The decision

Upper Tribunal
(Immigration and Asylum Chamber) Appeal Number: HU/19203/2016


Heard at Field House
Decision & Reasons Promulgated
On 20 April 2018
On 30 April 2018






For the Appellant: Mr Laurence Tarlow, Senior Home Office Presenting Officer
For the Respondent/Claimant: Mr A Malik, Counsel instructed by Aldgate Immigration

1. The Secretary of State appeals to the Upper Tribunal from the decision of the First-tier Tribunal (Judge Row sitting at Birmingham on 11 August 2017) in which he allowed, on Article 8 grounds, the claimant's appeal against the decision to refuse to grant him leave to remain as the spouse of a British national. The only issue in the appeal was whether the Secretary of State had made out her case under S-LTR.1.6 of Appendix FM - namely, that the presence of the applicant in the UK was not conducive to the public good because their conduct, character, associations or other reasons made it undesirable to allow him to remain in the UK. The Secretary of State alleged that the claimant had used a proxy to take two English language speaking tests on 19 October 2011 and 15 November 2011 at Burnley Training College, and had thereby obtained two TOEIC test certificates by deception.
The Reasons for the Grant of Permission to Appeal
2. On 7 February 2018, First-tier Tribunal Judge Michael Keane granted the Secretary of State permission to appeal for the following reasons: "For the reasons mentioned in the grounds the Judge made arguable errors of law but for which the outcome of the appeal might have been different. The Judge found at paragraph 7 of his decision that the respondent discharged the evidential burden of proof in respect of the appellant's alleged deception. The Judge arguably should have gone on to decide whether the appellant had raised an innocent explanation whereupon the burden would revert to the respondent to discharge the legal burden of proof. The Judge arguably did not observe such an approach established by authorities which were binding on the Judge."
Relevant Background Facts
3. The claimant is a national of Pakistan, whose date of birth is 2 December 1977. On 21 July 2016 he made a human rights application for leave to remain in the United Kingdom as the spouse of a settled person.
4. On 1 August 2016 the Secretary of State gave her reasons for refusing the application. On 19 October 2011 and 15 November 2011, he purported to take a TOEIC speaking test with ETS. ETS had a record of his speaking test. Using voice verification software, ETS was able to detect that a single person was undertaking multiple tests. ETS had undertaken a check of his test and had confirmed to the Secretary of State that there was sufficient evidence to conclude that his certificates were fraudulently obtained by the use of a proxy test-taker. His scores from the test taken on 19 October 2011 and 15 November 2011 at Burnley Training College had now been cancelled by ETS. On the basis of information provided to her by ETS, the Secretary of State was satisfied that his certificates were fraudulently obtained.
The Hearing Before, and the Decision of, the First-Tier Tribunal
5. Both parties were legally represented before Judge Row. The claimant gave oral evidence. He adopted as his evidence in chief his witness statement signed on 11 August 2017. In the statement he said that he failed the TOEIC test twice and he had only passed it on his third attempt. On his third attempt, he had obtained 170 marks, and he had thereby been able to obtain a student visa. He took the exam for the third time at a different location. The Home Office had not objected to his "third test results" or to other results which he had obtained when taking English language tests.
6. In his subsequent decision, the Judge set out his consideration of the evidence and his findings at paragraphs [7] to [12]. At paragraph [7], he held that the Secretary of State's case was established by producing the statements from Professor Peter French and Matthew Lister, and the standard statements used in the ETS cases of Peter Millington and Rebecca Collings: "It is established that this evidence provides a prima facie case that the tests were taken by proxy. The [claimant] can rebut this."
7. At paragraphs [8]-[11], the Judge gave his reasons for holding that the claimant had rebutted the Secretary of State's prima facie case; and for being satisfied that the claimant had genuinely sat for the speaking tests on 19 October 2011 and 15 November 2011, both of which he had failed.
8. The error of law challenge is supported by extensive citation of passages from SM & Qadir (ETS - evidence - burden of proof) UTIAC 21 April 2016, Shehzad & Another [2016] EWCA Civ 615 and R (On the application of Nawaz) -v- SSHD (ETS: review standard/evidential basis) [2017] UKUT 00288 (IAC).
9. This extensive citation of authority is directed at establishing the uncontentious proposition that the standard generic evidence relied on by the Secretary of State in ETS cases is sufficient to raise a prima facie case such that the evidential burden shifts to the claimant to provide an innocent explanation.
10. I consider that the Judge adequately acknowledged that this was his necessary starting point. The Judge expressly held at paragraph [7] of his decision that the Secretary of State had discharged the evidential burden of proof in respect of the claimant's alleged deception. Accordingly, there is no merit in the argument that the Judge failed to observe the approach established by the authorities. On the contrary, he did precisely that which he was required to do, which was to go on to decide whether the claimant had raised an innocent explanation which rebutted the prima facie case of deception.
11. The second issue raised in the grounds is whether the Judge gave adequate reasons for finding that the claimant had rebutted the prima facie case of deception. Thus, in paragraph 10 of the grounds, the following is pleaded: "In this case is it respectfully submitted that the FTT has failed to give adequate reasoning why the respondent has not met the legal burden, nor that there is any innocent explanation. The fact that the test was failed does not amount to an innocent explanation, neither does the fact that the test was not relied on."
12. Although this is not a point which is specifically taken in the grounds of appeal, the Judge's line of reasoning is susceptible to the criticism that he did not engage with the supplementary evidence provided by the Secretary of State to reinforce her prima facie case. The significance of the expert report of Professor Peter French is that it counters the expert evidence of Dr Harrison which was adduced in the case of SM & Qadir, and which was the foundation for the Tribunal's finding at paragraph [68] as follows: "As our analysis and conclusions in the immediately preceding section made clear, we have substantial reservations about the strength and quality of the Secretary of State's evidence. Its shortcomings are manifest. On the other hand, while bearing in mind that the context is one of alleged deception, we must be mindful of the comparatively modest threshold which an evidential burden entails. This calls for an evaluative assessment on the part of the Tribunal. By an admittedly narrow margin, we are satisfied that the Secretary of State has discharged this burden."
13. In reaching the conclusion that the standard generic evidence only discharged the evidential burden by a narrow margin, the Tribunal attached great weight to the expert evidence of Dr Harrison, who was of the opinion that ETS' method of analysing the TOEIC test data was deeply flawed and was capable of generating a very high number of false positives. Professor French's report is a direct riposte to the expert evidence of Dr Harrison. The lawyers for the Secretary of State made a late application for Professor French's report to be admitted into evidence at the hearing of the SM & Qadir, but the application was refused on the grounds of lateness.
14. In his report, Professor French was asked to give his opinion of whether, on the balance of probabilities, ETS's methodology was likely to result in any false positives (i.e. speaking comparison test results wrongly indicating that different speakers were the same person). If he considered false positives were likely, and he was asked to estimate how many.

15. His conclusion at Section 3.2 is that the ASR used by ETS is extremely likely to produce some false positives among the 58,464 matches identified by the software in respect of TOEIC test recordings. This is despite the threshold for producing matches having been set conservatively. The number of false positives produced by the ASR cannot however be estimated with any great degree of precision.

16. The error rate of 2% quoted for the pilot test was not broken down into false positives versus false negatives. But even if one assumes the worst case scenario, that all the errors were false positives, the the safety net system of having trained listeners assess all the matches shown up by the ASR would, in his opinion, have made a very large reduction to the overall number of false positives.

17. At paragraph 3.3.7, Professor French refers to the fact that of the 58,464 matches produced by the ASR, only 33,735 were confirmed by the listener pairs. In other words, only 57.7% of matches were accepted.

18. Professor French opines that the very high rejection rate of 42.3% is in part attributable to the stringent conditions laid down for match confirmation i.e. for acceptance of a match both listeners working independently had to confirm it, and the test for individual acceptance was that, "any doubt about the validity of the match will result in it being rejected", citing the witness statement of Peter Millington at paragraph 45.

19. His conclusions at paragraph 4 are that the conditions used for trained listener pair confirmation, in conjunction with the conservative threshold set for ASR match identification, would, in his view, have resulted in substantially more false rejections than false positives.

20. Even though there is material missing from the body of information called for by Dr Harrison, he is not convinced that the provision of such information could be used to establish a closely specified percentage of false positives.

21. If the 2% error rate established for the pilot recordings were to apply to the TOEIC recordings, he estimates that the number of false positives is likely to be substantially less than 1% after the process of assessment by trained listeners has been applied.
22. Accordingly, in the absence of expert evidence from Dr Harrison in rebuttal, a much stronger prima facie case of deception was thereby raised against this claimant than was raised against the claimants in SM & Qadir, and the evidential burden on this claimant to produce an innocent explanation for the invalid test results became correspondingly greater.
23. However, I am nonetheless wholly unpersuaded that the Judge failed to give adequate reasons for finding that the claimant had given an innocent explanation which rebutted the objectively strong prima facie case of deception.
24. The Judge gave four reasons for finding that the claimant had given a satisfactory innocent explanation which rebutted the prima facie case.
25. The first reason was that the claimant had failed the two speaking tests in which he was alleged to have used a proxy test-taker. The Judge observed in paragraph [10] as follows: "If the appellant had used a proxy it would have been odd to use someone who is not capable of passing the test. It would tend to defeat the object of the exercise. It would have been even odder to use someone in the second test who was also not capable of passing it."
26. The second reason was that the claimant had not relied upon either of the two failed tests in any application.
27. The third reason was that, having failed the first two tests, he then went on a course to improve his English and he subsequently passed the test; and the third test result was used to support his application for further leave to remain; and the validity of that test was not in question.
28. The fourth reason was contained in paragraph [11]: "He gave convincing detail about the test and his efforts to pass it. I saw him, heard him, and I believed what he had to say. I was satisfied that the appellant did take the two tests."
29. It is clear from the surrounding evidence that the claimant wished to obtain further leave to remain as a student, and for this purpose he required a speaking score of between 160 and 180, corresponding to level B2 CEFR. On the test he purportedly sat on 19 October 2011, he only obtained a speaking score of 150, which meant that he had not achieved level B2. In the speaking test purportedly undertaken on 15 November 2011, he only obtained a speaking score of 140, which meant that he was even further away from attaining the required level of competence. He was still in the level 6 scale score range, corresponding to level B1 CEFR. However, in the speaking test which he purportedly sat for at third time, on 13 December 2011, he obtained a speaking score of 170, which meant that he reached level B2 CEFR. It was not part of the Secretary of State's case that the claimant had used a proxy test-taker to obtain the third ETS speaking test result.
30. The fact that the claimant did not rely on the two "failed" tests in support of his application for further leave to remain as a student would not, in itself, negate the proposition that he had used a proxy test-taker on the first two occasions. Similarly, the fact that the first two tests did not yield a high enough result for the claimant's purposes does not necessarily entail that he did not use a proxy test-taker. However, these two reasons must not be viewed in isolation from the remaining two reasons put forward by the Judge, which are not alleged in the grounds of appeal to be defective.
31. It was unarguably open to the Judge to find that the striking sequence of events - a valid and successful third ETS speaking test in December 2011 which was preceded by two failed ETS speaking tests in October and November 2011 - pointed very strongly to the claimant having genuinely sat for the first two tests, but as not having at that stage sufficient training and expertise to obtain a speaking score of at least 160; and then genuinely achieving a speaking score of 170 on his third attempt, after further training. Equally, it was unarguably open to the Judge to find that it was inherently unlikely that the claimant would pay for a proxy test-taker to achieve a false speaking test result that was worthless to him - and that he would do this not just once, but twice.

Notice of Decision

The decision of the First-tier Tribunal did not contain an error of law, and accordingly the decision stands.

The Secretary of State's appeal to the Upper Tribunal is dismissed.

The First-tier Tribunal did not make an anonymity direction, and I do not consider that such a direction is required for these proceedings in the Upper Tribunal.

Signed Date 20 April 2018

Deputy Upper Tribunal Judge Monson