James Coombs v The Information Commissioner & Anor [2023] UKFTT 1026 (GRC)

First-tier Tribunal
General Regulatory Chamber
[2023] UKFTT 01026 (GRC)Date 18 March 2024
Information Rights
Appeal References: EA/2020/0310
Heard remotely (by CVP) on 27-28 November 2023
Deliberations in private on 29 November 2022
Before
Judge ANTHONY SNELSONTRIBUNAL MEMBER EMMA YATESTRIBUNAL MEMBER DAVE SIVERS
Between
JAMES COOMBSAppellantTHE INFORMATION COMMISSIONERRespondentTHE BUCKINGHAMSHIRE GRAMMAR SCHOOLSRespondentDecision
[1]The Appellant, Mr James Coombs, to whom we will refer by name, holds an undergraduate degree in Mathematics with a specialism in Statistics and a Masters degree in Data Science, and has recently been accepted by the Institute of Education at Reading University to undertake a quantitative research-based PhD. He is a committed supporter of comprehensive education and strongly opposed to selective education, including that provided in the state sector through grammar schools.[2]The Second Respondent, The Buckinghamshire Grammar Schools, which we will call TBGS, is a company limited by guarantee set up by Buckinghamshire’s 13 Grammar Schools to manage and administer the annual 11+ Examination, known as the Secondary Transfer Test (‘the Exam’).[3]GL Assessment Ltd (‘GLA’), a company registered in the UK and part of a large US-based education services group called Renaissance, has been party to a contract with TBGS since 2017 under which it designs and supplies test materials and provides numerous other services relating to the Exam.[4]The Exam for the 2020 Entry was held in Buckinghamshire on 12 September 2019. Problems arose because there were defects in the Verbal Reasoning element of the English & Verbal Reasoning paper. As a result, GLA and TBGS agreed upon a statistics-based solution involving artificially crediting all candidates with one raw mark in respect of each of the two defective questions and discounting the last six questions. Public statements put out by both organisations on 1 October 2019 explained the decision and defended it as a measure which was based on ‘detailed statistical analysis‘ and ensured fairness for ‘all’ the children involved.[5]On 13 October 2019, Mr Coombs submitted a request for information to TBGS, pursuant to the Freedom of Information Act 2000 (‘FOIA’) To which any section number mentioned below refers , in the following terms: 1. Please provide a copy of the “detailed statistical analysis” referred to in your letter dated 1 October 2019 … so that truly independent members of the public can satisfy themselves that the issue has been resolved in a fair manner for all children and that the results are indeed robust. Please also provide the following specific information if it is not included in the report. 2. The number and nature of the ‘subtests’ making up the overall assessment (e.g. Verbal Skills, Comprehension, Maths/Numeracy and Non-Verbal Reasoning) 3. For each subtest please provide the number of questions set and reliability when the tests are set and administered without any errors. 4. Specific to the recent errors, for each subtest please providea. The number of questions removed from the assessment andb. The revised reliability.[6]On 12 November 2019 TBGS responded. It disclosed the information requested at para 4a but refused the remainder, citing ss41 (confidential information) and 43 (commercial interests).[7]Mr Coombs challenged the response but, following an internal review, TBGS maintained its position.[8]Mr Coombs then complained to the Information Commissioner (‘the Commissioner‘) about the way in which his request for information had been handled. An investigation followed.[9]By a decision notice dated 22 September 2020 (‘the DN’), the Commissioner purported to hold (in the ‘Decision’ section of the document, at para 2) that TBGS had correctly applied ss 41 and 43(2). In the accompanying reasons, however, she dealt only with s41 and stated that the exemption under s43(2) had not been considered.[10]By a notice of appeal dated 24 October 2020 Mr Coombs appealed to the Tribunal.[11]In due course, TBGS was joined as Second Respondent to the appeal.[12]By a decision dated 22 February 2022 the Tribunal, in a constitution which included none of the three members of this Tribunal, dismissed the appeal.[13]Mr Coombs appealed to the Upper Tribunal. By a decision dated 5 July 2023 Upper Tribunal Judge Citron allowed the appeal, set aside the first-instance decision and remitted the matter for rehearing before a fresh Tribunal, subject to the qualification that it must treat it as established in fact that the statistician’s report referred to in the request was not held by TBGS (or any other person on its behalf) at times relevant to the appeal.[14]The remitted appeal came before us in the form of a ‘remote’ hearing, conducted by CVP, with three days allocated. Mr David Lawson, counsel, appeared on behalf of Mr Coombs. Ms Felicity McMahon and Ms Hannah Gilliland, both counsel, represented TBGS. We are most grateful to them for the helpful and cooperative way in which they presented their cases.[15]The Commissioner was not represented before us and was content to rely on his written case, which corresponded closely with that of TBGS.[16]We heard oral evidence from Mr Coombs and his two supporting witnesses, Mr Alan Parker and Mr Luke Knightly-Jones, and, on behalf of TBGS, Ms Sue Walton and Mr David Hilton.[17]In addition to witness evidence, we were presented with two voluminous bundles of open documents and a slim closed bundle. The paperwork was completed by the helpful skeleton arguments on both sides.[18]There was a brief discussion about whether or not to hold a closed hearing. Neither counsel pressed us on the matter one way or the other. In the absence of any evident need to depart from the principle of open justice (even briefly), we decided against holding a closed session.[19]Closing submissions were presented on the afternoon of day two, following which we reserved our decision. By that point, it was common ground that the information which was the subject of the appeal was limited to(a) the slides used for the PowerPoint presentation made by GLA and shared with Ms Walton and members of the Board of TBGS on 24 September 2019;(b) the number of questions in the Exam (although that information was contained within the PowerPoint presentation anyway); and(c) the information concerning the reliability of the Mathematics & Non-Verbal Reasoning elements of the Exam. As to (c), Ms McMahon pragmatically took the line that, although this information was not (on TBGS’s case) held at the time of the request or the refusal, the Tribunal should proceed on the footing that it was so held and determine the substance of the dispute concerning the applicability of the two exemptions cited.[20]We completed our deliberations in private on day three. In the result, we were divided. The majority, consisting of Tribunal Member Yates and Tribunal Member Sivers, took the view that the appeal against the Commissioner’s decision on s41 should be dismissed and that TBGS had also been entitled to withhold the information in reliance on s43(2). The minority member, Judge Snelson, would have allowed the appeal on the basis that TBGS had not been entitled to succeed under either of the exemptions cited.

The Statutory Framework

[21]FOIA, s1 includes: (1) Any person making a request for information to a public authority is entitled– (a) to be informed in writing by the public authority whether it holds information of the description specified in the request, and (b) if that is the case, to have that information communicated to him. ‘Information’ means information “recorded in any form” (s84).[22]The right under s1 is subject to exemptions. FOIA, s41 includes: (1) Information is exempt information if – (a) it was obtained by the public authority from any other person …, and (b) the disclosure of the information to the public (otherwise than under this Act) by the public authority holding it would constitute a breach of confidence actionable by that or any other person.[23]The exemption under s41 is absolute (see s2(3)(g)).[24]Counsel were agreed that the question of actionable breach of confidence is most conveniently addressed by applying the three-part test propounded by Megarry J in Coco v A N Clark (Engineers) Ltd [1966] RPC 41 at p47, where the learned judge said this: In my judgment, three elements are normally required if, apart from contract, a case of breach of confidence is to succeed. First, the information itself, in the words of Lord Greene MR in Saltman Engineering Co Ltd v Campbell Engineering Co Ltd,must “have the necessary quality of confidence about it”. Secondly, that information must have been imparted in circumstances importing an obligation of confidence. Thirdly, there must be an unauthorised use of that information to the detriment of the party communicating it.[25]Counsel also agreed that, where the Coco test is satisfied, a further question will arise, namely whether the (hypothetical) action for breach of confidence could withstand a public interest defence. It is well established that any public interest defence is for the party advancing it to make good. This being so, and the exemption being absolute, it follows that the burden falls upon the requester to show that a public interest defence to the breach of confidence action would succeed. In other words, there is an in-built public interest test, but here there is a presumption against disclosure whereas, in any case where a qualified exemption is relied upon, the presumption (under FOIA, s2(1)(b)) is in favour of disclosure. But that said, the presumption in each case is a mild one. The presumption wins the day where the issues are finely balanced, but it can readily be displaced by an overriding public interest the other way. Overriding does not mean overwhelming. Moreover, and for the avoidance of doubt, if and in so far as Ms McMahon sought to suggest that the presumption against disclosure under s41 (where the Coco test is satisfied) is somehow weightier or more powerful than the presumption in favour of disclosure under s2(1)(b) where any qualified exemption is under consideration, we reject that submission. We direct ourselves in accordance with the following guidance of the Court of Appeal in Brevan Howard Asset Management LLP v Reuters Ltd [2017] EMLR 28 (para 75): The only question which the Judge had to address, and which he did address, was whether the important public interest in the observation of obligations of confidence was outweighed by sufficiently significant matters of public interest in favour of publication. Unless his conclusion on that issue was one which no judge could properly reach, or he was swayed by matters he wrongly took into account or by failing to take into account matters he should have considered, his decision cannot be disturbed on appeal.[26]By s43(2), information is exempt if its disclosure under FOIA, ‘would, or would be likely to, prejudice the commercial interests of any person (including the public authority holding it).’[27]In assessing prejudice and/or the risk of prejudice for the purposes of s43, we direct ourselves in accordance with the decision of the FTT in Hogan and Oxford City Council v ICO (EA/2005/0026), which proposes three questions. First, what interest (if any) is within the scope of the exemption? Second, would or might prejudice in the form of a risk of harm to such interest(s) that was ‘real, actual or of substance’ be caused by the disclosure sought? Third, would such prejudice be ‘likely’ to result from the disclosure in the sense that it ‘might very well happen’, even if the risk falls short of being more probable than not? (Hogan is, of course, not binding on us but it draws directly on high authority In particular, on the meaning of ‘likely’, the judgment of Munby J in R (on the application of Lord) v Secretary of State for the Home Office [2003] EWHC 2073 (Admin). and has been specifically approved by the Court of Appeal: see Department of Work and Pensions v IC [2017] 1WLR 1.)[28]If a qualified exemption, such as that under s43, is shown to apply, determination of the disclosure request will turn on the public interest test under s2(1)(b), namely whether, ‘in all the circumstances of the case, the public interest in maintaining the exemption outweighs the public interest in maintaining the exemption’. The proper approach, as explained by the Upper Tribunal (‘UT’) in APPGER v IC [2013] UKUT 560 (para 149) 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 would (or would be likely to or may) cause or promote.[29]The relevant date for the purposes of applying any public interest balancing test and, it seems, determining the applicability of any exemption, is the date on which the request for information was refused, not the date of any subsequent review (see Montague v ICO and DIT [2022] UKUT 104 (AAC), especially at paras 47-90). The UT decision was overturned by the Court of Appeal (see [2023] EWCA Civ 1378), but not on this point.[30]The appeal is brought pursuant to the FOIA, s57. The Tribunal’s powers in determining the appeal are delineated in s58 as follows:(1) If on an appeal under section 57 the Tribunal consider – (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.(2) On such an appeal, the Tribunal may review any finding of fact on which the notice in question was based. The Key Facts Arrangements for grammar schools admissions in Buckinghamshire

The Key Facts

[31]In her third witness statement, Ms Walton provided a helpful summary of how the system operates in Buckinghamshire. We cannot do better than reproduce it in full. 4. There are two stages to grammar school admissions in Buckinghamshire:a. qualifying for grammar school andb. applying for a grammar school place. TBGS oversees the qualification process only. Qualifying for grammar school 5. To qualify for grammar school in Buckinghamshire, children need to sit the Buckinghamshire Secondary Transfer Test (STT). Unlike in other areas children attending Buckinghamshire primary schools all take the STT unless their parents decide to opt them out. Children outside of Buckinghamshire also apply to take the Buckinghamshire STT. Approximately ten thousand children take our STT each year. 6. The Buckinghamshire STTs are produced by GL Assessment with a new test provided each year. Each test comprises two test papers covering three areas: verbal skills, maths and non-verbal skills. Children write their answers on separate answer sheets. These are then machine-marked with 1 raw mark awarded for each correct answer. The raw marks for each of the three areas are then age-standardised to ensure younger children are not disadvantaged. The child’s three age-standardised scores are then added together to give the Secondary Transfer Test Score using the weightings: verbal skills 50%; maths 25%; non-verbal 25%. 7. In order to qualify for grammar school, children need to achieve a Secondary Transfer Test Score of 121 or above. Unlike most other selective areas, in Buckinghamshire we do not rank order children. Children just need to achieve a minimum score of 121 to qualify – there is no advantage in gaining a higher score. Results are released to parents in mid-October. Under the national School Admissions Code selective schools are required to provide results to parents ahead of 31 October which is the deadline for parents to finalise and submit their secondary school preferences for their child. 8. If children do not qualify for a place but parents think that their child should go to grammar school, they have two options for appealing. The first is Selection Review, and the second is via an appeal following the allocation of places, as explained below. 9. When parents apply for Selection Review, they have to provide reasons for why their child did not qualify in the test. The headteacher of their primary school also provides information including details about the child’s academic record. This information is considered by a Selection Review Panel of two grammar school headteachers and one primary school headteacher. In submitting their evidence parents are advised that: a. Strong academic evidence will be vital to the success of a review. b. The Selection Review Panel (SRP) will look for clear academic evidence to show that a grammar school would be best for your child.c. The Panel will also want to see evidence of any exceptional reasons to show why your child did not do as well as expected in the Secondary Transfer Test. 10. Having considered the evidence, the SRP can qualify a child. All children who have qualified automatically or via SRP are eligible to apply for a place at a Buckinghamshire grammar school. Applying for a grammar school place 11. The timeline for secondary school admissions is set centrally and applies to all schools in England. All parents must apply for secondary school places by 31 October of the year before their child is due to move up to secondary school. They do this via the Local Authority where they live. In Buckinghamshire parents can list up to 5 secondary schools in priority order. The list may be all grammar schools, all non-selective schools or a mix of the two. It may also include state schools in other Local Authorities. 12. School places are allocated by the Local Authority taking account of the preference order given by parents and then if there are more preferences for a particular school than places, the over-subscription criteria for the school are applied in the order specified in the school’s admission policy. The criteria vary slightly but will include priority for looked after children and disadvantaged children, school catchment area and how far the child lives from the school. Parents receive one offer for a secondary school on National Offer Day which is 1 March. If a child has not qualified for grammar school either automatically or via Selection Review the Local Authority automatically ignores any grammar school preferences the parents have listed. 13. If a parent thinks one or more grammar schools should have offered their child a place (whether the child qualified or not) they have the right to appeal that decision. School appeals must follow the national School Admission Appeals Code and all appeals must be heard by panels completely independent of the school. In Buckinghamshire most of the grammar schools buy into the independent appeals service administered by Buckinghamshire Council. The others make their own arrangements.

The Exam

[32]Some 10,000 children sit the Exam in Buckinghamshire each year. It takes the form of two multiple-choice tests: Verbal Skills, subdivided into two elements: English and Verbal Reasoning, and Mathematics and Non-Verbal Skills, also subdivided into two elements: Mathematics and Non-Verbal/Spatial Reasoning. One hour is allowed for each test. Allowing for built-in practice time, that leaves not more than 25 minutes for each element (20 in the case of Verbal Reasoning). Candidates are provided with a written paper and separate sheets on which to enter their answers. For each question, five possible answers are offered. A correct answer attracts one raw mark. A wrong answer scores nothing, but there is no penalty for a wrong answer.[33]The papers are machine-marked and the raw scores collected. Age-based factors are then applied to give standardised scores. The qualifying mark of 121 is applied to the standardised scores.

Past papers, practice materials and the familiarisation booklet

[34]Past papers for the Buckinghamshire 11+ are not published and the content of the Exam is treated as strictly confidential. Candidates are warned not to discuss the papers after sitting them.[35]GLA does, however, publish for gain and actively market ‘practice materials’ to assist putative candidates in preparing to sit the Exam. Many other organisations in the assessment and testing field also publish and market such materials.[36]GLA also publishes a ‘Familiarisation Booklet’ which gives practical guidance concerning the Exam and includes practice questions in all four disciplines, details of the correct answers together with briefly reasoned explanations, and sample answer sheets. Among other things, it advises candidates not to ‘spend ages’ on any question and, if stuck, to move on and come back to any unanswered problem at the end, if time allows.[37]We have no doubt that the published practice materials and the Familiarisation Booklet, which are intended to be informative and representative, fairly capture the broad nature of the Exam and the level of difficulty to be expected. No one suggested anything to the contrary.

Tutoring

[38]We were told that tutoring of putative candidates ahead of the Exam is widespread. Such support comes at a price and children of poorer families are likely not to receive it at all or, at best, to receive it less frequently than children of more affluent families.[39]No doubt for good reason, TBGS and GLA consider that tutoring may provide members of better-off families with an unfair advantage, and that nothing should be done to encourage tutoring for the Exam.

The business and contractual relationship between TBGS and GLA

[40]TBGS and GLA entered into the contract which governs the relationship between them in December 2017. Its initial term was for two years, but it has been extended and continues to run. Under it, GLA undertook for reward to manage and administer the Exam on behalf of TBGS. Across the annual cycle, its obligations are substantial, including designing the precise form and content of the Exam As already noted (see Ms Walton’s first statement, para 6, cited above), the content of the Exam is new every year. This is required under the contract: schedule 1, paragraph 12 stipulates that no ‘item’ (question) will have been used in any secondary selection test in Buckinghamshire or any neighbouring area. To ensure that the robustness of the Exam is maintained from year to year, GLA is also obliged to introduce new ‘item types’ from time to time (ibid). , making all necessary arrangements for ‘Test Day’ (the normal practice is for candidates to sit both parts of the Exam on the same day in two one-hour sessions), overseeing the raw marking and standardisation processes and delivering a full report to TBGS.[41]The contract between TBGS and GLA is subject to strict terms binding the parties The preponderance of the confidentiality obligations are upon GLA. to treat the ‘Test Materials’ as confidential. ‘Test Materials’ are defined to include, ‘answer sheets, question papers, test administration instructions (of whatever nature), test data, analyses and reports and such other material as may be required to enable the children to undertake the Tests’ (cl 1.1).

The errors, the Solution and related public communications

[42]In 2019 the Exam took place on 12 September. Soon after 9 a.m. Buckinghamshire Council started to receive reports about errors in the Verbal Reasoning element of the English & Verbal Reasoning paper. The errors were quite swiftly identified and messages sent to the various test centres asking them to reassure the children and instruct them to move on from the defective questions. In some cases, these reached test centres before the Exam began, with the result that pupils were forewarned. In other cases, the information reached the test centres (or at least those administering the test) when the Exam was already underway or even after it was over.[43]Not surprisingly, the errors caused a substantial degree of upset and consternation among pupils and their families.[44]Urgent consultation between TBGS and GLA followed. It was established that the errors were three-fold. First, for one Verbal Reasoning question, although the correct answer was shown on the answer sheet, three of the other answer options did not match those on the question paper. Second, for another Verbal Reasoning question, the correct answer was not included on the answer sheet. Third, although a practice question in the English element could be answered, one of the five other answer options did not match the answers on the question paper. Following investigation, GLA and TBGS agreed that the defective practice question had not affected candidates’ performance and that accordingly no remedial action was needed in relation to the English element of the paper.[45]Several possible solutions were canvassed. One was to ask the pupils to re-sit the Verbal Reasoning element (or perhaps the entire English & Verbal Reasoning test). Another was to devise a solution based on a statistical manipulation of the scores in order to take account of the errors and seek to nullify their effect. It is clear that TGBS and GLA were strongly opposed to the re-sit option. Ms Walton, speaking for TGBS, told us that this was because of the perceived need for a rapid resolution of the problem and the difficulty of accommodating a fresh sitting of part of the Exam in time for the national 31 October deadline for the delivery of secondary school preferences. Mr Hilton told us that GLA opposed a re-sit because of the added stress which it would place upon the affected children. Neither acknowledged any desire to avoid reputational damage or manage public relations as a factor Ms Walton’s explanation before us was also not consistent with a contemporary document put out by TBGS (Open Bundle, pp413-4), which said that the re-sit option was excluded because it would be too stressful for the children. . Mr Coombs took a more jaundiced view. It is not our place to enter into this disagreement.[46]What is beyond question is that, in short order, GLA came forward with a proposal which involved awarding a mark in respect of each of the two defective questions and discounting the last six Verbal Reasoning questions.[47]Immediately following a meeting with GLA on 20 September, TBGS sent a letter to parents ruling out the possibility of any re-sit. This was before any final decision had been taken as to how to deal with the errors and their consequences.[48]On 24 September 2019 Mr Alistair Durno, GLA’s Head of Admissions Testing, wrote to TBGS referring to a ‘remote’ meeting to be held later that day and warning that the information to be presented was strictly confidential and that disclosure of it would result in legal action. Mr Mark Sturgeon, Chair of TBGS, formally acknowledged that the information would be received on that understanding.[49]Later on 24 September 2019 the scheduled ‘remote’ hearing took place, at which the PowerPoint presentation was delivered to Mr Sturgeon and Ms Walton. In support of the proposed solution, the point was made that it had been approved by an ‘independent statistician’. Mr Sturgeon and Ms Walton were persuaded, subject to approval being given by the full Board of TBGS.[50]Mr Sturgeon and Ms Walton attended a meeting of the full Board of TBGS on 27 September 2019. They went through the salient points in the PowerPoint presentation. The members of the Board agreed to the proposed solution (hereafter, ‘the Solution’), although they did require confirmation about the independence of the ‘independent statistician’.[51]By letter from GLA’s HR Department dated 30 September 2019, Ms Walton was notified that the company’s records went back only to 2011 and that the ‘independent statistician’ had not worked for it since then. It seems that no further enquiry was made into any possible connection between him and GLA.[52]By letter of 1 October 2019, Mr Sturgeon wrote to the parents and carers of children who had sat the 2019 Exam. Having repeated TBGS’s earlier apology for the errors and acknowledged support from headteachers and the County Council, he continued: Along with this letter is a further letter from GL Assessment explaining what actions have been carried out in order to ensure fair and reliable results for all children. Detailed statistical analysis has been carried out and the solution proposed to and accepted by TBGS is robust. TBGS is confident that the issue has been resolved in a fair manner for all children and that the results for testing are robust. This outcome has been verified by an independent statistician. When results are released on 18 October, as in all previous years, you have the option of going to Selection Review or Appeal if your child has not qualified with the score of 121. … We understand that you may focus upon the testing error, however, with the statistical analysis and solution implementation this will not be sufficient grounds for reviews or appeals on their own. We understood from Ms Walton that the reference to the ‘testing error’ not being a sufficient ground for review or appeal ‘on their (sic) own’ was intended to exclude any challenge based on any complaint of unfairness in the Solution itself but to leave open the possibility of extenuating circumstances relating to the errors being relied upon (she gave the hypothetical example of a child with autism whose performance was affected by distress or confusion caused by the errors). But we were not told of any instance of an appeal being pursued, let alone entertained, on this basis.[53]The letter from GLA to which Mr Sturgeon had referred, also dated 1 October 2019, was signed by Mr Durno. Having explained that the errors had resulted from a ‘failure’ in GLA’s ‘quality assurance process’, he explained that the company’s overriding priority had been to ‘ensure all children get a fair result’ and that this had involved a focus on two particular matters: the reliability of the Exam and the particular impact of the errors on the ability of candidates to complete the Verbal Reasoning element. He continued: GL Assessment’s statisticians have reviewed the test performance in detail and passed their findings to an independent statistician, who has been approved by [TBGS]. … In relation to the reliability of the test to determine whether or not a child is suited to a grammar school, our analysis found that the overall reliability of the test has not been compromised and remains very good. However, we recognise the need to consider the impact on each individual child given the different instructions they may or may not have been given. We considered two specific possibilities: that the time spent by children on attempting to resolve the problem might mean they were less likely to complete the test; and that children might have been unsettled by being unable to find an answer to the two questions, so that their performance on later ones was affected. We also compared and analysed completion rates between the 2018 and 2019 papers in order to understand how children performed against a similar paper in similar circumstances. Our analysis found that, while the test reliability and completion rates as a whole remains statistically sound, there was evidence that completion rates started to drop during the last six questions of the Verbal Reasoning section of the paper. This evidence is consistent with some of the feedback we have received from parents. Our solution has therefore been based on this evidence. The solution The solution we have agreed with [TBGS] is: 1. All children will be awarded a mark for each of the two erroneous questions, thereby ensuring no advantage or disadvantage from these two questions; 2. in order to ensure no individual child is penalised for not being able to complete the test, the last six questions of the Verbal Reasoning section will not count towards the final mark. It is important to reiterate that the independent statistician has verified that the outcome of the test, without those questions, is still fair for all children, highly reliable and above the accepted conventions for admissions tests.[54]Shortly after 1 October 2019 a ‘FAQ’ document was published on the TBGS website. This confirms that the qualifying score remains 121. The next question is:
‘If you are removing six questions from the Verbal Reasoning section, how can you still have a qualifying score of 121?’
The answer offered is:
‘A child’s final score is an age-standardised score rather than a raw score. (A raw score is just the total of the correct marks on the papers.) The standardisation process has taken into account the removal of the last six Verbal Reasoning questions in order to maintain the qualifying score at 121.’
How the standardisation process takes account of the removal of the six questions is not explained. The next question is:
‘Won’t the solution penalise children whose strength is Verbal Reasoning?’
To this, the following reply is given: No, both because there are a high number of VR questions within the paper so each child will still be able to demonstrate their strengths in VR, and because the weightings of the sections of the test will not be changed. Verbal Skills will continue to have a weighting of 50%. It is important to reiterate that the independent statistician has looked in detail at how children performed in the Verbal Reasoning section of the paper, taking into account the different conditions under which the children sat the paper (i.e. the children who were advised not to attempt the questions, children who were told part-way through the paper, and children who were not informed). The independent statistician has verified that the outcome of the test, without those questions, is still fair for all children, highly reliable and above the accepted conventions for admissions tests. The contention that the child with a particular strength in Verbal Reasoning is not prejudiced by the removal of six questions in that discipline because the 50% weighting applied to the Verbal Skills part as a whole was not explained. The answer does not state that the removal of six Verbal Reasoning questions is statistically insignificant. Nor does it justify (as a matter of statistical analysis or otherwise) the apparently implicit argument that a child (indeed every child) who is strong in Verbal Reasoning will be equally strong in English. As far as we are aware, these unanswered questions arising out of the FAQ document are nowhere addressed in any of the evidence before the Tribunal.[55]The identity of the ‘independent statistician’ was not given in either of the letters of 1 October 2019. Eventually, after some pressure was exerted, he was identified as Dr Dougal Hutchison.[56]By letter of 11 November 2019 addressed ‘to whom it may concern’, which TBGS published, Dr Hutchison set out the relevant background and made (among others) the following observations: One possibility considered was that time effectively wasted by candidates on attempting to resolve the problem would mean that they were less likely to complete the test. … There was some evidence of differential completion rates starting to have a noticeable effect on the No Warning group on the last six questions in the Verbal Reasoning section of the Verbal Skills paper. … Another effect of the anomaly was that candidates might be sufficiently discomfited by being unable to find an answer to the anomalous items that their performance on later ones would be affected: this would be expected to be most pronounced in the No Warning group. Differential Item Functioning was investigated … Allowing for multiple comparisons, none of the differences were statistically significant. It was concluded that removing a proportion of the questions from the VR component of the Verbal Skills test, namely the final six questions from the end of the VR component as well as awarding a mark for each of the two anomalous questions, was the fairest approach to resolving this issue since it meant that all candidates were rated on the same items and it was not considered that the remaining items were differentially affected by the anomaly. … I am now able to confirm that the statistical analysis verifies the approach taken to resolve the issue and has produced a fair and robust solution to resolve the issue. In his letter, Dr Hutchison did not make the assertion attributed to him by Mr Sturgeon and Mr Durno in their letters of 1 October 2019 and repeated in the FAQ document that the Solution was ‘fair’ to ‘all’ candidates. Nor did he offer a view as to whether it would have been fairer to arrange a re-sit of the Verbal Reasoning element. Nor did he state at whose behest his letter was written or what points he had been asked to cover (or not cover). Nor did he explain the circumstances in which he had become involved in the issues arising out of the 2019 Exam.[57]In fact, as appears now to be common ground, although he was consistently presented as ‘independent‘, Dr Hutchison had a very long-standing professional association with GLA. He had worked for the company, then trading under a different name, for about 24 years ending in 2010. It seems that the connection was never disclosed by GLA or TBGS It is unclear when TBGS was first made aware of it. or Dr Hutchison, and was uncovered by Mr Coombs.[58]In an undated document headed ‘Issues with the 2020 entry Buckinghamshire Secondary Transfer Test’, published after 11 November 2019, TBGS set out a fresh narrative of the errors, the investigation and the Solution, and disclosed the new information that the reliability of the Verbal Skills paper, following application of the Solution, had been calculated at 0.903 (against a maximum of 1.0). The document did not disclose whether separate calculations had been made in respect of each of the two elements of the Verbal Skills paper. Nor did it state what measure of reliability had been taken.

Reliability

[59]We heard quite a lot of evidence about reliability. In the context of statistics, we understand the term to refer to the degree of consistency with which a test or method measures something. It is a subject which stirs strong passions, not only among statisticians. Fortunately for us, it is not our function to reach any view on the arguments about it which were canvassed before us, but it may be useful to outline the rival positions. In bare summary, TBGS says that the reliability calculation in respect of the Verbal Skills test was appropriately published in order to reassure all interested parties that the Solution had resolved the difficulties arising from the errors in a manner which was fair and statistically sound. It resists disclosure of the corresponding calculation in respect of the Mathematics and Non-Verbal Skills test on the ground that the information is confidential. When it refers to reliability, it means ‘internal consistency reliability’ which is, on its case, the only appropriate measure which can be applied to the Exam. Mr Coombs replies that ‘internal consistency reliability’ is inapplicable and the proper measure is ‘test/re-test reliability’. He further contends that it is elementary that shortening a test (for example by deleting eight questions) inevitably reduces its reliability. He considers that the published reliability number post-Solution is at best the result of incompetence and at worst designed to mislead.[60]One fact can be confidently stated: it is not the current practice nationally to publish reliability statistics in relation to 11+ tests. The reason for this is not clear. The same practice does not apply to other tests, such as GCSEs.

What the PowerPoint material contains

[61]In his first witness statement, para 19, Mr Hilton set out the nature of the confidential information contained in the PowerPoint slides, as follows (entries in square brackets added by the Tribunal): Slide 2: Details of how the completion rate analysis was conducted. Slide 3: Reliability figures by section and number of items … Slide 4: Details of how the completion rate analysis was conducted. Slide 5: Number of items per section, as well as completion rates for [English] section. Slide 6: Number of items per section and completion rates for [English] section. Slide 7: Number of items per section and completion rates for [English] section. Slide 8: Mean raw scores for sections, number of items for Verbal Reasoning and completion rates for section. Slide 9: Number of Verbal Reasoning items and completion rates. Slide 10: Number of Verbal Reasoning items and completion rates. Slide 11: Details of how the completion rate analysis was conducted. His summary is adequate for the purposes of identifying the main categories of information contained in the PowerPoint slides.[62]Mr Coombs’s case is that, without disclosure of the disputed information neither he, nor the public at large, will be in a position to make a comprehensive, independent, statistics-based assessment of the fairness of the Exam (post-Solution). We are agreed that he is plainly right about that. Ms McMahon did not argue to the contrary. The question is whether the exemptions relied upon by TBGS (or either of them) preclude him from doing so.

The Rival Arguments

[63]The valuable skeleton arguments prepared by counsel on both sides, which may be read alongside the draft lists of issues which both counsel produced and which (as counsel agreed) say more or less the same thing in slightly different words, can largely be left to speak for themselves. What follows is a brief summary of the central submissions advanced. Information provided in confidence[64]Mr Lawson submitted that s41 was not engaged because:(a) the disputed information did not have the necessary quality of confidence;(b) in context, it was not communicated to TBGS in circumstances importing an obligation of confidence; and(c) disclosure would not operate to the detriment of GLA and the detriment-based arguments advanced on its behalf (in particular the ‘Tutor Advantage’ argument and the ‘Competitor Advantage’ argument) were without substance. The arguments are considered in the Tribunal's analysis and conclusions below. For now it is enough to say that the Tutor Advantage argument is premised on the notion that disclosure of the information would operate to the advantage of tutors and candidates they were hired to prepare for the Exam, thereby disadvantaging candidates whose parents could not afford such assistance and so undermining the fairness of the Exam. The Competitor Advantage argument holds that disclosure of the information would present any current or future competitor of GLA with an unfair advantage in the market.[65]Mr Lawson further submitted in any event that disclosure of the information would not be ‘actionable’ because any claim by GLA based upon it would be defeated by a public interest defence.[66]Ms McMahon submitted We did not hear from Ms Gilliland. Purely for brevity, we will credit Ms McMahon with the submissions advanced, although no doubt Ms Gilliland contributed to their preparation. that s41 was fully engaged in that:(a) the information in issue did have the necessary quality of confidence;(b) it was communicated to TBGS in circumstances importing an obligation of confidence; and(c) disclosure would entail detriment to GLA, reliance being placed on the Tutor Advantage and Competitor Advantage arguments which, on her case, had compelling force.[67]Ms McMahon further resisted Mr Lawson’s contention that any action by GLA based on disclosure of the information would be met by a successful public interest defence. She submitted that the public interest was firmly in favour of maintaining the confidentiality of the disputed information. Commercial interests[68]Turning to s43(2), Mr Lawson, noting that the Commissioner had given no adjudication on that exemption, submitted first, that the disputed information, in particular that relating to the number of questions and the reliability calculation(s), did not amount to GLA’s (or TBGS’s) intellectual property and was not in any event information such as could engage the subsection. Secondly, he disputed GLA’s case on prejudice, contending that it fell well short of establishing any real or significant risk of harm. Thirdly, and in any event, he submitted that, even if the exemption was engaged, the public interest strongly favoured disclosure of the information.[69]Ms McMahon raised a frontal challenge to all three elements of Mr Lawson’s submission. Analysis and Conclusions – Majority Opinion

Information provided in confidence

[70]The first question posed by s41 is whether the disputed information was obtained by TBGS from GLA. It was, as the parties agree.[71]The second, and controversial, question is whether disclosure of the information by the public authority holding it (TBGS) to the public (otherwise than under FOIA) would constitute a breach of confidence actionable by GLA or any other person. As counsel agreed, this issue is most conveniently addressed by considering the three sub-questions identified by Megarry J in the Coco case (cited above).[72]The first sub-question is whether the information has the necessary quality of confidence. In the view of the majority, this requirement is fully satisfied. The information (not only the analysis and reasoning underpinning the Solution but also the number of questions in each part of the Exam and the reliability statistics) was important and sensitive. Although, self-evidently, TBGS already knew the number of questions, the information as a whole was communicated by GLA to TBGS on a strictly confidential basis and treated as such by TBGS (see further below). None of it had been made public before.[73]The second sub-question is whether the information was communicated to TBGS in circumstances importing an obligation of confidence. Again, the majority view is that this requirement is clearly met. As already noted, the contract between GLA and TBGS contained detailed terms imposing mutual obligations to ensure the confidentiality of the ‘Test Materials’, which include ‘test data, analyses and reports’. In addition, as we have mentioned, the letter of 24 September 2019 from Mr Durno of GLA to Mr Sturgeon, Chair of TBGS, referred explicitly to the PowerPoint presentation and other information about to be disclosed, stressing the strictly confidential nature of the material. Mr Sturgeon signed the letter to signal TBGS’s acceptance of its terms.[74]The third sub-question is whether unauthorised use of the information would be detrimental to the party communicating it (here GLA). As we have noted, two main contentions on detriment were advanced on behalf of TBGS: the ‘Tutor Advantage’ argument and the ‘Competitor Advantage’ argument. These will be considered in turn.[75]The majority view is that there is some, albeit limited, force in the Tutor Advantage argument. The following points are made.(1) Any release of specific information about the Exam, including the number of questions it contains, would be of some benefit to tutors offering their services to prospective Exam candidates.(2) In particular, knowledge of the number of questions would remove a point of uncertainty and might lead to tutors suggesting a target (average) time within which candidates should seek to complete each question. The information (at least relating to question numbers) might thus operate to the advantage of prospective candidates from more affluent families who could afford tuition, in circumstances where those from poorer families would enjoy no such advantage.(3) This would be to the detriment of GLA in that it would, or might, suffer reputational damage as a consequence of the fairness of the Exam, which it was responsible for devising and administering, being undermined or appearing to be undermined.(4) These things said, the majority accepts the limitations of the Tutor Advantage argument and recognises that, as a matter of common sense, tutors could at the time of Mr Coombs’s request (and can now) be expected to base their coaching of prospective candidates largely on priorities which would not be affected by knowledge of the precise number of questions in the Exam or any other part of the disputed information, namely (a) developing their relevant subject-matter skills through practice, (b) training them to answer the questions as quickly as is consistent with accuracy, and (c) impressing upon them the critical importance of tackling the entire paper, even if, owing to shortness of time, the last questions can be completed only by resorting to guesswork.(5) More specifically, the majority rejects the wider and somewhat surprising theories advanced by Mr Hilton and Ms Walton concerning the tactics which tutors might use if made aware of the number of questions in the 2019 Exam (which theories are discussed in the minority member’s analysis below).[76]The majority sees more force in the Competitor Advantage argument, for the following reasons.(1) At the time of the refusal of the request, there was a major competitor to GLA in the market. The fact that it has since withdrawn from the market is, strictly speaking, beside the point.(2) The absence now of a genuine competitor is, in any event, of limited significance. The Competitor Advantage argument is persuasive in relation to a potential competitor as well as to a competitor already in the field.(3) Release of the information to the competitor (or potential competitor) would expose GLA to unfair competition. The competitor would have the benefit of information in circumstances where its corresponding information would be shielded from public view and GLA would be denied any chance of scoring a commercial advantage from it.(4) The information is, by its nature, commercially sensitive, and the competitor (or potential competitor) would be in a position to exploit it for its own advantage. The information would provide a benchmark against which a competitor could seek to develop a rival test, which it could market as superior to the Exam and better value for money. Information about question numbers would present a competitor with the option to copy GLA’s model or, if it saw a benefit in doing so, deviate from it. If shown GLA’s reliability figures (for the Mathematics & Non-Verbal Reasoning elements of the Exam) a competitor might be incentivised to seek to devise a rival test yielding a higher reliability score. Whatever the use made of the disputed information, GLA would not enjoy a corresponding opportunity to view and exploit any rival’s information.(5) Release of the information relating to the number of questions would be likely to cause GLA a commercial disadvantage as against any competitor or potential competitor in that it would (or might) feel compelled to alter the structure of the Exam more frequently than hitherto in order to guard against the danger (or perceived danger) of tutors deriving an advantage from knowing the question numbers in the 2019 test.(6) The detriment under (5) would lie not only in the increased costs associated with amending the structure of the Exam but also in the relative commercial disadvantage consequential upon the fact that any competitor would not, or might not, be put to such expense.[77]Taking its analysis of the Tutor Advantage and Competitor Advantage arguments together, the majority holds that disclosure of the disputed information would occasion detriment to the party communicating it (GLA) and that accordingly the third sub-question must also be answered in the affirmative.[78]For the reasons stated, the majority arrives at the conclusion in principle that disclosure of the disputed information by TBGS (otherwise than under FOIA) would constitute a breach of confidence actionable by GLA.[79]This conclusion ‘in principle’ must be tested by the next question, namely whether a breach of confidence action would be met by a successful public interest defence. Whilst recognising the critical importance of education in our national life and the powerful public interest in fostering a well-informed debate on the subject generally and on the issue of selective education specifically, the majority is not persuaded that a public interest defence would succeed here, for the following reasons.(1) The structure of the legislation is such as to create a presumption in favour of protecting information communicated in confidence. It is for a party seeking to override that presumption to make out a good case for doing so.(2) The information in dispute is rightly seen by TBGS and GLA as genuinely sensitive and worthy of protection, and they have taken steps to protect it.(3) This is not a case in which TBGS or GLA can fairly be charged with misleading or attempting to mislead in its public pronouncements concerning the errors in the Exam or the fairness of the Solution.(4) It is not a telling point that (with GLA’s permission) TBGS has released some of the information requested. It acted reasonably in doing so, given the need to explain its decision-making and reassure those affected (and the wider public) that a considered and defensible resolution of the problems arising from the errors in the Exam had been achieved. In these circumstances, it is, if anything, a point against disclosure that a substantial amount of information relating to the errors and the Solution has been made public.(5) The fact that GLA’s competitor has left the market does not materially change matters: even if the Tribunal considered the public interest as at the date of the hearing, it would favour protection of the confidential information. The impact on future tendering of any breach of commercial confidence is an important public interest consideration leaning against disclosure (Jackley v Information Commissioner EA/2016/0082).[80]It follows that, in the opinion of the majority, the exemption under s41 is validly cited and Mr Coombs is not entitled to the information requested.

Commercial interests

[81]Given the majority view on s41, it is strictly unnecessary for the Tribunal to consider s43, but we think it appropriate for us to provide the parties with a decision which covers all disputed points.[82]The first question under s43(2) is whether disclosure of the information would, or would be likely to, prejudice the commercial interests of any person, including the public authority holding it. In the view of the majority, the question must be answered in the affirmative. In this context, as noted above, something is ‘likely’ if it ‘could well happen’. For the reasons given above in relation to the ‘detriment’ points under s41, the majority is satisfied that an appreciable risk of prejudice to the commercial interests of both GLA and TBGS would arise if the information were made public. In the case of the former, it is not necessary to add to the analysis already offered. In respect of TBGS, at least a degree of commercial harm could result if, as a result of disclosure of the number of the questions, it were judged necessary for the structure of the Exam to be changed more frequently than before and any associated costs were passed on by GLA.[83]Turning to the public interest balance under s2(1)(b), the majority relies on its observations above on the (hypothetical) public interest defence considered in relation to s41. Whilst acknowledging that, under s43(2), there is a mild presumption in favour, rather than against, disclosure, the majority takes the view that it is displaced by the factors tending the other way.[84]Accordingly, even if the exemption under s41 had not been engaged, the majority would have held against Mr Coombs in reliance on s43(2). Analysis and Conclusions – Minority View Information provided in confidence

Analysis and Conclusions – Minority View

[85]The minority member does not disagree with the majority on the first two elements of the three-part test proposed by Megarry J in the Coco case.[86]He does, however, differ from his colleagues on the third element, namely the question of detriment. In the first place, he sees no substance in any part of the Tutor Advantage argument. His main reasons are the following.(1) The argument depends on an unhealthy combination of mere assertion and creative speculation. It rests on no evidential basis.(2) The argument is in large part irrational and fanciful, in particular the theory that, if the number of questions (in any year’s Exam) and the mean ‘drop-off points’ (at which candidates ran out of time) were disclosed, tutors in later years would coach candidates to answer at speed up to the (assumed) drop-off point and thereafter to guess. The theory makes no sense and the spectacle of Mr Hilton grimly clinging to it in cross-examination it while attempting to repackage it by means of circumlocution of various sorts was less than edifying. By the end of Mr Hilton's evidence, it was unclear whether he was maintaining that tutors would advise candidates to start guessing at the putative ‘drop-off’ point, regardless of whether they were running out of time or not. If that was his position, the minority member is content to leave it to speak for itself. If his ultimate position was to align himself with Ms Walton (see her third witness statement, para 12) the theory was reduced to the assertion that the tutor would advise candidates to start guessing when they reached the putative ‘drop-off’ point if they were running out of time, but not otherwise. This, said Ms Walton, would give tutored children an unfair advantage. But that made no sense either. All witnesses before us agreed that the right strategy for all candidates was to attempt to complete the entire paper and that if they could not do so otherwise than by resorting in the latter stages to guessing, they should guess (there being no penalty for a wrong answer and a guess standing a 20% chance of being right). In other words, the right time to start guessing was when it became clear that completing the paper without guessing was impossible, not when, as a matter of statistical analysis, any particular proportion of candidates in a previous year had reached the ‘drop-off’ point.(3) The theory referred to at (b) is not only irrational to the point of absurdity, it also ignores the fact that the sole purpose of the Exam is to differentiate between children who achieve the pass mark of 121 and thus qualify for grammar school, and those who do not; the argument that coaching based on knowledge of the number of questions can give an advantage to a tutored pupil relative to one without tutoring makes little sense, since it does not matter by how much (if at all) the tutored pupil’s score exceeds that of his/her untutored peer if both achieve scores of not less than 121.(4) Equally irrational was Mr Hilton’s complaint (first witness statement, para 22) that, ‘If the mean raw scores were widely known, then tutors would use this information to coach students to aim for a certain raw score mark. This would show tutors how difficult the test is, thereby giving them a framework for coaching students.’ In the next paragraph he refers to tutors coaching ‘towards an optimal raw mark’. To state the obvious, tutors can be expected to coach candidates to score the highest marks possible. If a tutor thought that his/her candidate had reached a standard which appeared to correspond with the mean raw scores for 2019, that would not be the moment to stop coaching. The ‘optimal raw mark’ would be the best mark that the candidate could realistically be expected to achieve.(5) In so far as the argument applies to the issue of the number of questions, it also ignores the inevitably narrow range within which the number of questions is likely to fall given the general nature of the questions and the time allowed for each element of the Exam (both being information already in the public domain).(6) The Tribunal heard from at least one witness on behalf of TBGS a vague suggestion that disclosure of the number of questions and/or the reliability data in respect of the maths and non-verbal reasoning elements of the test would or might somehow enable tutors, by a mysterious ‘reverse-engineering’ process, to gain important insights which might inform their advice to prospective candidates. Here again, the Tribunal was presented with mere assertion wholly unsupported by any scientific or empirical analysis. The fact that such evidence was given merely served to undermine TBGS’s case.(7) Unlike that of Mr Hilton and Ms Walton, the evidence provided by Mr Knightly-Jones on the Tutor Advantage point came, in the view of the minority member, from someone with conspicuous learning and an impressive command of the subject matter. He provided detailed, cogent and well-informed reasons as to why disclosure of the disputed information could not result in any appreciable advantage to tutored as against untutored candidates. His analysis rested on empirical foundations and was in keeping with what the minority member regards as basic common sense.(8) The minority member does not share the view of the majority that ‘any’ disclosure of relevant information is liable to cause some degree of advantage to tutors. In his view, there is no evidential basis for that approach. If it is not shown on evidence that the information (or any part of it) can be used, and can realistically be expected to be used, to secure some discernible advantage for tutors and those tutored by them, there is, in the minority member’s view, nothing approaching a detriment of the sort required under the third element of the Coco test.[87]The minority member is also not persuaded by the Competitor Advantage argument, which, he notes, appears to have been accorded greater significance latterly, perhaps because of the exposure by the UT of the obvious difficulties confronting what appeared originally to be the central (if not only) plank of TBGS’s case on detriment, namely the Tutor Advantage argument. Again, he has several reasons.(1) The argument is hugely overstated. It ignores the simple fact that the request is very narrow in scope and is concerned only with information relating to one year.(2) The judge does not accept the assertion that disclosure of the number of questions in the Exam could materially advantage any competitor of GLA’s. As has already been pointed out, given the age of the candidates, the nature of the questions and the time allowed for the Exam, disclosure of the precise number of questions would not tell the competition (if any) anything of real significance. The number of questions in each element of each test will inevitably fall within a narrow range.(3) Moreover, even if it was of interest for a competitor to know the number of questions in the 2019 Exam, that information could, in any event, provide a competitor with no significant advantage in relation to subsequent years. As Ms Walton volunteers (third witness statement, para 6), the Exam content is new every year, and there would be no reason for any competitor to think that the precise numbers of questions in the four elements of the 2019 Exam, or the overall total, would be replicated in any later year.(4) Nor is there any evidence to suggest that disclosure of the question numbers would have caused any detriment in the form of an increase in the amount of work involved in setting the Exam for the years after 2019. Self-evidently, that work was always going to, and did, consist mainly of selecting new questions each year in place of those used in the previous year. If GLA saw fit to make a change at the same time to the number of questions within each element, that change would inevitably be of a modest order since, as already stated, its room for manoeuvre would be limited. Not surprisingly, the Tribunal was not favoured with any evidence about the likely annual cost of reviewing and (if so advised) ‘tweaking’ the numbers of questions in the Exam. For that matter, the Tribunal also received no evidence about the annual cost of changing the content of the Exam – a topic on which the witnesses called on behalf of TBGS were less than forthcoming. That cost would surely be negligible and, in the view of the minority member, no question of detriment or competitor advantage could possibly arise.(5) The minority member is also not persuaded by the contention that disclosure of the reliability calculations in respect of the Maths & Non-Verbal Reasoning elements of the Exam would disadvantage GLA in a competitive market, given that it has voluntarily published the corresponding calculations in respect of the English and verbal reasoning elements.[88]For the reasons stated, the minority member takes the view that the third limb of the Coco test is not satisfied and that accordingly, the exemption under s41 is not engaged.[89]Even if he had concluded that the requisite detriment was shown, the minority member would have found that an action for breach of confidence by GLA would have been defeated by a public interest defence and that, accordingly, s41 was in any event not correctly applied. While fully acknowledging the strong public interest in protecting information communicated in confidence, he is satisfied that there is here an even more compelling public interest in the disputed material being disclosed. His main reasons are the following.(1) The errors were of GLA’s making and highly embarrassing for GLA. The Solution was devised by GLA in a matter of days and immediately approved by TBGS. It is quite understandable that Mr Coombs and many disinterested members of the public should begin from a position of scepticism as to GLA’s motives for suppressing the disputed information and concern that a wish to avoid reputational damage may have influenced its choice of the Solution. There is an obvious public interest in that scepticism and that concern being tested through disclosure of the information.(2) There is also an obvious public interest in fostering a properly informed debate as to whether (regardless of GLA’s (or TBGS’s) motives) the Solution was appropriate or fair.(3) There is a further, and equally strong, public interest in facilitating a fully informed debate into the question whether TBGS issued misleading information by announcing, on 1 October 2019 and in the FAQ document issued shortly afterwards, that the outcome of the Exam following application of the Solution was ‘fair’ to ‘all children’ involved. As noted above, that claim was not echoed by the ‘independent statistician’ in his letter of 11 November 2019. Mr Coombs makes a powerful case for the proposition that that assertion could not be right as a matter of statistical analysis. The public interest strongly favours a fully informed inquiry into (a) whether the assertion was right, or at least defensible, and (b) the extent, if any, to which it was misleading, and (c) if and to the extent that it was misleading, why and in what circumstances it was put into the public domain.(4) The public interest referred to in (3) above is all the more obvious in circumstances where it appears to be undisputed that TBGS proclaimed that the Solution had been approved by an ‘independent statistician’ without making public the fact that the person concerned had worked for GLA (then trading under a different name) for some 24 years between 1986 and 2010.(5) The public interests referred to in (1)-(3) above could not be met without disclosure of the disputed information. The fairness of the Solution and the propriety of TBGS’s communications about it cannot be assessed in a transparent and fair debate without the disputed information being disclosed.(6) The points made by the minority member on the Tutor Advantage and Competitor Advantage arguments are repeated. The detriment-based grounds on which TBGS resists disclosure have no substance or, at best, very little. They offer no material counterweight to the compelling public interests favouring disclosure.(7) As the members in the majority rightly acknowledge, the public interests identified in (1)-(3) above must be seen in the context of the wider public interest in transparency generally over how school places in the state sector are allocated.(8) There is also a wider public interest in proper understanding and scrutiny of reliability calculations in relation to tests for admission to selective schools and the clarity (or lack of clarity) with which such calculations are published and explained. The disclosure of the reliability statistics in relation to the Mathematics & Non-Verbal Reasoning elements of the Exam would foster such understanding and scrutiny.[90]In the minority member’s view, these considerations combined dictate the conclusion that, if and in so far as there was a public interest in protecting the confidentiality of the disputed information, it was (despite the mild in-built presumption in favour of protecting confidence) comfortably outweighed by the public interest in disclosure.

Commercial interests

[91]For the reasons stated above in relation to the third limb of the Coco test, the minority member is satisfied that TBGS falls well short of establishing what is required to engage s43(2). TBGS signally fails to demonstrate that disclosure of the disputed information would or ‘could well’ give rise to a risk of harm (to it or to GLA) which was ‘serious, actual or of substance’.[92]Even if he had found s43(2) engaged, the minority member would have held that the public interest balance under s2(1)(b) came down clearly in favour of disclosure, for the reasons stated in relation to the public interest analysis under s41. In his view, the result here is all the more clear given the (mild) legal presumption in favour, rather than against, disclosure.

Disposal

[93]The majority view prevails. The appeal is dismissed.[94]For the avoidance of any doubt, the Tribunal would not wish its decision to be read as signalling any view (on the part of any of its three members) about the merits of wider FOIA requests to do with 11+ Exams or analogous subjects. The appeal was narrow in its scope and arose out of a unique set of circumstances. (Signed) Anthony Snelson Judge of the First-tier Tribunal Dated: 8 December 2023 Re-dated 16 March 2024

Cited in 2 later judgments