EPI Submission to Education Select Committee Inquiry on Primary Assessment

Today the Education Policy Institute has published its official submission to the House of Commons Education Select Committee inquiry on primary assessment. The inquiry examines:

• The purpose of primary assessment and how well the current system meets this
• The advantages and disadvantages of assessing pupils at primary school
• How the most recent reforms have affected teaching and learning
• Logistics and delivery of the SATs
• Training and support needed for teachers and senior leaders to design and implement effective assessment systems
• Next steps following the most recent reforms to primary assessment

You can read the submission in full below, or download a PDF version here.

Select Committee inquiry on Primary Assessment – Education Policy Institute response

Summary of submission

We have focused on three areas of the inquiry:

• The purposes of primary assessment;

• How well the current system meets these purposes; and

• Next steps following the recent reforms to primary assessment

The purpose of primary assessment 

1. The Education Policy Institute considers that there are three objectives in formally assessing pupils at the beginning and end of primary.

 Formative assessment

2. Attainment in the early years and at primary school is particularly important for disadvantaged pupils because we know that how well children do in their early and formative years can have a lasting effect. The Education Policy Institute’s Annual Report found that, by the end of secondary school, disadvantaged pupils are on average, 19 months behind their more affluent peers. But, crucially, the majority of the gap (60 per cent) between the most disadvantaged pupils and the rest had developed by the end of primary and 40 per cent of the gap was there before children started Key Stage 1. These stark findings highlight the need to identify pupils who are at risk of falling behind, diagnose them and intervene early on.

3. Teacher assessment is of course already widely used for these purposes and can help to support a child’s learning and development. But relying solely on unmoderated teacher assessment poses a risk of unconscious bias, particularly towards disadvantaged pupils. Research[1]published by Simon Burgess found that gender and class stereotypes can affect teacher assessments. His study found that, compared with national test scores, girls were over-assessed in English and under-assessed in maths. He also found that pupils on free school meals were consistently under-assessed by teachers compared with how well they achieved in national tests. In other words, ‘poor pupils systematically and significantly out-performed what their teachers thought they would achieve’.

4. And so the overriding purpose of assessment in education ought to be to give teachers and school leaders an understanding of each child’s attainment and progress so that they, in turn, can identify issues or delays (or, indeed, children who are working well above ‘expected standards’) and put in place effective and proven strategies to support that child.

5. This is particularly the case when considering the transition from primary to secondary school. For most secondary schools, their Year 7 intake will be unknown to them and this will be particularly the case for those with a large number of feeder primary schools. This means that it is important for secondary school teachers to have a robust and consistent assessment of a pupil’s attainment on entering the school, and for such assessment to be nationally consistent given the relatively large and varied catchment areas from which secondary schools draw their pupils.
    

Accountability 

6. Implemented correctly and proportionately, assessment outcomes can provide helpful data that should contribute to judgements about a school’s overall performance. Once again, unmoderated teacher-assessment (while valuable in many contexts) will inevitably vary from school to school. It cannot give a reliable national, or indeed local, comparison of how well a school is supporting children to progress from one phase to another.

7. The Key Stage 2 National Curriculum assessments are high-stake tests. If a school fails to meet government standards, it could be faced with intervention from either Ofsted or the Regional Schools Commissioner. This could result in the school facing intervention by its local authority or being forced to become an academy (if it is an LA maintained school) or, being transferred to a different Trust if it is already an academy.

 Funding

8. At present, local authorities can target funding (from the Dedicated Schools Grant) to pupils who have failed to meet certain benchmarks at the end of Reception and / or at the end of Key Stage 2. For pupils who have failed to meet the specified benchmarks at the end of Reception, additional money may be given to the infant or primary school and, for those who fail to meet the end of Key Stage 2 benchmarks, additional money may be given to their secondary, all-through or middle school (depending on which school the pupil(s) go on to attend). The purpose of this is to provide schools with adequate funding to support children who are ‘falling behind’.

9. Under the current funding system, local authorities (via their School Forums) have the discretion to decide whether to target funding to schools on the basis of prior attainment and the quantum of any funding that is targeted in this way. The Department for Education is currently consulting on a new national funding formula for schools[2]and their latest proposals would mean that this funding would be calculated by the Department for Education, rather than the relevant local authority.
    

How well does the current system meet these purposes? 

10. The EPI’s view is that the current system could be strengthened to support better all of the purposes set out above

 Formative Assessment challenges

11. Effective formative assessments require teachers to understand how much progress pupils have made over time and in comparison to their peers. This, in turn, requires a reliable baseline assessment. According to the DfE’s own research[3], many schools are already using baseline or ‘on entry’ assessments of some kind to identify the needs of children on entry to the primary school.

12. The Department for Education has recently had to abandon plans to roll-out new reception baseline tests. These tests were designed to provide an assessment of children on entry to primary school in order to identify any emerging needs as well as to judge the progress made by pupils from the beginning to the end of primary schools (and then hold schools to account if pupils did not make the expected levels of progress). But in April of this year, Ministers dropped plans to use the Reception baseline as the ‘starting point’ for the progress measure following a study[4]which found that the three baseline tests being piloted were not comparable with one another.

13. As a result, Key Stage 1 tests will continue to be used as the starting point to measure pupils’ progress over the course of primary school. In addition, the Early Years Foundation Stage Profile (EYFS), which is a teacher assessment and takes place over the spring term in a pupil’s reception year, will remain statutory up to and including the 2017-18 academic year. It is unclear what will happen beyond that point, but the Secretary of State has said that she will launch a consultation on primary assessment early in the new year.

 Accountability and Funding Challenges

 The differences between the old National Curriculum Levels measure and the new ‘expected standards’ 

14. The new ‘expected standard’ in Key Stage 2 assessments is intended to be equivalent to a Level 4b under the old National Curriculum levels. While, prior to 2015, floor standards were set based on the proportion of pupils achieving a Level 4 or higher, in practice, it has been more commonly asserted that the sub-level 4b (the middle third of those achieving Level 4) is a more valid predictor of secondary success, and this has formed the basis for the Government’s reforms.[5]

15. In 2016,53 per cent of pupils in English schools achieved the expected standard in all of reading, writing and mathematics. In 2015, 80 per cent achieved a Level 4 in Key Stage 2 assessments under the old National Curriculum system – a 27 percentage point-higher figure. The proportion achieving Level 4b in 2015 for reaching, writing and mathematics was reported by the DfE as 69 per cent. Because writing is not broken down into sub-levels, the 69 per cent includes those who achieved a Level 4b in reading and maths, but all of those who achieved a Level 4 in writing. The Education Policy Institute’s Annual Report estimated that the true proportion of pupils achieving a Level 4b across all three subjects was likely to be around 59 per cent in 2015 – 6 percentage points higher than the proportion reaching the Government’s expected standard in 2016.[6]

16. Given the difficulty of comparing figures, and the fact that this is the first year of these assessments, it is difficult to say at this stage whether the new standards are consistent or not with the previous Level 4b, but the shift is of the expected magnitude. Our Annual Report set a proposed goal for 85 per cent pupils to achieve what was a Level 4b by 2025. This would put us on track to ensure that the majority begin secondary school with the skills required to go on to succeed in their GCSEs, putting England among the highest performing developed countries by 2030.

17. However you look at it, the scale of the challenge is clear: these figures suggest that over a quarter of a million pupils not reaching the Government’s expected standard in reading, writing and mathematics.

18. Unsurprisingly, as the graph below shows, the reduction in the number of pupils achieving the overall respective benchmarks between 2015 and 2016 is reflected in changes in each of the subjects assessed.

19. The most interesting pattern is that only 66 per cent of pupils achieved the reading test threshold, lower than for the other subjects. As noted in the DfE’s statistical release[7], the new results reverse the pattern of recent years of reading having the highest proportion reaching the standard benchmark among reading, writing and mathematics.

20. In comparison, the proportion reaching the expected standard in the new grammar, punctuation and spelling test was only 8 percentage points lower than the proportion achieving the old Level 4 in 2015.

21. It is not possible to discern how far this change in pattern is a result of a conscious policy decision by the Department to encourage schools to focus more on reading than previously, or an unexpected outcome of the way the assessments have played out in practice between externally marked exams and teacher assessments.

22. More generally, there is some evidence of less consistency in pupils’ results across subjects. The graph below plots the proportion of pupils achieving the respective standards in maths, and whether they also achieved the respective standards across reading and writing in 2015 and 2016. In 2015, 92 per cent of pupils who achieved a Level 4 in maths, also did so in reading and writing. In 2016, only 76 per cent of those achieving the expected standard level in maths achieved it in both reading and writing as well.
    

23. This change could be a result of a number of potential factors, for instance:

• There is more variability in pupils’ assessments in the different subjects, and a bigger proportion of this being unrelated to actual ability;

• Schools have – consciously or otherwise – focused on or been more effective in adapting their teaching of some subjects than others in response to the new curriculum; or

• The new set of assessments and curricula inherently reward the performance of ‘all rounders’ less, and children have achieved greater scores in some subjects at the expense of focus on the others more than previously.

24. In our ‘Divergent Pathways’ Report[8], we found that, between 2009 and 2015, very little progress had been made in closing the socio-economic gap between Reception and the end of Key Stage 1. In 2009, the progress gap (that is, how much further disadvantaged children fall behind from one Key Stage to the next) was 2.3 months and had only narrowed to 2.1 months by 2015. This is in contrast to the progress gap between the end of Key Stage 1 and the end of Key Stage 2. Which was 1.8 months in 2009 and 1.2 months in 2015.

25. This suggests that the more rigorous accountability measures in place between Key Stages 1 and 2, is incentivising teachers and school leaders to focus on that upper phase of primary. In order to make similar gains in closing the gap in the earlier years of primary, consideration should be given to whether the progress measure should begin in Reception (rather than at the end of Key Stage 1) and whether Key Stage 1 assessments should be more robust and linked to accountability measures.

26. There is also little evidence that the phonics screening test is leading to improved results in reading by the end of Key Stage 1. Introduced in 2012 (for Year 1 pupils), with the intention of identifying children who need additional help in decoding words when reading, we would expect to see a marked improvement in pupils’ reading scores by the time they reached the end of Key Stage 1. However, since the first cohort of pupils who had the phonics screening took their Key Stage 1 assessments in 2014, we’ve seen little discernible difference to the proportion of children achieving either a Level 2 or 2b in reading[9].

The consequences for primary school accountability

 The Department has introduced new floor standards and a new coasting definition for primary schools, both of which will be determined by assessment data. Schools will be judged as below the floor standard if less than 65 per cent of pupils achieve the expected standard or it fails to achieve sufficient progress scores in all three subjects (at least -5 in reading, -5 in mathematics and -7 in writing).

28. In 2016, a school will be judged as coasting if, in each of 2014 and 2015, it had less than 85 per cent of pupils achieving a Level 4 in reading, writing and maths and was below the national median percentage of pupils making expected progress in all of those subjects; and if, in 2016, less than 85 per cent of pupils achieve the expected standard at the end of primary and average progress made by pupils is below -2.5 in reading or -2.5 in mathematics or -3.5 in writing.

29. Because, in 2016, only 54 per cent of pupils in state funded primary schools met the expected standard at the end of Key Stage 2, a large number of schools are therefore reliant on having met the progress criteria so as not to be deemed as being below the floor. In a recent statement to the House[10], Justine Greening confirmed that no more than 6 per cent of primary schools will be below the floor in 2016. Until primary school performance tables are published in December this year, we will not know which schools these are and how many will fall into the ‘coasting’ definition.

30. But, what we do know is that schools will be increasingly focused on ensuring that all pupils make maximum progress, because the attainment standards are far higher than in previous years.Historically, the attainment element of the floor standard has been set below national average. In 2016 it is above.

31. The use of assessment for both accountability and funding also risks creating perverse incentives. If a pupil scores highly on an EYFSP assessment, the relevant primary school would not attract any additional funding and neither would those positive scores (or the progress made by pupils between reception and the end of Key Stage 1) be reflected in any accountability measures, including school performance tables. And so the current system creates an incentive on primary schools to repress the scores of pupils at either the beginning or end of reception. This carries two significant risks. The first is that it could distort children’s true attainment levels on entry to formal schooling (thereby undermining the investment into universal early years education) and the second is that it means funding intended for pupils who are genuinely behind set benchmarks, risks being diluted and spread to a much larger group of children who do not necessarily require the targeted support for which this money is intended.

Next steps following the most recent reforms to primary assessment 

32. The Department for Education has committed to launching a consultation in the new year, which will look at primary assessment, accountability and the role of teacher assessment. The Education Policy Institute welcomes this.

33. In order to address the issues raised in this submission, the consultation, and subsequent reforms, will need to ensure:

• That a new baseline or early years assessment encourages schools to give a true account of children’s attainment in Reception, while minimising incentives to suppress results either to make progress measures look better or to secure more funding from government.

• Schools have enough notice of any further changes to curriculum or assessment. In January this year, a number of changes were made to Key Stage 2 assessments for summer 2016, leading to widespread criticism from the sector that there simply wasn’t enough time to embed these changes before assessment dates. While the Schools Minister then announced extra time for schools to submit Key Stage 1 and Key Stage 2 assessment data (for 2016 only), the Department must ensure that reforms have adequate lead-in times.

• That it recognises the inconsistencies of the new assessment outcomes and does not instigate intervention procedures based on these measures. The Secretary of State has said that no interventions will take place on the basis of 2016 data alone, but if the same assessment arrangements remain in place for 2017 (which looks likely), then schools will want a similar commitment that these results will also carry less weight when considering intervention options.

[1] https://cmpo.wordpress.com/2015/03/09/gender-teacher-assessment-and-stereotypes/

[2] https://consult.education.gov.uk/funding-policy-unit/schools-national-funding-formula/supporting_documents/Schools_NFF_consultation.pdf

[3] https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/444646/RR409_-_Reception_baseline_research.pdf

[4] https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/514581/Reception_baseline_comparability_study.pdf

[5] As explained in the Department for Education’s 2013 consultation on primary assessment, “In 2012, fewer than half the pupils who had only just reached the current expected standard in both of these subjects went on to achieve 5 A*-C GCSEs at 16, including English and mathematics. In contrast, 7 in 10 of those with a good level 4 in these subjects achieved this GCSE standard”. Department for Education (2013) Primary assessment and accountability under the new national curriculum.

[6] CentreForum (2016) Education in England: Annual Report 2016

[7] https://www.gov.uk/government/statistics/national-curriculum-assessments-key-stage-2-2016-provisional

[8] J.Hutchinson and J. Dunford, @Divergent Pathways’, July 2016

[9] https://www.gov.uk/government/statistics/phonics-screening-check-and-key-stage-1-assessments-england-2016 and https://www.gov.uk/government/statistics/phonics-screening-check-and-key-stage-1-assessments-england-2015

[10] https://hansard.parliament.uk/commons/2016-10-19/debates/16101944000006/PrimaryEducation