall - 120
  • Derbyshire - 160
  • Devon - 200
  • Essex - 125
  • Hampshire - 165
  1. In Hertfordshire there are many different categories of schools at the primary phase: first schools, infant only, junior only and junior, mixed and infants (JMI) schools. In order to ensure that schools were compared like with like and to ensure that schools were divided into groupings that would provide some idea of the number of cohorts,the analysis focuses on JMI schools in the first instance. This gave a total of 304 current JMI schools out of the 416 primary phase schools in Hertfordshire. Of the 304 schools, seven were not included in the Ofsted grade results because they are either new schools or amalgamated schools and haven't been inspected since they became JMI schools.

Categorisation of Hertfordshire schools used in this report

  1. For the purposes of the report the JMI schools were divided into the following four categories of size of school:
  • 100 or fewer pupils (approximately 4 or less classes per school)
  • 101 - 210 pupils (approximately 4 to 7 classes per school - 210 being equivalent to one form entry)
  • 211 - 300 pupils (approximately 7 to 10 classes per school)
  • 301+ pupils (approximately 10 classes or more per year - 420 being equivalent to two form entry)

According to the Pupil Level Annual School Census (PLASC) returns the national average size of primary schools in 2004 is 239 pupils. Over the last five years this figure has remained relatively stable and was 243 in 2000. The Hertfordshire average size of primary school for 2005-06 is 208 pupils.

  1. The number of Hertfordshire JMI schools within each category of school size will alter year on year as school rolls go up or down. Set out below, however, is information on the number and percentage of schools within each category of size according to the 2005 school roll data and also size of school roll at the time of Ofsted inspection.

School size category / Number of schools in each category according to their roll at inspection / Number of schools in each category according to 2005 school roll records
100 or fewer pupils / 28 (9.4%) / 32 (10.5%)
101 – 210 pupils / 96 (32.3%) / 99 (32.6%)
211 – 300 pupils / 119 (40.1%) / 114 (37.5%)
301+ pupils / 54 (18.2%) / 59 (19.4%)
Total number of schools / 297[4] / 304
  1. The secondary schools have been divided into the 3 categories of size which the Hertfordshire Management Information Unit uses for comparison purposes. The categories were set so that each of the three groupings had roughly the same number of schools in them. They are as follows:
  • 400 - 900 pupils
  • 901 - 1100 pupils
  • 1101 - 1500 pupils

School size category / Number of schools in each category according to their roll at inspection / Number of schools in each category according to 2005 school roll records
400 – 900 pupils / 26 / 24
900 – 1100 pupils / 33 / 28
1101 – 1500 pupils / 17 / 24
Total number of schools / 76 / 76

According to the PLASC returns the national average size of secondary schools in 2004 is 1025 pupils. It is interesting to note that over the last five years this average has risen year on year,unlike primary schools, and was 955 pupils in 2000. The Hertfordshire average size of secondary school for 2005-06 is 1017 students.

Hertfordshire policy on school size

  1. Hertfordshire is currently going through a number of geographical reviews of provision, mainly of primary schools. For primary areas these are as follows:
  • Hoddesdon
  • Hatfield (remainder)
  • Harpenden
  • Hemel Hempstead
  • Potters Bar
  • and Hertford and Ware in the secondary sector.
  1. Work has been done to develop a set of factors which will be used to inform these reviews. With regard to primary school size, the reviews will aim to produce schools normally not fewer than two forms of entry in size.
  1. At secondary level recent re-organisations have sought to provide larger secondary schools ranging from 6 forms-of-entry, in order to strengthen the scope for achieving breadth and balance in the curriculum. This compares with the smaller schools (typically 4 to 5 forms of entry) developed up to 1990.

Main findings

1Overall performance and value-added achievement

1.1`National research

  1. Nationally, there is no clear conclusion about the impact of school size and academic achievement in the primary phase,although, interestingly, the impact of small schools and academic achievement appears to be significantly less of an issue than proponents for small schools would have people believe.
  1. Ofsted produced a report in March 2000 entitled Small schools: how well are they doing?This was based on the data for primary schools from 1999 inspections and national test results. This concluded that the standards of achievement are higher in small schools (Ofsted's definition is fewer than 100 pupils) compared to all other schools but with the caveat that these schools have a disproportionately low number of pupils from disadvantaged backgrounds, and this has a positive impact on standards. The following key points are made:
  • The majority of small schools achieve standards which are higher than the national average at the age of 7 and 11, but these standards are lower than those achieved in larger schools when socio-economic background is taken into account.
  • While there are also many successful very small schools (50 pupils or less) this group is less effective in its provision for under fives and is more vulnerable to the adverse influences of weak teaching and/or weak leadership.
  1. In 2002 the DfES published data on Key Stage 1 & 2 results which indicated that for schools with a cohort of more than ten pupils, school size appeared to have little significant effect on pupil performance. For example, at level 1 for the mathematics test, the percentage of pupils achieving the expected level or above by size of school cohort (maintained schools only) was 92 per cent on average for a cohort of up to 10 pupils and 92 per cent on average for a cohort of 100 pupils and over.[2]
  1. The NfER's research on The Impact of School Size and Single-sex Education on Performance, published in 2002, examines the impact of school size and single-sex education on pupil performance and opportunities, using national value-added datasets (2001)which contain individual pupil data across 979 primary and 2,954 secondary schools.
  1. This research also indicates that size is not a significant factor in academic achievement of pupils during the primary phase.

1.2Kent research

  1. However, Dr. Ian Craig's research which was carried out in Kent to find evidence to support school planning and organisational proposals came to different conclusions. The project was designed as six separate but parallel case studies carried out between 1996 and 2000, three focused on two LEAs in the south-east of England and three using a wider English and Welsh LEA sample.
  1. The report claims that school effectiveness is influenced by school size. Dr. Ian Craig analysed the 1998/99 KS2 SATs level 4+ English and mathematics percentage scores by SEN 'family' separately for small schools[3] and ‘not small’ schools. Schools were grouped into families according to their levels of SEN and so family A had the lowest SEN levels while family J had the highest SEN levels.
  1. For both the ‘not small’and small schools in all three subjects, school 'families' with the least SEN achieve highest and schools with the most SEN achieve least at KS2 level 4+.
  1. For English the ‘not small’ schools perform marginally better than the small schools for SEN Families A and B and significantly better for Family C (Families A to C have the lowest SEN levels). However, for all the other SEN Families (D to J)performance was better in small schools.
  1. For mathematics school size appeared to have little impact on the results for SEN School Families with the lowest levels of SEN (A and B). However, the ‘not small’ schools’ performance was better than the small schools for SEN School Families C to E. This is reversed for those families with highest levels of SEN (F to J),with performance better in the small schools.
  1. For science performance in small schools seemed to be better than performance in the ‘not small’ schools across all SEN family groupings.
  1. It is interesting to note that for the two family groups with the lowest SEN levels (A and B) school size seems to have the least influence. Performance was similar for all three subjects.

Table showing the comparison of Kent LEA’s 1998/99 KS2 SATs Level 4 scores by 'small' and 'not small' schools and SEN families

SEN Family
A / B / C / D / E / F / G / H / I / J / ALL
KS2 English / Small / 82.20 / 76.17 / 62.33 / 92.00 / 73.50 / 79.50 / 59.00 / 69.33 / 48.33 / 56.00 / 69.84
Not small / 82.45 / 83.50 / 76.95 / 73.83 / 69.26 / 67.90 / 60.94 / 54.78 / 50.60 / 45.25 / 66.50
KS2 Maths / Small / 85.30 / 77.83 / 62.00 / 62.00 / 66.17 / 66.50 / 58.40 / 80.00 / 47.67 / 50.80 / 65.7
Not small / 81.58 / 77.60 / 73.58 / 71.43 / 68.03 / 62.80 / 58.47 / 50.33 / 52.10 / 50.56 / 64.60
KS2 Science / Small / 84.40 / 87.50 / 92.00 / 81.50 / 89.00 / 89.00 / 72.40 / 77.33 / 66.20 / 47.67 / 78.70
Not small / 85.40 / 85.18 / 81.49 / 76.91 / 74.95 / 74.95 / 69.65 / 63.83 / 61.60 / 53.94 / 72.80
  1. The Kent research also indicates that pupils with high levels of social and educational need tend to do better in smaller schools. However, the report also specifies that children with low levels of social and educational need generally thrive at any size of school and that there are possible advantages for them in the larger schools where the curriculum on offer is likely to be more varied and focused.
  1. Interestingly, its main conclusion is that other than in very exceptional circumstances primary schools should not fall below 100 pupils, or grow beyond 630 pupils (3 FE). There was overall agreement from all those involved in the project, that when all aspects of education are considered the disadvantages to operating outside this 'window', outweigh the advantages. Schools could be allowed to fall to, but not below 60 pupils, if four classes and a non-teaching headship were able be protected. There was, however, overwhelming agreement from respondents in the Main Study (in-depth interviews with sixty head teachers) that ideal school size should, if possible, be whole forms of entry, suggesting ideal school sizes of 200/210 pupils (1 FE) or 400/420 pupils (2 FE).

1.3Hertfordshire primary schools - overall effectiveness and value-added measures