EA/2021/0234
Case Reference: EA/2021/0234
First-tier Tribunal
General Regulatory Chamber
Information Rights
Heard by: Cloud Video Platform
Heard on: 23 March 2022
Written submissions received: 20 & 27 May 2022
Decision given on: 14 June 2022
Before
Judge O’Connor – Chamber President
Tribunal Member Matthews
Tribunal Member Yates
Between
Office of Qualifications and Examination Regulation
Appellant
and
Information Commissioner
Respondent
and
Secretary of State for Education
Interested Party
Representation:
Appellant: S. Kosmin and K. Hafesji of Counsel
First Respondent: K. Taunton of Counsel
Second Respondent: N. Patel of Counsel
Decision: The appeal is DISMISSED
1
EA/2021/0234
REASONS
Preamble
1. This appeal was heard remotely, without objection from the parties, using the
Cloud Video Platform. There was no indication from the parties during the
hearing, or thereafter, that the mode of hearing led to an inability to participate
fully and effectively in the proceedings.
2. This document is the open decision on the appeal and may be disseminated and
published without hindrance. Although there was a closed bundle before the
Tribunal, the hearing of the appeal on 23 March 2022 remained in open session for
its duration, and there is no closed annex to this decision.
Introduction
3. The Freedom of Information Act 2000 (“FOIA”) provides for a general right of
access to information held by public authorities. That right is subject to exceptions
and exemptions. It makes provision for its enforcement by the Information
Commissioner (“ICO”) and for a right of appeal from a decision of the ICO to the
General Regulatory Chamber of the First-tier Tribunal (“FtT”).
4. This decision considers an appeal brought by the Office of Qualifications and
Examination Regulation (“Ofqual”) against the ICO’s Decision Notice (referenced
FOI2021255) dated 5 August 2021, in which the ICO required Ofqual to disclose,
within 35 days of the date of the Decision Notice, information requested by Pedro
Hedro (“the Requested Information”). Although notified of these proceedings,
Pedro Hedro (“the requestor”) did not seek to join as a party and took no part in
this appeal.
5. The Requested Information comprises of a list of education centres in England and
shows the percentage of Centre Assessed Grades (“CAGs”) in each education
centre that were adjusted following application of what is known as, ‘the
Algorithm’ – a standardisation model. In particular, the list identifies the
percentage of CAGs that were adjusted down by 3 or more grades, down by 2
grades, down by 1 grade, up by 1 grade, up by 2 grades, and up by 3 or more
grades, as well as the percentage of grades where there was no adjustment.
6. The Appellant, Ofqual, is a non-ministerial governmental department, which
regulates qualifications, examinations, and assessments in England. Ofqual’s
statutory objectives are set out in section 128 of the Apprenticeships, Skills,
Children and Learning Act 2009 (“ASCL Act”). It reports directly to Parliament.
The interested party (“DfE”) is responsible for children’s services and education
and works closely with Ofqual in this regard.
Factual Background
7. The following factual matrix is not in dispute.
2
EA/2021/0234
8. On 18 March 2020, the then Secretary of State for Education, Gavin Williamson MP
(“the Secretary of State”), made a statement in the House of Commons on changes
to the operations of educational settings as a result of the coronavirus (COVID 19)
pandemic. He announced that, as from 20 March 2020, schools and colleges would
be closed to most students until further notice. He said: “I can confirm that we will
not go ahead with assessments or exams, and that we will not be publishing performance
tables for this academic year.”
9. The DfE published a press notice on 20 March 2020 setting out the Government’s
policy that Ofqual should develop and set out a process to provide a calculated
grade for each student. The press notice indicated that Ofqual and the exam
boards would discuss the approach with teachers’ representatives before finalising
it. It further stated that, “The Government will not publish any school or college level
educational performance data based on tests, assessments or exams for 2020”.
10. The Secretary of State made a written statement to the House of Commons on 23
March 2020, in which he set out more information as to how grades would be
awarded. At the end of the statement, Mr Williamson indicated that, “The
Government will not publish any school or college level educational performance data
based on tests, assessments or exams for 2020.” DfE restated this commitment on 8
April 2020, in published guidance entitled: “Coronavirus (COVID-19): school and
college performance measures” which stated that the DfE “will not hold schools and
colleges to account on the basis of exams and assessment data from Summer 2020 and that
data will not be used by others, such as Ofsted and local authorities, to hold schools and
colleges to account”. It was further stated that the DfE would not be publishing
various performance data - including school, college and multi-academy trust
level performance data based on summer 2020 tests, assessments and exams at any
phase, or performance tables relating to the 2019/2020 academic year. In addition,
it was said that institution level educational performance data from the 2020
exams and assessments would not be shared with schools or third-party
organisations.
11. On 20 March 2020, Ofqual published a statement in which it welcomed the content
of the DfE’s press notice and confirmed that it was working with exam boards and
teachers’ representatives to develop proposals on how grades would be provided.
12. On 31 March 2020, the Secretary of State issued a direction to Ofqual under section
129(6) of the ASCL Act. That direction stated that it was Government policy that
“students should be issued with calculated results based on their exam centres’ judgements
of their ability in the relevant subjects, supplemented by a range of other evidence”. The
direction also introduced the concept of grade standardisation across centres. This
was said to be a statistical standardisation model which predicted the distribution
of grades for each individual education centres based on the historical
performance of the centre in that subject (“the Algorithm”).
13. On 3 April 2020, Ofqual published: information for heads of centre and teachers
on the submission of Centre Assessed Grades (“CAGs”); guidance for teachers,
students, parents, and carers on the arrangements (the “Summer 2020 Grade
Guidance”); and an open letter to students. The guidance re-stated the
3
EA/2021/0234
Government’s policy on performance measures: “DfE has confirmed that it will not
hold schools/colleges to account on the basis of the exams and assessment data from
summer 2020, and that data should not be used by others, such as Ofsted, local authorities,
academy trusts, and so on to hold schools/colleges or teachers to account.”
14. The information published by Ofqual on 3 April 2020 explained that:
(a) Centres would be asked to determine a CAG for every student in each of their
subjects i.e., the grade they would have most likely have achieved had they sat
their exams and completed any non-examination assessment. Judgements
would need to balance various sources of evidence such as classwork, any non-
examination assessments, and mock results.
(b) Centres would be asked to rank order their students within each grade for each
subject. This information would be used in the statistical standardisation of a
centre’s judgements.
(c) The head of each centre would make a declaration when the CAGs were
submitted, confirming their understanding that the submitted CAGs would be
subject to statistical standardisation by the exam board and that, if the profile
of CAGs was substantially different to that which would have been expected
given the centre’s results in previous years and the prior attainment of the 2020
cohort, they would be adjusted to bring them into line with national standards.
Heads of centre would also confirm in their declarations that the CAGs and
rank order judgements:
i. had been checked for accuracy.
ii. had been reviewed by a second member of staff.
iii. represented objective and professional judgements.
iv. were an honest and fair representation of the grades the centre’s students
would most likely have achieved had they taken exams.
(d) The standardisation model, which would be the subject of a consultation,
would not change the rank order of students, nor would it assume that the
distribution of grades in each subject or centre would be the same. It would
recognise the past performance of centres. If grading judgements in some
centres appeared to be more severe or lenient than others, exam boards would
adjust the grades of some or all those students upwards or downwards
accordingly.
(e) Centres were not to tell students, their parents, or carers their CAGs.
15. Between 15 and 29 April 2020, Ofqual consulted on the details of the
arrangements, including on.
(a) Which students would be eligible to receive a CAG.
(b) The aims and operation of the standardisation model.
(c) Arrangements for private candidates.
4
EA/2021/0234
(d) Appeal arrangements; and,
(e) Arrangements as to who should be eligible to enter an exceptional autumn
exam series (in line with the Secretary of State’s direction).
16. Ofqual also sought views through consultation on the likely impact of the
proposals on students who share protected characteristics, and on the regulatory
burden of the proposed arrangements. The consultation included Ofqual’s initial
equality impact assessment and regulatory impact assessment.
17. Following an analysis of the consultation responses and consideration by the
Ofqual Board, the decisions taken following the consultation were published on 5
May 2020. The key decisions for the summer 2020 arrangements, were as follows:
(a) To incorporate into Ofqual’s regulatory framework for summer 2020 a
requirement for exam boards to collect information from centres on CAGs and
their student rank orders in line with the published information document.
(b) To confirm Ofqual’s expectation that exam boards would investigate as
potential malpractice the inappropriate disclosure of CAGs and rank order
information.
(c) To adopt the proposed aims of the standardisation process as publicly
consulted upon, subject to re-ordering each aim.
(d) To adopt a modified form of Ofqual’s proposal, through which the proposed
statistical model would “place more weight on historical evidence of centre
performance (given the prior attainment of students) than the submitted centre
assessment grades where that will increase the likelihood of students getting the grades
that they would most likely have achieved had they been able to complete their
assessments in summer 2020”.
(e) Not to allow students to challenge their CAGs, although a student could ask
their centre to check whether it made a mistake when submitting their CAG or
their rank order to the exam board.
(f) To allow a centre to appeal to an exam board on the grounds that the exam
board used the wrong data when calculating grades, and/or incorrectly
communicated the grades; and
(g) Not to allow appeals in respect of the operation of the statistical
standardisation model. The possibility and appropriateness of allowing a
centre to appeal where there was reliable evidence of a significant
demographic difference between the centre’s cohort and the historical data
used for statistical standardisation, was to be considered further (Ofqual
subsequently confirmed that this could be the basis for a centre to appeal to an
exam board on the grounds that it had used the wrong information).
18. The consultation outcome document included an updated equality impact
assessment and regulatory impact assessment.
19. On 22 May 2020, Ofqual issued an updated version of the Summer 2020 Grade
Guidance.
5
EA/2021/0234
20. Between 22 May 2020 and 1 June 2020, Ofqual completed a technical consultation,
with the exam boards which make available A level qualifications, in respect of the
proposed framework of Conditions and Requirements to implement the policy
decisions. Ofqual set and published the GQCovid-19 Framework of regulatory
requirements, on 11 June 2020. Condition GQCov5 provided for a centre to appeal
to an exam board on grounds which included that the exam board had used the
wrong data when applying the standardisation model.
21. Between 30 June 2020 and 14 July 2020, Ofqual consulted publicly on proposed
statutory guidance to Conditions GQCov4 (information to be provided to centres)
and GQCov5 (appeals). Further consultation in respect of the proposed guidance,
with the exam boards only, took place on 4 and 5 August 2020. Ofqual published
the final guidance on 6 August 2020.
22. Centres were required to submit to the relevant exam boards, CAGs, and rank
order information for their students by 12 June 2020. The exam boards then
applied the standardisation model to produce a grade (“the calculated grade”),
which was intended to be a candidate’s final A level grade. The standardisation
model was developed to (i) even out, in the interests of fairness to all students, the
likelihood that some teachers would have been more lenient than others in their
judgements of the grades their students would have achieved had exams taken
place and some more severe, and (ii) to make sure that, in line with Government
policy, the overall distribution of grades was similar to that of previous years.
23. On 21 July 2020, Ofqual published more information about how the
standardisation model was operating, together with a set of PowerPoint slides.
The PowerPoint slides included Ofqual’s preliminary equalities analysis. In these
slides Ofqual reported that: “Initial findings show that the differences in outcomes for
students with and without particular protected characteristics and of differing SES [social
economic status] are similar to those seen in previous years.”
24. The standardisation model was not applied in the same way to all centres or in the
same way to all subjects within some centres. It did not operate in relation to small
centres with fewer than five students in any subject (based on the harmonic mean
of the 2020, and historical, entries). A centre with a harmonic mean of entries
between five and fifteen students in all subjects would have had a tapered
application of the standardisation model. Centres with larger entries in all subjects
(leading to a harmonic mean greater than fifteen in all subjects) would have had
the standardisation model applied in full in all subjects. Centres with a mixture of
large, medium, and small entry subjects would have had the standardisation
model applied to a different extent (or not at all) between those subjects.
25. On 10 August 2020, Ofqual published a response to the media coverage of the
results in Scotland, in which it confirmed that: “Our early analysis shows that
students from all backgrounds – including disadvantaged and black, ethnic minority and
Asian students – have not been disadvantaged by this year’s awarding process…”.
26. On 11 August 2020, the DfE announced that it would ask Ofqual to allow students
to appeal on the grounds that they had achieved a higher result in their mock
exams than the grade with which they were issued. A ministerial direction
requiring Ofqual to have regard to this policy followed, on 13 August 2020.
6
EA/2021/0234
27. A level results day in England, Northern Ireland and Wales was 13 August 2020.
Results were sent to centres on 12 August 2020, with Ofqual publishing its interim
report in relation to awarding, on the same date.
28. However, the Ofqual Board decided, on 16 August 2020, that students should be
given the higher of their CAG or the calculated grade. This was announced by the
Secretary of State on 17 August in the following terms:
“Ofqual had consulted on and implemented a standardisation process for exam
results this summer, but the system has resulted in too many inconsistent and
unfair outcomes for A and AS level students. Over the last few days, it has
become clear that the algorithm has revealed a number of anomalies that had not
been anticipated by Ofqual and which severely undermined confidence in the
system.
Subsequently, the Government and Ofqual have jointly agreed to revert to centre
assessment grades, which are the grades which schools and colleges assessed
students were most likely to have achieved, had exams gone ahead. This was
deemed to be the fairest approach to avoid some students receiving grades that
did not reflect their prior performance.
Students awaiting their GCSE grades on Thursday will receive their centre
assessment grade from their school or college, and students who received their A
or AS level results last Thursday (13 August) will be reissued their centre
assessment grade. If students’ calculated grades were higher than the centre
assessment grade, their calculated grade will stand.”
29. Following the decision that students should be given the higher of their CAG or
calculated grade, A level results were re-issued to students.
30. On 2 September 2020, Julie Swan, together with the then Chair of Ofqual,
appeared before the Education Select Committee. The Committee questioned them
on the 2020 grading arrangements. On the same date, the then Chair of Ofqual
published a statement explaining the arrangements that had been put in place and
confirming that the arrangements were not biased either against, or towards,
particular groups of students.
31. In October 2020, Ofqual commenced the GRADE project, through which relevant
awarding data was to be made available to specified researchers through the ONS
Secure Researcher Service.
32. On 26 November 2020, Ofqual published student level equalities analyses for
GCSEs and A levels.
33. On 3 December 2020, DfE published updated accountability guidance titled
‘Coronavirus (COVID-19): school and college accountability 2020/21’ (“the 3 December
Accountability Guidance”), which assumed examinations would be going ahead in
summer 2021. DfE confirmed in its 3 December Accountability Guidance, that it
would produce and share institutional level performance data for Key Stage 2,
Key Stage 4 and 16-18 with schools, local authorities, Ofsted and DfE teams.
7
EA/2021/0234
However, the DfE committed not to publish data on school and college
performance tables, noting that:
“We need to make sure our accountability arrangements are appropriate in the
current public health context and that they take account of the challenges faced
by schools and colleges in both the last and current academic years. So, while
school and college level test, assessment and exams data will be made
available to Ofsted, and to schools and colleges themselves to support
improvement, we will not be publishing this data on school and college
performance tables.”
34. In December 2020, Ofqual also issued its final “Student-level equalities analyses for
GCSE and A level: Summer 2020” (“the Final Report”) and on 7 December 2020,
Ofqual published the code that was provided to examination boards to
demonstrate the method it required them to take and to support them in
developing their own code to standardise A level grades in 2020.
35. On 18 December 2020, Ofqual published three evaluation reports on, inter alia, the
differences between calculated grades and CAGs issued to students.
36. In January 2021, the Government considered that it was not possible or fair for
exams to go ahead as normal, in the summer of 2021. This was announced by the
Prime Minister in a televised address to the nation on 4 January 2021. On 15
January 2021, DfE and Ofqual undertook a joint consultation on alternative
arrangements for awarding A level grades in 2021.
37. The Government’s policy position was confirmed in a direction from the Secretary
of State to Ofqual’s then Chief Regulator, dated 23 February 2021. The direction set
out that it was Government policy that, in summer 2021, students should be
issued with Teacher Assessed Grades (TAGs). TAGs differed from CAGs,
including in the following ways:
(a) TAGs were based on a range of evidence of students’ performance on the
content which they had been taught in a subject. This contrasted with CAGs,
which were a teacher’s judgement of the grade the teacher believed a student
would most likely have achieved had the exams not been cancelled, which
did not need to be informed by specific evidence of a student’s performance,
and which assumed that students would have been taught all of the subject
content.
(b) TAGs were determined by the standard at which students were performing
when they were assessed. This contrasted with the 2020 CAG arrangements
when teachers were asked to look forward and predict the grade standard at
which a student would have performed had the exams not been cancelled.
(c) The TAGs were subject to internal quality assurance within centres and
external quality assurance undertaken by the exam boards. Each centre had
to develop and submit to the exam boards a policy that it would follow when
determining TAGs. Each centre also had to submit to an exam board, for a
sample of students chosen by the exam board, evidence of the work that had
determined those students’ TAGs in each subject. The exam boards looked at
a sample of this work and followed up with centres any concerns that the
8
EA/2021/0234
TAGs were not supported by the work. This contrasted with 2020 during
which the exam boards did not require centres to explain how they were
determining CAGs and did not review any students’ work.
38. On 25 February 2021, the DfE issued updated accountability guidance for the year
2020/2021, replacing that previously published on 3 December 2020 (‘the
2020/2021 Accountability Guidance’). The 2020/2021 Accountability Guidance
stated that, as in 2020, the DfE would “not publish institution level data based on 2021
Key Stage 1 and 2 assessments, tests, GCSEs, AS levels, A levels, other regulated general
qualifications, or vocational and technical qualifications”, nor would this data “be
available for others, such as Ofsted, Regional Schools Commissioners (RSCs) or local
authorities, to use to hold schools and colleges to account.” The Guidance further
specified that schools and colleges should not use the 2020 or 2021 results data as
part of their own teacher performance management processes
The Information Request
39. An initial information request was made by the Requestor on 15 August 2020,
stating:
“For each a level (sic.) provider in England, please provide the data you have for
the % of students who had their teacher assessed grades adjusted up or down
and by how many grades?”
40. On 15 September 2020, Ofqual responded to the request, confirming that it held
the information and referring the Requestor to publicly available documents.
41. On 24 September 2020, the Requestor further explained:
“The information I’m after is the Percent of A level grades that were marked Up,
down or kept the same For [sic] each centre By [sic] the algorithm.”
42. On 28 September 2020, Ofqual responded to the Requestor and referred him to the
Interim Report which it had published on 13 August 2020, - A level results day.
The Interim Report included a detailed student-level equalities analysis which
concluded:
“The analyses conducted show no evidence that this year’s process of awarding
grades has introduced bias. Changes in outcomes for students with different
protected characteristics and from different socio-economic backgrounds are
similar to those seen between 2018 and 2019.”
43. On 29 September 2020, the Requestor stated:
“...Im [sic] looking for a list of all the centres With [sic] the adjusted grades for
each centre, up, down or same. % is gr8”.
44. Ofqual responded by email on 8 October 2020, confirming that it held the
information requested but that it had applied the personal data exemption under
section 40 of FOIA in order to withhold the information in its entirety. Ofqual
9
EA/2021/0234
noted, however, that it intended to make some of the information sought by the
Requestor available through the Office for National Statistics (“ONS”) Secure
Research Service, and that it would release more information about this in due
course.
45. On 9 October 2020, the Requester responded to Ofqual, further explaining what
information he was seeking:
“Just centre name; % grades up 2 grades at that centre; % grades up 1 grade; %
the same grade; % -1 grade; %-2 grades; %-3 grades”.
46. Ofqual treated the correspondence of 9 October 2020 as a request for an internal
review. On 5 November 2020, Ofqual wrote to the Requestor confirming both that
it held the requested information and that it no longer relied upon section 40 of
FOIA but relied instead on the exemption identified in section 36(2)(c) of FOIA.
47. Ofqual and DfE submit that it is the Requestor’s correspondence of 9 October 2020
that is the relevant information request for the purposes of the instant appeal [see
[11] of ‘Ofqual’s Supplementary Submissions’ dated 27 May 2022, and DfE’s letter
of 27 May 2022]. This was also the position initially taken by the ICO [see [9] of the
ICO’s ‘Response’ to the appeal dated 2 September 2021]. However, in its
submissions of 20 May 2022, the ICO reserved its position on this issue, observing
the lack of general guidance as to what constitutes a public authority’s ‘original
refusal’. The ICO further submitted that in the instant appeal nothing turned on
the precise date of the information request, a position that the DfE and Ofqual
agreed with.
48. In our conclusion, the relevant information request was made on 9 October 2020,
this being the date of the correspondence from the Requestor in which the
information that it is now know was being sought, first crystallised.
The ICO’s Decision Notice
49. On 12 November 2020, the Requestor made a complaint to the ICO about the way
in which the Information Requests had been handled.
50. By a Decision Notice served on 5 August 2021, the Information Commissioner
concluded that section 36(2)(c) of FOIA was properly engaged on the facts, but
that the public interest balance required Ofqual to disclose the Requested
Information.
Notice of Appeal and Response
51. In its Notice of Appeal, dated 2 September 2021, Ofqual submit (in Grounds
running to 100 paragraphs, drawn up by Mr Kosmin) that the ICO: (a)
mischaracterised its submissions; (b) overstated the public interest in disclosure;
and, (c) adopted an insufficiently forensic approach to the consideration of
whether other public interests identified in favour of disclosure would be given
effect upon disclosure of the information sought by the Information Request.
10
EA/2021/0234
52. In particular, it was asserted that the ICO’s analysis of the public interest in
disclosure placed undue weight upon assertions that: (i) the Algorithm was
discriminatory; (ii) the operation of the Algorithm was in breach of Article 22
GDPR; (iii) disclosure would provide assistance to those seeking legal action
against Ofqual; (iv) disclosure would provide assistance to students who believe
they received an unfair grade in 2020 or 2021 when making complaints; and (v)
disclosure would increase accountability. It was further contended that the ICO
had significantly understated Ofqual’s submissions concerning: (i) the detrimental
impact on centres and students educated at those centres; (ii) the future conduct of
CAG or TAG grading; (iii) the DfE’s public commitments not to publish
educational data at centre level for 2020; and (iv) the importance of the
maintenance of good working relationships between centres and third parties
(including AOs). Ofqual additionally asserted that the ICO’s handling of the
Requestor’s complaint breached the basic principles of natural justice and
procedural fairness.
53. In its Response, drawn up by Ms Taunton, the ICO asserts, in summary, that the
submissions made in the Notice of Appeal do not disturb the ICO’s analysis of the
public interest balance set out in the Decision Notice. It is further averred that the
ICO’s investigation was not procedurally unfair but that, in any event, the FtT
exercises a full merits jurisdiction and consequently if there were any procedural
unfairness in the ICO’s investigation, Ofqual would not be prejudiced.
The issues for consideration by the Tribunal
54. The following issues fall for consideration by the Tribunal:
(1) Is section 36(2)(c) of FOIA engaged?
(2) If so, does the public interest in favour of maintaining the section
36(2)(c) exemption outweigh the public interest in disclosure?
The legislative background
55. Section 1(1)(b) of FOIA confers a duty on a public authority, in response to a
request, to provide information held by it. By virtue of section 2(2)(b) of FOIA, the
duty does not extend to information if it falls within an absolute exemption, or
within a qualified exemption and “the public interest in maintaining the exemption
outweighs the public interest in disclosing the information”.
56. In the instant matter, the relevant exemption claimed is that set out in section
36(2)(c) of FOIA, which is a qualified exemption.
57. Section 36 relevantly states:
“(1) This section applies to—
(a) information which is held by a government department or by the
Welsh Government and is not exempt information by virtue of
section 35…
(2) Information to which this section applies is exempt information if, in the
reasonable opinion of a qualified person, disclosure of the information
under this Act—
11
EA/2021/0234
(a) – (b) …
(c) would otherwise prejudice, or would be likely otherwise to
prejudice, the effective conduct of public affairs.
…
(4) In relation to statistical information, subsections (2) and (3) shall have effect
with the omission of the words “in the reasonable opinion of a qualified
person”.”
58. By section 50 of FOIA:
“(1) Any person (in this section referred to as “the complainant”) may apply to
the Commissioner for a decision whether, in any specified respect, a request
for information made by the complainant to a public authority has been dealt
with in accordance with the requirements of Part I.”
59. Section 57 of FOIA materially states:
“Appeal against notices served under Part IV
(1) Where a decision notice has been served, the complainant or the public
authority may appeal to the Tribunal against the notice.”
The Tribunal’s role
60. The role of the Tribunal is set out in section 58 of FOIA:
“If on an appeal under section 57 the Tribunal considers -
(a) that the notice against which the appeal is brought is not in
accordance with the law, or
(b) to the extent that the notice involved an exercise of discretion by the
Commissioner, that he ought to have exercised his discretion
differently,
the Tribunal shall allow the appeal or substitute such other notice as could
have been served by the Commissioner, and in any other case the Tribunal
shall dismiss the appeal.
On such an appeal, the Tribunal may review any finding of fact on which the
notice in question was based.”
61. The import of section 58 is that the right of appeal to the Tribunal involves a full
merits consideration of whether, on the facts and the law, the public authority’s
response to the FOIA Request is in accordance with Part I of FOIA (Information
Commissioner v Malnick and ACOBA [2018] UKUT 72 (AAC); [2018] AACR 29, at
paragraphs [45]-[46] and [90]).
Discussion
12
EA/2021/0234
62. We have before us an Open Bundle running to 1354 pages, an Addendum Bundle
consisting of 83 folios, a Closed Bundle which, unnecessarily, replicates the Open
Bundle and additionally includes the Requested Material, a bundle of authorities
(318 pages), detailed written skeleton arguments drawn by each of the three
parties and supplementary written submissions provided by the parties after the
completion of the oral hearing in response to the promulgation of the three judge
panel Upper Tribunal decision in Montague v Information Commissioner and DIT
[2022] UKUT 104 (AAC).
63. During the hearing of 23 March 2022, the Tribunal heard oral evidence from Julie
Swan – the Executive Director of General Qualifications at Ofqual, and Jenny
Oldroyd – the Director of the Qualifications Directorate at the DfE. In addition, the
parties each made oral submissions at the hearing.
64. In reaching our conclusion on this appeal, we have taken account of all of the
evidence and submissions before us, irrespective of whether such evidence or
submissions have been specifically alluded to during this decision.
Issue 1: Is section 36(2)(c) engaged?
65. This is an issue upon which the parties are ad idem, and we can see no reason to
depart from the agreed position.
66. We accept that the Requested Information is statistical information and,
consequently, that section 36(4) of FOIA is to be applied. We also accept that
disclosure of the Requested Information “would be likely otherwise to prejudice, the
effective conduct of public affairs..”
67. In accepting the parties agreed position on this issue, we had in mind, and
applied, the following words of the Information Tribunal in Hogan and Oxford
City Council v Information Commissioner [2011] 1 Info LR 588, which were
approved by the Court of Appeal in Department of Work and Pensions v ICO and
Zola [2016] EWCA Civ 75:
"29. First, there is a need to identify the applicable interest(s) within the relevant
exemption…
30. Second, the nature of the "prejudice" being claimed must be considered. An
evidential burden rests with the decision maker to be able to show that some
causal relationship exists between the potential disclosure and the prejudice and
that the prejudice is, as Lord Falconer of Thoroton has stated, "real, actual or of
substance" (Hansard HL, Vol. 162, April 20, 2000, col. 827). If the public authority
is unable to discharge this burden satisfactorily, reliance on "prejudice" should be
rejected. There is therefore effectively a de minimis threshold which must be met…
…
34. A third step for the decision maker concerns the likelihood or occurrence of
prejudice. A differently constituted division of this Tribunal in John Connor Press
Associates Limited v Information Commissioner (EA/2005/0005) interpreted the
phrase "likely to prejudice" as meaning that the chance of prejudice being
13
EA/2021/0234
suffered should be more than a hypothetical or remote possibility; there must
have been a real and significant risk. That Tribunal drew support from the
decision of Mr Justice Munby in R (on the application of Lord) v Secretary of State for
the Home Office [2003] EWHC 2073 (Admin), where the comparable approach was
taken to the construction of similar words in the Data Protection Act 1998. Mr
Justice Munby stated that "likely":
“connotes a degree of probability where there is a very significant and
weighty chance of prejudice to the identified public interests. The degree of
risk must be such that there "may very well" be prejudice to those interests,
even if the risk falls short of being more probable than not".
35. On the basis of these decisions there are two possible limbs on which a
prejudice-based exemption might be engaged. Firstly, the occurrence of prejudice
to the specified interest is more probable than not, and secondly there is a real
and significant risk of prejudice, even if it cannot be said that the occurrence of
prejudice is more probable than not."
68. Given the consensus between the parties on this issue, and what we say below, we
waste no further ink at this stage on providing reasons for the conclusions reached
above.
Issue 2: Does the public interest in favour of maintaining the section 36(2)(c)
exemption outweigh the public interest in disclosure?
69. Section 2(2)(b) of FOIA requires decision-makers, including the Tribunal, to
conduct a balancing exercise, weighing the factors in favour of maintaining the
exemption against the public interest factors that favour disclosure. There is
neither a presumption in favour of disclosure nor a presumption in favour of non-
disclosure. Adjudging the balance of public interest is a mixed question of law and
fact, not the exercise of discretion: Information Commissioner v Malnick [2018]
UKUT 72 at [45(5)]. Where the decision-maker concludes that the competing
interests are equally balanced, the decision-maker will not have concluded that the
public interest in maintaining the exemption outweighs the public interest in
disclosing the information – so that disclosure will be required (Department of
Health v Information Commissioner [2017] EWCA Civ 374, [2017] 1 WLR 3330, at
[46]).
70. In All Party Parliamentary Group on Extraordinary Rendition v Information
Commissioner [2013] UKUT 0560 (AAC), the Upper Tribunal said at [149]:
“When assessing competing public interests under FOIA the correct
approach is to identify the actual harm or prejudice that the proposed
disclosure would (or would be likely to or may) cause and the actual
benefits its disclosure would (or would be likely to or may) confer or
promote. This…requires an appropriately detailed identification, proof,
explanation and examination of both (a) the harm or prejudice, and (b)
benefits that the proposed disclosure of the relevant material in respect of
which the exemption is claimed would (or would be likely to or may) cause
14
EA/2021/0234
or promote.”
71. In accordance with the recent decision of the Upper Tribunal in Montague, at [86],
“…the public authority is not to be judged on the balance of competing interests on how
matters stand other than at the time of the decision on the request which it has been obliged
by Part I of FOIA to make.” In the instant appeal, Ofqual’s refused the Information
Request of 9 October 2020, on 5 November 2020 and it is as of this date that the
public interest balance must be assessed. Evidence about events after 5 November
2020 is only relevant in so far as it throws light on the position as it was on the 5
November 2020 (Montague, at [52])
72. We turn first to consider the public interest in favour of disclosure.
73. Ofqual accepts that there is a public interest in the promotion of the integrity of
awarding qualifications through publication of centre-level awarding information.
It contends, however, that limited weight should be attached to this public interest
in the instant appeal.
74. For reasons which we set out in more detail below, we accept the evidence of Julie
Swan that disclosure of the Requested Information will not of itself, or in
conjunction with other information that is currently in the public domain, shed
light on the behaviour and integrity of an educational centre’s determination of
CAGs. In such circumstances, on the facts of this case no weight can be attached to
the public interest identified in the preceding paragraph.
75. Ofqual further accepts that there is a general public interest in disclosure of
information relevant to the decision-making of a public authority but contends
that in the instant scenario the weight to be attached to this general public interest
is low.
76. The public interest in disclosure of information relevant to the decision-making of
a public authority reflects the principles underpinning FOIA, namely transparency
and openness. We agree that if the dual interests of transparency and openness are
viewed simply as a generality, then the weight attached to such public interest is
limited. However, the ICO alights not on the general public interest in
transparency and openness but on the specific public interest in open and
transparent information about the impact of the Algorithm. It is submitted, inter
alia, that disclosure of the Requested Information would add to the public debate
both as to the fairness of the Algorithm and as to whether the CAG model was a
fairer method of awarding A level grades in 2020 than the model used to obtain
the calculated grades.
77. It is to be recalled that, on 17 August 2020, the then Secretary of State for
Education publicly announced that use of the Algorithm to calculate A level
grades in 2020 was to be abandoned, save in those circumstances where its use
provided for an award of a grade higher than the CAG. In our view, it is of some
significance that the Secretary of State’s publicly stated reason for taking such a
course was that the Algorithm resulted in “too many inconsistent and unfair
outcomes” and that confidence in the system had been “severely undermined”. It was
further said that the change from calculated grades to CAGs “was deemed to be the
fairest approach to avoid some students receiving grades that did not reflect their prior
15
EA/2021/0234
performance.” Jenny Oldroyd did not seek to demur from the terms of the Secretary
of State’s announcement, during her evidence.
78. Ofqual’s position is that the Algorithm did operate fairly. In her evidence, Julie
Swan averred that the Algorithm “produced fair results overall” and that “analysis
showed that the standardisation model [the Algorithm] operated fairly”. Under cross-
examination, Julie Swan’s attention was drawn to the Secretary of State’s
announcement of 17 August 2020 and, in response to being asked whether she
accepted that the government had changed its approach to the grading of A levels
in 2020 because the “system” had produced “too many inconsistent outcomes”, she
stated that “Ofqual decided to change its approach because public confidence had fallen
away” This evidence is consistent with Ofqual’s publicly stated position: see, for
example, the Executive Summary of the Final Report (December 2020): “It became
apparent following the issue of A level results, that the above approach [the Algorithm]
did not command public confidence. Consequently, Ofqual instructed awarding
organisations to reissue A level results, awarding candidates the higher of their CAG and
their calculated grade.”
79. In our conclusion, the change in the method by which A level grades were
awarded in 2020 was a matter of personal interest to a large number of students
and teachers, and also attracted a significant amount of interest from the public at
large at the time. We accept that this was so not only in August 2020, when the
change was announced, but also in November 2020 at the time of Ofqual’s
response to the Information Request.
80. In addition to the public interest in better understanding the historical operational
fairness of the Algorithm and its implications for the 2020 A level grades, the
position as to the mechanism for the award of A level grades in 2021 had yet to be
concluded upon. As of November 2020, it remained a possibility that the 2021 A
level grades would not be awarded after open examinations and, consequently, it
remained a possibility that an alternative mechanism for awarding such grades
would be required. Ultimately, the 2021 A level examinations did not proceed ‘as
normal’ and an alternative mechanism for awarding grades was operated -
Teacher Assessed Grades (TAGs). There is no dispute that these significantly
differed from CAGs in that CAGs were notionally predictive of how students
would have performed had they been taught all of the course content and taken
exams, whereas TAGs assessed performance on the basis of the course content
which had already been taught. It is, also, noteworthy that the TAG system did
not operate a standardisation model. Nevertheless, in our conclusion a better
understanding of the fairness of the Algorithm in November 2020 could only have
enhanced the consideration and debate as to an appropriate alternative
mechanism for awarding A level grades in 2021, the need for which, as we have
already indicated, was a possibility at the relevant date.
81. Ofqual submits that the disclosure of the Requested Information would add little
if anything to the aforementioned public interest considerations. In support of this
contention, Julie Swan points to the scrutiny of the 2020 A level grades that has
already taken place in the public arena, including “extensive publications to date
concerning the Summer 2020 grades (including three evaluation reports), [and] its
participation in the Education Select Committee”. Amongst other things, Ofqual also
16
EA/2021/0234
directs attention to the ongoing “GRADE project” [GRADE being an acronym for
Grading and Admissions Data for England].
82. In response, the ICO observes that much of the information identified in this
regard by Julie Swan was published after the relevant date of 5 November 2020,
and therefore, it is said, cannot be used to detract from the strength of the public
interest in disclosing the Requested Information in November 2020. For example,
it was only on 7 December 2020 that Ofqual published the relevant code, and it
was not until 18 December 2020 that Ofqual published three evaluation reports on
the differences between calculated grades, CAGs and the final grades issued to
students.
83. In our view, the fact that there has subsequently been extensive publication of
material relating to the 2020 A level grading system, does little to detract from the
strength of the public interest in the disclosure of the Requested Information as it
stood on 5 November 2020. The very fact that the 2020 A level grading system has
attracted such scrutiny, supports our conclusion as to the strength of the public
interest in understanding whether the Algorithm operated fairly and whether the
change of method by which A level grades were awarded in 2020 produced fairer
grade awards. In any event, as asserted by the ICO, the analysis in the post
November 2020 information does not address the significant concern, and the
potential consequences flowing from such concern, that by standardising on the
basis of each centre’s past performance, students attending the worse performing
schools were more likely to have their CAG downgraded by the Algorithm than a
student in a better performing school. We accept the ICO’s contention that release
of the Requested Information would inform scrutiny of this issue.
84. In considering the strength to be attached to the aforementioned public interest
consideration, we have taken full account of all the information that was already
in the public domain as of 5 November 2020, such as the slide deck from July 2020
entitled “Summer Symposium 2020” which contained “initial findings” from an
earlier equalities analysis, the report of September 2020 “Data Protection Impact
Assessment: Summer 2020 Grading”, and the statement by the Ofqual Chair to the
Education Select Committee. We have also taken account of the fact that the
GRADE project had begun by 5 November and that an intention had been evinced
by that date to the effect that the Office for National Statistics Secure Researcher
Service would release information about the 2020 examinations, although it had
not done so at that time. In any event, although the GRADE project is a scheme
whereby a wide range of data from 2017 to 2020 has been, or will be, shared with
accredited researchers “undertaking ethical research projects with a public benefit”
which will allow those accredited researchers to undertake their own analysis of
the data from 2020 and previous examinations, the Requested Information is not
data that has been shared within the GRADE project, the shared data having been
“pseudonymised…to make sure that…centres cannot be identified.”
85. As already alluded to, a considerable amount of material about the 2020 A level
grading mechanism entered the public domain in December 2020, including the
code. Whilst we accept that there was an intention, as of 5 November 2020, to
publish this material and that the relevant material was in the process of being
collated and considered at that time, we do not accept that material which was not
17
EA/2021/0234
in the public domain on 5 November 2020 is material which bears weight in the
public intertest balance that we must perform.
86. Drawing all of this together, we find, contrary to the Ofqual’s submissions, that
the Requested Information would, as of 5 November 2020, have added significant
value to any debate regarding the 2020 A level grading system. We conclude, in
particular, that it would help inform the public interest in understanding whether
the Algorithm operated fairly. We find this public interest to be very strong.
87. We now turn to the elements of the public interest consideration that Ofqual
submit weigh in favour of maintaining the section 36(2)(c) exemption. These have
been summarised by Ofqual in the following terms:
(a) Prejudice to centres.
(b) Prejudice to students who were awarded their A level grades and GCSE
grades in 2020.
(c) Prejudice to Stakeholder relationships.
(d) Prejudice to the integrity of grading in future years.
88. It is evident that there is a degree of overlap between the four matters identified in
the preceding paragraph; nevertheless, it is prudent to take the submissions in
turn for the purpose of analysing the respective elements of the public interest.
89. First, Ofqual contends that educational centres would likely be prejudiced by
disclosure of the Requested Information because such disclosure “would present a
misleading impression of how individual centres awarded CAG’s.” In evidence, Julie
Swan explained, as a consequence of her experience in the field of qualifications
and grading, that she believes readers of the Requested Information would seek to
draw unjustified conclusions from the data therein which, it is said, would likely
lead to the professional judgments of some educational centres being called into
question; for example, on the basis that they were overly generous in the
assessment of their CAGs.
90. It is contended, inter alia, that the drawing, and/or subsequent reporting, of such
unjustified conclusions could lead to reputational damage for educational centres,
which in turn could lead to damage to such centres ability to attract students and
good staff and to staff morale, which may lead to an inability of a centre to retain
its teaching cohort. It is said that there would also be a consequential impact on a
centre’s resources, “because funding follows students”.
91. We accept that reliable conclusions about the integrity of an individual centre’s
determination of CAGs cannot be drawn solely from a consideration of the
Requested Information. This is obviously so because, as we have already
identified above, the standardisation model was not applied in the same way to all
centres or in the same way to all subjects within a centre. For example, centres
with a mixture of large, medium, and small entry subjects would have had the
standardisation model applied to a different extent between those subjects. In
18
EA/2021/0234
order to properly identify how the standardisation model operated in relation to a
particular centre, information as to the harmonic mean applied to subjects within
that centre would be required. This information is not currently in the public
domain.
92. Even assuming a reader of the Requested Information had sufficient relevant data
to enable a meaningful analysis and comparison between a centre’s CAGs and the
calculated grades, we accept that this would not assist in an analysis of the
integrity of a centre’s award of CAGs. Any integrity analysis which is undertaken
by way of a comparison with a centre’s calculated grades, requires an assumption
that the calculated grade is the gold standard against which the integrity of other
grade predictions should be judged. It is exactly this assumption that is the subject
of debate, and we observe, once again, that in his public announcement of 17
August 2020 the Secretary of State for Education identified unforeseen anomalies
in the application of the Algorithm and took the position that the fairest approach,
to avoid some students receiving grades that did not reflect their prior
performance, was, broadly, to award students their CAG rather than the
calculated grade. It is, also, trite to say that the Covid-19 pandemic affected each
school and college differently.
93. It is clear from what we have said above that any responsible reporting of the
Requested Information could not seek to use that information as a basis to
challenge the integrity of a centre’s CAGs. Ofqual’s concern does not, however,
relate to those who seek to responsibly report on the Requested Information, but
rather the reputational damage to centres caused, inter alia, by those whose
reporting is irresponsible and unjustified or those who draw unjustified
conclusions after considering the Requested Information.
94. The ICO submits that in order to mitigate against possible prejudice to educational
centres caused by publication of the Requested Information, Ofqual could publish
the information with a ‘health warning,’ advising against the drawing of
conclusions about a centres’ behaviour, and explaining how the standardisation
model operated. In response to this suggestion, Julie Swan asserts that the
publication of any such warning would not stop a significant number of readers
and commentators from drawing spurious conclusions because any warning
would have to be highly sophisticated in order to adequately correct the
misleading impression given by disclosure.
95. We accept that a highly sophisticated warning would be necessary if the intention
were to provide the reader of such a warning with ‘chapter and verse’ about why
inferences as to the integrity of an educational centre cannot be drawn from the
Requested Information, but we also see benefit in attaching a short and simple
warning to the effect that Requested Information cannot be used draw
comparisons between educational centres or as an assessment of the accuracy of
the CAGs, because of the nature of the process which led to the production of the
calculated grades. This would immediately alert any reader of the information to
19
EA/2021/0234
the issue, even if it did not permit the reader a full understanding of the reasons
behind the warning. In any event, we have approached our considerations on the
basis that there would be no such warning to compliment release of the Requested
Information.
96. Returning to the question as to what, if any, prejudice there would be to centres if
the Requested Information were to be disclosed; we recognise, as the Upper
Tribunal did in Department of Work and Pensions v Information Commissioner
[2015] UKUT 0535 “the trouble that can be caused by the media taking a selective
approach to what it publishes and putting its own spin on that material” (at [21]), and we
accept that irresponsible and unjustified reporting may have a negative impact on
the reputation of an educational centre in the uniformed eyes of those who read or
otherwise engage with such reporting, or those who read the Requested
Information directly. We also accept that this might lead to a diversion of centre
resources to respond to any criticism. We also concur with Ofqual’s view, that
reputational damage to a centre may affect the morale of teachers working there,
even if the damage is as a consequence of irresponsible and unjustified reporting
or assessment of the Requested Information.
97. We further accept the evidence from DfE, supported by letters from Awarding
Organisations, that stakeholders are not in favour of disclosing the Requested
Information, particularly as they consider it is likely to be misleading and
misconstrued.
98. The weight to be attached to the prejudice caused to centres by disclosure of the
Requested Information has, however, to be considered in its proper context. We
accept that there is likely to be some irresponsible reporting and unjustified
conclusions drawn from reading the Reported Information. The public are,
however, generally aware of the unique pressurised circumstances within which
the education sector was operating in 2020, as well as the fact that there was a
general inflationary effect on A level grades caused by the adoption of the CAG
mechanism. These matters were widely reported at the time.
99. Furthermore, whilst we cannot discount the possibility that students, prospective
students and parents of prospective students may read or otherwise take on board
irresponsible reporting of the Requested information that seeks to challenge the
integrity of an education centre, we take into account the fact that such persons
would also inevitably want to scrutinise the reporting and be as informed as
possible about an educational centre before making as choice as to whether to
attend, or retain attendance, there.
100. We turn next to consider the prejudice to students who were awarded their A
level, or GCSE, grades in 2020. Ofqual’s evidence and submissions on this issue
are summarised thus, in its skeleton argument:
20
EA/2021/0234
“[39] Ensuring the integrity of grades, and public confidence in grades, is of
paramount importance to Ofqual as this forms part of its statutory
objectives and furthers the public interest. Whilst individual students know
the extent to which their calculated grade varied from their CAG (as they
have been in receipt of or able to request both), they do not know the extent
of the variation across their centre, or between different centres (Swan W/S
§99). Disclosure of the withheld information in the way described above,
would be likely to call into question the integrity of the grades awarded to
students who attended particular centres in 2020. This is likely to
disproportionately affect students who attended centres where there
appeared to be greater variation between CAGs and calculated grades,
although insufficiently precise reporting of the information is likely to tar a
greater number of students who were awarded grades in 2020 (Swan W/S
§100). Ms Swan is bolstered by consistent and compelling evidence
recognising this serious risk of prejudice to students: see also W/S Oldroyd
§§76-77 and 80, [716], [719-720], and [723 -724].
[40] Damage to confidence in the awarding of A level CAGs in 2020 is also
likely to affect public confidence in the award by certain centres of the 2020
GCSE CAGs (Swan W/S §102) or the award of TAGs in 2021 (see [717]:
“Disclosure of this information could therefore create an ongoing risk that the
qualifications awarded to candidates from specific Centres will be devalued or
undermined”). In that sense, this form of prejudice was a serious and
significant factor in favour of maintaining the exemption at the date of the
internal review and remains a live issue.
[41] The Commissioner acknowledges the likelihood of the integrity of
grades being undermined as a result of disclosure. However, he asserts that
this would not result “in any further prejudice to the students in question who
will now have been placed into their respective employment and/or university and
from which they will be assessed on their merits of the past year” Ofqual rejects
this unduly narrow characterisation of the currency of examination results.
As is demonstrated by the AOs’ letters and references to the demand for
certificate replacement services, A level qualifications are “in effect the
currency used by learners to gain access to further education and jobs throughout
their lives.” [396]
[42] Disclosure of the withheld information was therefore likely to prejudice
students who were awarded A level grades in 2020 by unfairly
undermining the integrity of the grades awarded to them. …”
101. We accept, as we have above, that there is a risk that the Requested Information
will be misinterpreted by some, and that unjustified conclusions will be drawn
from the data therein as to the relative behaviour of different educational centres
in their determination of CAGs in 2020. We also accept that one possible
consequence of this is that a reader of the information may treat some CAGs as
less reliable than others, and that there is a possibility that readers of such
information may include employers and universities.
21
EA/2021/0234
102. Whilst we do not accept that “the risk that third parties such as employers would
compare an individual’s grade against the performance of their centre and make a
prejudicial assumption that their grades lack integrity to be a remote one.”, as contended
for by the ICO, neither do we accept Ofqual’s contention that great weight should
be attached to such prejudice. Consideration of the weight to be attached to the
possibility of prejudice by employers against some students, must be set against
the fact that it was known, prior to 5 November 2020, that the system used to
calculate grades in 2020 had an inflationary effect on those grades when compared
to grades from previous years, as was accepted by Julie Swan in her oral evidence.
We also have no doubt that it will be well understood by employers that the
Covid-19 pandemic operationally affected each school and college differently.
103. As to prejudice caused to students applying to universities, this we do find to be
remote. During her oral evidence, Julie Swan painted a detailed picture of the
sophisticated methods used by universities to ensure that offers are made to the
appropriate candidates. This picture of sophistication does not sit easily with the
submission that universities would use the Requested Information to draw
unjustified inferences from the data therein about the integrity of the results from
a particular centre, and then extrapolate that unjustified inference to a
consideration of the merit of the grades awarded to a particular applicant.
104. Moving on to Ofqual’s submission that disclosure of the Requested Information
would prejudice both its, and the Department of Education’s, relationship with
key stakeholders. In this regard, it is broadly contended that disclosure of the
Requested Information would undermine assurances given to centres in 2020 and
2021 regarding the confidentiality of centre-level grading and that this would be
likely to undermine trust in Ofqual and its relationships with its stakeholders.
105. Ofqual and the Secretary of State draw particular attention to the following
assurances. On 18 March 2020, the then Secretary of State for Education, Gavin
Williamson MP, made a statement in the House of Commons on changes to the
operations of educational settings as a result of the coronavirus (COVID 19)
pandemic. He announced that as from 20th March 2020, schools and colleges
would be closed to most students until further notice. He said: “I can confirm that
we will not go ahead with assessments or exams, and that we will not be publishing
performance tables for this academic year.”
106. The DfE subsequently published a press notice on 20 March 2020 setting out the
Government’s policy that Ofqual should develop and set out a process to provide
a calculated grade for each student. The press notice said that Ofqual and the
exam boards would discuss the approach with teachers’ representatives before
finalising it. The press notice stated: “The Government will not publish any school or
college level educational performance data based on tests, assessments or exams for 2020”.
107. Thereafter, the Secretary of State made a written statement to the House of
Commons on 23 March 2020 in which he set out more information as how grades
22
EA/2021/0234
would be awarded. The statement explained that students would be awarded
“calculated grades”: At the end of the statement, the Secretary of State once again
said that; “The Government will not publish any school or college level educational
performance data based on tests, assessments or exams for 2020.”
108. On 8 April 2020, the DfE published guidance entitled: “Coronavirus (COVID-19):
school and college performance measures” (“the Accountability Guidance”). The
DfE’s Accountability Guidance included the following:
“As part of steps taken to fight the spread of coronavirus (COVID-19), the
government announced that all exams due to take place in schools and
colleges in England in summer 2020 are cancelled and that it will not publish
any school or college level educational performance data based on tests,
assessments, or exams for 2020.
…
Changes to accountability arrangements
We will not hold schools and colleges to account on the basis of exams and
assessment data from summer 2020 and that data will not be used by others,
such as Ofsted and local authorities, to hold schools and colleges to account.
School, college or MAT level performance data
We will not be publishing school, college or multi-academy trust (MAT) level
performance data based on summer 2020 tests, assessments and exams at any
phase.
We will not be publishing, or sharing, school, college or MAT level
accountability measures, such as Progress 8 and level 3 value added, using
the summer 2020 data.
The performance tables that were due to be released in October and
December 2020, and in January and March 2021, will not go ahead.
We will also not publish any institution-level qualification achievement rates
in the national achievement rate tables for the 2019 to 2020 academic year.
…
Exams and assessments
We will not share the educational performance data from 2020 exams and
assessments with schools via Analyse school performance (ASP), or through
the ASP accredited service.
Schools and colleges will receive details of grades awarded directly from
awarding organisations.
We will not run the primary, secondary or 16 to 18 school and college
performance data checking exercises and will not return any pupil level or
institution level data on the 2020 grades to schools and colleges.
We will not use the 2020 exams data when setting criteria in other areas, such
23
EA/2021/0234
as teaching schools, English and maths hubs and free school applications.
Schools and colleges should not use the 2020 exams data as part of their
teacher performance management process.”
109. Julie Swan provides, inter alia, the following evidence as to why Ofqual considers
trust with its stakeholders to be of importance:
“[107] …stakeholder engagement remains important in more normal times.
In fact, it is essential for the fair administration of exams and assessments for
all students and so for the performance of Ofqual’s statutory functions.
[108] That relationship depends upon teachers and school and college leaders
having trust and confidence in Ofqual. Ofqual would be likely to lose that
trust and confidence if it disclosed centre level information about grades in
2020 when assurances have previously been given that centre-level
information would not be published, and which the teachers and leaders
themselves had agreed should be kept confidential. In particular, the
stakeholders would be likely to have little trust and confidence in future
public assurances given to them by Ofqual and the DfE. This would
materially undermine the effective conduct of public affairs given, as I have
explained, the importance of a trusting relationship with Ofqual’s key
stakeholders to its ability to effectively carry out its functions.”
110. Jenny Oldroyd states in evidence that: “The DfE’s view is that, were the Requested
Information to be released, this would be likely to weaken schools’ and colleges’ trust in
the government which would be likely to have adverse consequences for the delivery of
government education policy in the future.” This conclusion is further explained at
[68] and [71] of her witness statement, as follows:
“[68] Stakeholders point out that Covid-19 has affected schools and colleges
differently – with infection rates higher in some areas/schools and colleges
than others. The make-up of a school or college community, including the
proportions of disadvantaged students, ethnic minority groups (given that
those groups are more likely to be diagnosed with Covid-19) and the
clinically vulnerable are all factors that might lead to a school or college
being more or less impacted, and the degree to which students are affected
by lost learning. Teaching associations have argued that due to the uneven
impact of the pandemic, comparisons between schools and colleges based
on the performance data would not be fair or meaningful and would
potentially be damaging to schools and colleges. This is because it would be
unclear whether differences in performance were attributable to the
pandemic and its consequences, or to other factors (such as quality of
teaching, for example). School leaders and sector representatives have also
expressed concerns that publishing the data could have a damaging impact
on a school or college’s reputation, with negative consequences for
pupil/student and teacher recruitment. Schools have told the DfE that
published performance data influences their reputation, particularly in a
local area. If a school or college has a poor set of data, this can then lead to
fewer parents applying to send their children to the institution, which can
24
EA/2021/0234
have consequences for the amount of per pupil funding the school or
college receives. This in turn can affect their retention and recruitment of
teachers. …
[71] …it is clear from the 2019/2020 and 2020/2021 Accountability
Guidance, and the related statements, that the DfE intended that institution
level data, which could be used to make judgements about school and
college performance and enable performance comparisons between schools
and colleges, should not be published for those academic years or used to
hold those institutions to account. Ofqual’s decision not to release the
Requested Information, which is institution level data, and could be used in
those ways, is entirely consistent with the DfE’s Accountability Guidance.
111. We also have before us two letters written by examination boards to Ofqual,
indicating their serious concerns about the consequences of disclosure of the
Requested Information. Both examination boards highlight the damage that they
believe would be done to the relationships between teachers, DfE, Ofqual and/or
the examination boards, if the Requested Information were to be disclosed.
112. Having carefully considered this evidence, we acknowledge that there is in
general a significant public interest in Ofqual, the DfE and the examination boards
maintaining a relationship of trust with their respective stakeholders. When
considering the consequences of disclosure of the Requested Information to that
relationship, the evaluation that is required is that of a real-world assessment and,
in particular, how the commitments given by DfE would have reasonably been
understood by stakeholders, and what the consequences would be if the
commitments, as reasonably understood, were breached.
113. Whilst we have taken account, inter alia, of the letter from The National
Association of Head Teachers (“NAHT”) of 3 April 2020, which welcomes the
DfE’s announcement that it would not hold schools to account “on the summer’s
data”, and that there would be no “requirement on schools and colleges to collate or
provide evidence to awarding organisations to support their judgments”, and the
Association of Schools and College leaders (“ASCL”) letter of 3 December 2020,
the parties did not draw our attention to any evidence from individual
stakeholders or groups representing stakeholders, such as the NAHT or ASCL, as
to how the DfE’s commitment was more widely understood, or as to whether
disclosure of the Requested Information would be thought to have breached that
commitment and the consequences if it were thought to have done so.
114. We do, however, have evidence from Ofqual, DfE and the two examination boards
on this issue, which we have considered. We acknowledge, as with all other
aspects of our consideration, that the assurances made by the DfE to stakeholders
must be viewed in their proper context, which must include the rationale for the
assurances, summarised by Jenny Oldroyd as: “1) to help ensure that grades were
accurate; and 2) to ease schools’ and colleges’ accountability pressures so that they could
focus on responding to the huge Covid-19 challenges they faced.”
25
EA/2021/0234
115. The submissions before us focused on whether the Requested Information could
be considered as “performance data”. In particular, the ICO contended that
disclosure of the Requested Information would not breach any commitments
made by Ofqual or DfE to stakeholders, nor would it lead to a detrimental impact
on Ofqual’s ability to perform its regulatory functions, ostensibly because the
Requested Information was said to be, “of a different type to that which DfE
committed not to publish” as “it does not evidence a centre’s performance in 2020
assessments, nor could a reader of the information infer how well students are any
particular centre performed in the 2020 assessments.”
116. In the context of a real-world assessment, we do not accept that an over-technical
analysis of whether the Requested Information can be properly referred to as
“performance data” is helpful. In our conclusion, the accountability policy could
reasonably have been understood by stakeholders as meaning that institutional
level data, which could be used make comparisons between different centres,
would not be disclosed.
117. What is clear, is that the Requested Information does not disclose the grades
achieved at centre level, nor could a reader of the information infer how students
at an individual centre performed in the 2020 A level assessments. To that extent, a
meaningful performance comparison between centres cannot be undertaken as a
consequence of disclosure of the Requested Information. Insofar as a performance
indicator can be drawn from the Requested Information, it does not relate to the
success or lack of success of the students or centres, but rather to the extent, or
degree, to which a centre’s CAGs were adjusted by operation of the Algorithm for
the 2020 cohort. We accept, as we have done above, that this would permit of the
possibility of comparisons between educational centres as to the extent to which
the centre’s CAGs were adjusted by operation of the Algorithm but, as we have
also previously emphasised, such comparisons would not be meaningful. To this
limited extent, we accept Ofqual’s submission that stakeholders could view
disclosure of the Requested Information as breaching the commitments made by
DfE and Ofqual.
118. We also accept that disclosure of the Requested Information could have a negative
impact on the relationship between the stakeholders, Ofqual, the DfE and the
examination boards. However, given the limited nature of the material that would
be disclosed, knowledge that the data would not be used by any of the bodies
specifically referred to by the DfE and Ofqual when making the commitments, and
the limited and meaningless performance comparisons that could be undertaken
using such data, we find that any perceived breach of commitment would not, in
any substantial way, weaken schools and colleges trust in DfE, Ofqual or
Awarding Organisations, to the extent that it would be likely to have any weighty
adverse consequences for the delivery of government education policy in the
future. Although some weight should be attached to such prejudice in the public
interest analysis, we do not accept that it should be accorded significant weight, as
26
EA/2021/0234
contended for by Ofqual and the DfE.
119. We finally turn to Ofqual’s submission that disclosure of the Requested
Information would be prejudicial to the integrity of grading in future years. This
submission is founded on the possibility that, as of November 2020, the 2021
examinations would not go ahead as normal, and that teachers would once again
have an involvement in alternative arrangements for the awarding of grades. It is
contended by Ofqual that if the Requested Information had been disclosed on 5
November 2020, contrary to the assurances identified above, teachers would have
been unlikely to believe any future assurances relating to the publication of data
relating to grades determined by means other than by way of examinations. This,
it is said, would lead to additional pressures on teachers associated with the
publication of data which, in turn, could affect their approach to whatever
involvement they might have in alternative arrangements put in place for 2021.
120. In this regard we observe, in particular, the evidence provided by Julie Swan and
that set out in AQA’s letter to Ofqual, where AQA assert that the assurance that
teachers’ decisions would remain confidential and not open to scrutiny “enabled an
environment for teachers to participate fully and be objective, without having to be
concerned about their own situation, as well as that of their students”. It is further
asserted that if the Requested Information were to be disclosed, it would not be
possible to recreate such an environment.
121. We accept that, as of November 2020, it was not known whether in future years, in
particular in 2021, A level and GCSE examinations would go ahead as normal.
Consequently, we accept that there was, at that time, a possibility of teacher
involvement in the award of A level and GCSE grades in future years. We also
accept that had the Requested Information been disclosed in 2020 there would
have been a possibility that some teachers might have viewed this as a breach of
the 2020 assurances relating to the publishing of performance data, as discussed
above. Whilst there is a chasm between publishing data of the type that would
inexorably lead to a comparison between CAGs provided by an individual teacher
and the associated calculated grades, and data of the type in the Requested
Information, we, nevertheless, accept that there is a possibility that some teachers
may view the breach of the assurances occasioned by disclosure of the Requested
Information as being of such significance that the “undue pressure” imposed on
them as a consequence “could affect the accuracy of their alternative grading method”
[see Ofqual’s skeleton argument at [43]]. However, in all the circumstances, we
find the risk of this occurring, to be low. In addition, we accept the ICO’s
submission that it is likely that any future alternative system for awarding GCSE
and A level grades would have a ‘quality assurance’ designed to ensure the
integrity of the awarding system. For these reasons, we find that the risk of
prejudice to the integrity of grading in future years occasioned by disclosure of the
Requested Information, to be low and that only limited weight should be attached
to such matter in the public interest analysis.
27
EA/2021/0234
122. Given what we have said above, having considered the balance of public interests
as of 5 November 2020 we find that the public interest in maintaining the section
36 (2)(c) exemption in relation to the Requested Information, whether the matters
relied upon by Ofqual are taken separately or cumulatively, is outweighed by a
significant margin by the strong public interest in favour of disclosure.
123. Consequently, we find, as did the ICO, that Ofqual’s response to the FOIA Request
was not accordance with Part I of FOIA. The ICO’s Decision Notice is in
accordance with the law and Ofqual’s appeal is dismissed.
Signed: Dated: 4 June 2022
Judge O’Connor
Chamber President
28