Notes on: Gillborn, D., Demack, S., Rollock, N. and Warmington P. (2017). Moving the Goalposts: Education Policy and 25 years ofthe Black/White achievement gap. British Educational Research Journal Vol. 43, No. 5, October 2017, pp. 848–874DOI: 10.1002/berj.3297

Dave Harris

This looks at secondary analysis of official statistics examining the changing scale of inequality of achievement between white and black (black Caribbean) students, from the introduction of the GCSE in 1988 to the murder of Stephen Lawrence in 2013. This is the longest period analysed. The gap has never been consistently high on the agenda. Educational policy has changed it negatively, and whenever the key benchmark for achievement has been redefined, race inequity has been restored and gaps widened. Throughout the period white students 'were always at least 1 1/2 times more likely to achieve the dominant benchmark'(848).

A research project was designed to explore the impact on education of the Lawrence murder and subsequent changes to legislation especially in the McPherson Enquiry, so they examined the black-white gap for the 25 years leading up to and found that the gap has remained consistent despite changes in education policy especially how standards have been measured and debated. This framework is informed by CRT, 'that uses race as a social construction... definition and deployment (policy and practice) is highly complex, contingent and fluid'(849) [citing Delgado and Stephancic and others, with an emphasis on 'business as usual forms of racism... every day taken for granted processes and assumptions ...that shape the interest of people identified as white'].

Deficit assumptions to explain the achievement gap have been increasingly challenged, for Latino students as well as black ones [and drastically for social class gaps, although these are not mentioned]. They are using the term gap while avoiding racist stereotypes and argue gap is 'a creation of educational policymakers and their preferred accountability strategies' (850). They measure achievement in relation to benchmarks set by government and applied by various stakeholders including the media, employers, parents, academics and H/FEIs. These are  useful measures to show inequalities in education opportunity and they also have an influence on opportunities afterwards.

They want to analyse the relative attainments of students who 'self identify as white British or black Caribbean according to the dominant census categories' (851). This might be criticised because other minority groups perform better than white peers such as students of Indian ethnic heritage. However, the debate initially was about blacks and whites, and the black Caribbean group 'retains huge significance educationally and politically… They are one of the most politically active of minority communities and… consistently among those achieving those results overall and most likely to be permanently excluded' [reference to a 2015 study]. Studies of multicultural education were mostly concerned with their experiences, as were campaigns for justice. Focus on black Caribbean students is also suggested by CRT which is often criticised for focusing particularly upon black people as emblematic, and CRT has developed various offshoots to include Latinx Native American and Asian-American people which in turn has led to work on the specificity of blackness ['BlackCrit']. Black Caribbean education is still seen as one which can 'speak directly to deep structures and processes of racial injustice in education' and is critical in resisting white supremacy, possibly even the '"fulcrum of white supremacy"' [more justifiable in the American experience because it's linked to slavery? Reference is American?].

A racial lens is prioritised despite arguments that there are clearly alternatives affecting inequality such as class or gender or disability, intersectionality. We must 'remain equally aware of the need to avoid becoming trapped in an endless pursuit of more and more interlocking categories and forms of analysis' (851) and Delgado himself has observed that intersection can paralyse progressive work because someone can always come along and '"point out that you forgot something"' [analysis is always much more complex than politics – This is a clear example of where politics simplifies the matter]. Fortunately, there are also 'limits of space'. At least the concept of intersectionality was 'coined by Crenshaw' (852), another CRT theorist [who wanted to include more people, especially black women], but since then it has 'often eviscerated any critical content and even acting to close down race critical analyses' [with reference to Gillborn 2015]   [oh dear, analysis ran away with him].

Research on race and racism and inequality has always been controversial, seen in the early Rampton Report and the political activity by black parents that it produced. The report used a special survey of six LEAs with about half of the minority ethnic pupils and the results were quoted prominently. The stark difference in achievement was revealed in that only 3% of West Indian students reached the desired level compared with 16% of all the leavers [which also included Asian students]. The Report said this was clear evidence that West Indian children were failing in our education system and called for action to combat racism especially in the form of lower expectations. There was to be systematic ethnic monitoring, but the government disregarded the report in major respects. The follow-up Swann Report used a second DES survey of five LEAs to show inequality of achievement  (6% of West Indians gained five or more high-grade results compared to 19% of all other leavers ) (853).

These were one-off special investigations that did not focus on white students. There were several other studies of limited ranges of schools or LEA's. The Youth Cohort Study was nationally representative, and it confirmed the patterns above, and is still useful. In the 1990s, school effectiveness was the focus of research looking at the impact of different schooling and management practices. This rarely considered race  or social justice as a variable, however, until an OFSTED report to mark the 10th anniversary of Swann. That used not just results but the statistics and concluded that African Caribbean students had not '"shared equally in the increasing rates of educational achievement… In some areas there is a growing gap"' (853). That led to a new 10 point plan [with a Hammersley 2006 reference here] but that concerned only the possibility of gathering more data and an ad hoc group.

A later OFSTED review in response to Lawrence drew 118 statistical returns from LEAs that had been funded under the Ethnic Minority Achievement Grant. Ethnic monitoring is absent in many areas. One third of LEAs did not have any quantitative data [strange you buggers should be regretting that?]. Overall though, black Caribbean students were not achieving as highly on average as their white peers, but there were regional variations and 'pockets of high achievement by black students, '"in one in 10 authorities"' [I wonder if this was gendered?].

In 2002 it finally became a national requirement to gather ethnically based data on the achievements of students in compulsory education, after further soul-searching from the Lawrence enquiry and the subsequent Race Relations Act of 2000. This enabled more detailed analysis including intersection of various factors 'including ethnic origin, social class, gender, special educational needs, student aspirations and parental education' (854). The research has been ambitious and often uses multiple regression trying to quantify the separate independent influence these factors. [Surprisingly] 'such research has been critiqued by critical race scholars who identify shortcomings in the conceptualisation of race and racism in traditional quantitative approaches' (855) [for the first time presumably?]. Quantitative research tends to treat ethnic origin as a causal factor rather than a social identity already associated with discrimination. Apple has argued that race is not stable but contingent and historical, a construction arising from social relationships [so?] Quantitative models are crude and mechanistic [multiple regression?] Racism is seen as 'something that can only be identified after other relevant factors have been removed from the data'[but this is a technique not a political priority? I admit it might make racism a residual category?]. Race discrimination intersects with and expresses itself in prior attainment and disproportionate placement in the lower ranks [I thought it was he was going to say in social class too. Doesn't this newly sophisticated discussion prevent comparative analysis or does it apply to the earlier stuff as well which renders invalid all those early findings too?]

Recent research has slid away from minority ethnic achievement to look at things like the attainment of students in receipt of free school meals, 'often used as a crude proxy for economic disadvantage', and the lower achievements of white students, who 'become the principal point of concern to political parties' (855). There is no broader arguments about working class whites and white students as a whole, sometimes seen in Daily Mail headlines like the betrayal of white pupils [so we can dismiss this research as propaganda]. It would be easy to assume that tables are been reversed and that black students no longer experience inequality when compared with their white peers. However 'the contemporary reality strongly echoes the patterns that used to generate headlines decades ago', reveals a uniquely long timescale and an historical relationship to policy [well, only if the objections in the paragraph above don't apply].

There is a lot more statistical data on achievement, a general trend towards outcomes, performance and public management based on qualitative data. However the quality of the material still in doubt. Ethnic diversity rarely features on the educational agenda, say in national school performance tables, although it is now mandatory [since 2002, although school tables began in 1992]. We only have 'the best available national data' (856).

YCS is a series of longitudinal surveys following cohorts of young people for three years after compulsory schooling recording achievements and experiences and activities outside of education and employment. The first one began in 1985 and ran until the 2000 when a new survey was introduced at all the Longitudinal study of Young People in England [LE], which first ran alongside and has now replaced the old one. Now funded by the ESRC. They have summarised the changes in achievement at the end of compulsory schooling, following the introduction of GCSE [a table on page 856 shows sample size rising from 8000 to 14,000 --not sure why --bullshit?n].

They also looked at the National Pupil Database which has a range of statistical sources relating to school-age students and young people. They looked at those relating to students at the end of compulsory schooling, key stage 4. Data is available from 2004 onwards when a more detailed ethnic breakdown was provided, better than the YCS which used only a few ethnic categories. If we combine the data with YCS [assumptions here of course] we get the best available picture — but changes in ethnic classification mean that the groupings are not the same, for example the category black was modified to be more precise in later work. There are also differences in geographical and educational coverage, with early reference to England only, and differences in terms of whether private schools were attended.

They want to debate the difference between quantitative and qualitative data, especially question the measures by which standards are judged. They think changes in the way that achievement is benchmarked will have 'clear and iniquitous impact on the black/white gap'.

The GCSE already used the system of tiered exam papers, in effect a dual status system seemingly a single examination which few parents and students knew about. 'Tiering decisions tend to exacerbate social inequalities' (859), especially gender, and black students were overrepresented in foundation tiers. Academic success was usually viewed as five or more higher grade passes, A to C, later A* -C, and performance tables were ranked according to that criterion. Using the original benchmark, by 2013 the black-white gap had virtually disappeared, with 83% of white British students achieving it and 80% of Black Caribbeans, diminishing from 22 percentage points in 1993. However there was considerable fluctuation. It is not until 2006 that we see a steady reduction in the gap.

The most dramatic and consistent reductions in inequality coincide with the new benchmark [oddly they argue that this was when 'the measure itself became less important' (860)]. They admit that 'we cannot know whether the gap would have shrunk so dramatically had the original benchmark remained in place. However at least we know that 'there is no inherent reason why black students cannot attain on a par with their white peers'.

In 2005 the benchmark now included the five higher grades which must include success in English and maths — the '"gold standard"', intended to be more demanding and leading to a fall in overall achievement rates. Here there was inequality again with a black-white gap growing to 15 percentage points. Both English and maths used tiered examination papers which could have produced this effect. As schools adjusted, overall achievement increased year on year and 'both white British and black Caribbean students shared in this trend. Indeed, the rate of improvement was somewhat greater for black Caribbean students and as a result the black-white gap narrowed from 15 percentage points in 2006 to 11 percentage points in 2010' (861). A changing government meant another change in the benchmark and again an immediate negative impact on race equality.

Cameron's government expected a broad academic education to age 16, a new English baccalaureate, GCSEs at grade C or above in English maths sciences humanities and a foreign or ancient language. It was supposed to help disadvantaged students who were excluded from higher status GCSE subjects, including kids eligible for free school meals and ethnic communities. There are obvious difficulties with high status curriculum areas. Levels of attainment fell dramatically in 2011 for both white British and black Caribbean students — 'an Ebacc penalty' (862) and the greatest proportional impact was experienced by black Caribbean students.

So in terms of the original achievement measure the black-white gap almost disappeared, but the gap changed and changed in scale. We can look at the odds ratio calculations, comparing white students chances of achieving the benchmark in relation to black Caribbean peers, and despite fluctuations, the overall pattern is that 'white students are always at least one and 1/2 times more likely to attain the dominant benchmark' (864), ranging from an odds ratio 2.84 in 1993, and several times when the odds ratios was greater than two. So '"raising the bar"' widens race inequality, especially the introduction of the EBacc. This re-stored historic rates of disadvantage, 'seven years of improvement was wiped away'.

The impact of educational policy is difficult to estimate because it can take more than a decade for children to move through schooling [and, as we've seen, schools can respond]. Nevertheless, there is little support for the view that policies are having a positive impact on race inequality. There has been much optimistic 'gap talk' in official statements, and optimism based on 'apparently impressive percentage improvements (measured over a brief timescale)' (865) [these are quoted but I would have welcomed more information about them and where they came from — they seem to be Department for Education data]. There's even been a deterioration masked by 'marginal periodic gains', and in some case active widening.

Cameron's government abolished Equality Impact Assessments in 2012, just as his latest educational reform was widening the achievement gap. Their analysis 'suggests that this was not a one-off mistake or aberration' (867).

[Amusingly] 'These differences in achievement are not meaningless statistical artefacts; they indicate genuine inequities that will have lasting consequences for Black young people' (868). There have been huge improvements in educational attainment, in absolute terms for both white and black students — white students' chances of achieving five higher grade GCSEs are '11 times greater than in 1988; black Caribbean students are 18 times' [but this is the old measure]. New benchmarks have had a regressive effect. [Assuming they are taken seriously]. The overall conclusion is that 'negative impacts are much more certain and predictable'. It has been moving the goalposts that has maintained the black-white achievement gap.