United Kingdom Immigration and Asylum (AIT/IAC) Unreported Judgments

Upper Tribunal

(Immigration and Asylum Chamber) Appeal Number: IA/09123/2015

 

THE IMMIGRATION ACTS

Heard at Field House	Decision & Reasons Promulgated
On 19 December 2016	On 03 May 2017

 

Before

 

DEPUTY UPPER TRIBUNAL JUDGE MANDALIA

 

Between

 

 

THE SECRETARY OF STATE FOR THE HOME DEPARTMENT

Appellant

and

 

MR munshi pabel ahmed

(anonymity direction not made)

Respondent

 

Representation :

 

For the Appellant: Ms J Isherwood, Home Office Presenting Officer

For the Respondent: No appearance

 

DECISION AND REASONS

1.                   This is an appeal against the decision of First-tier Tribunal Judge Anstis promulgated on 23 ^rd June 2016. The appellant before me is the Secretary of State for the Home Department and the respondent to this appeal, is Mr Ahmed. However, for ease of reference, in the course of this determination I shall adopt the parties' status as it was before the First-tier Tribunal. I shall in this decision, refer to Mr Ahmed as the appellant, and the Secretary of State as the respondent.

2.                   The underlying decision that was the subject of the appeal before the First-tier Tribunal was the decision of the respondent dated 4 ^th February 2015. That was a Notice to a Person Liable to Removal in form IS.151A. The reason for the respondent's decision was that the respondent considered the appellant to be a person who had sought leave to remain in the UK by deception. The respondent claims that for purposes of an application dated 13 ^th March 2013, the appellant submitted a certificate from ETS to his sponsor in order for them to provide him with a Confirmation of Acceptance for Studies. ETS undertook a check of the test and confirmed to the respondent that there was significant evidence to conclude that the certificate was fraudulently obtained by use of a proxy test taker. The scores from the test taken on 6 ^th February 2013 at Queensway College, were cancelled by ETS.

3.                   There was no appearance by or on behalf of the appellant before me. The hearing of the appeal was called on at just after 12:15pm after the clerk to the Tribunal had tried to contact the representatives that are on record as acting for the appellant, JS Solicitors. The clerk to the Tribunal had been unable to establish contact with anyone from that firm by telephone. Notice of the hearing listed before me was sent to the parties on 29 ^th November 2016. It was sent to the appellant at an address in London, and although it seems that the appellant has returned to Bangladesh, a copy of the Notice was also sent to his representatives. I am satisfied that the appellant has received Notice of the Hearing. No explanation for the absence of the appellant's representatives has been provided to the Tribunal.

The decision of the First-tier Tribunal ("F tT")

4.                   The F tT Judge sets out the background to the appeal at paragraphs [1] to [5] of the decision. At paragraph [6], the Judge identifies the evidence, and at paragraphs [7] to [30], the Judge sets out the material evidence set out in the statements and expert report before the F tT. The Judge addresses the evidence at paragraphs [40] to [57] of the decision.

5.                   Although the Judge treats the evidence of Lesley Singh with some caution, for the reasons identified at paragraphs [40] and [41] of the decision, the Judge accepted her evidence that the respondent was notified by ETS that the appellant's test score had been treated at 'invalid'. Exhibited to the witness statement of Lesley Singh was an extract from a spreadsheet provided by ETS to the respondent confirming that ETS was treating the test completed by the appellant at Queensway College on 6 ^th February 2013, as invalid.

6.                   The Judge also had available, a document issued by the respondent entitled "Project Façade - Criminal inquiry into abuse of the TOIC" referring specifically to a criminal investigation into Queensway College. The Judge states:

"42. On the face of it, the note of the investigation into Queensway College suggests that there were very serious problems with the ETS tests at Queensway College in the period during which the Appellant took his test.."

7.                   At paragraphs [43] to [49] of the decision, the Judge sets out the perceived problems with the note, and at paragraph [49] concludes:

"49. Whilst apparently suggesting very serious problems at Queensway College, I find that this evidence raises more questions than answers. It is crying out for an explanation as to what ETS did or did not do, an update on what, if any, further steps have been taken in the investigation, and also whether there was any evidence suggesting honesty in testing at Queensway College. The lack of this causes me to treat this evidence with great caution."

8.                   At paragraphs [50] to [52] of the decision, the Judge addresses the generic evidence of Peter Millington that was considered by the Upper Tribunal in SM and Qadir. The Judge notes that that evidence was subject to some criticism in SM and Qadir and at paragraph [53] the Judge states:

"53. I do not consider Rebecca Collings's statement to contribute substantially to the Respondent's case. It contains helpful background and contextual information, but suffers from the same difficulties as Mr Millington's in that it relies largely on information and assurances given by ETS..."

9.                   At paragraphs [54], the Judge addresses the evidence of Professor French. The Judge accepted, in principle, his conclusion that if the process was as ETS said it was, then there would, after listener verification, have been more false rejections than false positives. However, the Judge noted that this does depend on matters being as ETS say they are, and there is much that remains unknown about exactly what training and audit processes were adopted by ETS.

10.               At paragraphs [55] to [57] of the decision, the Judge addresses the evidence of the appellant. The Judge accepts at paragraph [56], that the appellant's evidence does not contain detail about the circumstances in which he took the test, but did not think that the absence of that detail is something that undermines, or should otherwise lead the Judge to doubt his evidence. The Judge accepted what the appellant had said about his education in Bangladesh and about his educational achievements in the UK.

11.               The conclusions of the Judge are found at paragraphs [58] to [63] of the decision. Having weighed up all of the evidence, the Judge concluded that the appellant's specific and apparently honest and first-hand account that he did not cheat must prevail in this case against the generic and second-hand evidence concerning ETS's checking methods. The Judge noted at paragraph [63] that the evidence in relation to Queensway College certainly reveals what appear to be very suspicious circumstances, but the deficiencies in that evidence, mean that it does not weigh heavily against the appellant.

The appeal before me

12.               The respondent advances one ground of appeal. That is, the F tT Judge failed to make rational, adequate or adequately reasoned findings, and that the findings made, were not open to the Judge on the evidence produced. The respondent submits that the Judge failed to provide adequate reasons for the finding at paragraph [63] that the appellant's specific and apparently honest and first-hand account that he did not cheat, must prevail against the generic and second-hand evidence concerning ETS's checking methods. The respondent refers to the evidence that was before the Judge in the form of the witness statements from Lesley Singh, Peter Millington and Rebecca Collins. The respondent submits that read properly, the evidence of Peter Millington and Rebecca Collins establishes that a test score is not categorised as "invalid" until the test has gone through a computer programme analysing speech, and then two independent voice analysts. It is only if all three agree, that a proxy has been used, that the test is categorised as 'invalid'.

13.               The respondent refers to the evidence of Professor French who concluded that there would have been a large reduction in the number of false positives because of human verifiers who are able to assess variables excluded from analysis by the ASR software. The respondent also refers to the evidence of Professor French that the large number of rejected matches by trained listeners, indicates that the ETS system was robust. Only 57.7% of matches (of the 58,464 identified) were accepted by trained listeners, demonstrating that ETS employed the criminal standard, in assessing whether a match was invalid. Staff had to be satisfied beyond all reasonable doubt before declaring a match. The respondent submits that Professor French found that if the 2% error rate established for the TOEFL recordings were to apply to the TOEIC ones, then he would estimate it as less than 1%, after the assessment by trained listeners.

14.               The respondent recognises that in SM and Qadir, the Upper Tribunal concluded that every case belonging to the ETS/TOEIC stable will invariably be fact sensitive and that every appeal will be determined on the basis of the evidence adduced by the parties. The respondent submits that it in the case of the first appellant in SM and Qadir, there was full evidence of how the test was taken, and of the steps taken by the appellant on being told that the test was invalid. Here, there was simply a bare denial by the appellant.

15.               The respondent submits that it was incumbent upon the Judge to properly consider the expert evidence of Professor French as a whole, and his opinion as to the reliability of the conclusion reached by ETS, that the test was 'invalid'.

16.               Permission to appeal was granted by F tT Judge Page on 23 ^rd November 2016. The matter comes before me to consider whether the decision of the F tT involved the making of a material error of law, and if so, to remake the decision.

17.               Before me, Ms Isherwood submits that the decision of the F tT Judge discloses a material error of law. She adopts the matters set out in the grounds of appeal and submits that although a detailed decision, the Judge erred in stating at paragraph [59] of the decision that the respondent has not cast doubt on the appellant's general credibility, and thus treating the appellant as an apparently honest witness. She submits that the credibility of the appellant is at the core of the decision because the respondent's decision proceeds on the basis that the appellant is a person who had sought leave to remain in the UK by deception. That is a direct challenge by the respondent to the credibility of the appellant, and the appellant's very vague evidence, as set out at paragraph [28] of the decision, could not rationally have overcome the evidence relied upon by the respondent in this appeal. She submits that the error particularly manifests itself in the conclusions of the Judge at paragraph [63], in which it is said that " I consider that the Appellant's specific and apparently honest and first-hand account that he did not cheat must prevail in this case, against the generic and second-hand evidence concerning ETS's checking methods....". She submits that the Judge failed to properly consider the evidence of the expert, Professor French, who had prepared his report to give an opinion as to whether, on the balance of probabilities, ETS's methodology was likely to result in any false positives (i.e. speaker comparison test results indicating that different speakers are same person), and if false positives were likely, to estimate how many.

18.               It is now well established that w hat is required in a decision is that the reasons provided must give sufficient detail to show the parties and the appellate Tribunal, the principles upon which the lower Tribunal has acted, and the reasons that led it to its decision, so that they are able to understand why it reached its decision. The reasons need not be elaborate, and need not deal with every argument presented.

19.               In SM and Qadir, the Tribunal had before it the witness statements of Peter Millington and Rebecca Collings in the form that they were presented to the F tT here. A Presidential panel of the Tribunal in SM and Qadir, at [91], described the respondent's case against both appellants as " non-specific and generalised". This contrasts with the present appeal in which the production of further evidence, both general and specific, was available to the Judge. Here, there was, inter alia, clear prima facie evidence of what the Judge describes at paragraph [63] as " very suspicious circumstances" concerning the tests undertaken at Queensway College, London. The Judge also had the evidence of Lesley Singh. The Judge accepted, at [41], that the respondent was notified by ETS that the appellant's test had been treated as 'invalid' by ETS. Furthermore, the Judge had the expert evidence as to whether ETS's methodology was likely to result in false positives.

20.               In SM and Qadir, the following was stated by the Upper Tribunal;:

102. We take this opportunity to re-emphasise that every case belonging to the ETS/TOEIC stable will invariably be fact sensitive. To this we add that every appeal will be determined on the basis of the evidence adduced by the parties."

21.               This was confirmed in the statement of Beatson LJ in SSHD v Shehzad and Chowdhury [2016] EWCA Civ 615, at [23]:

"I do not address the question of what evidence will be sufficient to enable a Tribunal to conclude that there has been no deception. That is likely to be an intensely fact-specific matter."

22.               At [30] Beatson LJ said: regarding the issue of burden of proof in Shehzad and Chowdhury, at [30].

" ..... In circumstances where the generic evidence is not accompanied by evidence showing that the individual under considerations test was categorised as 'invalid', I consider that the Secretary of State faces a difficulty in respect of the evidential burden at the initial stage."

23.               At paragraph [58] of the decision, the Judge found that the respondent's evidence is sufficient to discharge the evidential burden. The Judge also found that the appellant has raised an innocent explanation that satisfies the minimum level of plausibility such as to cause the Judge to consider whether the respondent has shown, on the balance of probabilities, that the appellant cheated in his test.

24.               The appellant did not give evidence at the hearing before the F tT. The Judge refers to the evidence set out in the witness statement of the appellant at paragraph [28] of the decision. The question for the Judge was whether the respondent had discharged the burden on her to show, on the balance of probabilities, and when measured against the appellant's evidence, that he cheated. The Judge resolved that question in favour of the appellant, but did so, in the mistaken belief that the respondent did not cast doubt on the appellant's general credibility.

25.               The Judge notes that the evidence relied upon by the respondent highlights that there remain unanswered questions and uncertainties relating in particular to systems, processes and procedures concerning the ETS testing, and the subsequent testing of scores.

26.               However, the Judge treated the appellant as an honest witness whose general credibility was said not to be challenged by the respondent. The Judge concluded that the deficiencies in the respondent's evidence were such that the appellant's apparently honest, and first-hand account that he did not cheat, must prevail against the generic and second-hand evidence concerning ETS's checking methods. The Judge did so in the mistaken belief that the respondent did not cast doubt on the appellant's general credibility. I accept the submission made by Ms Isherwood that the appellant's general credibility was at the heart of the appeal. The Judge was required to assess the appellant's explanation in that light. Had the Judge done so, it is far from clear that the Judge would have reached the same conclusion and I find that the decision of the Judge is infected by a material error of law and I set aside the decision.

27.               The decision needs to be re-made. I remind myself of the correct approach to the issue of burden of proof and, in particular, the "burden of proof boomerang" discussed in Muhandiramge (Section S-LTR.1.7) [2015] UKUT 675 (IAC), at [9] - [11].

28.               At paragraph [58] of the decision, the Judge found that the respondent's evidence is sufficient to discharge the initial evidential burden. Whilst most of that which is set out on the witness statements of Peter Millington and Rebecca Collins, is generic, the evidence of Lesley Singh establishes that the test score of the appellant is being treated by ETS as invalid. The respondent is entitled to rely upon the evidence provided to her by ETS showing the appellant's ETS certificate to have been invalidated by ETS. Details of how that decision was reached is in witness statements relied upon by the respondent. There is evidence from ETS that the appellant himself has been individually identified as having an invalid test result. In this case, ETS has itself invalidated, following a thorough analysis, the appellants ETS test certificate. If the certificate is withdrawn or invalidated by ETS there is simply no basis for the respondent to simply disregard that withdrawal or invalidation.

29.               The evidence of Peter Millington establishes that the decision is reached with not only voice biometric technology being deployed, but also an independent check by two analysts, one of whom is experienced, working separately.

30.               The document relied upon by the respondent entitled "Project Façade - Criminal inquiry into abuse of the TOIC" refers specifically to a criminal investigation into Queensway College, the college at which the appellant claims to have sat his test. Project Façade is a nationwide criminal inquiry into the abuse of the TOEIC exam. As part of this inquiry, 21 separate criminal investigations are taking place into specific test centres which have been prioritised according to a number of factors. Queensway College is identified as one such test centre. The criminal inquiry has revealed that of the 2793 TOEIC speaking and writing tests completed at Queensway College between 20 ^th March 2012 and 5 ^th February 2014, 70% have been identified by ETS as invalid. The test relied upon by the appellant was completed during this period.

31.               The information relating to the criminal inquiry refers to three audits and a search carried out at Queensway College. Information relating to two of the audits is provided. Evidence of remote testing was seen during an audit on 16 ^th April 2013 and on 17 ^th September 2013, a secret room was identified where 'pilots' (imposters) were taking speaking and writing tests on behalf of the candidates that were located in the examination room. During a search on 25 ^th June 2014, documents relating to TOEIC exams were discovered relating to tests taken between 19 ^th March 2013 and 3 ^rd April 2013, that identified twenty-six candidates with further names alongside them that are labelled 'pilots' and 'agents'. The document also notes that ETS voice samples were compared with the voices of 13 candidates that had been interviewed under caution. None of them were the same. The voice analysis therefore shows evidence of widespread cheating at Queensway College.

32.               Professor French was instructed by the respondent to give an opinion as to whether, on the balance of probabilities, ETS's methodology was likely to result in any false positives i.e. speaker comparison test results indicating that different speakers are same person. If he considered false positives were likely, he was asked to estimate how many.

33.               Professor French begins from the premise that all automatic speaker recognition (ASR) systems make errors. He is of the view that the ASR used by ETS is extremely likely to have produced some false positives among the 58,464 matches identified to software in respect of TOEIC test recordings. He states that the number of false positives produced by the ASR, however, cannot be estimated with any great degree of precision. In his witness statement, Peter Millington refers to the testing of the technology by ETS in 2012/13 with representative data which deliberately included 285 pairs of repeat test takers where it was already known they were recordings of the same person. ETS wanted to know whether the technology could identify occasions where, within one test centre, there was evidence of multiple tests being taken by the same person (or people who had already been identified as imposters). The results of the pilot were that matching samples produced values that were higher than values from the non-matching samples the majority of the time, with a less than 2% error rate. In his report, Professor French notes that the error rate of 2% quoted for the pilot tests is not broken down into false positives versus false negatives. However, in his opinion, even if one assumes the worst case' scenario, that all the errors were false positives, then the 'safety net' system of having trained listeners assess all the matches thrown up by the ASR would have made a very large reduction to the overall number of false positives. The reasons why Professor French is of the opinion that it reviews of the ASR matches by trained listeners would result in a large reduction to the overall number of false positives, are set out at section 3.3.1 of his report.

34.               Since the witness statement of Peter Millington was first provided, ETS has confirmed that the ASR system identified 58,464 matches of which just 33,735 were confirmed by the listener pairs. That is, only 57.7% of the matches were accepted. This 57.7% referred to by Professor French is of course at odds with the much higher 80% referred to by Peter Millington in his statement. However, Peter Millington was working from the rough and ready figure of "over 33,00 possible matches", whereas Professor French refers to the much more specific and focused information that has come from ETS in response to questions raised by the respondent. Although the 57.7% set out by Professor French is considerably lower than the 80% set out in the generic statement of Peter Millington, Professor French presumes the very high rejection rate (42.3%) is in part attributable to the stringent conditions laid down for match confirmation by listeners. That is, 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 a match will result in it being rejected.

35.               Overall, Professor French concludes that the conditions used for trained listener pair confirmation, in conjunction with the (albeit unspecified) conservative thresholds set for ASR match identification, would, in his view, have resulted in substantially more false rejections than false positives. Furthermore, if the 2% error rate established for the TOEFL pilot recordings were to apply to the TOEIC recordings, then he would estimate the rate of false positives to be very substantially less than 1%, after the process of assessment by trained listeners had been applied.

36.               I acknowledge that the respondent's evidence if far from conclusive. There remain unanswered questions and uncertainties relating in particular to the systems, processes and procedures, concerning the TOEIC testing, and the subsequent allocation of scores and the later conduct and activities of ETS. However, I remind myself that the standard of proof is a simple balance of probabilities, and requires neither absolute certainty nor infallibility.

37.               I am satisfied that the respondent has discharged the evidential burden of establishing that the appellant procured his TOEIC certificate by deception.

38.               The appellant's evidence takes the form of a witness statement. That evidence falls to be considered in conjunction with the evidence adduced relied upon by the respondent.

39.               I have carefully considered the matters set out in the witness statement of the appellant. The appellant maintains that he had sat the ETS test and that another person or proxy test taker had not completed the test. The appellant's evidence is that from his childhood he was taught in an Oxford International School and that all the subjects were taught in English. He states that he has been studying in English and that there was no need for him to use a proxy. The material evidence of the appellant is set out at paragraphs [28] and [29] of the decision of the F tT.

40.               It is right to note that there are significant gaps in the appellant's witness statement and, related to this, a complete failure to provide even the most basic description of the tests that he completed including, when, where and in what circumstances. There is simply the barest of claims that the appellant had sat the ETS test, and that the appellant had been taught in English since his childhood in an Oxford International School.

41.               As the Tribunal has recognised, in the abstract, there are a range of reasons why persons otherwise proficient in English may engage in TOEIC fraud. These include, inter alia, lack of confidence, fear of failure, lack of time and commitment and contempt for the immigration system.

42.               The appellant has raised an innocent explanation but noting as I have, the significant gaps in the appellant's evidence in material respects, when weighed against the various strands of the evidence relied upon by the respondent, both generic and specific, I am satisfied that the respondent has discharged the burden upon her of establishing that the appellant had sought leave to remain in the UK by deception.

43.               It follows that I re-make the decision, and dismiss the appellant's appeal

Notice of Decision

44.               The decision of the First-tier Tribunal Anstis promulgated on 23 ^rd June 2016 involved the making of an error of law such that it is set aside.

45.               I re-make the decision and dismiss the appellant's appeal.

46.               No anonymity direction is made.

Signed Date: 18 ^th April 2017

 

Deputy Upper Tribunal Judge Mandalia

TO THE RESPONDENT

FEE AWARD

As I have set aside the decision of the F tT, I also set aside the fee award. I have gone on to re-make the decision and I dismiss the appeal. In the circumstances, no fee award is made by the Upper Tribunal.

 

Signed

 

Deputy Upper Tribunal Judge Mandalia Date: 18 ^th April 2017