The Judicial Assessment of ‘Expert Evidence’ in the United Kingdom’s Immigration and Asylum Chamber

: This paper examines the role of cultural evidence submitted by anthropologists and social scientists to assist individuals seeking asylum in the United Kingdom. Expert evidence is an essential element in the way that Immigration Judges decide asylum claims. The paper begins by looking at the role of experts and the limits of scientiﬁc evidence in the legal process. I set out the context in which expert evidence features in the First-Tier Tribunal of United Kingdom’s Immigration and Asylum Chamber (IAC). I then show how Immigration Judges in the Upper Tribunal of the IAC assess expert evidence in ‘country guidance’ cases decided between 2015 and 2019. Analysis reveals that experts submit a range of different types of evidence, that judges problematically assess this evidence and that there are serious defects in the judicial process. I conclude by suggesting ways to mediate between the very different roles, training of


Introduction
This paper examines the role of cultural expertise in the United Kingdom's asylum process. Cultural expertise has been defined as a 'special knowledge that enables socio-legal scholars, experts in laws and cultures, and cultural mediators . . . to locate and describe relevant facts, in light of the particular background of the claimants/litigants/accused and for the use of conflict resolution or the decision-making authority' (Holden 2020, p. 45). In the asylum process, expert evidence comes before the court in a written form and, in some cases, it is provided via oral testimony. This evidence may be provided by a range of social scientists who are asked to provide socio-cultural, contextual, linguistic and/or legal knowledge that would be useful to the court. The use of experts in asylum proceedings comes, however, with certain problems for all the parties involved in asylum litigation.
In a brief introduction in Section 1, I set out my methodology in selecting and analysing cases decided by the Upper Tribunal of the IAC which are analyzed in Section 4. In Section 2, I review the history of expert witnessing in the British courts which shows how judges in criminal and civil law have adopted procedural rules to 'hedge' and control expert evidence to maintain their power to decide cases. In Section 3, I discuss immigration law and practice and look at specific cases where IJs have evaluated expert reports which I provided to the First-Tier Tribunal of the IAC. In Section 4, I review cases decided by Senior Immigration Judges (SIJs) in the Upper Tribunal (UT) of the IAC which shows that their analysis of a wide range of qualitative and quantitative evidence submitted in 'Country Guidance' (CG) cases is highly problematic. In Section 5 I set out recommendations that could result in better decision-making by IJ's, and in Section 6 I conclude by discussing why IJs struggle to assess qualitative and statistical expert evidence. in a legal case. This situation contributed to a distorted view of science by the judiciary, namely that scientists should provide evidence that was impartial, disinterested and neutral. However, when expert evidence deviated from this ideological construct, judges used procedural rules to 'dismiss expert evidence as the weakest kind of testimony'.
The 'stick' devised by the judiciary to 'beat the expert witness' took the form of new rules of evidence and procedural/admissibility rules which 'forced experts to fit their evidence into an artefact of the law's design', namely a written report which requires experts to address the issues identified by the lawyer. Experts have an 'overriding duty' to assist the court-which overrides their obligations to the party which instructed them-and their reports must conform to a specific format. In addition, experts must not advocate on behalf of the claimant. 2 The rules give the lawyer who submits the evidence maximum control over the expert and her evidence, and they give judges the power to admit or refuse expert evidence. Furthermore, the adversarial system relies on the parties to question the integrity of an expert and undermine the value of his/her evidence. In effect, experts are co-opted by the legal process. It is important to note, however, that co-optation can be countered, in part, when experts use their research strategically to resist judicial understandings about the 'objective' facts of a case and when they ethnographically analyse 'the dialectical logic' of law as a mode of regulation (Goodale 2017, chp. 5).
Jones interprets the judiciaries concern over the role of experts as arising from a 'professional power struggle' over who has special knowledge and, necessarily, who has the right-the jury, the expert or the judge-to decide a legal dispute (p. 122). Jones argues that: 'The structural placement of experts as witnesses means that they still remain vulnerable to accusations of partisanship. This provides the law with the classical strategy of divide and rule. It sets an expert to catch an expert' (p. 126). Redmayne's (2001, chp. 4) assessment of 'adversarial experts' raises an additional issue, namely the presentation and assessment of quantified data/statistics. This is an important issue because, as the cases examined below demonstrate, IJs 'prefer' 3 quantified/statistical data to qualitative/ethnographic evidence. However, as I argue below, the presentation of statistical data to IJs is just as problematic as in criminal trials because IJs are unable to adequately assess the quantitative evidence that is submitted.
On the basis of empirical research on the sociology of science and law, Jasanoff (2006) observes that 'when science is used for legal purposes, it cannot be taken for granted that the same institutional imperatives apply'. She argues that the imperatives of law 'are not necessarily well suited to discriminating between good and bad scientific claims' (p. 329). She observed that when science 'enters the court room', it is 'not in the form of bare facts or claimed truths about the world, but as evidence' which must meet criteria specific to the law which are that it 'must be relevant, it must be timely, and it must meet the Daubert tests of admissibility' (the emphasis is hers).
She reviews the history of DNA profiling as an acknowledged and decisive form of evidence and demonstrates why the belief that DNA evidence is one of the most reliable kinds of evidence is fundamentally mistaken. Indeed, she shows that judges do not recognize biases or flaws in the conduct and interpretation of forensic science 4 (p. 331). Jasanoff argues that 'what counts as true for the law need not count as true for science, and in exceptional cases even scientific truths may not be accepted as valid for legal purposes. Three dimensions of difference are worthy of note, each reflecting underlying nor-mative concerns that differentiate science from legal practice: first, the divergent roles of fact-finding in science and law; second, the unequal need for certainty in scientific and legal contexts; and third, the disparate ethical constraints framing the production and use of knowledge in the two institutional settings' (p. 333).
Jasanoff shows that 'the standard of proof which litigants have to meet in order to win their case is different from the standard of certainty needed to establish the truth of a scientific fact' and that 'the idea that science can be relied upon to produce fool proof verdicts . . . is both ethically and practically questionable' (p. 334). The situation arises because of flaws and normative pressures in law enforcement/forensic science and because court decisions which arrive at a decision based on 'scientific evidence' cannot be challenged by new evidence and because the adversarial system co-opts experts and instrumentalises their evidence to tell a particular 'legal story'. In short, law is not interested in the 'truth'; it is a contest in which both sides in a dispute attempt to spin a particular story that is favourable to their client. 5 Two key issues emerge from this literature. The first issue is procedural: it is necessary to understand how IJs use procedural rules to regulate experts. The second issue is substantive and concerns the 'rules of evidence' imposed on experts: it is necessary to understand how judges assess the reliability of evidence. In what follows, I examine how both issues affect the way that IJs assess evidence and decide cases.

The UK's Asylum and Immigration System and Cultural Expertise
Refugee law differs from other legal jurisdictions in the sense that: (i) one party is the claimant while the other party is the state (specifically the United Kingdom's Visa and Immigration Department, UKVI); (ii) an asylum claim focuses on the political conditions in the state where the applicant originates; (iii) the focus of judicial decision making is with the future rather than the past, specifically what harms might a claimant be subjected to if her claim for protection is refused and she is forcibly returned to her country of origin; and (iv) judges take it upon themselves to assess the evidence (without a jury).
Because the British asylum system has been discussed by numerous authors (Good 2004(Good , 2007Gill and Good 2019), here I provide an outline of its work based on two years of ethnographic research (Campbell 2017). It is important to point out that the key actors-IJs, UKVI officials, lawyers and experts-exercise very different degrees of power/authority, and they exhibit quite different levels of professional hubris reflecting their radically different view of the world, their place in it and their right to act in particular ways. If we are to understand how asylum litigation and policy unfold, it is necessary to understand how the adversarial legal system shapes the work of all legal actors and expert witnesses (Campbell 2020a).
The first actor to consider are Immigration Judges, most of whom were solicitors with limited knowledge of asylum and immigration law before they joined the Tribunal. While substantial information exists about judges in other jurisdictions in the UK, there is little information about the background or training of IJs. Once they are recruited, IJs receive limited legal training. Reports by the Judicial College contain no information about the content of their training, though it is clear that most training takes place in events lasting one day or less. Newly appointed IJs are provided with a three-to five-day intensive training programme which introduces them to asylum and immigration law and legislation, case law, 'primary purpose' and family law, aspects of 'judge-craft' (for example how to conduct oneself in court), procedural matters, demonstrating impartiality, how to write determinations and how to handle appeals relating to work permits, visitors, marriage and so on. Annual training is in the form of one-day 'updates' and two-day training events and is provided by Senior IJs (SIJs).
The authority exercised by IJs is set out in the AIT's Procedural Rules (PRs) and Practice Directions (PDs). 6 As defined by PR no. 4, an IJ's 'overriding objective' is 'to secure that proceedings before the Tribunal are handled as fairly, quickly and efficiently as possible'. Efficiency is achieved by preventing disruption in the flow of appeals. Indeed, the key role of the Rules and Directions is to regulate appellants, their lawyers and experts.
Unlike in other jurisdictions, the Tribunal exercises a wide discretion regarding the 'evidence' which can be submitted to it. PD no. 29 states that: (1) 'The Tribunal may allow oral, documentary or other evidence to be given of any fact which appears to be relevant to an appeal . . .
(2) The Tribunal may not compel a party or witness to give any evidence or produce any document which he could not be compelled to give or produce at the trial of a civil claim . . . ' This Practice Direction is double edged. On the one hand, 'it is for the appellant to prove that the fact asserted is true', which is to say that the burden of proof (PR 31), though low, lies with the appellant. However, while the 'burden of proof' may be low, it is nevertheless the case that a very large percent of asylum claims are wrongly decided on points of law, which means many claims have to be reheard. In addition, an applicant's evidence is often deemed to lack credibility by an IJ. However, research by Byrne (2007) and others has found that adverse credibility findings are due in large part to a failure by the IAT to adopt appropriate evidentiary principles regarding how to assess testimonial evidence. In short, IJs wrongly assess testimonial evidence and end up refusing many meritorious claims.
In the words of a former SIJ, the Tribunal believes that it possesses 'its own level of expertise as a specialist tribunal, not only in the legal issues . . . but also in its knowledge of country situations and, to a lesser extent perhaps, in consideration and evaluation of medical reports' (Barnes 2004, p. 349;Thomas 2011). 7 This claim is disputed by experts. For instance, Good (2004, pp. 129-31) refuted the positivist attitude of IJs regarding 'objective' evidence. He argues that lawyers and IJs 'are trained to think in radically different ways' than social scientists, which reflects their reliance on the 'purportedly neutral principles' of law which are 'believed to transcend human variation'. For this reason, notions of 'subjectivity' and 'objectivity' are in effect inverted in sharp contrast to the 'analogic thinking' of anthropologists. Social scientists who provide cultural expertise to the courts adopt different theoretical perspectives and reach conclusions not by deduction or syllogistic reasoning but by 'dialectical reasoning' which produces qualified statements rather than objective facts.
Tribunal PDs also define the responsibilities of expert witnesses. Thus, PD no. 10 reiterates the Civil Procedure Directions (CPD) 8 which state that expert reports 'should be the independent product of the expert uninfluenced by the pressures of litigation', that experts should provide 'objective, unbiased opinion on matters within his or her expertise, and [that they jc] should not assume the role of an advocate'. At the same time, experts are required to report adverse judicial decisions on their work and to reveal their sources. As will be seen, however, experts address the responsibilities set out in the Rules in different 6 Tribunal practice directions can be found at: https://www.judiciary.uk/wp-content/uploads/2018/12/ practice-directions-iac-dated-18-dec-2018.pdf (accessed on 10 August 2019); Tribunal Procedural Rules can be found at: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/ file/701695/Consolidated_Tribunal_Procedure__FTT___Immigration_and_Asylum_Chamber__Rules_2014 .pdf (accessed on 10 August 2019). 7 The importance of medical evidence in assessing asylum claims that require separate treatment in diagnosing scarring, trauma, PTSD and other medical issues in asylum claim has been repeatedly upheld in the courts. ways just as they conduct research using different methods and interpret their empirical findings using different theoretical constructs. The role exercised by IJs in legal proceedings enhances the power of UKVI because UKVI is not required to produce its own evidence. Instead, Home Office Presenting Officers (HOPOs, junior Home Office caseworkers) attempt to discredit experts who are instructed by legal counsel for appellants. If expert evidence carried significant weight in the Tribunal, as it does in criminal proceedings, one might understand why IJs regulate experts. However, the status of experts in the Tribunal is limited to providing hearsay/opinion evidence, and PDs give IJs the power to decide how much weight, if any, to attach to an expert's evidence.
In the vast majority of asylum appeals heard in the FTT, the principal type of evidence which is submitted are generic reports from a unit in the UKVI, the Country of Origin Information Service (COIS) and international human rights organizations. It is often the case that reports by human rights organizations are rejected by UKVI and IJs as being 'too generalized', insufficiently 'objective' or 'partisan'. In the FTT, it is rare for experts to submit expert evidence.
The second actor to consider is UKVI. As Gibb and Good (2013), Good (2015) and Thomas (2011) have found, COIS 'Country of Origin' reports vary considerably in terms of their independence from UKVI policy, with regard to the way that staff conduct 'research', their selective use of specific types of sources and how they compile and cite sources. Indeed, in 2012, the Independent Chief Inspector (2017) of Borders and Immigration which reviews UKVI's work found that its reports were biased because they blurred policy and facts. The Chief Inspector also found that their work fails to abide by the guidelines for COI research set out by the European Asylum Support Organization (EASO 2012) regarding the quality and reliability of the COI that should be submitted in the Refugee Status Determination process.
A third actor that influences expert evidence are immigration lawyers. Elsewhere (Campbell 2017) I have described the role played by immigration caseworkers and barristers. Caseworkers, with assistance from interpreters, translate an asylum applicant's story of flight into a claim of persecution. Caseworkers possess variable abilities depending upon the level of their accreditation and whether they are supervised by an experienced solicitor. The quality of their work, including how they identify and instruct experts, varies immensely. Barristers, who have little contact with asylum applicants, are instructed by caseworkers to litigate the case in the Tribunal. As with caseworkers, the quality of their work also varies: younger and less experienced barristers can provide poor advice and may lose an appeal due to a limited knowledge of case law and poor advocacy skills.
The final actor to consider are country experts who can play an important role by providing evidence about a case to IJs. In asylum and immigration law, 'experts' are identified and instructed by legal counsel and may be drawn from medicine, law, anthropology, history, linguistics or journalism. 9 Individuals from different disciplines adopt different methods and use diverse theories; even so, their work should provide nuance and background information that is relevant to an asylum appeal. The key issue for an instructing solicitor is whether an expert possesses an in-depth knowledge of an appellant's country of origin, his/her language or culture and whether s/he can provide a persuasive view of the issues raised by the case. Likewise, the key issues for experts are the need to address the issues raised by the appellant's lawyer, s/he must possess a substantial knowledge of conditions in the appellant's country or origin-either from published 'object evidence' or from his own research-and s/he should not 'advocate' on behalf of an appellant. Finally, and most importantly, in providing their evidence to the court experts expect IJs to adequately engage with and understand their evidence.
Medical evidence is also considered indispensable for deciding whether asylum applicants have been subjected to torture and, if they have, whether the trauma they suffered may affect their memory and thus their account of persecution. These assessments are provided by Medical Practitioners (GPs), psychiatrists, clinical psychologists and other specialists whose reports are based on examining and interviewing asylum applicants. As Good (2007) has noted, however, even though medical professionals use the internationally recognized standards set out in the Istanbul Protocol (UNHCR 2004), their assessments are routinely rejected by IJs who argue that they rely on 'what the applicant told them' which, in their view, means that they have been 'hoodwinked' and are 'biased' in favour of applicants. Medical evidence is also rejected by IJs because 'there are nearly always alternative explanations for an appellant's injuries' (Good 2007, p. 208).
To illustrate how IJs assess evidence, I draw on my personal experience as an anthropologist and country expert in cases in which I have submitted written reports to the FTT. Over the past twenty-eight years, I have provided six-hundred-plus reports to the IAC on behalf of asylum applicants from Ethiopia and Eritrea. I have also provided oral testimony in appeals heard by the FTT and in Country Guidance cases decided by the UT. In my experience, IJs can adopt different approaches in evaluating expert evidence. Three brief examples will have to suffice.
In case PA/00790/2017, UKVI's failure to respond to a fresh application for asylum by an Eritrean applicant triggered a legal appeal in which his lawyer instructed me to provide a report. At the appeal, the IJ summarized my evidence as follows, ' . . . in Dr Campbell's opinion the appellant has provided . . . a consistent and plausible account of events in Eritrea which is partially corroborated by the evidence of Ms. H; the appellant does not fall into any of those categories of persons who might have been allowed to leave Eritrea lawfully; . . . his account of his illegal exit is plausible . . . and the evidence of the appellant in his appeal includes new details (regarding in particular his treatment whilst in the military, his account of his escape from the military camp, and events in Italy) which Dr Campbell finds plausible and consistent with the accounts of other escapees . . . Dr Campbell has also considered the appellants 'sur place' activities and the evidence of Mr. Y, and concludes that public demonstrations referred to therein are likely to have been monitored by the Eritrean embassy, with reports being sent back to Asmara. He also supports the view that dissidents in this country are routinely spied upon by government supporters . . . It is Dr Campbells opinion that upon return the appellant would in consequence be arrested and detained. He comments that desertion from a military post would attract a three to five-year sentence under the Eritrea Penal Code'.
At the end of the hearing, the IJ granted the appellant asylum because he had a well-founded fear of persecution in Eritrea.
In PA/05762/2017, UKVI did not accept that the appellant 'was Eritrean but considered that he was Ethiopian, though it did accept that he had been trafficked out of Ethiopia'. UKVI also argued that the appellant would be protected by the Ethiopian authorities if he were returned there. The IJ heard submissions from both parties and oral evidence by the appellant, and he considered my expert report. During the appeal, UKVI altered its position to argue that the appellant was a national of Eritrea, but the HOPO argued that he could qualify for Ethiopian nationality through his grandmother. 10 The key issue was the nationality of the applicant. In his assessment of the evidence, the IJ concluded: 'In contrast [to UKVI], I have been provided with an expert opinion by the appellant as to the scope of the law and its current implementation. Dr Campbell was in fact the same expert who gave evidence before the tribunal in ST (Ethnic Eritrean-nationality-return) Ethiopia CG [2011] UKUT 00252(IAC). According to the text of the law a person shall be of Ethiopian nationality by descent where both or either parent is Ethiopian . . . Dr Campbell considers this law... He considers that the mechanism by which stateless Eritreans could register operated for a twoweek period in 2004 and since then no efforts have been made by the Ethiopian authorities to register the persons which Ethiopia had made stateless. They exclude all individuals whom they view as being Eritrean nationals or Ethiopian born ethnic Eritreans (which in fact is what the appellant claims to be) . . . In his expert opinion the appellant cannot acquire Ethiopian nationality'.
The IJ continued, 'I must have regard to the opinions of experts when interpreting foreign law. I have not been referred to any contrary authority to that of Dr Campbell as to the appellant's ability to acquire Ethiopian nationality through his mother. He faces the added difficulty that his mother has now been dead for a considerable period of time and he has no documentation to support her Ethiopian nationality'.
In part because the Home Office failed to rebut my evidence, the IJ concluded that 'the appellant had a well-founded fear of persecution for Convention reasons in Ethiopia, and therefore he is a refugee'.
PA/13871/2018 was an appeal which also hinged on a determination of the nationality of the appellant. Following a substantial attack on my evidence by UKVI contained in two separate reports which I submitted in this case, at ¶74, the IJ concluded that 'Dr Campbell, at paragraph 30 of his first report, is critical of Judge Burnett . . . It was not for Dr Campbell to purport to act in an appellate capacity when considering Judge Burnett's decision. The Upper Tribunal have refused permission to appeal the decision that Judge Burnett made'. And at ¶76, the IJ concluded that 'Much of the remainder of Dr Campbell's reports addressed the issues surrounding the appellant's visit to the Ethiopian Embassy and applications for a laissez-passer. He has said that the Home Office have misunderstood his report and that the country guidance cases are not binding on the appellant. Paragraphs 9, 10, 11, 12 and 13 is more advocacy than expertise and we agree with the submission made by Mr. Holt [the HOPO] that this expert has not been impartial, or does not appear to have been impartial given the tone of his two reports'. Having dismissed my evidence, the IJ then refused the claim.
The principal strategy of UKVI in these proceedings is to attack the qualifications of an expert and/or undermine his or her research and evidence (this will become clear in Section 4). UKVI submits COIS reports, but it never provides expert evidence, and its officers and legal counsel always argue that asylum claims should be refused because the claimant has no right to asylum in the country or because it is 'safe' to forcibly return the claimant to their country of origin, i.e., that the claimant is not subject to refoulement. The decisions of IJs, on the other hand, vary. They may decide that an expert has wrongly assumed the role of an advocate for an appellant and reject his/her evidence, IJs may decide that the evidence experts submit is wrong or insufficient and refuse the claim, or they may decide that an expert is qualified to provide evidence and accept their evidence. In all cases, expert evidence is not the sole or even the most important element of an asylum claim because the principle rationale driving decision making appears to be a desire to uphold immigration enforcement to prevent 'migrants'-including the majority of those seeking asylum-from entering the country rather than arriving at a fair decision (Hambly 2019).

The Upper Tribunal and Country Guidance Cases
The statutory basis for hearing country guidance cases in the UK was established in 2004, and since that time, the number of country guidance appeals has rapidly expanded. The process begins with a case management hearing by SIJs who select asylum appeals that raise similar issues; following this meeting, they convene a hearing. The principal focus of cases is to assess the risk facing individuals whom UKVI seeks to deport to their country of origin. At these hearings, a much wider range of evidence is requested by SIJs than occurs in the FTT including written reports and oral evidence from country experts. The rationale for country guidance cases is set out in Tribunal Practice Direction 12.2 11 which states: 'A reported determination of the Tribunal, the AIT or the IAT bearing the letters "CG" shall be treated as an authoritative finding on the country guidance issue identified in the determination, based upon the evidence before the members of the Tribunal . . . As a result, unless it has been expressly superseded or replaced by any later "CG" determination, or is inconsistent with other authority that is binding on the Tribunal, such a country guidance case is authoritative in any subsequent appeal, so far as that appeal: (a) relates to the country guidance issue in question; and (b) depends upon the same or similar evidence'. Box S1.
In other words, CG decisions set a precedent which binds IJ's deciding subsequent appeals which raise the same issues. By the time the UT hears an appeal, many years will have passed since the asylum applicant entered the UK and claimed asylum. This means that country guidance decisions attempt to anticipate future harms which might arise from the possible refoulement of a failed asylum seeker and/or a 'Foreign National Prisoner' (FNP) back to their country of origin.
Case 1. AA (Art 15C) Iraq CG (2015) was convened over two days in May 2015, and a decision was issued four months later. The UT's task was to review evidence on internal armed conflict in Iraq to assess whether there are grounds to believe that a civilian could be returned there without facing a real risk of being subjected to 'indiscriminate violence' amounting to a violation of Art. 15C of the European Qualification Directive. 12 The evidence submitted by the parties consisted of: (i) a written report, a response to supplementary questions by UKVI and oral evidence by country expert Dr Fateh; (ii) one hundred fifty-six reports from international organizations (many were updates on displacement, violence, etc.); and (ii) seventy-eight media reports. The UT were impressed with the expert who, in addition to having considerable first-hand knowledge of Iraq, had written two thousand expert reports. His written evidence was summarized at length, though the Tribunal paid particular attention to his oral testimony regarding Iraqi Civil Status Identity Documents (CSIDs) which purported to register all Iraqis and to his information about the economic and political situation in Iraq. The SIJs found Dr Fateh to be 'an impressive and authoritative witness, in all circumstances we attach significant weight to his evidence' ( ¶90). However, they did not accept his argument about the level of violence in Bagdad or his argument that a returnee would need a 'social network' or 'family connections' to survive. The SIJs dismissed a report by Amnesty International which argued that Bagdad city was 'the most dangerous in the world', and they dismissed as too general and outdated a 2014 UNHCR 'position paper' which advised that States should not forcibly return individuals to Iraq.
The UT decided that while the level of indiscriminate violence had risen in Iraq, the situation was different in Bagdad. They arrived at this conclusion largely on the basis of their analysis of statistics on 'body counts', housing, employment, living conditions, access to health services and attacks or incidents of violence in specific locales such as Bagdad city, the 'Bagdad Belts' and so on. Tellingly, however, and despite expressing a 'note of caution' about relying on 'any particular survey' ( ¶98), at ¶196, the SIJs accepted an assessment published by the International Organization on Migration which allowed them to conclude that the level of violence in Bagdad would not breach the appellant's rights under Art. 15C. It is worth quoting the UT on this point: ' ¶196. Given the precision of these statistics, and the body producing them, we accept they provide a reasonably accurate picture of the living arrangements in Baghdad for persons displaced there-although it is also not difficult to accept that there will be some IDPs who are not captured within these statistics' (my emphasis).
To arrive at their decision, the SIJs divided the number of incidents of violence in Bagdad city by the total population to arrive at an average 13 that was below the number of incidents outside the city ( ¶126). This calculation allowed them to conclude that if a returnee is properly documented by the Iraqi embassy in London prior to being returned to Iraq, then the level of violence in Bagdad did not constitute 'indiscriminate violence'.
An assessment of the level of violence in Bagdad/Iraq is clearly a difficult task which requires an analysis of qualitative and quantitative data as well as a consideration of the personal circumstances of the individuals who are to be returned to Iraq. In so doing, the SIJs relied on statistics to conclusively assess 'risk', but, unlike their 'analysis' of expert qualitative evidence, they did not assess the methodology used by IOM to gather and analyse data, and thus they did not understand the limitations and reliability of IOM's data. This reasoning is more problematic than the UT's reasoning in AS (Safety of Kabul) Afghanistan CG [2018]. A basic consideration of this data would have shown that the statistics were based on 'estimates' of massive population flows which were often too large to track; that officials had limited access to certain routes or settlements due to political instability; and that many displaced persons either were unwilling to register or were under pressure not to register with the officials who provided the data to IOM (Verhulst and Young 2018). 14 In short, IOM data seriously underestimated the level, extent and nature of population displacement in Bagdad which meant that its decision to return civilians to Iraq was wrong. The UT had no more reason to trust statistical data from IOM than other data, including qualitative data.
The court record does not make it clear which party submitted the IOM report. What is clear, however, is that the SIJs did not 'test' this evidence in the same way that they assessed qualitative evidence. As Redmayne (2001, p. 112) noted, 'Most judges . . . cannot be expected to be especially scientifically literate'. 15 Nevertheless, all the evidence submitted to the court should be 'tested'; there is no reason why statistical data should be excluded from this requirement.
Case 2. MST and Others (national service-risk categories) Eritrea CG [2016]. The hearing was convened over five days in June 2016, and the decision was issued four months later. The task facing the UT was to review current evidence regarding the risk on return to Eritrea of individuals who absconded from military service and others who had left the country illegally. This called for an assessment of 'the categories of lawful exit' and whether the 13 A similar line of reasoning was reached by the UT in AS (Safety of Kabul) Afghanistan CG [2018], in which SIJs set aside expert evidence that returnees would need a social network to obtain accommodation or employment in Kabul. Their reasoning on this issue can be found at ¶152-167 and at ¶194-205. This line of reasoning allowed them to set aside evidence from two experts and conclude that social networks 'were not essential for a person to obtain accommodation or employment in every case' ( ¶207). In 'AS', the UT concluded that failed asylum seekers and others could safely be returned to Afghanistan. However, 'AS' was appealed to the Court of Appeal on the basis of the UT's reliance on statistics, and in [2019] EWCA Civ 873, the Court of Appeal found at ¶26 that: 'It is evident that the Tribunal . . . expressed the risk on that basis as being not 0.1% (1 in 1000), as he had put it, but as 0.01% (1 in 10,000). It is common ground before us that that is simply wrong-or, to put it more formally, that a conclusion that the percentage risk was 0.01% was not reasonably available to the Tribunal on the evidence before it'. The UT's decision was based on an error of law, and the case was remitted back for a fresh decision. 14 See: 'IOM forced migration or displacement' (accessed on 12 August 2019) at: https://migrationdataportal. org/themes/forced-migration-or-displacement. 15 There is every likelihood that the calculations underpinning the SIJs' analysis of IOM data are based on a statistical fallacy; see: 'Common Statistical Fallacies and Paradoxes' at: https://www.realclearscience.com/ articles/2017/04/05/common_statistical_fallacies_and_paradoxes_110241.html (accessed on 12 April 2020).
Eritrean system of military service amounted to 'slavery, servitude or forced or compulsory labour' contrary to Art. 4 of the European Convention of Human Rights. 16 The evidence consisted (i) of five written expert reports, answers to written questions from UKVI and the oral testimony of Professor Gaim Kibreab 17 ; (ii) one hundred twenty-six country reports (some were generic, but many specifically addressed reports by the Danishand UK-government Fact Finding Missions); (iii) reports of four European Fact-Finding Missions; (iv) sixteen academic publications; and (v) eighty-six media reports. The UT acknowledged that 'views about the nature and extent of risk awaiting Eritreans faced with forcible return are extremely polarized' and that a good part of this disagreement took the form of 'strong disputes over the methodology [used] by sources' including UK government reports and the country expert. A fair resolution of the case was not helped by the role adopted by COIS.
A notable aspect of this case was that both parties agreed that Eritrea was a closed, one-party state which had imposed a state of emergency. There was no disagreement that the Constitution was suspended; that there was widespread and systematic torture and an absence of 'rule of law'; that national service was indefinite; and that it was a criminal offence to exit the country illegally or to evade national service ( ¶245). Nevertheless, the parties fundamentally disagreed about whether there were significant improvements 'that bear on the issue of risk on return'.
Many of the organizations and individuals whose evidence was submitted to the UT fell into one of two positions and argued in diametrically opposed terms. Legal counsel for UKVI argued that failed asylum seekers faced no or very limited risk if they were forcibly returned to Eritrea, whereas the country expert and the United Nations High Commission for Refugees-which intervened in the case-argued that such persons would face a level of risk which violated their rights under the Refugee Convention.
The SIJs explicitly acknowledged the importance of the criteria identified by the Court of Appeal in 'TK 2009' required them to carefully evaluate evidence to assess its reliability, etc. 18 Indeed, they found that reports from the Danish-and UK-government Fact-Finding Missions provided 'an incomplete snapshot even of key items of evidence at that time and since' ( ¶11). During the hearing, the country expert was subjected to three days of intense cross-examination by counsel for the Home Office who sought to prove that he was biased, that 'his independence was compromised' and that his reports were methodologically unsound and could not be relied upon. 19 The SIJs noted that expert reports 'may sometimes (if not often) amount in the end to just one among other items of evidence which have to be weighed in the balance' ( ¶160). In fact, they criticized all the organizations whose evidence was submitted for their reliance on 'anonymous sources' which limited the amount of weight which the Tribunal could 16 Article 4 concerns the 'Prohibition of slavery and forced labour' and states: '1. No one shall be held in slavery or servitude. 2. No one shall be required to perform forced or compulsory labour. 3. For the purpose of this Article the term "forced or compulsory labour" shall not include: (a) any work required to be done in the ordinary course of detention imposed according to the provisions of Article 5 of this Convention or during conditional release from such detention; (b) any service of a military character or, in case of conscientious objectors in countries where they are recognised, service exacted instead of compulsory military service; (c) any service exacted in case of an emergency or calamity threatening the life or wellbeing of the community; (d) any work or service which forms part of normal civic obligations'. See: https: //www.echr.coe.int/Documents/Convention_ENG.pdf (accessed on 16 February 2022). 17 Professor Kibreab, who is an Eritrean national, has written hundreds of expert reports and had given oral evidence in numerous country guidance cases dealing with Eritrea. 18 The UT adopted a similar position in OO (Gay Men) Algeria CG [2016], in which the SIJs found that one of the experts had provided a 'paucity of evidence' regarding prosecutions of gay men. They also complained about the absence of statistical information. However, the SIJs failed to appreciate that the lack of evidence about discrimination/persecution of gay men might be due to the state's refusal to allow independent organizations to monitor how homosexuals are treated. 19 I attended the hearing and took notes; the pressure placed on Professor Kibreab is not accurately reflected in the written decision. Questions by counsel for the Home Office sought to undermine his credibility; they did not attempt to understand his research methods.
place on the reports (this criticism extended to reports by the UN and other organizations whose reports were not produced for the court; ¶163-164). Counsel for the appellant and the Intervenor (UNHCR) argued that: ' . . . the D[anish] FFM and UK FFM reports on Eritrea have verged on generic a priori arguments to the effect that such missions are inherently compromised because they are only needed when the country of origin in question is repressive and it being in the nature of repressive regimes to be closed societies, they are likely to feature wrongful reliance on anonymous sources, an unrepresentative range of sources and on government or pro-government sources' ( ¶165).
The SIJs set aside the 'generic reports' provided by the Danish and British Fact-Finding missions 20 and stated: 'It is the settled practice of the Tribunal to treat such evidence as of potential value, whether it emanates from a governmental agency or from an international body or an NGO' ( ¶166). Their assessment of the UK's 2016 FFM report was that COIS officials had relied on Eritrean government interpreters and on focus groups and interviews with anonymised sources which were not independent of the Eritrean Government ( ¶199-201).
The SIJs also chastised the expert. For instance, they claimed that he had become 'too personally involved in the DFFM controversy' and that 'the language employed in his critique . . . [was] untypical' of his past reports because he failed to reference certain material and because there was 'an element of one-sidedness' in his treatment of certain NGO sources. A more fundamental critique was that there were 'some significant shortcomings in his patterns of research' and carelessness in how he sourced his material, and the SIJs 'did not always find it easy to follow how he reached certain of his key conclusions'. Indeed, under cross-examination, he was forced to retract some of his evidence and qualify certain conclusions and evidence contained in his written reports.
The UT decided that most Eritreans who left Eritrea after 1991 had done so illegally and that the categories defining lawful exit had remained largely unchanged since 2011. The UT also decided that individuals who are of conscription age and are returned are 'likely to be perceived as a deserter/evader' and would 'face a real risk of persecution, serious harm or ill-treatment contrary to Articles 3 or 4 of the ECHR because they would be required to undergo indefinite military service which is similar to 'slavery".
The UT ignored the fact that UKVI had published biased COIS reports and that its Fact-Finding Mission report had attempted to undermine 'due process' by manufacturing evidence in an attempt to ensure that Eritrean asylum applicants would be refused asylum. The Upper Tribunal's treatment of the evidence manufactured by UKVI as compared to its criticisms of the expert makes it clear that the principal rationale of the court is to enforce official immigration policy rather than arrive at a decision based on a fair assessment of all the evidence.
Once again, we find SIJs arguing that they possess the ability to critically assess social science research. The Tribunal procedural rules allowed SIJs to selectively adopt the evidence they prefer while simultaneously sustaining a very uneven playing field allowing UKVI to undermine the appellants' evidence and the evidence of independent country experts. 20 Following COIS' integration into UKVI's Policy Unit, its reports have increasingly blurred the difference between objective information and UK government policy. This became clear in late 2015 when the Independent Chief Inspector of Borders and Immigration convened a meeting with UKVI to discuss COIS reports on Eritrea which the Inspector had contracted me to review (Campbell 2015). I was highly critical of the two recent country reports on Eritrea for their selective citation of objective evidence, for COIS' failure to observe and follow established guidelines in conducting research and reporting objective evidence for the Refugee Determination System (as set out by EASO 2012) and because the reports clearly reflected the policy of the British state by failing to provide accurate information. UKVI strongly objected to my assessment of their reports, and immediately after the meeting, COIS arranged a Fact-Finding Mission to Eritrea which took place in February 2016 (UK 2016). Its report on this visit provided information which supported its 2015 country reports (which the UT subsequently found to be of 'limited value'). The issue here is that UKVI 'manufactured' evidence to support its own position to undermine the judicial process.

Recommendations
Is the situation described in this paper regarding the analysis and interpretation of expert evidence inevitable or is there a possible way through some of the problems which might result in better decision making? The principal issue is that lawyers, judges and social scientists perceive the world, and the nature of 'facts', quite differently. This situation occurs in an institutional context in which the law is required to address the existence of diverse normative orders (e.g., state-centred v legal pluralism) and a plurality of legal sources (e.g., culture, customary law, foreign and religious law, etc.) to adjudicate asylum claims (Holden 2020, p. 42-f). It is precisely because the behaviour of asylum applicants is influenced by competing norms and discourses that judges need to better understand how social science data are generated and how this type of data should be evaluated. Similarly, experts need to develop a much clearer appreciation of how the adversarial legal system operates in order to work effectively as country experts. 21 A logical first step, but one which the UKVI will surely refuse to accept given its interest in undermining expert evidence, is for the Court to appoint a single expert to examine the issues raised by both parties. This proposal would go a long way towards meeting judges' concerns about the reliability of expert evidence; it would also substantially reduce the need for lengthy cross-examinations of experts and encourage qualified social scientists to provide evidence to the courts. Such a move would need to be accompanied by setting out criteria and a process to identify qualified experts who can be called to provide evidence (as exists in other jurisdictions).
Alternatively, the Ministry of Justice could amend the Procedural Rules to address five issues (see Redmayne 2001, conclusion). First, the court should not permit the aggressive cross-examination of experts aimed at uncovering their supposed 'bias'; cross-examination should focus on allowing experts to explain their research and research methods and seek to clarify their evidence. Second, the rules should allow experts to comment on the quality/nature of the instructions they are given by lawyers and on other issues which they feel the court should consider. Third, the Ministry of Justice should insist that experts are given sufficient time to prepare their reports (frequently, experts are contacted shortly before the hearing which leaves them insufficient time to prepare their reports). Fourth, a register of qualified experts should be created with an explicit code of practice. Finally, at the end of their cross-examination, experts should be asked by the judge 'whether there is anything they wish to say' to clarify issues raised by cross-examination.
It would be useful for experts, lawyers and judges to discuss the plurality of data sources available to decide cases and how best to evaluate qualitative/ethnographic data and statistics. Such a dialogue would hopefully enable lawyers and judges to move away from binary forms of thinking-subjective v objective, qualitative v quantitative, hearsay v objective-to avoid deductive forms of reasoning and to realise the futility of searching for 'the truth'. If the judiciary operates with a 'distorted view of science', it must be said that many anthropologists/experts hold a simplistic view of how the law works.

Conclusions
Country experts provide different types of evidence, some of which is cultural (e.g., family and kinship, language, politics and culture, sexuality, etc.), some of which provides background information on a country (e.g., on political conditions, availability of health services, etc.) and some of which is legal (e.g., on foreign law, nationality, etc.). When expert evidence is submitted in legal cases, their evidence is challenged and instrumentalised to support a story spun by one of the parties in the dispute.
An analysis of the way that IJs in the FTT and UT decide cases shows that they use procedural rules to hedge and control expert evidence and that their assessment of the 21 Over the past fifteen years, several experts in the UK have taught short courses and/or tutored professionals who are interested in becoming expert witnesses. The Royal Anthropological Institute is currently planning to host an online training event for this purpose and it has introduced an accredited training programm for forensic anthropologicsts.
evidence is often flawed. This is particularly the case with qualitative data which are treated as unreliable hearsay evidence, but problems also occur with their preference for, but inability to adequately assess/test, statistical data. I am not arguing that the evidence of country experts should be seen as more valuable or objective than statistical data. However, it should be clear that IJs lack the training to adequately assess/test testimonial, qualitative and statistical evidence. The problem is that they do not realize their limitations and end up preferring evidence which they assume is more objective or scientific while setting aside cultural evidence on language, culture, kinship and the importance of social relations and social networks. Legal conceptions of objective evidence used by IJs, as, for example, set out in 'TK v SSHD 2009', do not assist them to assess social science research because the criteria focus on the analysis of texts (Latour 2011). Given the treatment of experts by IJs, it should be clear that the co-optation of experts by the law is virtually the same as occurs in criminal and civil law. Experts in all legal jurisdictions are 'vulnerable to accusations of partisanship' by judges who are content to regulate experts and to allow opposing counsel to undermine expert evidence.
Judges in the IAC have adopted an approach to social science that is analogous to the 'scientization of law' in criminal law identified by Jasanoff, namely IJs are unable to discriminate 'between good and bad scientific claims' as is evident in the way they assess and disparage social science research and research methodology in the cases discussed in this paper. The IJs' power to regulate experts and dismiss their evidence makes it easy for them to substitute their own 'expertise' to decide asylum claims. In this regard, we have seen that IJs in the UT prefer quantified/statistical data as being more robust than qualitative types of evidence. However, closer inspection shows that the quantitative evidence preferred by IJs has not been tested for its accuracy, reliability, etc.
The procedural and substantive irregularities identified in my analysis of IJ decision making illustrate the way that the judicial process works to assist the two sides in a case to use evidence to spin an effective story. 22 However, in asylum litigation, UKVI is not just spinning a story, it is actively manufacturing 'evidence' in an attempt to undermine independent expert evidence. Indeed, in asylum-related litigation, the Secretary of State for the Home Office (SSHD) 23 has the power to define government asylum policy, and it oversees UKVI which regularly attempts to manufacture 'objective' evidence to undermine the expert evidence adduced by appellants. This situation indicates the 'predominant' role of UKVI in litigation which is at the expense of those seeking asylum (Nader 2002). This is a worrying phenomenon, and it indicates that neither the SSHD nor UKVI believe that they are bound by the UK's legal commitment to the Refugee Convention or to domestic law which requires its staff to fairly assess the applications of asylum seekers. My argument concerning the ease with which IJs dismiss qualitative research in favour of statistics (whether adequately tested or not) also reflects the views of immigration barristers whom I have interviewed who have told me that 'senior IJs were unwilling to take decisions which might be understood by the Government as 'impeding the operation of immigration control" (Campbell 2020b). For this reason, IJs fail to adequately reconcile international legal frameworks with British law in case their decisions challenge government policy. In order to 'square the circle' without overturning case law, barristers felt that IJs diminished the weight attached to expert evidence.
Finally, an acknowledgement of the problems identified in this paper requires us, following Cloatre and Cowan (2019, p. 438), 'to rethink the ways in which law should be understood' by undertaking a critical, empirical analysis of the legal process aimed at laying bare the social practices-here, the assumptions and practices of Immigration Judges, UKVI officials, lawyers and experts-whose work is central to the adversarial legal system. Only empirical research-enabling us to move beyond official and simplistic 22 In the UK, lawyers are not allowed to 'coach' witnesses; see: https://www.barcouncil.org.uk/media/438420 /witness_preparation.pdf (accessed on 16 February 2022). 23 The Secretary of State for the Home Department is a government minister who defines asylum policy and oversees the work of the UKVI and the Country of Origin Information Office (Campbell 2017, chp. 3).
representations of law-will help us understand 'the meanings, sources of authority, and cultural practices that are commonly recognized as legal, regardless of who employs them or for what ends '. 24 Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/laws11020032/s1, Box S1: Country Guidance Cases Examined.
Funding: This research received no external funding.

Institutional Review Board Statement:
The study on which this research was initially conducted was approved by the Institutional Review Board of the School of Oriental Studies, University of London and the Economic and Social Research Council, United Kingdom which funded prior published work.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
No publicly archived datasets were generated during the study.