Beware the ‘short head’: PISA’s Resilient Students’ Measure

 

This post takes a closer look at the PISA concept of ‘resilient students’ – essentially a measure of disadvantaged high attainment amongst 15 year-olds – and how this varies from country to country.

7211284724_f3c5515bf7_mThe measure was addressed briefly in my recent review of the evidence base for excellence gaps in England but there was not space on that occasion to provide a thoroughgoing review.

The post is organised as follows:

  • A definition of the measure and explanation of how it has changed since the concept was first introduced.
  • A summary of key findings, including selected international comparisons, and of trends over recent PISA cycles.
  • A brief review of OECD and related research material about the characteristics of resilient learners.

I have not provided background about the nature of PISA assessments, but this can be found in previous posts about the mainstream PISA 2012 results and PISA 2012 Problem Solving.

 

Defining the resilient student

In 2011, the OECD published ‘Against the Odds: Disadvantaged students who succeed in school’, which introduced the notion of PISA as a study of resilience. It uses PISA 2006 data throughout and foregrounds science, as did the entire PISA 2006 cycle.

There are two definitions of resilience in play: an international benchmark and a country-specific measure to inform discussion of effective policy levers in different national settings.

The international benchmark relates to the top third of PISA performers (ie above the 67th percentile) across all countries after accounting for socio-economic background. The resilient population comprises students in this group who also fall within the bottom third of the socio-economic background distribution in their particular jurisdiction.

Hence the benchmark comprises an international dimension of performance and a national/jurisdictional dimension of disadvantage.

This cohort is compared with disadvantaged low achievers, a population similarly derived, except that their performance is in the bottom third across all countries, after accounting for socio-economic background.

The national benchmark applies the same national measure relating to socio-economic background, but the measure of performance is the top third of the national/jurisdictional performance distribution for the relevant PISA test.

The basis for determining socio-economic background is the PISA Index of Economic, Social and Cultural Status (ESCS).

‘Against the Odds’ describes it thus:

‘The indicator captures students’ family and home characteristics that describe their socio-economic background. It includes information about parental occupational status and highest educational level, as well as information on home possessions, such as computers, books and access to the Internet.’

Further details are provided in the original PISA 2006 Report (p333).

Rather confusingly, the parameters of the international benchmark were subsequently changed.

PISA 2009 Results: Overcoming Social Background – Equity in Learning Opportunities and Outcomes Volume II describes the new methodology in this fashion:

‘A student is classified as resilient if he or she is in the bottom quarter of the PISA index of economic, social and cultural status (ESCS) in the country of assessment and performs in the top quarter across students from all countries after accounting for socio-economic background. The share of resilient students among all students has been multiplied by 4 so that the percentage values presented here reflect the proportion of resilient students among disadvantaged students (those in the bottom quarter of the PISA index of social, economic and cultural status).’

No reason is given for this shift to a narrower measure of both attainment and disadvantage, nor is the impact on results discussed.

The new methodology is seemingly retained in PISA 2012 Results: Excellence through Equity: Giving every student the chance to succeed – Volume II:

‘A student is class­ed as resilient if he or she is in the bottom quarter of the PISA index of economic, social and cultural status (ESCS) in the country of assessment and performs in the top quarter of students among all countries, after accounting for socio-economic status.’

However, multiplication by four is dispensed with.

This should mean that the outcomes from PISA 2009 and 2012 are broadly comparable with some straightforward multiplication. However the 2006 results foreground science, while in 2009 the focus is reading – and shifts on to maths in 2012.

Although there is some commonality between these different test-specific results (see below), there is also some variation, notably in terms of differential outcomes for boys and girls.

 

PISA 2006 results

The chart reproduced below compares national percentages of resilient students and disadvantaged low achievers in science using the original international benchmark. It shows the proportion of resilient learners amongst disadvantaged students.

 

Resil 2006 science Capture

Conversely, the data table supplied alongside the chart shows the proportion of resilient students amongst all learners. Results have to be multiplied by three on this occasion (since the indicator is based on ‘top third attainment, bottom third advantage’).

I have not reproduced the entire dataset, but have instead created a subset of 14 jurisdictions in which my readership may be particularly interested, namely: Australia, Canada, Finland, Hong Kong, Ireland, Japan, New Zealand, Poland, Shanghai, Singapore, South Korea, Taiwan, the UK and the US. I have also included the OECD average.

I have retained this grouping throughout the analysis, even though some of the jurisdictions do not appear throughout – in particular, Shanghai and Singapore are both omitted from the 2006 data.

Chart 1 shows these results.

 

Resil first chart Chart 1: PISA resilience in science for selected jurisdictions by gender (PISA 2006 data)

 

All the jurisdictions in my sample are relatively strong performers on this measure. Only the United States falls consistently below the OECD average.

Hong Kong has the highest percentage of resilient learners – almost 75% of its disadvantaged students achieve the benchmark. Finland is also a very strong performer, while other jurisdictions achieving over 50% include Canada, Japan, South Korea and Taiwan.

The UK is just above the OECD average, but the US is ten points below. The proportion of disadvantaged resilient students in Hong Kong is almost twice the proportion in the UK and two and a half times the proportion in the US.

Most of the sample shows relatively little variation between their proportions of male and female resilient learners. Females have a slight lead across the OECD as a whole, but males are in the ascendancy in eight of these jurisdictions.

The largest gap – some 13 percentage points in favour of boys – can be found in Hong Kong. The largest advantage in favour of girls – 6.9 percentage points – is evident in Poland. In the UK males are ahead by slightly over three percentage points.

The first chart also shows that there is a relatively strong relationship between the proportion of resilient students and of disadvantaged low achievers. Jurisdictions with the largest proportions of resilient students typically have the smallest proportions of disadvantaged low achievers.

In Hong Kong, the proportion of disadvantaged students who are low achievers is 6.3%, set against an OECD average of 25.8%. Conversely, in the US, this proportion reaches 37.8% – and is 26.7% in the UK. Of this sample, only the US has a bigger proportion of disadvantaged low achievers than of disadvantaged resilient students.

 

‘Against the Odds’ examines the relationship between resiliency in science, reading and maths, but does so using the national benchmark, so the figures are not comparable with those above. I have, however, provided a chart comparing performance in my sample of jurisdictions.

 

Resil second chart

Chart 2: Students resilient in science who are resilient in other subjects, national benchmark of resilience, PISA 2006

 

Amongst the jurisdictions for which we have data there is a relatively similar pattern, with between 47% and 56% of students resilient in all three subjects.

In most cases, students who are resilient in two subjects combine science and maths rather than science and reading, but this is not universally true since the reverse pattern applies in Ireland, Japan and South Korea.

The document summarises the outcomes thus:

‘This evidence indicates that the vast majority of students who are resilient with respect to science are also resilient in at least one if not both of the other domains…These results suggest that resilience in science is not a domain-specific characteristic but rather there is something about these students or the schools they attend that lead them to overcome their social disadvantage and excel at school in multiple subject domains.’

 

PISA 2009 Results

The results drawn from PISA 2009 focus on outcomes in reading, rather than science, and of course the definitional differences described above make them incompatible with those for 2006.

The first graph reproduced below shows the outcomes for the full set of participating jurisdictions, while the second – Chart 2 – provides the results for my sample.

Resil PISA 2009 Capture

 

Resil third chart

Chart 3: PISA resilience in reading for selected jurisdictions by gender (PISA 2009 data)

 

The overall OECD average is pitched at 30.8% compared with 39% on the PISA 2006 science measure. Ten of our sample fall above the OECD average and Australia matches it, but the UK, Ireland and the US are below the average, the UK undershooting it by some seven percentage points.

The strongest performer is Shanghai at 75.6%, closely followed by Hong Kong at 72.4%. They and South Korea are the only jurisdictions in the sample which can count over half their disadvantaged readers as resilient. Singapore, Finland and Japan are also relatively strong performers.

There are pronounced gender differences in favour of girls. They have a 16.8 percentage point lead over boys in the OECD average figure and they outscore boys in every country in our sample. These differentials are most marked in Finland, Poland and New Zealand. In the UK there is a difference of 9.2 percentage points, smaller than in many other countries in the sample.

The comparison with the proportion of disadvantaged low achievers is illustrated by chart 3. This reveals the huge variation in the performance of our sample.

 

Resil fourth chart

Chart 4: Comparing percentage of resilient and low-achieving students in reading, PISA 2009

At one extreme, the proportion of disadvantaged low achievers (bottom quartile of the achievement distribution) is virtually negligible in Shanghai and Hong Kong, while around three-quarters of disadvantaged students are resilient (top quartile of the achievement distribution).

At the other, countries like the UK have broadly similar proportions of low achievers and resilient students. The chart reinforces just how far behind they are at both the top and the bottom of the attainment spectrum.

 

PISA 2012 Results

In 2012 the focus is maths rather than reading. The graph reproduced below compares resilience scores across the full set of participating jurisdictions, while Chart 4 covers only my smaller sample.

 

Resil PISA 2012 Capture

resil fifth chart Chart 5: PISA resilience in maths for selected jurisdictions by gender (PISA 2012 data)

 

Despite the change in subject, the span of performance on this measure is broadly similar to that found in reading three years earlier. The OECD average is 25.6%, roughly five percentage points lower than the average in 2009 reading.

Nine of the sample lie above the OECD average, while Australia, Ireland, New Zealand, UK and the US are below. The UK is closer to the OECD average in maths than it was in reading, however, and is a relatively stronger performer than the US and New Zealand.

Shanghai and Hong Kong are once again the top performers, at 76.8% and 72.4% respectively. Singapore is at just over 60% and South Korea at just over 50%. Taiwan and Japan are also notably strong performers.

Within the OECD average, boys have a four percentage point lead on girls, but boys’ relatively stronger performance is not universal – in Hong Kong, Poland, Singapore and South Korea, girls are in the ascendancy.  This is most strongly seen in Poland. The percentage point difference in the UK is just 2.

The comparison with disadvantage low achievers is illustrated in Chart 5.

 

Resil sixth chart

Chart 6: Comparing percentage of resilient and low-achieving students in maths, PISA 2012

 

Once again the familiar pattern emerges, with negligible proportions of low achievers in the countries with the largest shares of resilient students. At the other extreme, the US and New Zealand are the only two jurisdictions in this sample with a longer ‘tail’ of low achievers. The reverse is true in the UK, but only just!

 

Another OECD Publication ‘Strengthening Resilience through Education: PISA Results – background document’ contains a graph showing the variance in jurisdictions’ mathematical performance by deciles of socio-economic disadvantage. This is reproduced below.

 

resil maths deciles Capture

The text adds:

‘Further analysis indicates that the 10% socio-economically most disadvantaged children in Shanghai perform at the same level as the 10% most privileged children in the United States; and that the 20% most disadvantaged children in Finland, Japan, Estonia, Korea, Singapore, Hong Kong-China and Shanghai-China compare favourably to the OECD average.’

One can see that the UK is decidedly ‘mid-table’ at both extremes of the distribution. On the evidence of this measure, one cannot fully accept the oft-repeated saw that the UK is a much stronger performer with high attainers than with low attainers, certainly as far as disadvantaged learners are concerned.

 

The 2012 Report also compares maths-based resiliency records over the four cycles from PISA 2003 to PISA 2012 – as shown in the graph reproduced below – but few of the changes are statistically significant. There has also been some statistical sleight of hand to ensure comparability across the cycles.

 

resil comparing PISA 2003 to 2012 capture

Amongst the outcomes that are statistically significant, Australia experienced a fall of 1.9 percentage points, Canada 1.6 percentage points, Finland 3.3 percentage points and New Zealand 2.9 percentage points. The OECD average was relatively little changed.

The UK is not included in this analysis because of issues with its PISA 2003 results.

Resilience is not addressed in the main PISA 2012 report on problem-solving, but one can find online the graph below, which shows the relative performance of the participating countries.

It is no surprise that the Asian Tigers are at the top of the league (although Shanghai is no longer in the ascendancy). England (as opposed to the UK) is at just over 30%, a little above the OECD average, which appears to stand at around 27%.

The United States and Australia perform at a very similar level. Canada is ahead of them and Poland is the laggard.

 

resil problem solving 2012 Capture

 

Resilience in the home countries

Inserted for the purposes of reinforcement, the chart below compiles the UK outcomes from the PISA 2006, 2009 and 2012 studies above, as compared with the top performer in my sample for each cycle and the appropriate OECD average. Problem-solving is omitted.

Only in science (using the ‘top third attainer, bottom third disadvantage’ formula) does the UK exceed the OECD average figure and then only slightly.

In both reading and maths, the gap between the UK and the top performer in my sample is eye-wateringly large: in each case there are more than three times as many resilient students in the top-performing jurisdiction.

It is abundantly clear from this data that disadvantaged high attainers in the UK do not perform strongly compared with their peers elsewhere.

 

Resil seventh chart

Chart 7: Resilience measures from PISA 2006-2012 comparing UK with top performer in this sample and OECD average

 

Unfortunately NFER does not pick up the concept of resilience in its analysis of England’s PISA 2012 results.

The only comparative analysis across the Home Countries that I can find is contained in a report prepared for the Northern Ireland Ministry of Education by NFER called ‘PISA 2009: Modelling achievement and resilience in Northern Ireland’ (March 2012).

This uses the old ‘highest third by attainment, lowest third by disadvantage’ methodology deployed in ‘Against the Odds’. Reading is the base.

The results show that 41% of English students are resilient, the same figure as for the UK as a whole. The figures for the other home countries appear to be: Northern Ireland 42%; Scotland 44%; and Wales 35%.

Whether the same relationship holds true in maths and science using the ‘top quartile, bottom quartile’ methodology is unknown. One suspects though that each of the UK figures given above will also apply to England.

 

The characteristics of resilient learners

‘Against the Odds’ outlines some evidence derived from comparisons using the national benchmark:

  • Resilient students are, on average, somewhat more advantaged than disadvantaged low achievers, but the difference is relatively small and mostly accounted for by home-related factors (eg. number of books in the home, parental level of education) rather than parental occupation and income.
  • In most jurisdictions, resilient students achieve proficiency level 4 or higher in science. This is true of 56.8% across the OECD. In the UK the figure is 75.8%; in Hong Kong it is 88.4%. We do not know what proportions achieve the highest proficiency levels.
  • Students with an immigrant background – either born outside the country of residence or with parents were born outside the country – tend to be under-represented amongst resilient students.
  • Resilient students tend to be more motivated, confident and engaged than disadvantaged low achievers. Students’ confidence in their academic abilities is a strong predictor of resilience, stronger than motivation.
  • Learning time – the amount of time spent in normal science lessons – is also a strong predictor of resilience, but there is relatively little evidence of an association with school factors such as school management, admissions policies and competition.

Volume III of the PISA 2012 Report: ‘Ready to Learn: Students’ engagement, drive and self-beliefs’ offers a further gloss on these characteristics from a mathematical perspective:

‘Resilient students and advantaged high-achievers have lower rates of absenteeism and lack of punctuality than disadvantaged and advantaged low-achievers…

….resilient and disadvantaged low-achievers tend to have lower sense of belonging than advantaged low-achievers and advantaged high-achievers: socio-economically disadvantaged students express a lower sense of belonging than socio-economically advantaged students irrespective of their performance in mathematics.

Resilient students tend to resemble advantaged high-achievers with respect to their level of drive, motivation and self-beliefs: resilient students and advantaged high-achievers have in fact much higher levels of perseverance, intrinsic and instrumental motivation to learn mathematics, mathematics self-efficacy, mathematics self-concept and lower levels of mathematics anxiety than students who perform at lower levels than would be expected of them given their socio-economic condition…

….In fact, one key characteristic that resilient students tend to share across participating countries and economies, is that they are generally physically and mentally present in class, are ready to persevere when faced with challenges and difficulties and believe in their abilities as mathematics learners.’

Several research studies can be found online that reinforce these findings, sometimes adding a few further details for good measure:

The aforementioned NFER study for Northern Ireland uses a multi-level logistic model to investigate the school and student background factors associated with resilience in Northern Ireland using PISA 2009 data.

It derives odds ratios as follows: grammar school 7.44; female pupils 2.00; possessions – classic literature 1.69; wealth 0.76; percentage of pupils eligible for FSM – 0.63; and books in home – 0-10 books 0.35.

On the positive impact of selection the report observes:

‘This is likely to be largely caused by the fact that to some extent grammar schools will be identifying the most resilient students as part of the selection process. As such, we cannot be certain about the effectiveness or otherwise of grammar schools in providing the best education for disadvantaged children.’

Another study – ‘Predicting academic resilience with mathematics learning and demographic variables’ (Cheung et al 2014) – concludes that, amongst East Asian jurisdictions such as Hong-Kong, Japan and South Korea, resilience is associated with avoidance of ‘redoublement’ and having attended kindergarten for more than a year.

Unsurprisingly, students who are more familiar with mathematical concepts and have greater mathematical self-efficacy are also more likely to be resilient.

Amongst other countries in the sample – including Canada and Finland – being male, native (as opposed to immigrant) and avoiding ‘redoublement’ produced stronger chances of resilience.

In addition to familiarity with maths concepts and self-efficacy, resilient students in these countries were less anxious about maths and had a higher degree of maths self-concept.

Work on ‘Resilience Patterns in Public Schools in Turkey’ (unattributed and undated) – based on PISA 2009 data and using the ‘top third, bottom third’ methodology – finds that 10% of a Turkish sample are resilient in reading, maths and science; 6% are resilient in two subjects and a further 8% in one only.

Resilience varies in different subjects according to year of education.

resil Turkey Capture

There are also significant regional differences.

Odds ratios show a positive association with: more than one year of pre-primary education; selective provision, especially in maths; absence of ability grouping; additional learning time, especially for maths and science; a good disciplinary climate and strong teacher-student relations.

An Italian study – ‘A way to resilience: How can Italian disadvantaged students and schools close the achievement gap?’ (Agasisti and Longobardi, undated) uses PISA 2009 data to examine the characteristics of resilient students attending schools with high levels of disadvantage.

This confirms some of the findings above in respect of student characteristics, finding a negative impact from immigrant status (and also from a high proportion of immigrants in a school). ‘Joy in reading’ and ‘positive attitude to computers’ are both positively associated with resilience, as is a positive relationship with teachers.

School type is found to influence the incidence of resilience – particularly enrolment in Licei as opposed to professional or technical schools – so reflecting one outcome of the Northern Irish study. Other significant school level factors include the quality of educational resources available and investment in extracurricular activities. Regional differences are once more pronounced.

A second Italian study – ‘Does public spending improve educational resilience? A longitudinal analysis of OECD PISA data’ (Agasisti et al 2014) finds a positive correlation between the proportion of a country’s public expenditure devoted to education and the proportion of resilient students.

Finally, this commentary from Marc Tucker in the US links its relatively low incidence of resilient students to national views about the nature of ability:

‘In Asia, differences in student achievement are generally attributed to differences in the effort that students put into learning, whereas in the United States, these differences are attributed to natural ability.  This leads to much lower expectations for students who come from low-income families…

My experience of the Europeans is that they lie somewhere between the Asians and the Americans with respect to the question as to whether effort or genetic material is the most important explainer of achievement in school…

… My take is that American students still suffer relative to students in both Europe and Asia as a result of the propensity of the American education system to sort students out by ability and assign different students work at different challenge levels, based on their estimates of student’s inherited intelligence.’

 

Conclusion

What are we to make of all this?

It suggests to me that we have not pushed much beyond statements of the obvious and vague conjecture in our efforts to understand the resilient student population and how to increase its size in any given jurisdiction.

The comparative statistical evidence shows that England has a real problem with underachievement by disadvantaged students, as much at the top as the bottom of the attainment distribution.

We are not alone in facing this difficulty, although it is significantly more pronounced than in several of our most prominent PISA competitors.

We should be worrying as much about our ‘short head’ as our ‘long tail’.

 

GP

September 2014

 

 

 

 

 

 

Advertisements

Closing England’s Excellence Gaps: Part 2

This is the second part of an extended post considering what we know – and do not know – about high attainment gaps between learners from advantaged and disadvantaged backgrounds in England.

512px-Bakerloo_line_-_Waterloo_-_Mind_the_gap

Mind the Gap by Clicsouris

Part one provided an England-specific definition, articulated a provisional theoretical model for addressing excellence gaps and set out the published data about the size of excellence gaps at Key Stages 2,4 and 5, respectively.

Part two continues to review the evidence base for excellence gaps, covering the question whether high attainers remain so, international comparisons data and related research and excellence gaps analysis from the USA.

It also describes those elements of present government policy that impact directly on excellence gaps and offers some recommendations for strengthening our national emphasis on this important issue.

 

Whether disadvantaged high achievers remain so

 

The Characteristics of High Attainers

The Characteristics of high attainers (DfES 2007) includes investigation of:

  • whether pupils in the top 10% at KS4 in 2006 were also high attainers at KS3 in 2004 and KS2 in 2001, by matching back to their fine grade points scores; and
  • chances of being a KS4 high attainer given a range of pupil characteristics at KS2 and KS3.

On the first point it finds that 4% of all pupils remain in the top 10% throughout, while 83% of pupils are never in the top 10% group.

Some 63% of those who were high attainers at the end of KS2 are still high attainers at the end of KS3, while 72% of KS3 high attainers are still in that group at the end of KS4. Approximately half of high attainers at KS2 are high attainers at KS4.

The calculation is not repeated for advantaged and disadvantaged high attainers respectively, but this shows that – while there is relatively little movement between  the high attaining population and other learners (with only 17% of the overall population falling within scope at any point) – there is a sizeable ‘drop out’ amongst high attainers at each key stage.

Turning to the second point, logistic regression is used to calculate the odds of being a KS4 high attainer given different levels of prior attainment and a range of pupil characteristics. Results are controlled to isolate the impact of individual characteristics and for attainment.

The study finds that pupils with a KS2 average points score (APS) above 33 are more likely than not to be high attainers at KS4, and this probability increases as their KS2 APS increases. For those with an APS of 36, the odds are 23.73, meaning they have a 24/25 chance of being a KS4 high attainer.

For FSM-eligible learners though, the odds are 0.55, meaning that the chances of being a KS4 high attainer are 45% lower amongst FSM-eligible pupils, compared to  their non-FSM counterparts with similar prior attainment and characteristics.

The full set of findings for individual characteristics is reproduced below.

Ex gap Capture 7

 

An appendix supplies the exact ratios for each characteristic and the text points out that these can be multiplied to calculate odds ratios for different combinations:

The odds for different prior attainment levels and other characteristics combined with FSM eligibility are not worked through, but could easily be calculated. It would be extremely worthwhile to repeat this analysis using more recent data to see whether the results would be replicated for those completing KS4 in 2014.

 

Sutton Trust

In 2008, the Sutton Trust published ‘Wasted talent? Attrition rates of high achieving pupils between school and university’ which examines the attrition rates for FSM-eligible learners among the top 20% of performers at KS2, KS3 and KS4.

A footnote says that this calculation was ‘on the basis of their English and maths scores at age 11, and at later stages of schooling’, which is somewhat unclear. A single, unidentified cohort is tracked across key stages.

The report suggests ‘extremely high rates of ‘leakage’ amongst the least privileged pupils’. The key finding is that two-thirds of disadvantaged top performers at KS2 are not amongst the top performers at KS4, whereas 42% advantaged top performers are not.

 

EPPSE

Also in the longitudinal tradition ‘Performing against the odds: developmental trajectories of children in the EPPSE 3-16 study’ (Siraj-Blatchford et al, June 2011) investigated through interviews the factors that enabled a small group of disadvantaged learners to ‘succeed against the odds’.

Twenty learners were identified who were at the end of KS3 or at KS4 and who had achieved well above predicted levels in English and maths at the end of KS2. Achievement was predicted for the full sample of 2,800 children within the EPPSE study via multi-level modelling, generating:

‘…residual scores for each individual child, indicating the differences between predicted and attained achievement at age 11, while controlling for certain child characteristics (i.e., age, gender, birth weight, and the presence of developmental problems) and family characteristics (i.e., mothers’ education, fathers’ education, socio-economic status [SES] and family income). ‘

The 20 identified as succeeding against the odds had KS2 residual scores for both English and maths within the highest 20% of the sample. ‘Development trajectories’ were created for the group using a range of assessments conducted at age 3, 4, 5, 7, 11 and 14.

The highest job level held in the family when the children were aged 3-4 was manual, semi-skilled or unskilled, or the parent(s) had never worked.

The 20 were randomly selected from each gender – eight boys and 12 girls – while ensuring representation of ‘the bigger minority ethnic groups’. It included nine students characterised as White UK, five Black Caribbean, two Black African and one each of Indian (Sikh), Pakistani, Mixed Heritage and Indian (Hindu).

Interviews were conducted with children, parents and the teacher at their [present] secondary school the learners felt ‘knew them best’. Teacher interviews were secured for 11 of the 20.

Comparison of development trajectories showed significant gaps between this ‘low SES high attainment’ group and a comparative sample of ‘low SES, predicted attainment’ students. They were ahead from the outset and pulled further away.

They also exceeded a comparator group of high SES learners performing at predicted levels from entry to primary education until KS2. Even at KS3, 16 of the 20 were still performing above the mean of the high SES sample.

These profiles – illustrated in the two charts below – were very similar in English and maths respectively. In either case, Group 1 are those with ‘low SES, high attainment’, while Group 4 are ‘high SES predicted attainment’ students.

 

Supp exgap Eng Capture

Supp exgap Maths Capture

 

Interviews identified five factors that helped to explain this success:

  • The child’s perceived cognitive ability, strong motivation for school and learning and their hobbies and interests. Most parents and children regarded cognitive ability as ‘inherent to the child’, but they had experienced many opportunities to develop their abilities and received support in developing a ‘positive self-image’. Parenting ‘reflected a belief in the parent’s efficacy to positively influence the child’s learning’. Children also demonstrated ability to self-regulate and positive attitudes to homework. They had a positive attitude to learning and made frequent use of books and computers for this purpose. They used school and learning as distractions from wider family problems. Many were driven to learn, to succeed educationally and achieve future aspirations.
  • Home context – effective practical and emotional support with school and learning. Families undertook a wide range of learning activities, especially in the early years. These were perceived as enjoyable but also valuable preparation for subsequent schooling. During the primary years, almost all families actively stimulated their children to read. In the secondary years, many parents felt their efforts to regulate their children’s activities and set boundaries were significant. Parents also provided practical support with school and learning, taking an active interest and interacting with their child’s school. Their parenting style is described as ‘authoritative: warm, firm and accepting of their needs for psychological autonomy but demanding’. They set clear standards and boundaries for behaviour while granting extra autonomy as their children matured. They set high expectations and felt strongly responsible for their child’s education and attitude to learning. They believed in their capacity to influence their children positively. Some were motivated by the educational difficulties they had experienced.
  • (Pre-)School environment – teachers who are sensitive and responsive to the child’s needs and use ‘an authoritative approach to teaching and interactive teaching strategies’; and, additionally, supportive school policies. Parents had a positive perception of the value of pre-school education, though the value of highly effective pre-school provision was not clear cut with this sample. Moreover ‘very few clear patterns of association could be discerned between primary school effectiveness and development of rankings on trajectories’. That said both parents and children recognised that their schools had helped them address learning and behavioural difficulties. Success was attributed to the quality of teachers. ‘They thought that good quality teaching meant that teachers were able to explain things clearly, were enthusiastic about the subject they taught, were approachable when things were difficult to understand, were generally friendly, had control over the class and clearly communicated their expectations and boundaries.’
  • Peers providing practical, emotional and motivational support. Friends were especially valuable in helping them to respond to difficulties, helping in class, with homework and revision. Such support was often mutual, helping to build understanding and develop self-esteem, as a consequence of undertaking the role of teacher. Friends also provided role models and competitors.
  • Similar support provided by the extended family and wider social, cultural and religious communities. Parents encouraged their children to take part in extra-curricular activities and were often aware of their educational benefits. Family networks often provided additional learning experiences, particularly for Caribbean and some Asian families.

 

Ofsted

Ofsted’s The most able students: Are they doing as well as they should in our non-selective secondary schools? (2013) defines this population rather convolutedly as those:

‘…starting secondary school in Year 7 attaining level 5 or above, or having the potential to attain Level 5 and above, in English (reading and writing) and or mathematics at the end of Key Stage 2.’ (Footnote p6-7)

There is relatively little data in the report about the performance of high-attaining disadvantaged learners, other than the statement that only 58% of FSM students within the ‘most able’ population in KS2 and attending non-selective secondary schools go on to achieve A*-B GCSE grades in English and maths, compared with 75% of non-FSM pupils, giving a gap of 17 percentage points.

I have been unable to find national transition matrices for advantaged and disadvantaged learners, enabling us to compare the proportion of advantaged and disadvantaged pupils making and exceeding the expected progress between key stages.

 

Regression to the mean and efforts to circumvent it

Much prominence has been given to Feinstein’s 2003 finding that, whereas high-scoring children from advantaged and disadvantaged backgrounds (defined by parental occupation) perform at a broadly similar level when tested at 22 months, the disadvantaged group are subsequently overtaken by relatively low-scoring children from advantaged backgrounds during the primary school years.

The diagram that summarises this relationship has been reproduced widely and much used as the centrepiece of arguments justifying efforts to improve social mobility.

Feinstein Capture

But Feinstein’s finding were subsequently challenged on methodological grounds associated with the effects of regression to the mean.

Jerrim and Vignoles (2011) concluded:

‘There is currently an overwhelming view amongst academics and policymakers that highly able children from poor homes get overtaken by their affluent (but less able) peers before the end of primary school. Although this empirical finding is treated as a stylised fact, the methodology used to reach this conclusion is seriously flawed. After attempting to correct for the aforementioned statistical problem, we find little evidence that this is actually the case. Hence we strongly recommend that any future work on high ability–disadvantaged groups takes the problem of regression to the mean fully into account.’

On the other hand, Whitty and Anders comment:

‘Although some doubt has been raised regarding this analysis on account of the potential for regression to the mean to exaggerate the phenomenon (Jerrim and Vignoles, 2011), it is highly unlikely that this would overturn the core finding that high SES, lower ability children catch up with their low-SES, higher-ability peers.’

Their point is borne out by Progress made by high-attaining children from disadvantaged backgrounds (June 2014) suggesting that Vignoles, as part of the writing team, has changed her mind somewhat since 2011.

This research adopts a methodological route to minimise the impact of regression to the mean. This involves assigning learners to achievement groups using a different test to those used to follow their attainment trajectories and focusing principally on those trajectories from KS2 onwards.

The high attaining group is defined as those achieving Level 3 or above in KS1 writing, which selected in 12.6% of the sample. (For comparison, the same calculations are undertaken based on achieving L3 or above in KS1 maths.) These pupils are ranked and assigned a percentile on the basis of their performance on the remaining KS1 tests and at each subsequent key stage.

The chart summarising the outcomes in the period from KS1 to KS4 is reproduced below, showing the different trajectories of the ‘most deprived’ and ‘least deprived’. These are upper and lower quintile groups of state school students derived on the basis of FSM eligibility and a set of area-based measures of disadvantage and measures of socio-economic status derived from the census.

 

Ex gap 8 Capture

The trajectories do not alter significantly beyond KS4.

The study concludes:

‘…children from poorer backgrounds who are high attaining at age 7 are more likely to fall off a high attainment trajectory than children from richer backgrounds. We find that high-achieving children from the most deprived families perform worse than lower-achieving students from the least deprived families by Key Stage 4. Conversely, lower-achieving affluent children catch up with higher-achieving deprived children between Key Stage 2 and Key Stage 4.’

Hence:

‘The period between Key Stage 2 and Key Stage 4 appears to be a crucial time to ensure that higher-achieving pupils from poor backgrounds remain on a high achievement trajectory.’

In short, a Feinstein-like relationship is established but it operates at a somewhat later stage in the educational process.

 

International comparisons studies

 

PISA: Resilience

OECD PISA studies have recently begun to report on the performance of what they call ‘resilient’ learners.

Against the Odds: Disadvantaged Students Who Succeed in Schools (OECD, 2011) describes this population as those who fall within the bottom third of their country’s distribution by socio-economic background, but who achieve within the top third on PISA assessments across participating countries.

This publication uses PISA 2006 science results as the basis of its calculations. The relative position of different countries is shown in the chart reproduced below. Hong Kong tops the league at 24.8%, the UK is at 13.5%, slightly above the OECD average of 13%, while the USA is languishing at 9.9%.

Ex Gap Capture 9

The findings were discussed further in PISA in Focus 5 (OECD 2011), where PISA 2009 data is used to make the calculation. The methodology is also significantly adjusted so that includes a substantially smaller population:

‘A student is classified as resilient if he or she is in the bottom quarter of the PISA index of economic, social and cultural status (ESCS) in the country of assessment and performs in the top quarter across students from all countries after accounting for socio-economic background. The share of resilient students among all students has been multiplied by 4 so that the percentage values presented here reflect the proportion of resilient students among disadvantaged students (those in the bottom quarter of the PISA index of social, economic and cultural status.’

According to this measure, the UK is at 24% and the US has leapfrogged them at 28%. Both are below the OECD average of 31%, while Shanghai and Hong Kong stand at over 70%.

The Report on PISA 2012 (OECD 2013) retains the more demanding definition of resilience, but dispenses with multiplication by 4, so these results need to be so multiplied to be comparable with those for 2009.

This time round, Shanghai is at 19.2% (76.8%) and Hong Kong at 18.1% (72.4%). The OECD average is 6.4% (25.6%), the UK at 5.8% (23.2%) and the US at 5.2% (20.8%).

So the UK has lost a little ground compared with 2009, but is much close to the OECD average and has overtaken the US, which has fallen back by some seven percentage points.

I could find no commentary on these changes.

NFER has undertaken some work on resilience in Northern Ireland, using PISA 2009 reading results (and the original ‘one third’ methodology) as a base. This includes odds ratios for different characteristics of being resilient. This could be replicated for England using PISA 2012 data and the latest definition of resilience.

 

Research on socio-economic gradients

The Socio-Economic Gradient in Teenagers’ Reading Skills: How Does England Compare with Other Countries? (Jerrim 2012) compares the performance of students within the highest and lowest quintiles of the ISEI Index of Occupational Status on the PISA 2009 reading tests.

It quantifies the proportion of these two populations within each decile of  achievement, so generating a gradient, before reviewing how this gradient has changed between PISA 2000 and PISA 2009, comparing outcomes for England, Australia, Canada, Finland, Germany and the US.

Jerrim summarises his findings thus:

‘The difference between advantaged and disadvantaged children’s PISA 2009 reading test scores in England is similar (on average) to that in most other developed countries (including Australia, Germany and, to some extent, the US). This is in contrast to previous studies from the 1990s, which suggested that there was a particularly large socio-economic gap in English pupils’ academic achievement.

Yet the association between family background and high achievement seems to be stronger in England than elsewhere.

There is some evidence that the socio-economic achievement gradient has been reduced in England over the last decade, although not amongst the most able pupils from advantaged and disadvantaged homes.’

Jerrim finds that the link in England between family background and high achievement is stronger than in most other OECD countries, whereas this is not the case at the other end of the distribution.

He hypothises that this might be attributable to recent policy focus on reducing the ‘long tail’ while:

‘much less attention seems to be paid to helping disadvantaged children who are already doing reasonably well to push on and reach the top grades’.

He dismisses the notion that the difference is associated with the fact that  disadvantaged children are concentrated in lower-performing schools, since it persists even when controls for school effects are introduced.

In considering why PISA scores show the achievement gap in reading has reduced between 2000 and 2009 at the lower end of the attainment distribution but not at the top, he cites two possibilities: that Government policy has been disproportionately successful at the lower end; and that there has been a more substantial decline in achievement amongst learners from advantaged backgrounds than amongst their disadvantaged peers. He is unable to rule out the latter possibility.

He also notes in passing that PISA scores in maths do not generate the same pattern.

These arguments are further developed in ‘The Reading Gap: The socio-economic gap in children’s reading skills: A cross-national comparison using PISA 2009’ (Jerrim, 2013) which applies the same methodology.

This finds that high-achieving (top decile of the test distribution) boys from the most advantaged quintile in England are two years and seven months ahead of high-achieving boys from the most disadvantaged quintile, while the comparable gap for girls is slightly lower, at two years and four months.

The chart reproduced below illustrates international comparisons for boys. It shows that only Scotland has a larger high achievement gap than England. (The black lines indicate 99% confidence intervals – he associates the uncertainty to ‘sampling variation’.)

Gaps in countries at the bottom of the table are approximately half the size of those in England and Scotland.

Ex gap 10 capture

 

One of the report’s recommendations is that:

‘The coalition government has demonstrated its commitment to disadvantaged pupils by establishing the Education Endowment Foundation… A key part of this Foundation’s future work should be to ensure highly able children from disadvantaged backgrounds succeed in school and have the opportunity to enter top universities and professional jobs. The government should provide additional resources to the foundation to trial interventions that specifically target already high achieving children from disadvantaged homes. These should be evaluated using robust evaluation methodologies (e.g. randomised control trials) so that policymakers develop a better understanding of what schemes really have the potential to work.’

The study is published by the Sutton Trust whose Chairman – Sir Peter Lampl – is also chairman of the EEF.

In ‘Family background and access to high ‘status’ universities’ (2013) Jerrim provides a different chart showing estimates by country of disadvantaged high achieving learners. The measure of achievement is PISA Level 5 in reading and the measure of disadvantage remains quintiles derived from the ISEI index.

Ex Gap 12 Capture 

The underlying figures are not supplied.

Also in 2013, in ‘The mathematical skills of school children: how does England compare to the high-performing East Asian jurisdictions?’ Jerrim and Choi construct a similar gradient for maths, drawing on a mix of PISA and TIMSS assessments conducted between 2003 and 2009, so enabling them to consider variation according to the age at which assessment takes place.

The international tests selected are TIMSS 2003, 4th grade; TIMSS 2007, 8th grade and PISA 2009. The differences between what these tests measure are described as ‘slight’. The analysis of achievement relies on deciles of the achievement distribution.

Thirteen comparator countries are included, including six wealthy western economies, three ‘middle income’ western economies and four Asian Tigers (Hong Kong, Japan, Singapore and Taiwan).

This study applies as the best available proxy for socio-economic status the number of books in the family home, comparing the most advantaged (over 200 books) with the least (under 25 books). It acknowledges the limitations of this proxy, which Jerrim discusses elsewhere.

The evidence suggests that:

‘between primary school and the end of secondary school, the gap between the lowest achieving children in England and the lowest achieving children in East Asian countries is reduced’

but remains significant.

Conversely, results for the top 10% of the distribution:

‘suggest that the gap between the highest achieving children in England and the highest achieving children in East Asia increases between the end of primary school and the end of secondary school’.

The latter outcome is illustrated in the chart reproduced below

Ex gap 11 Capture

 

The authors do not consider variation by socio-economic background amongst the high-achieving cohort, presumably because the data still does not support the pattern they previously identified for reading.

 

US studies

In 2007 the Jack Kent Cooke Foundation published ‘Achievement Trap: How America is Failing Millions of High-Achieving Students from Low Income Backgrounds’ (Wyner, Bridgeland, Diiulio) The text was subsequently revised in 2009.

This focuses exclusively on gaps attributable to socio-economic status, by comparing the performance of those in the top and bottom halves of the family income distribution in the US, as adjusted for family size.

The achievement measure is top quartile performance on nationally normalised exams administered within two longitudinal studies: The National Education Longitudinal Study (NELS) and the Baccalaureate and Beyond Longitudinal Study (B&B).

The study reports that relatively few lower income students remain high achievers throughout their time in elementary and high school:

  • 56% remain high achievers in reading by Grade 5, compared with 69% of higher income students.
  • 25 percent fall out of the high achiever cohort in high school, compared with 16% of higher income students.
  • Higher income learners who are not high achievers in Grade 1 are more than twice as likely to be high achievers by Grade 5. The same is true between Grades 8 and 12.

2007 also saw the publication of ‘Overlooked Gems: A national perspective on low income promising learners’ (Van Tassel-Baska and Stambaugh). This is a compilation of the proceedings of a 2006 conference which does not attempt a single definition of the target group, but draws on a variety of different research studies and programmes, each with different starting points.

An influential 2009 McKinsey study ‘The Economic Impact of the Achievement Gap in America’s Schools’ acknowledges the existence of what it calls a ‘top gap’. They use this term with reference to:

  • the number of top performers and the level of top performance in the US compared with other countries and
  • the gap in the US between the proportion of Black/Latino students and the proportion of all students achieving top levels of performance.

The authors discuss the colossal economic costs of achievement gaps more generally, but fail to extend this analysis to the ‘top gap’ specifically.

In 2010 ‘Mind the Other Gap: The Growing Excellence Gap in K-12 Education’ (Plucker, Burroughs and Song) was published – and seems to have been the first study to use this term.

The authors define such gaps straightforwardly as

‘Differences between subgroups of students performing at the highest levels of achievement’

The measures of high achievement deployed are the advanced standards on US NAEP maths and reading tests, at Grades 4 and 8 respectively.

The study identifies gaps based on four sets of learner characteristics:

  • Socio-economic status (eligible or not for free or reduced price lunch).
  • Ethnic background (White versus Black and/or Hispanic).
  • English language proficiency (what we in England would call EAL, compared with non-EAL).
  • Gender (girls versus boys).

Each characteristic is dealt with in isolation, so there is no discussion of the gaps between – for example – disadvantaged Black/Hispanic and disadvantaged White boys.

In relation to socio-economic achievement gaps, Plucker et al find that:

  • In Grade 4 maths, from 1996 to 2007, the proportion of advantaged learners achieving the advanced level increased by 5.6 percentage points, while the proportion of disadvantaged learners doing so increased by 1.2 percentage points. In Grade 8 maths, these percentage point changes were 5.7 and 0.8 percentage points respectively. Allowing for changes in the size of the advantaged and disadvantaged cohorts, excellence gaps are estimated to have widened by 4.1 percentage points in Grade 4 (to 7.3%) and 4.9 percentage points in Grade 8 (to 8.2%).
  • In Grade 4 reading, from 1998 to 2007, the proportion of advantaged learners achieving the advanced level increased by 1.2 percentage points, while the proportion of disadvantaged students doing so increased by 0.8 percentage points. In Grade 8 reading, these percentage point changes were almost negligible for both groups. The Grade 4 excellence gap is estimated to have increased slightly, by 0.4 percentage points (to 9.4%) whereas Grade 8 gaps have increased minimally by 0.2 percentage points (to 3.1%).

They observe that the size of excellence gaps are, at best, only moderately correlated with those at lower levels of achievement.

There is a weak relationship between gaps at basic and advanced level – indeed ‘smaller achievement gaps among minimally competent students is related to larger gaps among advanced students’ – but there is some inter-relationship between those at proficient and advanced level.

They conclude that, whereas No Child Left Behind (NCLB) helped to narrow achievement gaps, this does not extend to high achievers.

There is no substantive evidence that the NCLB focus on lower achievers has increased the excellence gap, although the majority of states surveyed by the NAGC felt that NCLB had diverted attention and resource away from gifted education.

In 2011 ‘Do High Fliers Maintain their Altitude?’ (Xiang et al 2011) provides a US analysis of whether individual students remain high achievers throughout their school careers.

They do not report outcomes for disadvantaged high achievers, but do consider briefly those attending schools with high and low proportions respectively of students eligible for free and reduced price lunches.

For this section of the report, high achievement is defined as ‘those whose math or reading scores placed them within the top ten per cent of their individual grades and schools’. Learners were tracked from Grades 3 to 5 and Grades 6 to 8.

It is described as exploratory, because the sample was not representative.

However:

‘High-achieving students attending high-poverty schools made about the same amount of academic growth over time as their high-achieving peers in low-poverty schools…It appears that the relationship between a school’s poverty rate and the growth of its highest-achieving students is weak. In other words, attending a low-poverty school adds little to the average high achiever’s prospects for growth.’

The wider study was criticised in a review by the NEPC, in part on the grounds that the results may have been distorted by regression to the mean, a shortcoming only briefly discussed in an appendix..

The following year saw the publication of Unlocking Emergent Talent: Supporting High Achievement of Low-Income, High-Ability Students (Olszewski-Kubilius and Clarenbach, 2012).

This is the report of a national summit on the issue convened in that year by the NAGC.

It follows Plucker (one of the summit participants) in using as its starting point,the achievement of advanced level on selected NAEP assessments by learners eligible for free and reduced price lunches.

But it also reports some additional outcomes for Grade 12 and for assessments of civics and writing:

  • ‘Since 1998, 1% or fewer of 4th-, 8th-, and 12th-grade free or reduced lunch students, compared to between 5% and 6% of non-eligible students scored at the advanced level on the NAEP civics exam.
  • Since 1998, 1% or fewer of free and reduced lunch program-eligible students scored at the advanced level on the eighth-grade NAEP writing exam while the percentage of non-eligible students who achieved advanced scores increased from 1% to 3%.’

The bulk of the report is devoted to identifying barriers to progress and offering recommendations for improving policy, practice and research. I provided an extended analysis in this post from May 2013.

Finally, ‘Talent on the Sidelines: Excellence Gaps and America’s Persistent Talent Underclass’ (Plucker, Hardesty and Burroughs 2013) is a follow-up to ‘Mind the Other Gap’.

It updates the findings in that report, as set out above:

  • In Grade 4 maths, from 1996 to 2011, the proportion of advantaged students scoring at the advanced level increased by 8.3 percentage points, while the proportion of disadvantaged learners doing so increased by 1.5 percentage points. At Grade 8, the comparable changes were 8.5 percentage points and 1.5 percentage points respectively. Excellence gaps have increased by 6.8 percentage points at Grade 4 (to 9.6%) and by 7 percentage points at Grade 8 (to 10.3%).
  • In Grade 4 reading, from 1998 to 2011, the proportion of advantaged students scoring at the advanced level increased by 2.6 percentage points, compared with an increase of 0.9 percentage points amongst disadvantaged learners. Grade 8 saw equivalent increases of 1.8 and 0.9 percentage points respectively. Excellence gaps are estimated to have increased at Grade 4 by 1.7 percentage points (to 10.7%) and marginally increased at Grade 8 by 0.9 percentage points (to 4.2%).

In short, many excellence gaps remain large and most continue to grow. The report’s recommendations are substantively the same as those put forward in 2010.

 

How Government education policy impacts on excellence gaps

Although many aspects of Government education policy may be expected to have some longer-term impact on raising the achievement of all learners, advantaged and disadvantaged alike, relatively few interventions are focused exclusively and directly on closing attainment gaps between advantaged and disadvantaged learners – and so have the potential to makes a significant difference to excellence gaps.

The most significant of these include:

 

The Pupil Premium:

In November 2010, the IPPR voiced concerns that the benefits of the pupil premium might not reach all those learners who attract it.

Accordingly they recommended that pupil premium should be allocated directly to those learners through an individual Pupil Premium Entitlement which might be used to support a menu of approved activities, including ‘one-to-one teaching to stretch the most able low income pupils’.

The recommendation has not been repeated and the present Government shows no sign of restricting schools’ freedom to use the premium in this manner.

However, the Blunkett Labour Policy Review ‘Putting students and parents first’ recommends that Labour in government should:

‘Assess the level and use of the Pupil Premium to ensure value for money, and that it is targeted to enhance the life chances of children facing the biggest challenges, whether from special needs or from the nature of the background and societal impact they have experienced.’

In February 2013 Ofsted reported that schools spending the pupil premium successfully to improve achievement:

‘Never confused eligibility for the Pupil Premium with low ability, and focused on supporting their disadvantaged pupils to achieve the highest levels’.

Conversely, where schools were less successful in spending the funding, they:

‘focused on pupils attaining the nationally expected level at the end of the key stage…but did not go beyond these expectations, so some more able eligible pupils underachieved.’

In July 2013, DfE’s Evaluation of Pupil Premium reported that, when deciding which disadvantaged pupils to target for support, the top criterion was ‘low attainment’ and was applied in 91% of primary schools and 88% of secondary schools.

In June 2013, in ‘The Most Able Students’, Ofsted reported that:

‘Pupil Premium funding was used in only a few instances to support the most able students who were known to be eligible for free school meals. The funding was generally spent on providing support for all underachieving and low-attaining students rather than on the most able students from disadvantaged backgrounds.’

Accordingly, it gave a commitment that:

‘Ofsted will… consider in more detail during inspection how well the pupil premium is used to support the most able students from disadvantaged backgrounds.’

However, this was not translated into the school inspection guidance.

The latest edition of the School Inspection Handbook says only:

‘Inspectors should pay particular attention to whether more able pupils in general and the most able pupils in particular are achieving as well as they should. For example, does a large enough proportion of those pupils who had the highest attainment at the end of Key Stage 2 in English and mathematics achieve A*/A GCSE grades in these subjects by the age of 16?

Inspectors should summarise the achievements of the most able pupils in a separate paragraph of the inspection report.’

There is no reference to the most able in parallel references to the pupil premium.

There has, however, been some progress in giving learners eligible for the pupil premium priority in admission to selective schools.

In May 2014, the TES reported that:

‘Thirty [grammar] schools have been given permission by the Department for Education to change their admissions policies already. The vast majority of these will introduce the changes for children starting school in September 2015…A small number – five or six – have already introduced the reform.’

The National Grammar Schools Association confirmed that:

‘A significant number of schools 38 have either adopted an FSM priority or consulted about doing so in the last admissions round. A further 59 are considering doing so in the next admissions round.’

In July 2014, the Government launched a consultation on the School Admissions Code which proposes extending to all state-funded schools the option to give priority in their admission arrangements to learners eligible for the pupil premium. This was previously open to academies and free schools via their funding agreements.

 

The Education Endowment Foundation (EEF)

The EEF describes itself as:

‘An independent grant-making charity dedicated to breaking the link between family income and educational achievement, ensuring that children from all backgrounds can fulfil their potential and make the most of their talents.’

The 2010 press release announcing its formation emphasised its role in raising standards in underperforming schools. This was reinforced by the Chairman in a TES article from June 2011:

‘So the target group for EEF-funded projects in its first couple of years are pupils eligible for free school meals in primary and secondary schools underneath the Government’s floor standards at key stages 2 and 4. That’s roughly 1,500 schools up and down the country. Projects can benefit other schools and pupils, as long as there is a significant focus on this core target group of the most needy young people in the most challenging schools.’

I have been unable to trace any formal departure from this position, though it no longer appears in this form in the Foundation’s guidance. The Funding FAQs say only:

‘In the case of projects involving the whole school, rather than targeted interventions, we would expect applicants to be willing to work with schools where the proportion of FSM-eligible pupils is well above the national average and/or with schools where FSM-eligible pupils are under-performing academically.’

I can find no EEF-funded projects that are exclusively or primarily focused on high-attaining disadvantaged learners, though a handful of its reports do refer to the impact on this group.

 

Changes to School Accountability Measures

As we have seen in Part one, the School Performance Tables currently provide very limited information about the performance of disadvantaged high achievers.

The July 2013 consultation document on primary assessment and accountability reform included a commitment to publish a series of headline measures in the tables including:

‘How many of the school’s pupils are among the highest-attaining nationally, by…showing the percentage of pupils attaining a high scaled score in each subject.’

Moreover, it added:

‘We will publish all the headline measures to show the attainment and progress of pupils for whom the school is in receipt of the pupil premium.’

Putting two and two together, this should mean that, from 2016, we will be able to see the percentage of pupil premium-eligible students achieving a high scaled score, though we do not yet know what ‘high scaled score’ means, nor do we know whether the data will be for English and maths separately or combined.

The October 2013 response to the secondary assessment and accountability consultation document fails to say explicitly whether excellence gap measures will be published in School Performance Tables.

It mentions that:

‘Schools will now be held to account for (a) the attainment of their disadvantaged pupils, (b) the progress made by their disadvantaged pupils, and (c) the in-school gap in attainment between disadvantaged pupils and their peers.’

Meanwhile a planned data portal will contain:

‘the percentage of pupils achieving the top grades in GCSEs’

but the interaction between these two elements, if any, remains unclear.

The March 2014 response to the consultation on post-16 accountability and assessment says:

‘We intend to develop measures covering all five headline indicators for students in 16-19 education who were in receipt of pupil premium funding in year 11.’

The post-16 headline measures include a new progress measure and an attainment measure showing the average points score across all level 3 qualifications.

It is expected that a destination measure will also be provided, as long as the methodology can be made sufficiently robust. The response says:

‘A more detailed breakdown of destinations data, such as entry to particular groups of universities, will continue to be published below the headline. This will include data at local authority level, so that destinations for students in the same area can be compared.’

and this should continue to distinguish the destinations of disadvantaged students.

Additional A level attainment measures – the average grade across the best three A levels and the achievement of AAB grades with at least two in facilitating subjects seem unlikely to be differentiated according to disadvantage.

There remains a possibility that much more excellence gap data, for primary, secondary and post-16, will be made available through the planned school portal, but no specification had been made public at the time of writing.

More worryingly, recent news reports have suggested that the IT project developing the portal and the ‘data warehouse’ behind it has been abandoned. The statements refer to coninuing to deliver ‘the school performance tables and associated services’ but there is no clarification of whether this latter phrase includes the portal. Given the absence of an official statement, one suspects the worst.

 

 

The Social Mobility and Child Poverty Commission (SMCPC)

The Commission was established with the expectation that it would ‘hold the Government’s feet to the fire’ to encourage progress on these two topics.

It publishes annual ‘state of the nation’ reports that are laid before Parliament and also undertakes ‘social mobility advocacy’.

The first annual report – already referenced in Part one – was published in November 2013. The second is due in October 2014.

The Chairman of the Commission was less than complimentary about the quality of the Government’s response to its first report, which made no reference to its comments about attainment gaps at higher grades. It remains to be seen whether the second will be taken any more seriously.

The Commission has already shown significant interest in disadvantaged high achievers – in June 2014 it published the study ‘Progress made by high-attaining children from disadvantaged backgrounds’ referenced above – so there is every chance that the topic will feature again in the 2014 annual report.

The Commission is of course strongly interested in the social mobility indicators and progress made against them, so may also include recommendations for how they might be adjusted to reflect changes to the schools accountability regime set out above.

 

Recommended reforms to close excellence gaps

Several proposals emerge from the commentary on current Government policy above:

  • It would be helpful to have further evaluation of the pupil premium to check whether high-achieving disadvantaged learners are receiving commensurate support. Schools need further guidance on ways in which they can use the premium to support high achievers. This should also be a focus for the pupil premium Champion and in pupil premium reviews.
  • Ofsted’s school inspection handbook requires revision to fulfil its commitment to focus on the most able in receipt of the premium. Inspectors also need guidance (published so schools can see it) to ensure common expectations are applied across institutions. These provisions should be extended to the post-16 inspection regime.
  • All selective secondary schools should be invited to prioritise pupil premium recipients in their admissions criteria, with the Government reserving the right to impose this on schools that do not comply voluntarily.
  • The Education Endowment Foundation should undertake targeted studies of interventions to close excellence gaps, but should also ensure that the impact on excellence gaps is mainstreamed in all the studies they fund. (This should be straightforward since their Chairman has already called for action on this front.)
  • The Government should consider the case for the inclusion of data on excellence gaps in all the headline measures in the primary, secondary and post-16 performance tables. Failing that, such data (percentages and numbers) should be readily accessible from a new data portal as soon as feasible, together with historical data of the same nature. (If the full-scale portal is no longer deliverable, a suitable alternative openly accessible database should be provided.) It should also publish annually a statistical analysis of all excellence gaps and the progress made towards closing them. As much progress as possible should be made before the new assessment and accountability regime is introduced. At least one excellence gap measure should be incorporated into revised DfE impact indicators and the social mobility indicators.
  • The Social Mobility and Child Poverty Commission (SMCPC) should routinely consider the progress made in closing excellence gaps within its annual report – and the Government should commit to consider seriously any recommendations they offer to improve such progress.

This leaves the question whether there should be a national programme dedicated to closing excellence gaps, and so improving fair access to competitive universities. (It makes excellent sense to combine these twin objectives and to draw on the resources available to support the latter.)

Much of the research above – whether it originates in the US or UK – argues for dedicated state/national programmes to tackle excellence gaps.

More recently, the Sutton Trust has published a Social Mobility Manifesto for 2015 which recommends that the next government should:

‘Reintroduce ring-fenced government funding to support the most able learners (roughly the top ten per cent) in maintained schools and academies from key stage three upwards. This funding could go further if schools were required to provide some level of match funding.

Develop an evidence base of effective approaches for highly able pupils and ensure training and development for teachers on how to challenge their most able pupils most effectively.

Make a concerted effort to lever in additional support from universities and other partners with expertise in catering for the brightest pupils, including through creating a national programme for highly able learners, delivered through a network of universities and accessible to every state-funded secondary school serving areas of disadvantage.’

This is not as clear as it might be about the balance between support for the most able and the most able disadvantaged respectively.

I have written extensively about what shape such a programme should have, most recently in the final section of ‘Digging Beneath the Destination Measures’ (July 2014).

The core would be:

‘A light touch framework that will supply the essential minimum scaffolding necessary to support effective market operation on the demand and supply sides simultaneously…

The centrepiece of the framework would be a structured typology or curriculum comprising the full range of knowledge, skills and understanding required by disadvantaged students to equip them for progression to selective higher education

  • On the demand side this would enable educational settings to adopt a consistent approach to needs identification across the 11-19 age range. Provision from 11-14 might be open to any disadvantaged learner wishing it to access it, but provision from 14 onwards would depend on continued success against challenging attainment targets.
  • On the supply side this would enable the full range of providers – including students’ own educational settings – to adopt a consistent approach to defining which knowledge, skills and understanding their various programmes and services are designed to impart. They would be able to qualify their definitions according to the age, characteristics, selectivity of intended destination and/or geographical location of the students they serve.

With advice from their educational settings, students would periodically identify their learning needs, reviewing the progress they had made towards personal targets and adjusting their priorities accordingly. They would select the programmes and services best matched to their needs….

…Each learner within the programme would have a personal budget dedicated to purchasing programmes and services with a cost attached. This would be fed from several sources including:

  • Their annual Pupil Premium allocation (currently £935 per year) up to Year 11.
  • A national fund fed by selective higher education institutions. This would collect a fixed minimum topslice from each institution’s outreach budget, supplemented by an annual levy on those failing to meet demanding new fair access targets. (Institutions would also be incentivised to offer programmes and services with no cost attached.)
  • Philanthropic support, bursaries, scholarships, sponsorships and in-kind support sourced from business, charities, higher education, independent schools and parents. Economic conditions permitting, the Government might offer to match any income generated from these sources.’

 

Close

We know far too little than we should about the size of excellence gaps in England – and whether or not progress is being made in closing them.

I hope that this post makes some small contribution towards rectifying matters, even though the key finding is that the picture is fragmented and extremely sketchy.

Rudimentary as it is, this survey should provide a baseline of sorts, enabling us to judge more easily what additional information is required and how we might begin to frame effective practice, whether at institutional or national level.

 

GP

September 2014

Closing England’s Excellence Gaps: Part One

This post examines what we know – and do not know – about high attainment gaps between learners from advantaged and disadvantaged backgrounds in England.

Mind the Gap by Clicsouris

Mind the Gap by Clicsouris

It assesses the capacity of current national education policy to close these gaps and recommends further action to improve the prospects of doing so rapidly and efficiently.

Because the post is extremely long I have divided it into two parts.

Part one comprises:

  • A working definition for the English context, explanation of the significance of excellence gaps, description of how this post relates to earlier material and provisional development of the theoretical model articulated in those earlier posts.
  • A summary of the headline data on socio-economic attainment gaps in England, followed by a review of published data relevant to excellence gaps at primary, secondary and post-16 levels.

Part two contains:

  • A distillation of research evidence, including material on whether disadvantaged high attainers remain so, international comparisons studies and research derived from them, and literature covering excellence gaps in the USA.
  • A brief review of how present Government policy might be expected to impact directly on excellence gaps, especially via the Pupil Premium, school accountability measures, the Education Endowment Foundation (EEF) and the Social Mobility and Child Poverty Commission (SMCPC). I have left to one side the wider set of reforms that might have an indirect and/or longer-term impact.
  • Some recommendations for strengthening our collective capacity to quantify address and ultimately close excellence gaps.

The post is intended to synthesise, supplement and update earlier material, so providing a baseline for further analysis – and ultimately consideration of further national policy intervention, whether under the present Government or a subsequent administration.

It does not discuss the economic and social origins of educational disadvantage, or the merits of wider policy to eliminate poverty and strengthen social mobility.

It starts from the premiss that, while education reform cannot eliminate the effects of disadvantage, it can make a significant, positive contribution by improving significantly the life chances of disadvantaged learners.

It does not debate the fundamental principle that, when prioritising educational support to improve the life chances of learners from disadvantaged backgrounds, governments should not discriminate on the basis of ability or prior attainment.

It assumes that optimal policies will deliver improvement for all disadvantaged learners, regardless of their starting point. It suggests, however, that intervention strategies should aim for equilibrium, prioritising gaps that are furthest away from it and taking account of several different variables in the process.

 

A working definition for the English context

The literature in Part two reveals that there is no accepted universal definition of excellence gaps, so I have developed my own England-specific working definition for the purposes of this post.

An excellence gap is:

‘The difference between the percentage of disadvantaged learners who reach a specified age- or stage-related threshold of high achievement – or who secure the requisite progress between two such thresholds – and the percentage of all other eligible learners that do so.’

This demands further clarification of what typically constitutes a disadvantaged learner and a threshold of high achievement.

In the English context, the measures of disadvantage with the most currency are FSM eligibility (eligible for and receiving free school meals) and eligibility for the deprivation element of the pupil premium (eligible for and receiving FSM at some point in the preceding six years – often called ‘ever 6’).

Throughout this post, for the sake of clarity, I have given priority to the former over the latter, except where the former is not available.

The foregrounded characteristic is socio-economic disadvantage, but this does not preclude analysis of the differential achievement of sub-groups defined according to secondary characteristics including gender, ethnic background and learning English as an additional language (EAL) – as well as multiple combinations of these.

Some research is focused on ‘socio-economic gradients’, which show how gaps vary at different points of the achievement distribution on a given assessment.

The appropriate thresholds of high achievement are most likely to be measured through national assessments of pupil attainment, notably end of KS2 tests (typically Year 6, age 11), GCSE and equivalent examinations (typically Year 11, age 16) and A level and equivalent examinations (typically Year 13, age 18).

Alternative thresholds of high achievement may be derived from international assessments, such as PISA, TIMSS or PIRLS.

Occasionally – and especially in the case of these international studies – an achievement threshold is statistically derived, in the form of a percentile range of performance, rather than with reference to a particular grade, level or score. I have not allowed for this within the working definition.

Progress measures typically relate to the distance travelled between: baseline assessment (currently at the end of KS1 – Year 2, age 7 – but scheduled to move to Year R, age 4) and end of KS2 tests; or between KS2 tests and the end of KS4 (GCSE); or between GCSE and the end of KS5 (Level 3/A level).

Some studies extend the concept of progress between two thresholds to a longitudinal approach that traces how disadvantaged learners who achieve a particular threshold perform throughout their school careers – do they sustain early success, or fall away, and what proportion are ‘late bloomers’?

 

Why are excellence gaps important?

Excellence gaps are important for two different sets of reasons: those applying to all achievement gaps and those which apply more specifically or substantively to excellence gaps.

Under the first heading:

  • The goal of education should be to provide all learners, including disadvantaged learners, with the opportunity to maximise their educational potential, so eliminating ‘the soft bigotry of low expectations’.
  • Schools should be ‘engines of social mobility’, helping disadvantaged learners to overcome their backgrounds and compete equally with their more advantaged peers.
  • International comparisons studies reveal that the most successful education systems can and do raise attainment for all and close socio-economic achievement gaps simultaneously.
  • There is a strong economic case for reducing – and ideally eradicating – underachievement attributable to disadvantage.

Under the second heading:

  • An exclusive or predominant focus on gaps at the lower end of the attainment distribution is fundamentally inequitable and tends to reinforce the ‘soft bigotry of low expectations’.
  • Disadvantaged learners benefit from successful role models – predecessors or peers from a similar background who have achieved highly and are reaping the benefits.
  • An economic imperative to increase the supply of highly-skilled labour will place greater emphasis on the top end of the achievement distribution. Some argue that there is a ‘smart fraction’ tying national economic growth to a country’s stock of high achievers. There may be additional spin-off benefits from increasing the supply of scientists, writers, artists, or even politicians!
  • The most highly educated disadvantaged learners are least likely to confer disadvantage on their children, so improving the proportion of such learners may tend to improve inter-generational social mobility.

Excellence gaps are rarely identified as such – the term is not yet in common usage in UK education, though it has greater currency in the US. Regardless of terminology, they rarely receive attention, either as part of a wider set of achievement gaps, or separately in their own right.

 

Relationship with earlier posts

Since this blog was founded in April 2010 I have written extensively about excellence gaps and how to address them.

The most pertinent of my previous posts are:

I have also written about excellence gaps in New Zealand – Part 1 and Part 2 (June 2012) – but do not draw on that material here.

Gifted education (or apply your alternative term) is amongst those education policy areas most strongly influenced by political and ideological views on the preferred balance between excellence and equity. This is particularly true of decisions about how best to address excellence gaps.

The excellence-equity trade-off was identified in my first post (May 2010) as one of three fundamental polarities that determine the nature of gifted education and provide the basis for most discussion about what form it should take.

The Gifted Phoenix Manifesto for Gifted Education (March 2013) highlighted their significance thus:

‘Gifted education is about balancing excellence and equity. That means raising standards for all while also raising standards faster for those from disadvantaged backgrounds.

Through combined support for excellence and equity we can significantly increase our national stock of high level human capital and so improve economic growth…

…Excellence in gifted education is about maximising the proportion of high achievers reaching advanced international benchmarks (eg PISA, TIMSS and PIRLS) so increasing the ‘smart fraction’ which contributes to economic growth.

Equity in gifted education is about narrowing (and ideally eliminating) the excellence gap between high achievers from advantaged and disadvantaged backgrounds (which may be attributable in part to causes other than poverty). This also increases the proportion of high achievers, so building the ‘smart fraction’ and contributing to economic growth.’

More recently, one of the 10 draft core principles I set out in ‘Why Can’t We Have National Consensus on Educating High Attainers?’ (June 2014) said:

‘We must pursue simultaneously the twin priorities of raising standards and closing gaps. We must give higher priority to all disadvantaged learners, regardless of their prior achievement. Standards should continue to rise amongst all high achievers, but they should rise faster amongst disadvantaged high achievers. This makes a valuable contribution to social mobility.’

 

This model provisionally developed

Using my working definition as a starting point, this section describes a theoretical model showing how excellence and equity are brought to bear when considering excellence gaps – and then how best to address them.

This should be applicable at any level, from a single school to a national education system and all points in between.

The model depends on securing the optimal balance between excellence and equity where:

  • Excellence is focused on increasing the proportion of all learners who achieve highly and, where necessary, increasing the pitch of high achievement thresholds to remove unhelpful ceiling effects. The thresholds in question may be nationally or internationally determined and are most likely to register high attainment through a formal assessment process. (This may be extended so there is complementary emphasis on increasing the proportion of high-achieving learners who make sufficiently strong progress between two different age- or stage-related thresholds.)
  • Equity is focused on increasing the proportion of high-achieving disadvantaged learners (and/or the proportion of disadvantaged learners making sufficiently strong progress) at a comparatively faster rate, so they form a progressively larger proportion of the overall high-achieving population, up to the point of equilibrium, where advantaged and disadvantaged learners are equally likely to achieve the relevant thresholds (and/or progress measure). This must be secured without deliberately repressing improvement amongst advantaged learners – ie by introducing policies designed explicitly to limit their achievement and/or progress relative to disadvantaged learners – but a decision to do nothing or to redistribute resources in favour of disadvantage is entirely permissible.

The optimal policy response will depend on the starting position and the progress achieved over time.

If excellence gaps are widening, the model suggests that interventions and resources should be concentrated in favour of equity. Policies should be reviewed and adjusted, or strengthened where necessary, to meet the desired objectives.

If excellence gaps are widening rapidly, this reallocation and adjustment process will be relatively more substantial (and probably more urgent) than if they are widening more slowly.

Slowly widening gaps will demand more reallocation and adjustment than a situation where gaps are stubbornly resistant to improvement, or else closing too slowly. But even in the latter case there should be some reallocation and adjustment until equilibrium is achieved.

When excellence gaps are already closing rapidly – and there are no overt policies in place to deliberately repress improvement amongst high-achieving advantaged learners – it may be that unintended pressures in the system are inadvertently bringing this about. In that case, policy and resources should be adjusted to correct these pressures and so restore the correct twin-speed improvement.

The aim is to achieve and sustain equilibrium, even beyond the point when excellence gaps are eliminated, so that they are not permitted to reappear.

If ‘reverse gaps’ begin to materialise, where disadvantaged learners consistently outperform their more advantaged peers, this also threatens equilibrium and would suggest a proportionate redistribution of effort towards excellence.

Such scenarios are most likely to occur in settings where there are a large proportion of learners that, while not disadvantaged according to the ‘cliff edge’ definition required to make the distinction, are still relatively disadvantaged.

Close attention must therefore be paid to the distribution of achievement across the full spectrum of disadvantage, to ensure that success at the extreme of the distribution does not mask significant underachievement elsewhere.

One should be able to determine a more precise policy response by considering a restricted set of variables. These include:

  • The size of the gaps at the start of the process and, associated with this, the time limit allowed for equilibrium to be reached. Clearly larger gaps are more likely to take longer to close. Policy makers may conclude that steady improvement over several years is more manageable for the system than a rapid sprint towards equilibrium. On the other hand, there may be benefits associated with pace and momentum.
  • The rate at which overall high achievement is improving. If this is relatively fast, the rate of improvement amongst advantaged high achievers will be correspondingly strong, so the rate for disadvantaged high achievers must be stronger still.
  • The variance between excellence gaps at different ages/stages. If the gaps are larger at particular stages of education, the pursuit of equilibrium suggests disproportionate attention is given to those so gaps are closed consistently. If excellence gaps are small for relatively young learners and increase with age, priority should be given to the latter, but there may be other factors in play, such as evidence that closing relatively small gaps at an early stage will have a more substantial ‘knock-on’ effect later on.
  • The level at which high achievement thresholds are pitched. Obviously this will influence the size of the gaps that need to be closed. But, other things being equal, enabling a higher proportion of learners to achieve a relatively high threshold will demand more intensive support. On the other hand, relatively fewer learners – whether advantaged or disadvantaged – are likely to be successful. Does one need to move a few learners a big distance or a larger proportion a smaller one?
  • Whether or not gaps at lower achievement thresholds are smaller and/or closing at a faster rate. If so, there is a strong case for securing parity of progress at higher and lower thresholds alike. On the other hand, if excellence gaps are closing more quickly, it may be appropriate to reallocate resources away from them and towards lower levels of achievement.
  • The relative size of the overall disadvantaged population, the associated economic gap between advantage and disadvantage and (as suggested above) the distribution in relation to the cut-off. If the definition of disadvantage is pitched relatively low (ie somewhat disadvantaged), the disadvantaged population will be correspondingly large, but the economic gap between advantage and disadvantage will be relatively small. If the definition is pitched relatively high (ie very disadvantaged) the reverse will be true, giving a comparatively small disadvantaged population but a larger gap between advantage and disadvantage.
  • The proportion of the disadvantaged population that is realistically within reach of the specified high achievement benchmarks. This variable is a matter of educational philosophy. There is merit in an inclusive approach – indeed it seems preferable to overestimate this proportion than the reverse. Extreme care should be taken not to discourage late developers or close off opportunities on the basis of comparatively low current attainment, so reinforcing existing gaps through unhelpfully low expectations. On the other hand, supporting unrealistically high expectations may be equally damaging and ultimately waste scarce resources. There may be more evidence to support such distinctions with older learners than with their younger peers. 

 

How big are England’s headline attainment gaps and how fast are they closing?

Closing socio-economic achievement gaps has been central to English educational policy for the last two decades, including under the current Coalition Government and its Labour predecessor.

It will remain an important priority for the next Government, regardless of the outcome of the 2015 General Election.

The present Government cites ‘Raising the achievement of disadvantaged children’ as one of ten schools policies it is pursuing.

The policy description describes the issue thus:

‘Children from disadvantaged backgrounds are far less likely to get good GCSE results. Attainment statistics published in January 2014 show that in 2013 37.9% of pupils who qualified for free school meals got 5 GCSEs, including English and mathematics at A* to C, compared with 64.6% of pupils who do not qualify.

We believe it is unacceptable for children’s success to be determined by their social circumstances. We intend to raise levels of achievement for all disadvantaged pupils and to close the gap between disadvantaged children and their peers.’

The DfE’s input and impact indicators  – showing progress against the priorities set out in its business plan – do not feature the measure mentioned in the policy description (which is actually five or more GCSEs at Grades A*-C or equivalents, including GCSEs in English and maths).

The gap on this measure was 27.7% in 2009, improving to 26.7% in 2013, so there has been a small 1.0 percentage point improvement over five years, spanning the last half of the previous Government’s term in office and the first half of this Government’s term.

Instead the impact indicators include three narrower measures focused on closing the attainment gap between free school meal pupils and their peers, at 11, 16 and 19 respectively:

  • Impact Indicator 7 compares the percentages of FSM-eligible and all other pupils achieving level 4 or above in KS2 assessment of reading, writing and maths. The 2013 gap is 18.7%, down 0.4% from 19.1% in 2012.
  • Impact Indicator 8 compares the percentages of FSM-eligible and all other pupils achieving A*-C grades in GCSE maths and English. The 2013 gap is 26.5%, up 0.3% from 26.2% in 2012.
  • Impact Indicator 9 compares the percentages of learners who were FSM-eligible at age 15 and all other learners who attain a level 3 qualification by the end of the academic year in which they are 19. The 2013 gap is 24.3%, up 0.1% from 24.2% in 2012.

These small changes, not always pointing in the right direction, reflect the longer term narrative, as is evident from the Government’s Social Mobility Indicators which also incorporate these three measures.

  • In 2005-06 the KS2 L4 maths and English gap was 25.0%, so there has been a fairly substantial 6.3 percentage point reduction over seven years, but only about one quarter of the gap has been closed.
  • In 2007-08 the KS4 GCSE maths and English gap was 28.0%, so there has been a minimal 1.5 percentage point reduction over six years, equivalent to annual national progress of 0.25 percentage points per year. At that rate it will take another century to complete the process.
  • In 2004-05 the Level 3 qualification gap was 26.4%, so there has been a very similar 2.1 percentage point reduction over 8 years.

The DfE impact indicators also include a set of three destination measures that track the percentage of FSM learners progressing to Oxford and Cambridge, any Russell Group university and any university.

There is a significant time lag with all of these – the most recent available data relates to 2011/2012 – and only two years of data have been collected.

All show an upward trend. Oxbridge is up from 0.1% to 0.2%, Russell Group up from 3% to 4% and any university up from 45% to 47% – actually a 2.5 percentage point improvement.

The Oxbridge numbers are so small that a percentage measure is a rather misleading indicator of marginal improvement from a desperately low base.

It is important to note that forthcoming changes to the assessment regime will impose a different set of headline indicators at ages 11 and 16 that will not be comparable with these.

From 2014 significant methodological adjustments are being introduced to School Performance Tables that significantly restrict the range of qualifications equivalent to GCSEs. Only the first entry in each subject will count for Performance Table purposes, this applying to English Baccalaureate subjects in 2014 and then all subjects in 2015.

Both these factors will tend to depress overall results and may be expected to widen attainment gaps on the headline KS4 measure as well as the oft-cited 5+ GCSEs measure.

From 2016 new baseline assessments, the introduction of scaled scores at the end of KS2 and a new GCSE grading system will add a further layer of change.

As a consequence there will be substantial revisions to the headline measures in Primary, Secondary and Post-16 Performance Tables. The latter will include destination measures, provided they can be made methodologically sound.

At the time of writing, the Government has made negligible reference to the impact of these reforms on national measures of progress, including its own Impact Indicators and the parallel Social Mobility indicators, though the latter are reportedly under review.

 

Published data on English excellence gaps

The following sections summarise what data I can find in the public domain about excellence gaps at primary (KS2), secondary (KS4) and post-16 (KS5) respectively.

I have cited the most recent data derivable from Government statistical releases and performance tables, supplemented by other interesting findings gleaned from research and commentary.

 

Primary (KS2) 

The most recent national data is contained in SFR51/2013: National Curriculum Assessments at Key Stage 2: 2012 to 2013. This provides limited information about the differential performance of learners eligible for and receiving FSM (which I have referred to as ‘FSM’), and for those known to be eligible for FSM at any point from Years 1 to 6 (known as ‘ever 6’ and describing those in receipt of the Pupil Premium on grounds of deprivation).

There is also additional information in the 2013 Primary School Performance Tables, where the term ‘disadvantaged’ is used to describe ‘ever 6’ learners and ‘children looked after’.

There is comparably little variation between these different sets of figures at national level. In the analysis below (and in the subsequent section on KS4) I have used FSM data wherever possible, but have substituted ‘disadvantaged’ data where FSM is not available.  All figures apply to state-funded schools only.

I have used Level 5 and above as the best available proxy for high attainment. Some Level 6 data is available, but in percentages only, and these are all so small that comparisons are misleading.

The Performance Tables distinguish a subset of high attainers, on the basis of prior attainment (at KS1 for KS2 and at KS2 for KS4) but no information is provided about the differential performance of advantaged and disadvantaged high attainers.

In 2013:

  • 21% of all pupils achieved Level 5 or above in reading, writing and maths combined, but only 10% of FSM pupils did so, compared with 26% of others, giving an attainment gap of 16%. The comparable gap at Level 4B (in reading and maths and L4 in writing) was 18%. At Level 4 (across the board) it was 20%. In this case, the gaps are slightly larger at lower attainment levels but, whereas the L4 gap has narrowed by 1% since 2012, the L5 gap has widened by 1%.
  • In reading, 44% of all pupils achieved Level 5 and above, but only 21% of FSM pupils did so, compared with 48% of others, giving an attainment gap of 21%. The comparable gap at Level 4 and above was eight percentage points lower at 13%.
  • In writing (teacher assessment), 31% of all pupils achieved level 5 and above, but only 15% of FSM pupils did so, compared with 34% of others, giving an attainment gap of 19%. The comparable gap at Level 4 and above was three percentage points lower at 16%.
  • In grammar, punctuation and spelling (GPS), 47% of all pupils achieved Level 5 and above, but only 31% of FSM pupils did so, compared with 51% of others, giving an attainment gap of 20%. The comparable gap at Level 4 and above was two percentage points lower at 18%.
  • In maths, 41% of pupils in state-funded schools achieved Level 5 and above, up 2% on 2012. But only 24% of FSM pupils achieved this compared with 44% of others, giving an attainment gap of 20%. The comparable gap at level 4 and above is 13%.

Chart 1 shows these outcomes graphically. In four cases out of five, the gap at the higher attainment level is greater, substantially so in reading and maths. All the Level 5 gaps fall between 16% and 20%.

 

Ex gap table 1

Chart 1: Percentage point gaps between FSM and all other pupils’ attainment at KS2 L4 and above and KS2 L5 and above, 2013 

 

It is difficult to trace reliably the progress made in reducing these gaps in English, since the measures have changed frequently. There has been more stability in maths, however, and the data reveals that – whereas the FSM gap at Level 4 and above has reduced by 5 percentage points since 2008 (from 18 points to 13 points) – the FSM gap at Level 5 and above has remained between 19 and 20 points throughout. Hence the gap between L4+ and L5+ on this measure has increased in the last five years.

There is relatively little published about KS2 excellence gaps elsewhere, though one older Government publication, a DfES Statistical Bulletin: The characteristics of high attainers (2007) offers a small insight.

It defines KS2 high attainers as the top 10%, on the basis of finely grained average points scores across English, maths and science, so a more selective but wider-ranging definition than any of the descriptors of Level 5 performance above.

According to this measure, some 2.7% of FSM-eligible pupils were high attainers in 2006, compared with 11.6% of non-FSM pupils, giving a gap of 8.9 percentage points.

The Bulletin supplies further analysis of this population of high attainers, summarised in the table reproduced below.

 

EX Gap Capture 1 

  

Secondary (KS4) 

While Government statistical releases provide at least limited data about FSM performance at high levels in end of KS2 assessments, this is entirely absent from KS4 data, because there is no information about the achievement of GCSE grades above C, whether for single subjects or combinations.

The most recent publication: SFR05/2014: GCSE and equivalent attainment by pupil characteristics, offers a multitude of measures based on Grades G and above or C and above, many of which are set out in Chart 2, which illustrates the FSM gap on each, organised in order from the smallest gap to the biggest.

(The gap cited here for A*-C grades in English and maths GCSEs is very slightly different to the figure in the impact indicator.)

 

Ex gap table 2

Chart 2: Percentage point gaps between FSM and all other pupils’ attainment on different KS4 measures, 2013

 

In its State of the Nation Report 2013, the Social Mobility and Child Poverty Commission included a table comparing regional performance on a significantly more demanding ‘8+ GCSEs excluding equivalents and including English and maths’ measure. This uses ‘ever 6’ rather than FSM as the indicator of disadvantage.

The relevant table is reproduced below. It shows regional gaps of between 20 and 26 percentage points on the tougher measure, so a similar order of magnitude to the national indicators at the top end of Chart 2.

 

ExGap 2 Capture

 

Comparing the two measures, one can see that:

  • The percentages of ‘ever 6’ learners achieving the more demanding measure are very much lower than the comparable percentages achieving the 5+ GCSEs measure, but the same is also true of their more advantaged peers.
  • Consequently, in every region but London and the West Midlands, the attainment gap is actually larger for the less demanding measure.
  • In London, the gaps are much closer, at 19.1 percentage points on the 5+ measure and 20.9 percentage points on the 8+ measure. In the West Midlands, the gap on the 8+ measure is larger by five percentage points. In all other cases, the difference is at least six percentage points in the other direction.

We do not really understand the reasons why London and the West Midlands are atypical in this respect.

The Characteristics of High Attainers (2007) provides a comparable analysis for KS4 to that already referenced at KS2. In this case, the top 10% of high attainers is derived on the basis of capped GCSE scores.

This gives a gap of 8.8 percentage points between the proportion of non-FSM (11.2%) and FSM (2.4%) students within the defined population, very similar to the parallel calculation at KS2.

Other variables within this population are set out in the table reproduced below.

 

ExGap Capture 3

Finally, miscellaneous data has also appeared from time to time in the answers to Parliamentary Questions. For example:

  • In 2003, 1.0% of FSM-eligible learners achieved five or more GCSEs at A*/A including English and maths but excluding equivalents, compared with 6.8% of those not eligible, giving a gap of 5.8 percentage points. By 2009 the comparable percentages were 1.7% and 9.0% respectively, resulting in an increased gap of 7.3 percentage points (Col 568W)
  • In 2006/07, the percentage of FSM-eligible pupils securing A*/A grades at GCSE in different subjects, compared with the percentage of all pupils in maintained schools doing so were as shown in the table below (Col 808W)
FSM All pupils Gap
Maths 3.7 15.6 11.9
Eng lit 4.1 20.0 15.9
Eng lang 3.5 16.4 12.9
Physics 2.2 49.0 46.8
Chemistry 2.5 48.4 45.9
Biology 2.5 46.8 44.3
French 3.5 22.9 19.4
German 2.8 23.2 20.4

Table 1: Percentage of FSM-eligible and all pupils achieving GCSE A*/A grades in different GCSE subjects in 2007

  • In 2008, 1% of FSM-eligible learners in maintained schools achieved A* in GCSE maths compared with 4% of all pupils in maintained schools. The comparable percentages for Grade A were 3% and 10% respectively, giving an A*/A gap of 10 percentage points (Col 488W)

 

Post-16 (KS5)

The most recent post-16 attainment data is provided in SFR10/2014: Level 2 and 3 attainment by young people aged 19 in 2013 and SFR02/14: A level and other level 3 results: academic year 2012 to 2013.

The latter contains a variety of high attainment measures – 3+ A*/A grades;  AAB grades or better; AAB grades or better with at least two in facilitating subjects;  AAB grades or better, all in facilitating subjects – yet none of them distinguish success rates for advantaged and disadvantaged learners.

The former does includes a table which provides a time series of gaps for achievement of Level 3 at age 19 through 2 A levels or the International Baccalaureate. The measure of disadvantage is FSM-eligibility in Year 11. The gap was 22.0 percentage points in 2013, virtually unchanged from 22.7 percentage points in 2005.

In (How) did New Labour narrow the achievement and participation gap (Whitty and Anders, 2014) the authors reproduce a chart from a DfE roundtable event held in March 2013 (on page 44).

This is designed to show how FSM gaps vary across key stages and also provides ‘odds ratios’ – the relative chances of FSM and other pupils achieving each measure. It relies on 2012 outcomes.

The quality of the reproduction is poor, but it seems to suggest that, using the AAB+ in at least two facilitating subjects measure, there is a five percentage point gap between FSM students and others (3% versus 8%), while the odds ratio shows that non-FSM students are 2.9 times more likely than FSM students to achieve this outcome.

Once again, occasional replies to Parliamentary Questions provide some supplementary information:

  • In 2007, 189 FSM-eligible students (3.7%) in maintained mainstream schools (so excluding sixth form colleges and FE colleges) achieved 3 A grades at A level. This compared with 13,467 other students (9.5%) giving a gap of 5.8 percentage points (Source: Parliamentary Question, 26 November 2008, Hansard (Col 1859W)
  • In 2008, 160 students (3.5%) eligible for FSM achieved that outcome. This compares with 14,431 (10.5%) of those not eligible for FSM, giving a gap of 7.0 percentage points. The figures relate to 16-18 year-olds, in maintained schools only, who were eligible for FSM at age 16. They do not include students in FE sector colleges (including sixth form colleges) who were previously eligible for FSM. Only students who entered at least one A level, applied A level or double award qualification are counted. (Parliamentary Question, 6 April 2010, Hansard (Col 1346W))
  • Of pupils entering at least one A level in 2010/11 and eligible for FSM at the end of Year 11, 546 (4.1%) achieved 3 or more GCE A levels at A*-A compared with 22,353 other pupils (10.6%) so giving a gap of 6.5 percentage points. These figures include students in both the schools and FE sectors. (Parliamentary Question, 9 July 2012, Hansard (Col 35W)) 

 In September 2014, a DfE response to a Freedom of Information request provided some additional data about FSM gaps at A level over the period from 2009 to 2013. This is set out in the table below, which records the gaps between FSM and all other pupils, presumably for all schools and colleges, whether or not state-funded.

Apart from the atypical result for the top indicator in 2010, all these percentages fall in the range 6.0% to 10%, so are in line with the sources above.

 

2009 2010 2011 2012 2013
3+ grades at A*/A or applied single/double award 9.0 12.8 9.3 8.7 8.3
AAB+ grades in facilitating subjects 6.3 6.2
AAB+ grades at least 2 in facilitating subjects 9.8

 

Additional evidence of Key Stage excellence gaps from a sample born in 1991

In Progress made by high-achieving children from disadvantaged backgrounds (Crawford, Macmillan and Vignoles, 2014) provides useful data on the size of excellence gaps at different key stages, as well as analysis of whether disadvantaged high achievers remain so through their school careers.

The latter appears in Part two, but the first set of findings provides a useful supplement to the broad picture set out above.

This study is based on a sample of learners born in 1991/1992, so they would presumably have taken end of KS2 tests in 2002, GCSEs in 2007 and A levels in 2009. It includes all children who attended a state primary school, including those who subsequently attended an independent secondary school.

It utilises a variety of measures of disadvantage, including whether learners were always FSM-eligible (in Years 7-11), or ‘ever FSM’ during that period. This summary focuses on the distinction between ‘always FSM’ and ‘never FSM’.

It selects a basket of high attainment measures spread across the key stages, including:

  • At KS1, achieving Level 3 or above in reading and maths.
  • At KS2, achieving Level 5 or above in English and maths.
  • At KS4, achieving six or more GCSEs at grades A*-C in EBacc subjects (as well as five or more).
  • At KS5, achieving two or more (and three or more) A levels at grades A-B in any subjects.
  • Also at KS5, achieving two or more (and three or more) A levels at grades A-B in facilitating subjects.

The choice of measures at KS2 and KS5 is reasonable, reflecting the data available at the time. For example, one assumes that A* grades at A level do not feature in the KS5 measures since they were not introduced until 2010).

At KS4, the selection is rather more puzzling and idiosyncratic. It would have been preferable to have included at least one measure based on performance across a range of GCSEs at grades A*-B or A*/A.

The authors justify their decision on the basis that ‘there is no consensus on what is considered high attainment’, even though most commentators would expect this to reflect higher grade performance, while few are likely to define it solely in terms of breadth of study across a prescribed set of ‘mainstream’ subjects.

Outcomes for ‘always FSM’ and ‘never FSM’ on the eight measures listed above are presented in Chart 3.

Ex gap Table 3

Chart 3: Achievement of ‘always FSM’ and ‘never FSM’ on a basket of high attainment measures for pupils born in 1991/92

 

This reveals gaps of 12 to 13 percentage points at Key Stages 1 and 2, somewhat smaller than several of those described above.

It is particularly notable that the 2013 gap for KS2 L5 reading, writing and maths is 16 percentage points, whereas the almost comparable 2002 (?) gap for KS2 English and maths amongst this sample is 13.5%. Even allowing for comparability issues, there may tentative evidence here to suggest widening excellence gaps at KS2 over the last decade.

The KS4 gaps are significantly larger than those existing at KS1/2, at 27 and 18 percentage points respectively. But comparison with the previous evidence reinforces the point that the size of the gaps in this sample is attributable to subject mix: this must be the case since the grade expectation is no higher than C.

The data for A*/A performance on five or more GCSEs set out above, which does not insist on coverage of EBacc subjects other than English and maths, suggests a gap of around seven percentage points. But it also demonstrates big gaps – again at A*/A – for achievement in single subjects, especially the separate sciences.

The KS5 gaps on this sample range from 2.5 to 13 percentage points. We cited data above suggesting a five percentage point gap in 2012 for AAB+, at least two in facilitating subjects. These findings do not seem wildly out of kilter with that, or with the evidence of gaps of around six to seven percentage points for AAA grades or higher.

 

Overall pattern 

The published data provides a beguiling glimpse of the size of excellence gaps and how they compare with FSM gaps on the key national benchmarks.

But discerning the pattern is like trying to understand the picture on a jigsaw when the majority of pieces are missing.

The received wisdom is capture in the observation by Whitty and Anders that:

‘Even though the attainment gap in schools has narrowed overall, it is largest for the elite measures’

and the SMCPC’s comment that:

‘…the system is better at lifting children eligible for FSM above a basic competence level (getting 5A*–C) than getting them above a tougher level of attainment likely to secure access to top universities.’

This seems broadly true, but the detailed picture is rather more complicated.

  • At KS2 there are gaps at L5 and above of around 16-20 percentage points, the majority higher than the comparable gaps at L4. But the gaps for core subjects combined are smaller than for each assessment. There is tentative evidence that the former may be widening.
  • At KS4 there are very significant differences between results in individual subjects. When it comes to multi-subject indicators, differences in the choice of subject mix – as well as choice of grade – make it extremely difficult to draw even the most tentative conclusions about the size of excellence gaps and how they relate to benchmark-related gaps at KS4 and excellence gaps at KS2.
  • At KS5, the limited evidence suggests that A level excellence gaps at the highest grades are broadly similar to those at GCSE A*/A. If anything, gaps seem to narrow slightly compared with KS4. But the confusion over KS4 measures makes this impossible to verify.

We desperately need access to a more complete dataset so we can understand these relationships more clearly.

This is the end of Part one. In Part two, we move on to consider evidence about whether high attainers remain so, before examining international comparisons data and related research, followed by excellence gaps analysis from the USA.

Part two concludes with a short review of how present government policy impacts on excellence gaps and some recommendations for strengthening the present arrangements.

 

GP

September 2014

What Happened to the Level 6 Reading Results?

 

Provisional 2014 key stage 2 results were published on 28 August.

500px-Japanese_Urban_Expwy_Sign_Number_6.svgThis brief supplementary post considers the Level 6 test results – in reading, in maths and in grammar, punctuation and spelling (GPS) – and how they compare with Level 6 outcomes in 2012 and 2013.

An earlier post, A Closer Look at Level 6, published in May 2014, provides a fuller analysis of these earlier results.

Those not familiar with the 2014 L6 test materials can consult the papers, mark schemes and level thresholds at these links:

 

Number of Entries

Entry levels for the 2014 Level 6 tests were published in the media in May 2014. Chart 1 below shows the number of entries for each test since 2012 (2013 in the case of GPS). These figures are for all schools, independent as well as state-funded.

 

L6 Sept chart 1

Chart 1: Entry rates for Level 6 tests 2012 to 2014 – all schools

 

In 2014, reading entries were up 36%, GPS entries up 52% and maths entries up 36%. There is as yet no indication of a backlash from the decision to withdraw Level 6 tests after 2015, though this may have an impact next year.

The postscript to A Closer Look estimated that, if entries continue to increase at current rates, we might expect something approaching 120,000 in reading, 130,000 in GPS and 140,000 in maths.

Chart 2 shows the percentage of all eligible learners entered for Level 6 tests, again for all schools. Nationally, between one in six and one in five eligible learners are now entered for Level 6 tests. Entry rates for reading and maths have almost doubled since 2012.

 

L6 Sept chart 2

Chart 2: Percentage of eligible learners entered for Level 6 tests 2012 to 2014, all schools

 

Success Rates

The headline percentages in the SFR show:

  • 0% achieving L6 reading (unchanged from 2013)
  • 4% achieving L6 GPS (up from 2% in 2013) and
  • 9% achieving L6 maths (up from 7% in 2013).

Local authority and regional percentages are also supplied.

  • Only in Richmond did the L6 pass rate in reading register above 0% (at 1%). Hence all regions are at 0%.
  • For GPS the highest percentages are 14% in Richmond, 10% in Kensington and Chelsea and Kingston, 9% in Sutton and 8% in Barnet, Harrow and Trafford. Regional rates vary between 2% in Yorkshire and Humberside and 6% in Outer London.
  • In maths, Richmond recorded 22%, Kingston 19%, Trafford, Harrow and Sutton were at 18% and Kensington and Chelsea at 17%. Regional rates range from 7% in Yorkshire and Humberside and the East Midlands to 13% in Outer London.

Further insight into the national figures can be obtained by analysing the raw numbers supplied in the SFR.

Chart 3 shows how many of those entered for each test were successful in each year. Here there is something of a surprise.

 

L6 Sept chart 3

Chart 3: Percentage of learners entered achieving Level 6, 2012 to 2014, all schools

 

Nearly half of all entrants are now successful in L6 maths, though the improvement in the success rate has slowed markedly compared with the nine percentage point jump in 2013.

In GPS, the success rate has improved by nine percentage points between 2013 and 2014 and almost one in four entrants is now successful. Hence the GPS success rate is roughly half that for maths. This may be attributable in part to its shorter history, although the 2014 success rate is significantly below the rate for maths in 2013.

But in reading an already very low success rate has declined markedly, following a solid improvement in 2013 from a very low base in 2012. The 2014 success rate is now less than half what it was in 2012. Fewer than one in a hundred of those entered have passed this test.

Chart 4 shows how many learners were successful in the L6 reading test in 2014 compared with previous years, giving results for boys and girls separately.

 

L6 Sept chart 4

Chart 4: Percentage of learners entered achieving Level 6 in reading, 2012 to 2014, by gender

 

The total number of successful learners in 2014 is over 5% lower than in 2012, when the reading test was introduced, and down 62% on the success rate achieved in 2013.

Girls appear to have suffered disproportionately from the decline in 2014 success rates. While the success rate for girls is down 63%, the decline for boys is slightly less, at 61%. The success rate for boys remains above where it was in 2012 but, for girls, it is about 12% down on where it was in 2012.

In 2012, only 22% of successful candidates were boys. This rose to 26% in 2013 and has again increased slightly, to 28% in 2014. The gap between girls’ and boys’ performance remains substantially bigger than those for GPS and maths.

Charts 5 and 6 give the comparable figures for GPS and maths respectively.

In GPS, the total number of successful entries has increased by almost 140% compared with 2013. Girls form a slightly lower proportion of this group than in 2013, their share falling from 62% to 60%. Boys are therefore beginning to close what remains a substantial performance gap.

 

L6 Sept chart 5

Chart 5: Percentage of learners entered achieving Level 6 in GPS, 2012 to 2014, by gender

 

In maths, the total number of successful entries is up by about 40% on 2013 and demonstrates rapid improvement over the three year period.

Compared with 2013, the success rate for girls has increased by 43%, whereas the corresponding increase for boys is closer to 41%. Boys formed 65% of the successful cohort in 2012, 61% in 2013 and 60% in 2014, so girls’ progress in narrowing this substantial performance gap is slowing.

 

L6 Sept chart 6

Chart 6: Percentage of learners entered achieving Level 6 in maths, 2012 to 2014, by gender

 

Progress

The SFR also provides a table, this time for state-funded schools only, showing the KS1 outcomes of those successful in achieving Level 6. (For maths and reading, this data includes those with a non-numerical grade in the test who have been awarded L6 via teacher assessment. The data for writing is derived solely from teacher assessment.)

Not surprisingly, over 94% of those achieving Level 6 in reading had achieved Level 3 in KS1, but 4.8% were at L2A and a single learner was recorded at Level 1. The proportion with KS1 Level 3 in 2013 was higher, at almost 96%.

In maths, however, only some 78% of those achieving Level 6 were at Level 3 in KS1. A further 18% were at 2A and almost 3% were at 2B. A further 165 were recorded as 2C or 1. In 2013, over 82% had KS1 L3 while almost 15% had 2A.

It seems, therefore, that KS1 performance was a slightly weaker indicator of KS2 level 6 success in 2014 than in the previous year, but this trend was apparent in both reading and maths – and KS1 performance remains a significantly weaker indicator in maths than it is in reading.

 

Why did the L6 reading results decline so drastically?

Given that the number of entries for the Level 6 reading test increased dramatically, the declining pass rate suggests either a problematic test or that schools entered a higher proportion of learners who had relatively little chance of success. A third possibility is that the test was deliberately made more difficult.

The level threshold for the 2014 Level 6 reading test was 24 marks, compared with 22 marks in 2013, but there are supposed to be sophisticated procedures in place to ensure that standards are maintained. We should be able to discount the third cause.

The second cause is also unlikely to be significant, since schools are strongly advised only to enter learners who are already demonstrating attainment beyond KS2 Level 5.There is no benefit to learners or schools from entering pupils for tests that they are almost certain to fail.

The existing pass rate was very low, but it was on an upward trajectory. Increasing familiarity with the test ought to have improved schools’ capacity to enter the right learners and to prepare them to pass it.

That leaves only the first possibility – something must have been wrong with the test.

Press coverage from May 2014, immediately after the test was administered, explained that it contained different rules for learners and invigilators about the length of time available for answering questions.

The paper gave learners one hour for completion, while invigilators were told pupils had 10 minutes’ reading time followed by 50 minutes in which to answer the questions. Schools interpreted this contradiction differently and several reported disruption to the examination as a consequence.

The NAHT was reported to have written to the Standards and Testing Agency:

‘…asking for a swift review into this error and to seek assurance that no child will be disadvantaged after having possibly been given incorrect advice on how to manage their time and answers’.

The STA statement says:

‘We apologise for this error. All children had the same amount of time to complete the test and were able to consult the reading booklet at any time. We expect it will have taken pupils around 10 minutes to read the booklet, so this discrepancy should not have led to any significant advantage for those pupils where reading time was not correctly allotted.’

NAHT has now posted the reply it received from STA on 16 May. It says:

‘Ofqual, our regulator, is aware of the error and of the information set out below and will, of
course, have to independently assure itself that the test remains valid. We would not
expect this to occur until marking and level setting processes are complete, in line with
their normal timescales.’

It then sets out the reasons why it believes the test remains valid. These suggest the advantage to the learners following the incorrect instructions was minimal since:

  • few would need less than 10 minutes’ reading time;
  • pre-testing showed 90% of learners completed the test within 50 minutes;
  • in 2013 only 3.5% of learners were within 1 or 2 marks of the threshold;
  • a comparative study to change the timing of the Levels 3-5 test made little difference to item difficulty.

NAHT says it will now review the test results in the light of this response.

 

 

Who is responsible?

According to its most recent business plan, STA:

‘is responsible for setting and maintaining test standards’ (p3)

but it publishes little or nothing about the process involved, or how it handles representations such as that from NAHT.

Meanwhile, Ofqual says its role is:

‘to make sure the assessments are valid and fit for purpose, that the assessments are fair and manageable, that the standards are properly set and maintained and the results are used appropriately.

We have two specific objectives as set out by law:

  • to promote assessment arrangements which are valid, reliable and comparable
  • to promote public confidence in the arrangements.

We keep national assessments under review at all times. If we think at any point there might be a significant problem with the system, then we notify the Secretary of State for Education.’

Ofqual’s Chair has confirmed via Twitter that Ofqual was:

‘made aware at the time, considered the issues and observed level setting’.

Ofqual was content that the level-setting was properly undertaken.

 

 

I asked whether, in the light of that, Ofqual saw a role for itself in investigating the atypical results. I envisaged that this might take place under the Regulatory Framework for National Curriculum Assessments (2011).

This commits Ofqual to publishing annually its ‘programme for reviewing National Assessment arrangements’ (p14) as well as ‘an annual report on the outcomes of the review programme’ (p18).

However the most recent of these relates to 2011/12 and appeared in November of that year.

 

 

I infer from this that we may seem some reaction from Ofqual, if and when it finally produces an annual report on National Curriculum Assessments in 2014, but that’s not going to appear before 2015 at the earliest.

I can’t help but feel that this is not quite satisfactory – that atypical test performance of this magnitude ought to trigger an automatic and transparent review, even if the overall number of learners affected is comparatively small.

If I were part of the system I would want to understand promptly exactly what happened, for fear that it might happen again.

If you are in any doubt quite how out of kilter the reading test outcomes were, consider the parallel results for Level 6 teacher assessment.

In 2013, 5,698 learners were assessed at Level 6 in reading through teacher assessment – almost exactly two-and-a-half times as many as achieved Level 6 in the test.

In 2014, a whopping 17,582 learners were assessed at Level 6 through teacher assessment, around 20 times as many as secured a Level 6 in the reading test.

If the ratio between test and teacher assessment results in 2014 had been the same as it was in 2013, the number successful on the test would have been over 7,000, eight-fold higher than the reported 851.

I rest my case.

 

The new regime

In February 2013, a DfE-commissioned report Investigation of Key Stage 2 Level 6 Tests recommended that:

‘There is a need to review whether the L6 test in Reading is the most appropriate test to use to discriminate between the highest ability pupils and others given:

a) that only around 0.3 per cent of the pupils that achieved at least a level 5 went on to achieve a level 6 in Reading compared to 9 per cent for Mathematics

b) there was a particular lack of guidance and school expertise in this area

c) pupil maturity was seen to be an issue

d) the cost of supporting and administering a test for such a small proportion of the school population appears to outweigh the benefits.’

This has been overtaken by the decision to withdraw all three Level 6 tests and to rely on single tests of reading GPS and maths for all learners when the new assessment regime is introduced from 2016.

Draft test frameworks were published in March 2014, supplemented in July by sample questions, mark schemes and commentary.

Given the imminent introduction of this new regime, together with schools’ experience in 2014, it seems increasingly unlikely that 2015 Level 6 test entries in reading will approach the 120,000 figure suggested by the trend.

Perhaps more importantly, schools and assessment experts alike seem remarkably sanguine about the prospect of single tests for pupils demonstrating the full range of prior attainment, apart from those assessed via the P-Scales. (The draft test frameworks are worryingly vague about whether those operating at the equivalent of Levels 1 and 2 will be included.)

I could wish to be equally sanguine, on behalf of all those learners capable of achieving at least the equivalent of Level 6 after 2015. But, as things stand, the evidence to support that position is seemingly non-existent.

In October 2013, Ofqual commented that:

‘There are also some significant technical challenges in designing assessments which can discriminate effectively and consistently across the attainment range so they can be reported at this level of precision.’

A year on, we still have no inkling whether those challenges have been overcome.

 

GP

September 2014