Oxford and Cambridge RSA Examinations v The Information Commissioner

This is an appeal against the Commissioner’s decision notice IC-233103-D4R0 of 21 July 2023 which held that Oxford and Cambridge RSA Examinations (OCR) was not entitled to rely on section 43(2) (commercial interests) of the Freedom of Information Act 2000 (FOIA) to withhold the requested information. The Commissioner required OCR to disclose the information withheld under section 43(2).

There are redactions to paragraphs 52 and 53 because they referred to evidence given in the closed session. That part of the reasoning is, at present, contained in a closed annex. The tribunal was unsure if that reasoning needed to remain closed and has ordered OCR to provide submissions on this issue.

OCR is not directly in receipt of public funding and is a commercial organisation. OCR is subject to FOIA by virtue of being wholly owned by the University of Cambridge. No other non-Cambridge examination board is subject to the Act.

Students’ Non-Examined Assessment (NEA/Coursework) is marked by the school/centre. A sample is then sent to OCR where a moderator reviews the schools marks. This is the part of the process referred to as ‘moderation’. If the moderator feels the school have been too harsh or too lenient they adjust the marks and the students are awarded the adjusted marks. This is the ‘adjustment’ part of the process, and can result in marks being increased or decreased according to the moderator assessment.

The request in issue relates to the moderation of A-Level exam results by OCR in Art and Design in 2022 and 2019. 2022 was a year in which examinations remained affected by Covid. Students did not study the full course and were not assessed on all the content. Component 2, the externally set task was removed for 2022 and students only had to complete component 1.

The request asks for the number of centres and the number of candidates that were awarded adjusted marks after moderation broken down by endorsement. A ‘centre’ is usually a school or a college. ‘Endorsement’, in this context means one of the specialisms within the Art and Design framework, for example Fine Art, Textile Design or Critical and Contextual Studies.

This appeal concerns a request made by Claire Fields on 22 January 2023:

“I am seeking the following information in relation to A level Art and Design results for 2022 and 2019 and hope that you will be able to assist me with this request.

1.

The number of centres that were awarded adjusted marks after moderation, broken down by endorsement e.g. Fine Art, Photography etc. in 2022 and 2019.

2.

The number of candidates that were awarded adjusted marks after moderation, broken down by endorsement e.g. Fine Art, Photography etc. in 2022 and 2019.

3.

The number of centres that were awarded adjusted marks after moderation for the first time in 2022 and 2019. How many centres had never had their centre/teacher marks adjusted before and were awarded adjusted marks for the first time in 2022 and 2019.”

Only parts 1 and 2 of the request are in issue in this appeal.

OCR replied on 14 February 2023. OCR confirmed that they held information requested in parts 1 and 2 and withheld that information relying on section 43(2) FOIA. OCR refused the request in part 3 under section 12 (cost limit) on the basis that they did not hold the information in that format and it would take time to extract and compile the requested information.

Ms Fields applied for an internal review in relation to OCR’s reliance on section 43(2).

Ms Fields referred the matter to the Commissioner on 29 April 2023. The Commissioner accepted the complaint without requiring an internal review. During the Commissioner’s investigation OCR reviewed and upheld its response.

In a decision notice dated 21 July 2023 the Commissioner decided that section 43(2) was not engaged. The Commissioner noted that the central argument from OCR appeared to be the fact that it was subject to FOIA and its competitors are not. The Commissioner accepted that this argument does have relevance where it is clear that disclosing certain information will cause detriment to the commercial interests of an organisation when that same information does not have to be disclosed by its competitors. In this instance, the Commissioner did not accept that OCR had convincingly explained the causal relationship between the disclosure of this specific information and the resulting prejudice to its commercial interests. He concluded that the suggestion that competitors could target customers from the disclosure of the requested information was tenuous. Consequently, the Commissioner did not accept that the exemption was engaged.

The Commissioner did not go on to consider the public interest balance.

The grounds of appeal are:

The Commissioner accepts that the supplying of public examination courses in competition with other examinations boards is a commercial interest.

The Commissioner submits that the appellant’s explanation fails to make a causal link between release of the information and prejudice to the appellant’s commercial interests. If a competitor obtains the requested information, it will not be able to approach specific centres about its comparative moderation statistics but would only be able to inform every centre without distinction. A competitor can already achieve the same result by approaching any centre with its own moderation statistics.

The Commissioner submits that even if the prejudice was made out, it is unlikely that a competitor could successfully use the information to undermine the appellant’s business. The information - anonymised scores moderated in 2019 and 2022 – is inadequately detailed to draw any conclusions as to the appellant’s competitive position in the market.

The Commissioner submits that disclosure of the information would not be in breach of section 2 of the Competition Act 1998 because release would not be the result of any agreement, decision, or concerted practice between the Appellant and other competing undertakings, whether overt or implied.

The Commissioner submits that the public interest favours disclosure.

In relation to ground 2 the Commissioner submits that the burden lies with the appellant to prove why there would be prejudice to commercial interests.

Section 43(2) provides:

“Information is exempt information if its disclosure under this Act, would, or would be likely to prejudice the commercial interests of any person (including the public authority holding it)” 

‘Commercial interests’ should be interpreted broadly. The ICO Guidance states that a commercial interest relates to a person’s ability to participate competitively in a commercial activity.  

The exemption is prejudice based. ‘Would or would be likely to’ means that the prejudice is more probable than not or that there is a real and significant risk of prejudice. The public authority must show that there is some causative link between the potential disclosure and the prejudice and that the prejudice is real, actual or of substance. The harm must relate to the interests protected by the exemption.  

Section 43 is a qualified exemption, so that the public interest test has to be applied. 

In considering the factors that militate against disclosure the primary focus should be on the particular interest which the exemption is designed to protect.

The APPGER case gives guidance on how the balancing exercise required by section 2(2)(b) of FOIA should be carried out:

“… 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 of, 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 or promote.”

The issues we have to determine under section 43(2) are:

The tribunal’s remit is governed by s.58 FOIA. This requires the tribunal to consider whether the decision made by the Commissioner is in accordance with the law or, where the Commissioner’s decision involved exercising discretion, whether he should have exercised it differently. The Tribunal may receive evidence that was not before the Commissioner, and may make different findings of fact from the Commissioner.

We read an open and a closed bundle of documents. The material in the closed bundle either contained or referred to the content of the withheld information and therefore it was necessary for this information to remain closed to avoid defeating the purpose of the proceedings and undermining the rule 14 order in place.

We heard open and closed evidence from Peter Canning, Director of Product at OCR. The following is a gist of the closed evidence:

“Dr Canning addressed five issues under cross-examination in the closed session. First, he explained the aspects of the withheld information that were beyond the scope of request, along with the aspects of the information that failed to provide the information requested. He accepted that aspects of the withheld information would need to be amended if the information was released for the purposes of accuracy. Second, Dr Canning discussed the benchmarking process. In particular, Dr Canning discussed a further instance whereby confidential information relating to moderations was revealed. He also explained the value of the relevant quantitative information in the closed exhibits for forming a benchmark, and whether it could be distinguished from the other types of information that circulate in the industry. Third, Dr Canning provided further detail on how a competitor may use the information to target a school using OCR for its Art and Design course in order to convince the school to use the competitor’s services instead. Fourth, Dr Canning addressed the particular features of the Art and Design A-Level, and relevant reputational factors that could follow on from the release of the information as to moderation. Fifth, Dr Canning explained how the information might reveal commercial information about the popularity of the relevant modules within Art and Design.

This was followed by questions from the Tribunal. Four broad issues were addressed. First, Dr Canning explained the extent to which examination boards can change their moderation numbers in response to feedback or obtaining information about their position in the market. Second, Dr Canning discussed the beneficial aspects of aggressive moderation, on a hypothetical basis. Third, Dr Canning described the relevant thresholds of moderation percentages that might encourage OCR to review its moderation practices. Fourth, Dr Canning provided a view on the degree to which moderation can occur upwards, along with downwards.”

We heard open submissions from both parties.

The following is a gist of Mr. Whelan’s closed submissions:

“Counsel for the Information Commissioner made three submissions in the closed session, which relied on the evidence obtained in the closed session. First, the withheld information could not form a proper basis for competitors to benchmark their results, due to the inaccuracies and omissions in the document that contained the information requested by the Complainant. Second, OCR’s closed evidence as to any possible reputational consequences should be disregarded as contrary to the purpose of FOIA. Third, the information relating to the relative popularity of modules should be disregarded as irrelevant.”

The applicable interest under section 43(2) is avoiding prejudice to a body’s commercial interests. We accept OCR’s ability to compete with other exam boards for the custom of centres (schools and colleges) is a commercial interest capable of protection under section 43(2).

The claimed prejudice in this appeal is that competitors could well use the requested information to advance their commercial interest, to the relative disadvantage of OCR by:

We accept that this claimed prejudice, if it occurred, would be real, actual and of substance.

In the tribunal’s view the questions of whether there is a causal link between disclosure and the claimed prejudice and whether the prejudice would be likely to happen (in the sense of there being a real and significant risk of prejudice) need to be considered in the light of (a) the nature of the withheld information and (b) the nature of the market in which OCR are operating.

The requested information consists of the number of centres and candidates that were awarded adjusted marks after moderation, broken down by endorsement e.g. Fine Art, Photography etc. in 2022 and 2019.

We accept that using information published by Ofqual in its Annual Qualifications Market Reports would allow a competitor to attempt to the work out the level of adjustment made by OCR in A-level Art and Design as whole. That is because Ofqual publishes the total number of certificates awarded by OCR in A-level Art and Design. The figure would not be accurate because the number of certificates entered is not the same as the number of certificates awarded due to some candidates withdrawing from the course or otherwise failing to achieve a certificate.

The tribunal notes that the publicly available statistics in data table 1 supporting that Ofqual report, available on the page linked from exhibit PC9 also give the number of certificates awarded by OCR by individual endorsement. (Footnote: ¹) For example table 1 shows that in 2018/2019 in OCR awarded 2,405 certificates H601 (Fine Art). A similar calculation can also therefore be made in relation to individual endorsements.

On that basis we accept that, using information available in the public domain, it would be possible to calculate what a competitor would assume to be a reasonably accurate, but not completely accurate figure representing the level of adjustment in relation to each endorsement an Art and Design overall by OCR in 2019 and 2022.

At the date of the response to the request in February 2023 the figures from 2019 were approximately 2.5 years old. The figure for 2022 is recent, but related to a year that can fairly be described as unrepresentative. It was a year impacted by Covid and students were only assessed on one of the usual two components. Exams were graded more generously compared to a normal exam year. Dr Canning gave evidence that now and then adjustment on moderation statistics might be particularly high in a subject in a particular year because of external market factors which might cause, for example, a lot of new teachers to come to a particular subject. The 2019 figure might be representative of OCR’s normal figures in 2023 but it might not be. It is impossible to tell with a one-off figure from over 2 years ago.

In our view this combination of factors makes the information almost worthless in terms of an indication of what OCR’s ‘normal’ adjustment statistics would be in 2023. It provides a one-off figure from over 2 years ago and a figure relating to an entirely unrepresentative year. Without revealing what the withheld figures are, it is possible to say that the figures in each year are very different from each other.

Mr. Canning stated that an exam board could carry out market research and use the approximate percentages derived from the withheld information to see if they lined up with the percentages in the survey to see if the people in the survey are representative of teachers who teach that subject. However, as the percentages in each year are very different, one is two years out of date and one is atypical, the withheld figures themselves are not representative and therefore would not assist in calibrating the survey figures.

In terms of the nature of the market, on the basis of Mr. Canning’s evidence we find that this a market where there is what he described as a ‘slow churn’. That means that schools and colleges are extremely unlikely to move from one exam board to another. On the basis of Mr. Canning’s evidence we find there is about a 3% churn rate, although this may be drifting upwards. Teachers and centres are reluctant to change exam boards because the change is significant in terms of teaching preparation and marking.

The three main drivers in relation to a move were said by Mr. Canning to be (i) a head of department moving schools (the head of department traditionally making the decision as to the exam board) (ii) unhappiness with moderation or exam papers and (iii) a multi-academy trust trying to consolidate to use fewer exam boards.

Moderation is said to form part of (ii), but the tribunal notes that it has been taken together with ‘unhappiness with exam papers’. We do not know the relative weight of those two factors, nor do we know the relative weight of each of the three drivers.

Art and Design accounts for about 3.5% of OCR’s enrolments, although OCR has what Mr. Canning calls ‘a long tail’ in that it has a large number of subjects with a small percentage that together contribute significantly to the running costs of the organisation.

Having outlined the nature of the withheld information and the market in which OCR operate, we turn to considering the question of whether OCR has established that there is a causative link between disclosure and a real and significant risk of prejudice that is real, actual and of substance.

OCR argue first that competitors could use the OCR’s rates of adjustment to benchmark their own moderation adjustment rates. Mr. Canning states that ‘OCR would effectively become the benchmark against which all of its competitors could measure the level of adjustment of their A Level Art and Design qualification on moderation.’

There are a number of limitations on the usefulness of the withheld information as a benchmark:

For those reasons, in our view the information is too limited for there to be a real and significant risk that it would be used for benchmarking by a competitor.

We accept that Dr Canning is a position of expertise compared to the tribunal, but we were not satisfied on the basis of his evidence that he had properly taken into account the limited nature of the information when giving his evidence on the potential uses of that information.

If the competitor’s rate was higher than OCR’s, assuming that somehow the competitor could deduce that it was its own rate that was too high rather than OCR’s rate that was too low, the steps that Mr. Canning suggests a competitor might take are as follows:

In essence, Mr. Canning’s evidence was that if an exam board found out that its moderation adjustment rates were too high, it would take steps to [Redacted – in the closed annex] thus ‘improving its performance’ in relation to moderation).

It must be remembered that although the other organisations are private companies, they perform together a vital function within the education system, which marks the culmination of many students’ journeys through school. The quality of the teaching and marking of A-levels is of huge and life changing importance to the lives of those students who go through the system. It is also of extreme importance to universities and to prospective employers that a student is given a mark that accurately reflects their performance.

Mr. Canning recognised in his written evidence that examination boards taking steps to improve the performance of their assessments ‘could be seen’ to have a positive impact on students. In oral evidence he confirmed that those improvements were ‘a good thing’. OCR’s fundamental submission was that it was not fair, because nobody else had to disclose similar information because they were not subject to FOIA and therefore ‘the potential benefits from such disclosure would solely be available to OCR’s competitors’. The tribunal disagrees. The potential benefits of an improvement in performance by an exam board as described above are not solely available to OCR’s competitors. The beneficiaries are the teachers, the students, Universities and prospective employers. The country as a whole benefits from improvements to its public examination system.

Further, we are not satisfied that even if the steps above were taken by a competitor would be likely to adversely impact OCR’s financial interests. If OCR’s figures are being used as a ‘benchmark’ then it is still able to compete at the same level as its competitors. The possibility that somehow a competitor which discovered that it had inappropriately high figures of adjustment on moderation would improve its training materials and examinations to such an extent that it managed to bring its rate of adjustment below what it considered to be the benchmark (and therefore presumably to an inappropriately low level) and then somehow tempt away OCR’s customers is remote.

We turn now to the second claimed prejudice: better targeting of a competitor’s own market research and marketing activities, including by engaging in marketing activity targeted at the level of individual schools and colleges.

We accept, on the basis of Mr. Canning’s evidence that market research is an important part of the process by which awarding organisations compare themselves against each other and that OCR and its competitors all invest substantial resources in this process in order to gather information as to how competitors are performing and how they may improve their share in the market.

We accept that information which enabled a competitor to short-circuit the early stages of market research and proceed straight to carrying out targeted and focused research on a specific course offered by a particular competitor would be highly advantageous and a considerable cost-saving. We do not accept that this information enables that to be done. First, we repeat the limitations set out above:

In relation to focussing market research, the figures are also limited by only relating to one subject. OCR’s adjustment figures might be worse, or might be better, in any or indeed all the other subjects offered by OCR. Or they might be the same. The release of figures relating to one subject alone gives no useful guide as to which subjects it is best to focus on in market research. The only reason that there is a focus on Art and Design is because that is what Ms Fields asked for. That is not a rational basis for a competitor to focus their market research on Art and Design.

The second aspect of marketing relied on by the Appellant is the engaging in marketing activity targeted at the level of individual schools and colleges. We accept that most exam boards are broadly aware of which exam boards are used by which centres, particularly in the larger subjects. Further, many centres indicate the exam boards used for each subject on their website. It would probably not be too difficult then for a competitor to identify a large number of schools/colleges who use OCR for Art and Design.

However, Mr. Canning accepted that these figures would not assist competitors in identifying which schools and colleges might have been subject to adjustment on moderation. The competitor will know, approximately, what percentage of centres have been subject to adjustment on moderation but there is no way of knowing which ones from the figures. Nor do the figures enable a competitor to know which of those schools that have had adjusted marks after moderation are unhappy about this. The figures do not therefore allow marketing to be targeted at particular schools or colleges.

Mr. Canning gave evidence that if a competitor was visiting a particular school anyway, then these figures might give them a reason to talk to the head of department in Art and Design, during which conversation they could ask about moderation.

For the reasons set out above in relation to market research, the tribunal does not agree that these figures provide a rational basis for focussing on Art and Design rather than other subjects. The figures cannot be used to deduce that OCR’s adjustment figures are unusually high for Art and Design. They do not therefore provide a rational basis on which a competitor could decide allocate part of the limited time available in a school to talking to the Head of Department in Art and Design. There is no reason, on the basis of these figures, that a conversation with a Head of Art and Design is more likely to be fruitful than a conversation with anyone else.

Even if the figures could be used to deduce that OCR’s adjustment percentage was unusually high for Art and Design, we find that there is not a real and significant risk of such an approach being to the competitive disadvantage of a OCR. For OCR to be disadvantaged, the following would need to occur. First that the school happened to be in the small subset of those whose marks had been adjusted. Second, of those whose marks had been adjusted, that the school happened to be in the subset of those who were unhappy with this. Third, of those that were unhappy, the school happened to be one of the small percentage who would move to a new exam board as a result, in a market where there is only about 3% churn every year.

Finally, given the limitations on the figures detailed above, we do not think that they could usefully and persuasively be presented by competitors to schools/customers as a representative figure of OCR’s levels of adjustment either in Art and Design or as a whole.

Moving to damage to reputation, it is submitted that the figures could be used to undermine the reputation of OCR’s assessments, because it might be inferred from the levels of adjustments on moderation that the syllabuses are not performing as they should or are not otherwise easy to implement and understand by centres. As the other competitors are not subject to FOIA, adverse inferences can only be drawn against OCR.

First, we are considering the release of information relating only to Art and Design, a course which accounts for a small percentage of OCR’s portfolio. Second, these figures are being released into a vacuum. There are no other figures in the public domain which could enable the public, or teachers or centres to make a comparison to determine whether these figures are high or low or about right. The figures themselves have the limitations set out above, including the fact that the 2019 figures are out of date and the 2022 figures relate to a covid year. For those reasons we are not persuaded there is a real and significant risk of prejudice to commercial interests as a result of reputational damage being cause by release of the information.

OCR submitted that there was a risk that disclosing the information could provide OCR’s competitors with an insight into which of the Art and Design modules are most popular. The publicly available data tables supporting the Ofqual report (available on the page linked from exhibit PC9) break down the number of certificates awarded by OCR by individual endorsement or module. (Footnote: ²) For example table 1 shows that in 2018/2019 in OCR awarded 2,405 certificates H601 (Fine Art). This gives a competitor a much better insight into the relative popularity of each OCR endorsement, and the withheld information adds nothing of value to the information already in the public domain in this respect. (Footnote: ³) For those reasons we do not accept that there is a causative link between disclosure and the prejudice claimed in relation to knowledge of the relative popularity of modules.

For all those reasons we do not accept that disclosure of the withheld information would be likely to impact on the revenue stream for OCR, and in turn have a negative impact on OCR’s charitable objectives’ as set out in paragraph 36 of Dr Canning’s witness statement.

In summary we are not persuaded that there is a causative link between disclosure of the information and the claimed prejudice and we do not accept that there is a real and significant risk of that prejudice occurring. For those reasons we find that the exemption is not engaged.

We have gone on to consider what we would have decided if we had had to consider the balance of public interest. If we had decided that the exemption was engaged, for all the reasons set out above we would have concluded that the risk of prejudice was very low – no more than a real and significant risk - and that the level of prejudice caused was likely to be very small.

We accept that there is a public interest in avoiding prejudice to commercial interests and a public interest in avoiding distortion to a competitive market. In this appeal because the risk of prejudice is very low and the level of prejudice likely to be small this carries limited weight in the public interest balance.

In our view this would be outweighed by the public interest in transparency. There is not only a general public interest in transparency in relation to any information held by public authorities but also a specific public interest in transparency in relation to the way that A levels are assessed. We also think there is a specific public interest in the difference in adjustment figures between 2019 and 2022 given the impact of Covid on the assessment regime. Given the limited nature of the information (for all the reasons highlighted above) we accept that disclosure only contributes to these specific public interests to a fairly limited extent, so these add only a little to the general public interest in transparency.

In addition, we take the view that benchmarking is a remote possibility for the reasons set out above, but if we are wrong, and it would be likely to happen, resulting in an improvement in the way in which exams were assessed by other boards, in our view any competitive disadvantage to OCR by dint of the fact that it would be deprived of the opportunity to benchmark against a competitor would be heavily outweighed by the public interest in such an improvement.

Further, if we had concluded disclosure would be likely to lead to a risk of commercially harmful reputational damage because the figures revealed that OCR’s adjustment on moderation statistics were inappropriately high, we would have held that the public interest in this being known, given the public importance in the quality of assessment of A-levels, would outweigh the public interest in avoiding prejudice to OCR’s commercial interests.