The Department for Education released this press release on 20th March 2020 (click here).

Synopsis

2020’s summer exam series, including A levels, GCSEs and other qualifications, and all primary assessments, have been cancelled as we fight to stop the spread of coronavirus.

The exam regulator, Ofqual, and exam boards will work with teachers to provide grades to students whose exams have been cancelled this summer, following our actions to slow the spread of coronavirus.

What is the plan? GCSE, A and AS level examinations will be awarded based on the work that pupils and students have already put in; this will done through a process of teacher assessment and evidence gathering, as the DfE state:

“The exam boards will be asking teachers, who know their students well, to submit their judgement about the grade that they believe the student would have received if exams had gone ahead.”

"The exam boards will be asking teachers, who know their students well, to submit their judgement about the grade that they believe the student would have received if exams had gone ahead." Click To Tweet

Teachers will be taking into account the range of evidence they have at their disposal including,

• Coursework submitted
• Teachers Assessment (judgement/prediction)
• Mock Examinations/Other internal Assessments

Then exam boards will look at this data along with the prior attainment (Key stage 2 SATS) and calculate a grade for each student. The calculated grades are due out before the end of July. These grades and certificates will be indistinguishable from other years so that this year’s students do not face a systemic disadvantage as a consequence of these extraordinary circumstances.

Students that feel they have been treated unfairly as a result of the procedures may appeal through a set procedure, and if they wish sit examination early when schools return in September 2021. From a school’s point of view, the Government will not publish any school or college level educational performance data based on tests, assessments or exams for 2020.

What are Stereotypes?

Stereotyping is the tendency to draw on overly simple beliefs about groups to make judgments about individuals. Confirmation bias is the tendency to perceive and seek out information that confirms one’s pre-existing beliefs, and avoid information that conflicts with those beliefs (Nelson, 2014).

Sterotyping is described by Kahnemann as ‘Anchoring’; This occurs when value is determined for an ‘unknown quantity’ before estimating that quantity. Anchoring is a ubitquitous and a perfectly natural human response; unfortunately this is an implicit process and as a result when it emerges through assessment it can be extremely problematic (Kahneman, 2013). I would reiterate here that anchoring and bias on the whole are completely natural processes.

Flow chart

 

A stereotype is introduced. Black and poorer people are less intelligent, aggressive and lazy when compared to the majority white population or richer people – these types of stereotypes appear during the first five years of life. This introduces two cognitive biases, the focalism (also know as the anchoring bias) and this is where the confirmation bias starts.

Focalism or the anchoring bias occurs when people are wedded to the constructs they encounter first. If I were to offer to sell you a board marker for £10 today, tomorrow I attempt to sell you the same product for £15. You would think that I am ripping you off. If I offer you the same board marker at £20 today, and then tomorrow I offer it to you for £15, you would think that you were receiving an enhanced price.

The only difference here is the order in which you hear the information. The price you received has not changed. Hence you anchor yourself depending on the first bit of information. In this case it is regardless of any other information i.e. is the board marker good value for money? What is the quality of the board marker?

In confirmation bias, once a bias is introduced people actively seek to confirm this bias. For example: Black people are more aggressive so toxic associations between black people and aggression, weapons, etc are formed. As you can see from the flow chart, the whole process is cyclical and concentric.

Everyone can be prone to these biases including academics. Both confirmation bias and stereotyping leads to over simplification such as ‘women are more risk adverse’. These assumptions are not well-informed in an empirical sense, but have been described repeatedly in economic literature as robust (Nelson, 2014).

“Many are grounded in the premises that stereotypes comprise invariant, homogenous, evaluative judgements of a given group (e.g. income, gender or ethnic group), and that stereotypes enable judgements of group members to be made quickly and with cognitive ease” (Hilton and von Hipple, 1996; McGarty et al., 2002. In Campbell 2015, p1).

Bias is commonly refered to as a habit of the mind; your brain is designed to skip information and rely on information and subsequently associations are quickly made. Therefore it is important to recognise that stereotypes are not an individual process rather they are systemic in nature and the consequences are systematic. Stereotyping processes respond to systematic principles that generalize across different specific instances of stereotypes, these processes are consistent over time, place and out group. (Fiske et al 2002) This is not about you as an individual, this is about how you have appropriated stereotypes through associations.

Bias is commonly refered to as a habit of the mind; your brain is designed to skip information and rely on information and subsequently associations are quickly made. Therefore it is important to recognise that stereotypes are not an… Click To Tweet

Like all dimensions of human psychology and the social sciences, none of this should be reduced to anecdotes or simple rules. Yes, people of colour are disadvantaged through our society however the intersections of complexity should not be ignored. This brings in inter-group dynamics which are often observed through people holding positive and negative stereotypes – pity targets the warm but not competent subordinates (rich to poor); envy targets the competent but not warm competitors (White to American Asian); contempt is reserved for out-groups deemed neither warm nor competent. (Fiske et al 2002).

Teaching and Stereotypes

By stereotyping, teachers can make judgements of pupils quickly and with cognitive efficiency through preconceived associations about the ability/attainment of lower socio-economic background, SEN, Black Caribbean pupils, and so on.

“The possibility, therefore, is that among the English teaching profession there exist normalised notional templates of pupil attainment, which are premised on pupil characteristics, inform judgements of each child, and skew assessments in line with these characteristics.” (Campbell, 2014. p519).

Teacher assessment, on the whole, is not reliable. Let’s start with some examples. This piece will concentrate mainly on race, however, the same pattern exists for disadvantaged pupils and the intersections between identities (for another day). Educators will be familiar with the terms disadvantaged, BAME, gender groups, FSM and multitude of other labels; the rhetoric in last decade has been about closing the attainment gap between these. However, through all of the key stages and phases, have educators been looking at the gap from a false angle?

Educators will be familiar with the terms disadvantaged, BAME, gender groups, FSM and multitude of other labels; the rhetoric in last decade has been about closing the attainment gap between these. Click To Tweet

That attainment indicators depend so heavily on teacher assessment invites the question of whether these apparent achievement gaps may to some extent be an artefact of the measurement method used. There is an enduring body of evidence which indicates that teacher assessments are subject consistently to a large and significant level of error (Brookhart, 2013; Eckert et al., 2006; Harlen, 2005) … and, more importantly, research also indicates that some of this error may be systematic (Harlen, 2005; Robinson and Lubienski, 2011) (Campbell, 2015, p518).

Burgess and Greaves (2009) look at the teacher assessment versus actual attainment of external exams of 11-year old across 16557 schools, 3 subjects and 4 years. This showed that the past performance of a specific ethnic group directly impacted on the current teacher assessment.

It is worth stating before we go on that key stage 1 SATs are completely based on teacher assessment and are marked in school. Key stage 2 and 3 SATs are, and were marked externally through a process formally known as quasi-blind i.e. with the names of the pupils known to the assessor.

Key stage 2 and 3 SATs are, and were marked externally through a process formally known as quasi-blind i.e. with the names of the pupils known to the assessor. Click To Tweet

The way society is structured you would expect external markers’ implicit bias to lower the marks and outcomes of pupils of colour. This may skew results positively towards white pupils and those with euro-centric names as they do not have the benefit of an intimate daily knowledge of the pupils themselves.  

However, precisely the opposite was found; pupils of certain groups were found to be assessed lower than others by their own teachers rather than external examiners.

English Key Stage 2

Ethnic group	TA < External Exam	Difference compared to White pupils	Percentage discrepancy compared to White pupils
White	12.4%	0.00%	0.0%
Black Caribbean	17.2%	4.80%	38.7%
Black African	18.3%	5.90%	47.6%
Pakistani	20.2%	7.80%	62.9%
Bangladeshi	18.1%	5.70%	46.0%
Indian	13.8%	1.40%	11.3%
Chinese	13.3%	0.90%	7.3%

In English, all pupils who do not ascribe to the white label have a higher percentage of teachers assessing them lower than via the external test. The rate at which Pakistani pupils are under-assessed (vs their external SATS grade) when compared relatively with white pupils is at a rate of 62.9%. That means you are 62.9% more likely for your teacher to think you are working at a level below which you are actually are.

Campbell (2015) found that pupils had a lower probability of their teachers rating them ‘above average’ for reading for lower income pupils, boys, pupils with SEN, pupils of all Bangladeshi, Pakistani, Black Other, Black African and Black Caribbean, male and EAL pupils.

Campbell (2015) found that pupils had a lower probability of their teachers rating them ‘above average’ for reading for lower income pupils, boys, pupils with SEN, pupils of all Bangladeshi, Pakistani, Black Other, Black African and… Click To Tweet

This has obvious implications for pupils of colour, the impact of the Golem effect being pertinent; day to day their teachers have lower expectations of them. Similar patterns (but exactly the same ethnic grouping) with Maths and Science were also found (please do head to this previous piece for further analysis).

What is the impact of Key Stage 2 predictions on the pupils?

Wilson et al (2006) show the difference in Key Stage 2 teacher assessment extrapolated to GCSE, versus actual grades at GCSE using white pupils as the comparison. This is impact is significant as the figures for most group are vastly different and with Black Caribbean British and Black Other British pupils bucking the trend. Which is interesting in itself.

This leads us to the secondary schools setting.

“Research not only shows some persistent patterns of poor educational outcomes for pupils from low socio-economic groups, minority ethnic groups and boys, studies also indicate differentiated experiences of schooling and the over-representation of these groups in low attainment sets” (Gillborn and Mirza 2000; Demie 2001; Singh Ghuman 2002; Mamon 2004; Connolly 2006) (in Muiji and Dunne, 2010, p393).

Muijs and Dunne (2010) analysed setting in Maths and English across 12 local authorities in England and randomly sampled 100 schools and completed a quantitative analysis of the composition of their sets.

Muijs and Dunne (2010) analysed setting in Maths and English across 12 local authorities in England and randomly sampled 100 schools and completed a quantitative analysis of the composition of their sets. This is what they found? Click To Tweet

They found that white students are statistically significantly over-represented in high sets and conversely that Black student are statistically under-represented in higher sets. Across the Asian subset the only significant result is Bangladeshi pupils being under represented in the high sets.

	Sample	Low sets	Middle sets	High sets
White	55.3	57.4	48.9	61.5
Bangladeshi	18.3	18.7	23.9	12.1
Pakistani	10.3	9.1	10.2	11.2
Indian	7.1	6.6	6.4	8.2
Black Caribbean	6.0	6.9	6.5	4.3

Wilson et al (2006) presented work from the CLASS thinktank. There are various reasons that black boys spike at 16 years and some of them are covered later – I would be remiss if I were not to point out that the GCSE exams are the first time that pupils are assessed completely anonymously.

Answers on a postcard. Why is there a spike at the 16 for Black pupils? Click To Tweet

University Predictions

16% of A Level pupils achieve their predicated grades and 75% of pupils are over-predicted, with the average student being over-predicted by 1.7 grades. When we look at socio-economic status, pupils from the lowest group who achieve the same grades are under-predicted in comparison, as are higher achieving pupils from state schools in comparison with private school pupils. (Murphy and Wyness 2020).

After describing the impact of the Pygmalion effect earlier, being under-assessed has a day to day impact on pupils with a low socio-economic status (Campbell, 2015; Murphy and Wyness, 2020; Burgess and Greaves, 2005).

After describing the impact of the Pygmalion effect earlier, being under-assessed has a day to day impact on pupils with a low socio-economic status (Campbell, 2015; Murphy and Wyness, 2020; Burgess and Greaves, 2005). Click To Tweet

High achieving, but under-predicted candidates are 10 percentage points less likely to apply to the most selective universities, and 6.9 percentage points more likely to enrol in a university in which they are over-qualified (have grades higher than their fellow peers). It should be noted that the nature of the university attended has been shown to be linked to a students’ eventual earnings (Belfield et al, 2018, in Murphy and Wyness, 2020).

Overall, Black pupils had the lowest percentage accuracy with only 39.1% of grades accurately predicted. This group also had the highest over- and under-prediction rates (53.8% and 7.1% respectively). This under-predicted statistic may seem interesting and to a certain extent seem to cancel out the effects of under prediction, we have to remember you cannot over predict a pupil who achieves an A grade (and now an A*).

Black British pupils of all three categories, Caribbean, African or Other (within the 2009 chort) are predicted A grades a minimum 12% percentage less than white students (this trend is similar for Bangladeshi and Pakistani pupils). This is also mirrored in achievement, with Black British Caribbean and Black African students gaining the lowest rates of A grade (in 2009) at 15.2% and 17.9% with similar rates for Bangladeshi and Pakistani pupils. (Everett and Papageorgiou, 2011).

This is also mirrored in achievement, with Black British Caribbean and Black African students gaining the lowest rates of A grade (in 2009) at 15.2% and 17.9% with similar rates for Bangladeshi and Pakistani pupils. (Everett and… Click To Tweet

Looking at socio economic background, Everett and Papageoriou (p. 20) found that:

‘The Higher managerial group had the highest percentage of accurate predictions of grade A at 69.5% while the Routine group had the lowest percentage at 53.4%. Grade C was the least accurately predicted for all social classes, apart from the Routine group which saw B grade prediction to be the most inaccurate (35.8%).’

This level of inaccuracy seems to pervade age as shown in Muijs and Dunne (2010) study of Year 7 pupils. When looking at attainment via setting, they found that pupils from higher socio-economic status (analysed through the use of ACORN and free school meals measures) are more likely to be assigned to higher sets and less likely to be assigned to lower sets.

The model of pupil self efficacy, pygamalion and teacher bias through stereotype seems to be an unending cycle. Pupils from lower socio-economic backgrounds are less likely to attend a high tariff university than their richer counterparts, even when they have similar prior academic attainment (Wyness 2017). After accounting for degree class, graduates from richer socio-economic backgrounds are more likely to become a professional and consequently higher earnings growth. (Crawford et al, 2016).

What are the possible reasons for this, other than the stereotype model (Behaviour and perception, confirmation and anchoring bias) suggested by Burgess and Greaves 2009? Well, within the same study Burgess and Greaves note that when teachers who assess pupils are the same teachers that teach them,  this will cause issues such as teaching to the test. This leads to teachers focusing on certain pupils, as they know that teacher-pupil day to day interactions can improve the external grade with the greater transmission of knowledge (T Campbell, 2015).

Sievertsen (2019) in Murphy and Wyness (2020) studied a reform which led to some students’ grades being recoded and ultimately their GPA (Grade Point Average) being downgraded. They found that students who were downgraded by the recoding performed better on subsequent assessments, indicating the importance of incentives. The only reason to keep the current system is that these students react more positively to predicted grades, although this incentivising could be problematic on a number of levels. (Murphy and Wyness 2020). This does not fit stereotype model but heralds questions around high stakes accountability around exams for teachers and leaders alike.

Wilson et al (2006) investigate non-stereotype factors on two national cohorts after delivering and analysis across four key stages:

1. Ethnic minority pupils go to worse schools than their white counterparts.
2. Language acquisition of EAL pupils, meaning pupils catch up as their language does.
3. Does the disadvantage start in the early years so school has little impact on the result?

Ethnic Minorities go to Worse schools

Wilson et al’s analyses of their results were found to be pervasive (across the whole school cohort). Ethnic minorities do not segregate into any schools nationally.

‘Over the country as a whole, attendance at substantially mono-ethnic schools is not the norm for members of the non-white groups (though it is for whites in many areas). Half of all non-white secondary students in England attended schools where more than 75 per cent of the total enrolment comprised whites’ (Burgess et al, 2004, p1)

As pupils of colour go to wide range of schools in the main this not a huge factor in the phenomenon.

Language Acquisition

This may be a factor, due to the fact that as pupils who speak English as an additional language become more and more proficient, their attainment in school assessment increases. Demie and Strand (2006) found in their paper which looked at 10 secondary schools in the inner London borough of Lambeth that bilingual pupils who are proficient in English performed better than their than solely English speaking peers on average (although their finding was not statistically significant).

Both Demie and Strand (2006) and Wilson et al (2006) both found that pupils make better than expected progress during key stage 3 and 4. Wilson et al interestingly found that pupils of colour with English as an additional language were found to have on average around 0.9 extra GCSE points; both EAL and non EAL pupils of colour gain 3 GCSE points when compared to the white population. Wilson et al (2006) suggest that the impact of language acquisition accounts for up to a third of ethnic minority rise at Key Stage 4. There are still 2 thirds unaccounted for.

Ethnocentric Testing

Yes, there are arguments around ethnocentric testing typically against Black students and poor students (Gipps, 1992; Murphy & Pardaffy, 1989 in Burgess 2009). I wholly agree there are issues with the content of external assessment, however, this would lead to teachers inflating pupils over the external examination. This is not the case as the converse is true.

Behaviour and the Perception of Behaviour

Muijs and Dunne (2010) completed a quantitative analysis of the composition of classes. Leaders were presented with a qualitative survey about the rationale behind their setting choices.

English

Words in Rationale	Percentage of Surveyed
Ability	22.7%
Attainment or Test Results	72.8%
Attitude and Behaviour (+Other)	4.5%

Maths

Words in Rationale	Percentage of Surveyed
Ability	11%
Attainment or Test Results	88.7%
Attitude and Behaviour (+Other)	2.3%

Even within schools there are inter-department variations on the basis of setting. This raises questions about bias being introduced through an ideological basis. Bearing in mind that ability (through teacher assessment), Test results and attitude and behaviour (teacher perception) are all prone to bias. Attitude and behaviour are critical in the day to day experiences of pupils and teachers in the classroom.

Can the behaviour of pupils impact on teacher assessment? We know that a disproportionate amount of British Black Afro Caribbean pupils are diagnosed with SEMH (Social, Emotional and Mental Health needs); a social construct which judges pupils’ behaviour against a standard set of expectations.

“A frequently proposed explanation for the over-representation of Black pupils with SEMH/MLD is an inappropriate interpretation of ethnic and cultural differences including teacher racism, low expectations and a failure of schools to provide quality instruction or effective classroom management” (e.g. Artiles et al, 2010; Waitoller et al, 2010 in Strand and Lindorf 2018).

Even when controlling for socioeconomic difference (this only explained 50% of the disparency) Strand and Lindorff (2018) found that Black Caribbean students have an odds ratio of 2.29 and mixed White and Black British Caribbean student have a slightly lower ratio of 1.94; this means that if you are a Black Britsh Caribbean student you are 2.29 times more likely to be diagnosed with a SEMH than a white British student. 

“Is it that these young people from this [Black British Caribbean] ethnic group are more confrontational with their teachers because of gang culture or is it a perception of their behaviour?”

(Professor Strand, BBC, 2019.)

"Is it that these young people from this [Black British Caribbean] ethnic group are more confrontational with their teachers because of gang culture or is it a perception of their behaviour?" (Professor Strand, BBC, 2019.) Click To Tweet

Behaviour (well, perception of behaviour) is impacted along racial lines and this is likely through the process of toxic association and positive prototypes (more here). Is there evidence that this impacts on the teacher assessment?

‘We divide the 61 neighbourhood types into the poorest third, middle (omitted category) and least poor thirds, and introduce indicators for these in the analysis. The results suggest the same factors at work’. (Burgess and Lindoff, 2009, p. 16).

This is reinforced when pupil bodies groups in schools are in smaller numbers, as this means that the association has not been built from numerous interactions but from wider society. These stereotypes are standalone and intersectional; hardly surprising considering the nature of toxic stereotypes.

What behavioural factors are involved in under-assessment looking at self reporting from pupils?

1. Reporting praise from your teachers is significantly negatively correlated with the probability of under assessment in all subjects.
2. Pupils who report working hard and like school also less likely to be underassessed (but there is not a statistical significance here)
3. Causing trouble in more than half of their classes leaves you 3.5% more likely to be underassessed.

Suggested from non-age specific analysis (Burgess, 2009).

The survey data suggest that non-academic factors impact teacher assessment and adverse behaviours have a greater impact on teacher assessment.

“In summary, whilst the survey data shows that student behaviours and attitudes do have an influence on the likelihood of under-assessment, such adverse behaviours are if anything more common among white pupils”.(Burgess 2009, p.23).

“In summary, whilst the survey data shows that student behaviours and attitudes do have an influence on the likelihood of under-assessment, such adverse behaviours are if anything more common among white pupils”.(Burgess 2009, p.23). Click To Tweet

There is no open and shut answer. But is there a systematic underassessment of pupils of colour and those of lower socio-economic background? Yes. Is there a bias when assessing pupil behaviour? Yes; of course the factors involved here are complicated, and Burgess describes the conclusions as having a two fold impact:

1. Pupils disengaging from school as a result of teacher interactions as they feel under valued

2. This undervalue reciprocating by under valuing education and qualifications (as per the earlier flow chart).

Critics of setting, point to the possibility of harming a student’s self-concept when that student is put into a lower set, to the fact that it is extremely difficult for teachers to have high expectations of low set classes, and to the loss of opportunities for lower-achieving peers to be peer-taught by higher achievers in the subject (Muijs and Reynolds, 2005. in Muij and Dunne 2010). 

The stereotype model is based on a wider view of the factors and as you can see both of these factors are cyclical. I would as a practitioner like to add that teacher expectations can contribute hugely to pupil achievement and so can the converse, as per the Rosenthal and Golem effect again; all feeding into a cycle where certain pupils do not feel valued, teacher don’t expect as much, the pupils disengage and the teacher is vindicated in their original stereotype through an anchoring and confirmation bias.

That is hard data, research and literature on teacher assessment being biased heavily, and this is maybe caused by the perception of behaviour. This already has an impact on the self efficacy of pupils leading to disproportiante exclusions.

Intersectionality – Multiple Disadvantages

1. Special Educational Needs
2. Eligibility for Free School Meals within the past 6 years
3. Maternal qualifications
4. Parental engagement in the young person’s education (based on whether they discussed school reports, attended parents’ evenings and talked about the young person’s plans for future studies).
5. A measure of the relationship between the main parent and child based on the frequency of arguments between them.
6. Whether the young person had access to an internet connected computer.
7. Ofsted rating of the young person’s school

These were defined as the linear factors by the EEF in their report into multiple disadvantages. The impact of each factor is shown below.

There is a linear relationship between the factors, so having more means you have a greater attainment penalty as a result. Although race is not disaggregated into the report, it does state that,

‘Young people whose ethnic group is Pakistani were more likely than White students to have three or more disadvantages and two or more disadvantages. Young people from the Bangladeshi, Black African and Black Caribbean ethnic minority groups were also more likely than White students to have two or more disadvantages.’

It could be argued that every factor could be exacerbated by a teacher stereotype model.

What is in Current Legislation?

The broad range of elements include teacher assessment and prior attainment. These are both biased against pupils of colour and disadvantaged students. What does the law actually say about discrimination? Well, the Equalities Act 2010 states that race is a protected characteristic, and any public sector body is bound by the public sector equalities duty. This means that any public sector body should not discriminate directly, indirectly, harassment or victimisation against groups which hold the protected characteristics. Indirect discrimination is normally enacted through the medium of policy, criterion or procedure (PCP).

Socio-economic status is not a protected characteristic, however, section 1.1 of the Equalities Act 2010 clearly states that.

“An authority to which this section applies must, when making decisions of a strategic nature about how to exercise its functions, have due regard to the desirability of exercising them in a way that is designed to reduce the inequalities of outcome which result from socio-economic disadvantage.”

Unfortunately, consecutive governments have failed to bring the section into commencement.

I know I may be jumping the gun here and being overly cynical as the DfE guidance suggests that a fair and robust process will be developed. The fact this guidance was released late afternoon on Friday, when the Secretary of State said:

‘We have a preferred method for examinations but that will be released tomorrow on Radio 4 on Thursday 19^th March 2020.’

Was that delay to stem out roar? To stop a coordinated counter protest? Is the guidance deliberately vague for a reason? One thing is for sure, this conversation needs to start.

The Government Guidance and what can School leaders do?

“How will you address the fact that students from disadvantaged backgrounds are more likely to have their grades under-predicted?”

We are not awarding students their predicted grades. Ofqual, the independent qualifications regulator, will develop a fair and robust process that takes into account a broad range of evidence, including teacher assessment and prior attainment. Ofqual will make every effort to ensure that the process agreed does not disadvantage any particular group of students.

Pupils who do not feel their calculated grade reflects their performance will have the opportunity to sit an exam, as soon as is reasonably possible after schools and colleges open again.”

The immediate concern is the grading for cohorts in year 11 and 13 and this is an urgent concern. The only controllable factor is teacher assessments, as predicted grades, schools and pupils’ prior data is out of the locus of control.

I am currently in discussion around assessment models and universities admissions around an equitable approach. I’ll keep you updated.

If you enjoy this piece of writing please do consider support us here.

References

https://www.bbc.co.uk/news/education-47240580

http://www.education.ox.ac.uk/wp-content/uploads/2018/08/Executive-Summary_2018-12-20.pdf

https://www.gov.uk/government/publications/coronavirus-covid-19-cancellation-of-gcses-as-and-a-levels-in-2020/coronavirus-covid-19-cancellation-of-gcses-as-and-a-levels-in-2020#what-will-happen-to-those-who-have-already-done-some-non-exam-assessment

https://gala.gre.ac.uk/id/eprint/16148/19/16148%20HARTLEY_Stereotype_Threat_Account_2013.pdf

http://eprints.lse.ac.uk/6245/1/The_Dynamics_of_School_Attainment_of_England%27s_Ethnic_Minorities.pdf

School segregation in multi-ethnic England Burgess et al 2004

http://www.bristol.ac.uk/media-library/sites/cmpo/migrated/documents/wp92.pdf Published in Ethnicities 4 (2) 237-265 (2004)

Demie and Strand 2006

https://www.researchgate.net/profile/Steve_Strand2/publication/237820937_English_language_acquisition_and_attainment_in_secondary_schools/links/0046353459922637e3000000/English-language-acquisition-and-attainment-in-secondary-schools.pdf

Daniel Muijs & Mairead Dunne (2010) Setting by ability – or is it? A quantitative study of determinants of set placement in English secondary schools, Educational Research, 52:4, 391-407, DOI: 10.1080/00131881.2010.524750

Campbell, T. (2015) Stereotyped at Seven? Biases in Teacher Judgement of Pupils’ Ability and Attainment. Journal of Social Policy, 44:3, pp 517-547. http://dx.doi.org/10.1017/S0047279415000227

Richard Murphy and Gill Wyness Working Paper No. 20-07 March 2020

Minority Report: the impact of predicted grades on university admissions of disadvantaged groups, Centre for Education Policy and Equalising Opportunities (CEPEO)

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/32412/11-1043-investigating-accuracy-predicted-a-level-grades.pdf

Department for Business, innovation and skills. BIS, RESEARCH PAPER NUMBER 37Investigating the Accuracy of Predicted A Level Grades as part of 2009 UCAS Admission Procese JUNE 2011

Everett & Papageorgiou (2011), ‘Investigating the Accuracy of Predicted A Level Grades as part of 2009 UCAS Admission Process’

Claire Crawford, Paul Gregg, Lindsey Macmillan, Anna Vignoles, Gill Wyness, Higher education, career opportunities, and intergenerational inequality, Oxford Review of Economic Policy, Volume 32, Issue 4, WINTER 2016, Pages 553–575, https://doi.org/10.1093/oxrep/grw030

Julie A. Nelson (2014) The power of stereotyping and confirmation bias to overwhelm accurate assessment: the case of economics, gender, and risk aversion, Journal of Economic Methodology, 21:3,211-231, DOI: 10.1080/1350178X.2014.939691

Kahneman, D. (2013), Thinking Fast and Slow.

Lessof et al (2019), Multiple disadvantage and KS4 attainment: evidence from LSYPE2,

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/808591/LSYPE2_multiple_disadvantage_and_KS4_attainment.pdf accessed 29th March 2020

Wyness (2017), ‘Disadvantaged students and the university admissions process’

https://www.suttontrust.com/wp-content/uploads/2017/12/Rules-of-the-Game.pdf accessed 29th March 2020