Psychological research evidence in refugee status determination Open Access

Abstract

Introduction

Most decisions to grant or deny refugee protection hinge on the credibility of the claimant’s allegations (Kagan 2003; UNHCR 2004, 2013; Kagan 2015; Van Veldhuizen et al. 2017a). In other areas of law, decision-makers judging the truth of a person’s account can typically look to external evidence. Criminal proceedings usually involve witnesses and may include forensic evidence (e.g. the results of fingerprint, fibre, or DNA analysis); personal injury claims will invariably be supported by medical documentation. In a claim for refugee protection, however, while decision-makers routinely consult general ‘country of origin’ information (COI) to learn about the conditions in the claimant’s home country, they often have little to rely on beyond the claimant’s testimony when it comes to assessing the claimant’s credibility. Some claimants provide corroborating documents, but the authenticity of these documents is often difficult to verify. Some provide medical reports confirming their injuries, but mistreatment may leave no physical traces. Decision-makers are very often left to decide whether the claimant is telling the truth based simply on their impressions of the story that the claimant tells and the way that the claimant tells it.

These impressions invariably rest on the decision-makers’ ideas about how people think and act: what kinds of details people remember; whether and to what extent people’s memories will remain consistent over time; how people respond to danger; what claimants understand about the asylum system. In some cases, claimants present expert evidence about their own individual psychological state that may help a decision-maker to answer some of these questions. All too often, however, where no individual report is available or where this report does not speak to the point in issue, decision-makers answer these questions using only their own common sense.

If common sense were the only available source of insight into human cognition, decision-makers could hardly be faulted for relying on it. But decades’ worth of relevant and reliable evidence is close at hand. Indeed, the United Nations High Commissioner for Refugees (UNHCR) has stressed that insights from ‘many disciplinary fields, including neurobiology, psychology, gender and cultural studies, anthropology, and sociology’ must inform refugee status decision-making: ‘it is necessary that the whole credibility assessment is duly informed by the substantial body of relevant empirical evidence that exists in these fields’ (UNHCR 2013: 56).

This article considers how refugee status decision-makers use one such body of evidence: evidence from the field of psychology. Part I argues that this evidence is notably absent from credibility assessment in refugee status decision-making. UNHCR makes an unprincipled distinction between COI and psychological evidence. Whereas decision-makers are encouraged to rely on both kinds of COI—‘claim specific’ information that speaks to a particular claimant, and ‘general’ information about what is broadly known about the circumstances in their home country—they are advised to seek out only ‘claim specific’ psychological evidence. Moreover, a body of research that suggests that, in the absence of psychological evidence, refugee status decision-makers rely on common sense assumptions about human cognition—and that while some of these assumptions are in line with established social scientific research, many are not. Part II presents a range of primary and secondary studies whose findings could provide better guidance for decision-makers’ credibility judgments than these flawed assumptions. The intention is not to review these studies in this article, merely to demonstrate that relevant, published, peer-reviewed studies exist, many of which are applied specifically to refugee status determination (RSD). Part III suggests ways of disseminating relevant high-quality scientific research findings to refugee status decision-makers, as COI is currently disseminated, to help support fair and sound credibility assessments.

As many critical scholars have noted, refugee law scholarship has long focused heavily if not exclusively on doctrinal questions, doing law for law’s sake and on law’s own terms. Chimni and others criticize this approach for its failure to consider socio-political factors: international refugee law ‘has been dominated by a positivist tradition which limits the possibility of engagement with politics’ (Chimni 1998; Costello et al. 2021: 5). While such critiques tend to focus on this one important consequence of this analytical orientation, the fundamental limitation of a positivist outlook is not that it ignores socio-political factors per se, but rather that it ignores everything that is not law. This article exposes another weakness of refugee law’s predominant doctrinal focus. Such an approach not only forestalls political engagement; it also isolates decision-makers from crucial non-legal sources of insight into good decision-making.

Part I: The Absence of General Psychological Evidence in RSD Credibility Assessment¹

As UNHCR has noted, ‘The information needed to assess a claim for asylum is both general and case specific’ (UNHCR 2004: para 9). If a claimant alleges that they were arrested during a demonstration, for example, general COI may establish that the demonstration took place, while an expert report may affirm that the claimant shares the political profile of those most often targeted for arrest. Similarly, in deciding whether a claim to a religious identity is credible, UNHCR notes that both kinds of COI can help. Decision-makers facing this task need general ‘country- or region-specific as well as branch- or sect-specific information’. They ‘could also benefit from the assistance of independent experts with particularised knowledge,’ whose expertise the decision-makers will understand within the context of the general COI (UNHCR 2019: 129). In short, UNHCR welcomes both kinds of COI as long as the evidence is ‘accurate, up-to-date and as comprehensive as possible’: ‘[a]ny information which is of potential assistance to the decision-maker in carrying out what can sometimes be the extraordinarily difficult task of assessing a claimant's credibility is to be welcomed’ (UNHCR 2004: paras 9–10).

When it comes to understanding a claimant’s psychological state, in contrast, UNHCR directs decision-makers to obtain case specific evidence ‘wherever possible’ (UNHCR 2019: para 208). An expert report will best allow the decision-maker to evaluate what UNHCR refers to as a ‘mentally disturbed’ claimant. UNHCR nowhere considers whether general evidence could further assist the decision-maker—and indeed whether, as in the case of COI, it might be needed to put the case specific evidence into context.

This single-minded focus on expert reports as the sole source of psychological evidence is dangerously impractical. Expert reports are often unavailable. Cost and the lack of available experts are two potentially significant barriers (see e.g. Wilson-Shaw et al. 2012; Refugee Council 2022: 48 in the UK context). Moreover, refugee lawyers or RSD decision-makers may not recognize the need for expertise to help them to understand the claimant’s psychological state if they feel that their own common sense and experience has equipped them adequately. A small study of lawyers’ decisions to commission expert psychological reports in the UK (Wilson-Shaw et al. 2012) showed that, while lawyers felt that clear signs of severe posttraumatic stress or psychosis were reasons to commission an expert report, other conditions, such as depression, were not, despite well-established general evidence that depression affects cognitive function and thus a claimant’s capacity to present their testimony (e.g. Shilyansky et al. 2016). The lawyers’ own tolerance of others’ distress was also relevant to the decision to commission a report, although obviously irrelevant to the claimant’s needs.

Even when an expert report is available, the expert may merely provide an opinion as to the claimant’s current psychological condition, leaving the decision-maker to work out its implications. The expert may diagnose the claimant as suffering from depression and anxiety, for example, and may even explain that these conditions will likely affect the claimant’s cognitive function. The decision-maker will want to know, however, whether and to what extent these impacts may help to account for specific deficiencies in the claimant’s testimony. General psychological evidence could help the decision-maker to interpret this expert report in the context of what is known generally about the effects of anxiety on a person’s concentration, for example, or about the link between depression and limitations on a person’s ability to give a detailed account (e.g. Graham et al. 2014).

In short, the general body of psychological evidence can help a decision-maker to understand case specific evidence. More fundamentally, since case specific evidence is often simply beyond reach, general evidence—about the consistency of human memory, about risk perception, about responses to trauma—may be the only expert information that decision-makers can rely on to help them to make good judgments about how people think and act.

As it stands, rather than considering this evidence, decision-makers draw on their own experiences in judging the behaviour and motivations of claimants, their families, their persecutors, their traffickers, and any other actors in the claim. Many of the assumptions that they rely on run counter to a significant body of published scientific evidence.

Three studies (Herlihy et al. 2010; Dowd et al. 2018; Skrifvars et al. 2022) have established a methodology for examining the psychological assumptions that underlie the reasoning in refugee status decisions. Herlihy et al. (2010) derived a definition of assumptions from a sample of 20 UK first tier appeal determinations (dating from 2001 to 2007): an assumption is any explanation given by the judge of the behaviour, intentions or motivations of claimants or other people involved in a claim. Using this definition, 117 assumptions were then identified in 10 further determinations (examples of all the parts of the definition were found in these 10 further determinations; no new types of assumption arose). These data were submitted to a thematic analysis (as described by Braun and Clarke 2006), in order to describe the themes present in the data which were: assumptions related to behaviours in the alleged country of origin (‘There’); in the country of asylum and during the claim process, including knowledge of the asylum system (‘Here’); or in the claimant’s presentation of themselves and their account (‘Now’).² Two subsequent studies have replicated and extended these findings in the Australian Refugee Review Tribunal (Dowd et al. 2018) and the Finnish Immigration Service (Skrifvars et al. 2022).

Some of the assumptions identified in these studies are in line with scientific evidence, while others run counter to it. For example, Herlihy et al. (2010) found that,

in line with empirical knowledge about autobiographical memory (Pillemer 1998), a rich account is assumed to be more likely to be an account of events that actually happened. However, there were also assumptions in line with the lay assumption that traumatic material is always clearly remembered. [For example], “given that rape is such a serious thing to happen to any woman, I would have expected a raped person to know when they were raped. This is not the type of event which I would expect a person to forget about or confuse”. (Herlihy et al. 2010: 361)

The latter assumption stands in stark contrast to scientific research findings on the effects of traumatic experiences on memory (Brewin et al. 1996; Herlihy et al. 2002; Deffenbacher et al. 2004; Bögner et al. 2007; Valentine and Mesout 2009; Brewin 2011; Herlihy et al. 2012).

No examples were found in these studies of decision-makers using scientific evidence to evaluate their assumptions. An Australian tribunal member writes, for example: ‘I do not accept [that the claimant’s] husband and children were threatened or in fear, as I expect that if this was [the] case they would not stay living in the same house until a year ago and then live in the same house from that date until now’ (Dowd et al. 2018: 91, our emphasis). To assess whether this expectation had merit, the decision-maker would have needed to consider the research findings on the cultural and psychological factors that affect people’s risk perception, assessment, and management (see Evans Cameron 2008)—which they did not do. A recent Canadian study of legal inferences in refugee status decision-making examined 162 decisions in which a decision-maker found that the claimant’s reaction to a dangerous situation was implausible. Despite an extensive literature on human responses to risk (Slovic 1987; Gierlach et al, 2010; Hogarth et al. 2011; Thompson et al. 2011; Yıldırım et al. 2021; Bernabe-Valero et al. 2021), only one decision-maker addressed psychological evidence (and dismissed it as irrelevant). In no case did the decision-maker look to social scientific sources to inform their judgments (Evans Cameron 2023).

These three studies of assumptions make no comment as to the legal merits of the decisions that they analysed—that is a matter for the courts. Nonetheless, if decision-makers disregard established scientific findings, it seems likely that their conclusions will be neither robust nor reliable. Indeed, both Dowd et al. (2018) and Evans Cameron (2023) note that the reasoning provided for the credibility judgments in the decisions that they analysed was often thin: ‘It was common for bald conclusions to be drawn’ (Dowd et al. 2018); ‘In more than one decision in ten (14%, 22/162), the adjudicators conclude that the claimant’s risk response undermines their credibility in five sentences or fewer.’ (Evans Cameron 2023). Moreover, some of the inaccurate assumptions noted in these decisions led to adverse credibility findings in some cases and to positive findings in others, showing the potential for unprincipled inconsistency in this kind of reasoning (Herlihy et al. 2010; Dowd et al. 2018).

The following section provides examples of social scientific research that tests the assumptions commonly used in refugee status decisions. It includes both primary hypothesis-led studies and secondary sources that compare these assumptions to the key insights arising from the large body of relevant psychological literature.

Part II: The Relevance of General Psychological Evidence to RSD Credibility Assessment

In a 2004 paper, then Vice President of the UK Immigration Appeal Tribunal (IAT) John Barnes suggests a rationale for treating COI and psychological evidence differently (Barnes 2004). Barnes distinguishes COI from medical evidence, including psychological evidence, on the basis that there is a ‘breadth of evidence’ for the former that is lacking for the latter:

In the case of country evidence, the expert is not the sole source of that evidence before the court. There will almost always be other evidence going to similar issues even if not as focussed on the claimant’s account as the expert report is likely to be. The expert evidence can therefore be evaluated against other material, much of which although of more general application will have been produced by other experts in the field…. In contrast, there will be no similar breadth of evidence to assist in the evaluation of expert medical evidence. (Barnes 2004: 354; emphasis added)

On the contrary, many decades of scientific study have yielded a breadth of general evidence about human cognition and behaviour. The sources set out below are included here, not as a systematic review of their findings, but rather to demonstrate the scope of the available peer-reviewed evidence in the psychological literature that is relevant to the questions at issue in refugee status decision-making.

In common with the aims of the special issue on transdisciplinary research, this article focuses on studies with an empirical approach, studies that use data-driven hypothesis-testing methodologies developed in the physical sciences. It also includes studies from the field of psychiatry within the umbrella term ‘psychological’, in so far as both relate to conditions of the mind, emotion, and behaviour.

Primary Studies

In asylum systems across the globe, refugee claimants provide their information through some combination of written statements and face-to-face interviews (including now online, see Given-Wilson and Memon 2022). The first substantive statement that the claimant gives typically then forms the baseline for any subsequent disclosures or answers to further questioning. Problems can arise when later interviews or hearings elicit different details (e.g. Evans Cameron forthcoming).

One assumption found across all of the studies outlined above (Herlihy et al. 2010; Dowd et al. 2018; Skrifvars et al. 2022) is that discrepancies in claimants’ accounts indicate fabrication. Herlihy et al. (2002) conducted two interviews with a sample of UNHCR programme refugees in the UK (these refugees had not been required to present individual asylum claims, nor did they expect to) and found that 32% of details given changed from one interview to the next. Moreover, they found that in people with higher levels of posttraumatic stress, the number of peripheral discrepancies increased with longer delays between interviews. This demonstrates that not only should discrepancies be questioned as a reliable indicator of deception in traumatized people, but also that when weight is attached to this feature, it actively discriminates against people with more severe traumatic stress symptoms. Moreover, this study found that the recall of peripheral information was more likely to be inconsistent over time than central information noting, importantly, that the centrality of each detail to the account was defined by the interviewee themselves. This difference between central and peripheral detail is likely to be important if a decision-maker makes the more subtle assumption that a liar is likely to recall that they claimed a history of torture in a certain year and place but cannot recall the same minor (peripheral) details they reported at their first interview, and therefore concludes (wrongly) that discrepant peripheral detail suggests fabrication. Inconsistencies in answers to repeated questions about a traumatic event have also been examined in unaccompanied asylum-seeking children (12–18 years), finding that almost 60% of them had two or more errors of omission or commission (Spinhoven et al. 2006).

As a more general indication that assumptions about how people present themselves and recall the details of their claim must be used with caution and in the context of claimants’ psychological state, Rogers et al. (2015) showed that behavioural indications of anxiety and posttraumatic stress disorder (PTSD) or depression overlap with the ways we think we can identify liars (Vrij 2008), leading to more adverse analogue credibility assessments when those emotional difficulties were present.

Another assumption found in the analyses concerns specificity and detail. For example, one Finnish determination contained the reasoning, ‘your statements have remained narrow and only contained few details that are typical for self-experienced accounts’ (Skrifvars et al. 2022: 12). In examining this assumption, Graham et al. (2014) extended an established finding that people with PTSD and depression are less able to provide specific memories from their past (e.g. Williams et al. 2007), to refugees and asylum seekers from diverse cultural backgrounds. This study also raises the issue of memory recall styles in different cultures, some studies showing that those from individualistic cultures are more likely to give specific detailed accounts of their history than those from more collectivistic cultures, where narratives play a different role (e.g. Wang et al. 2000; see Herlihy et al. 2012 for a further exploration of this). Both psychological state and cultural differences affect the likelihood that claimants will give detailed, specific accounts.

Another assumption in credibility assessments is that delay in disclosure of full details of a claim is an indication of fabrication. Exploring this assumption, Bögner et al. (2007) interviewed and tested people using standardized measures of PTSD, Depression, Shame and Dissociation following their substantive UK Home Office interviews, asking them also to rate (on a five point scale) how easy they had found it to disclose all that had happened to them. Half of them had been subject to sexual violence and not surprisingly this group had higher levels of PTSD and Shame. However, all participants in the study reported dissociation with respect to their interviews and across the whole group dissociation was highly associated with their inability to disclose personal information. One participant said,

I tried to talk, but my mind kept wandering off and I kept thinking about the trauma and my family that I lost. Everything seemed unreal to me, I felt like I was dreaming. I found it hard to focus on the interview and answer questions. (Bögner et al. 2007: 78)

Suggesting that the reporting of traumatic events is not only difficult for the claimant themselves, 10 of the 27 participants in this study reported that they wanted to disclose their story but had been prevented by the interviewer. For example, one participant said,

I wanted to explain properly, but they just stopped me. They ask you to make it short and give yes or no answers. You don’t get a chance to say much or explain to them. Therefore I did not go into much detail. But that affected me later [at the court] when I was asked why I did tell not them in the [Home Office] interview. (Bögner et al. 2007: 79)

In order to explore the asylum interview Skrifvars et al. (2020) examined refugee interview transcripts in Finland and report that of 7343 questions, only 12.2% were open questions. Almost half of the questions were directive, requiring a specific answer, and a further 34.2% required a yes/no answer. This points to likely validity problems if the completeness of information provided in this interview is then relied upon to assess credibility (Bull 2011; Fisher et al. 2011). One of the limitations of the reliance on closed questions is that they typically uncover only a subset of individual experiences. Failure to ask follow up questions such as ‘please tell me about any other times when you felt persecuted?’ means that from the outset, there will be strong reasons to doubt any conclusions drawn about subsequent omissions in or following this interview (see Vrij et al. 2014 on factors contributing to effective investigative interviews).

There is robust evidence that interview responses can be affected by suggestibility (e.g. Gudjonsson and Clark 1986; Gudjonsson 2018). In summary, experiments have shown that personal factors such as intellectual ability (e.g. Gudjonsson 1988; Gudjonsson and Joyce 2011), state anxiety (e.g. Wolfradt and Meyer 1998) and a high number of negative life events (e.g. Drake 2010a, 2010b; Gudjonsson 2018) can make a person more likely than a control group (‘normal controls’) to yield to leading questions and ‘shift’ their answers if questions are repeated. Anxiety and a high number of negative life events are factors that are likely to apply to refugees (Henkelmann et al. 2020).

Childs et al. (2021) used the gold standard measure of suggestibility, the Gudjonsson suggestibility scale (GSS2) to compare unaccompanied asylum-seeking minors, with a group of disadvantaged UK youth. They found that, compared to their UK peers, the young asylum seekers were significantly more likely to change their answers in response to interviewer feedback. Negative life events and lower non-verbal functioning were significantly more common among the young asylum seekers, increasing their vulnerability to this response pattern.

Other studies in general (non-asylum seeking) samples have shown that the conduct of the interviewer can affect yield and shift, including whether the interviewer is ‘friendly’ or ‘firm’, the latter eliciting more changes of answers (Baxter et al. 2006; Dukala and Polczyk 2014). Being ‘friendly’ in context of an asylum interview is not just human decency; these studies suggest that it can improve the validity of information obtained (Skrifvars 2020).

Another important theme of the assumptions found in credibility assessments concerns what people are expected to know, including what is important in the asylum process. For example, Dowd et al. cite a finding that a claimant, ‘would have been well aware of the need for him to put forward his claimed experiences in Pakistan to enable [his] application to succeed’ (Dowd et al. 2018: 97). Other types of knowledge—of the alleged country of origin—are also tested in order to assess the credibility of the claim. Van Veldhuizen et al. (2017b) designed a study to test the sensitivity (correctly identifying truth-tellers) and specificity (correctly identifying liars) of questions typically asked in asylum interviews to establish asylum seekers’ claims to come from a particular town. They told their study participants (who were not asylum seekers) that the study was about persuasiveness, and that they would be rewarded if they could persuade an interviewer that they had until recently lived in Tilburg in The Netherlands. One third of them did indeed live in Tilburg (‘truth-tellers’), one third were from Maastricht (same country; ‘partial liars’) and one third were from Gothenburg in Sweden (‘full liars’). Half of each group was given 20 minutes to prepare for the interview with access to the internet. The interviewer then asked each participant 10 questions identified previously as typical questions used in asylum interviews. The authors conclude, ‘Even though the results did provide some support for the validity of assessing claims about origin by asking knowledge questions, the differences between the groups were modest, and it was impossible to correctly identify all truth-tellers and liars.’

In sum, ample empirical research offers decision-makers a wide ‘breadth of evidence’ concerning human behaviour, which could help support well-considered decisions. Since critically assessing social scientific research is outside of the remit of most RSD decision-makers, work that collates and critically assesses this research is vital. The following section presents a similar smorgasbord of reviews, guidelines and manuals that provide more tailored forms of the scientific evidence for use in the RSD process, particularly in credibility assessment.

Secondary Sources: Reviews, Guidelines, and Manuals

There are many reviews of memory in the psychological literature, but for the last 20 years at least, authors have applied the science of memory directly to the requirements of RSD. In 2010, Evans Cameron examined the experimental and theoretical literature on memory for time (dates, duration, frequency, and sequence); common objects; repeated events; peripheral information; names; and verbatim memory, relating ‘failures’ in each of these areas to actual RSD cases (Evans Cameron 2010). Additionally, Evans Cameron examines the methodologies that memory researchers employ in order to study consistency including the thresholds commonly deemed necessary to call memory ‘consistent’ in research settings and how these throw light on the Canadian Immigration and Refugee Board (IRB)’s methods (e.g. repeated interviewing with different cues) and thresholds:

A subject [in a memory study] demonstrates a ‘high degree of consistency’ when she directly contradicts only 20 per cent of her previous testimony, and is doing ‘relatively well’ when she misremembers only 20 per cent of her most memorable personal event dates from within the last ten weeks. Such a claimant would be judged a liar by many IRB Members. (Evans Cameron 2010: 510)³

Notably, Evans Cameron’s paper does not focus on the effects of traumatic events on memory. The ‘failures’ of memory in the studies presented apply to us all.

Herlihy et al. (2012) address the role in RSD of autobiographical memory—a conscious ‘memory of an event that occurred in a specific time and place in one’s personal past’ (Nelson and Fivush 2004). This review starts by examining the recall of personal events, before moving on to the additional effects of emotion, notably anxiety (which might occur at the time of the event or at the time of recall), and then the effects of emotional disorder, such as PTSD. The authors also outline literature on cultural differences in autobiographical memories. Each section is related back to the RSD process, outlining how the literature cited might help to inform decision-making.

Some judicial and policy guidelines reference the best available psychological scientific evidence. For example, the multidisciplinary author team of Dowd et al. (2018) reviewed the Australian guidelines (Migration Review Tribunal/Refugee Review Tribunal 2012) and found them consistent with a ‘significant body of literature on credibility evaluation’.⁴ In Europe the Judicial Analysis on Evidence and Credibility Assessment in the context of the Common European Asylum System (European Union Agency for Asylum 2023), used for training judges, includes a chapter on the ‘Multidisciplinary approach to the assessment of evidence and credibility’, which draws together social sciences research alongside the legal considerations of assessment. UK guidance to first instance (state) decision-makers also stipulates that they must consider ‘variations in the capacity of human memory’ (Home Office 2022: 46), although there is little further guidance on what these comprise.

A more didactic approach is offered by the manuals which were produced by the CREDO project (UNHCR 2013)—a UNHCR European collaboration with the International Association for Refugee Law Judges, now the International Association of Refugee and Migration Judges (IARMJ) and other NGOs.⁵ The project analysed credibility assessment in the European context in order to highlight and formalize best practice (UNHCR 2013). Chapters in the subsequent training manuals, written by academics and practitioners from the relevant areas of the social and psychological sciences, provide a tailored, frontline decision-making focused training programme. Volume I covers memory, traumatic memory, culture, and factors affecting the decision-maker; it is available in Portuguese, Russian, Serbian and Spanish as well as the original English. Volume II, available in the same languages, addresses the issues of language, gender, sexuality and gender identity, and children in credibility assessment.

Another training manual was published by the International Centre for Migration Policy Development (ICMPD) following a series of workshops for decision-makers and policy makers in central and eastern Europe. The workshops, delivered by a multidisciplinary team including judges and a psychology researcher, focused on the use of existing jurisprudence in training, also adopting the ‘multidisciplinary approach’ including sections of scientific evidence regarding memory and other psychological processes, as advocated by the CREDO project (see preceding paragraph). This training manual is freely available in English and Russian (Prague Process 2016).

These few examples of how assumptions in refugee status decision-making have been investigated through scientific research methods counter the claim that there is an insufficient ‘breadth of evidence’ to address the psychological questions that arise in RSD. General psychological evidence could and should inform refugee status decision-making as generalized COI does. However, much of this information is published in the medical and psychological academic literature, and is therefore often not immediately available to RSD decision-makers.

Refugee status decision and policy makers require sources of relevant, collated and reviewed scientific evidence, like the many sources of general COI that are available online. The Judicial Practical Guide on Country Of Origin Information (European Asylum Support Office 2018) lists and discusses the range of sources, from intergovernmental bodies’ reports to social media. The final section of this article considers how social science evidence could be made available to improve refugee status decision-making.

Part III: Future Directions

RSD systems might look to two nearby models to develop a procedure for collating high quality, relevant research findings for decision-makers. Firstly an examination of how COI is prepared and disseminated could give guidance in developing models for the inclusion in RSD of psychological evidence. Secondly, another parallel area where decision-makers need the latest scientific evidence in order to make binary decisions is medicine.

Country of Origin Information in RSD

In 2006 a worldwide working party of highly knowledgeable and experienced practitioners and academics published the Judicial Criteria for Assessing Country of Origin Information (IARMJ 2006). They list: relevance; temporal relevance (being up to date); the quality of the sources including methodology; impartiality; independence; and balance.

This list applies equally to the compilation of psychological evidence. Studies need not necessarily be conducted with refugees (for example studies of human memory processes) but should be relevant to the needs of the RSD process; the findings cited should be regularly updated; and careful consideration should be given to the quality of the publications (blind peer review being a gold standard) and the methodology of the study, including the impartiality and independence of the authors and funders. Well-constructed reviews of areas should be prioritized as they can ensure a balanced conclusion where study findings may conflict.

The Research Directorate of the Canadian IRB, for example, produces a collection of COI ‘for every country from which refugee claims originate’ (IRB 2019), each of which is updated yearly (IRB 2022). These National Documentation Packages (NDPs) consist of public documents from international academic, NGO, governmental and media sources that report on country conditions such as political, social, cultural, economic, and human rights conditions. Additional information is gathered during IRB-led fact-finding missions. COI is loosely organized under broad headings, but it is not assessed or synthesized. It includes only publicly available materials for transparency and disclosure purposes (Yeates 2018).

In addition to this collection of general COI, the Board’s Research Directorate also produces Response to Information Requests (RIRs), ‘[f]ocused research reports that answer specific COI questions’ using publicly available information (Daskalova and Sprung 2016: 6). According to Board officials, in preparing RIRs, Research Officers use a methodology that ‘compares, contrasts and corroborates information using multiple sources,’ ‘[r]esearching a variety of publicly available sources to represent multiple points of view’. Officers assess a source’s reliability using the following criteria: ‘Currency; Objectivity; Transparency; Reputation; Mandate/mission of the source; Qualifications and background of the author; Information gathering and reporting methodology; Quality of the writing and presentation; Source funding’ (Daskalova and Sprung 2016: 12–15). Preparing an RIR takes ‘typically…between 3 to 15 days’ and the RIRs ‘undergo an extensive review and analysis.’ (Daskalova and Sprung 2016: 11). Such an approach might usefully be adapted for the provision of psychological evidence in order to similarly ‘answer specific questions’.

Non-governmental organizations that provide COI for use in asylum systems have also elaborated procedural standards. For example, the European Country of Origin Information Network states that it ‘aims at contributing to fair and efficient refugee status determination procedures by securing easy and fast access to high-quality and up-to-date country of origin information (COI) for all actors involved in asylum cases’, committing to provide, ‘information regardless of its positive or negative character’ (https://www.ecoi.net/en/about/about-ecoi.net/). Their ‘Frequently Asked Questions’ address the same standards of quality assurance, transparency of sources, being up to date and how to cite any information used.

It is beyond the scope of this article to examine in detail different national systems’ approaches to the acquisition and use of COI. These various COI processes have their critics and are not without their problems (e.g. Chelvan 2020; Good 2021; see also European Asylum Support Office 2018 on the importance of distinguishing policy from information; see also Yeates 2018 on the need to ensure that COI is helpfully ‘synthesized for decision makers’). Yet, in stark contrast to how general psychological evidence is neglected in RSD, these processes exist: they provide tailored information, available to all actors in the process and they are open to scrutiny.

Evidence-Based Decision-Making in Medicine

Asylum systems may also want to consider another model for providing scientific evidence to expert independent decision-makers, one developed outside of the asylum context. The ‘UK NICE Guidelines in Medicine’ (National Institute for Health and Care Excellence 2022) were developed for the use of health care providers. Doctors and nurses, like judges, have to make crucial, fast, independent decisions, combining their experience with the best available relevant evidence. Each guideline document, comprising treatment guidelines for a specific disease, is prepared by an expert working group, representing different interests, under the coordination of the National Institute for Health and Care Excellence. Based on a thorough scientific and statistically valid review of the evidence in a field, the working group drafts a series of statements that can be supported by the evidence, including negatives as well as positives. Alternatively, in Canada, where there is no national equivalent to the UK NICE Guidelines, health providers make use of an online service behind which medical experts and researchers review and collate evidence pertaining to specific conditions, one that markets itself as providing ‘evidence-based clinical decision support that is clear, actionable, and rich with real-world insights’.⁶

Asylum systems could follow the NICE approach: such a working group would need to include expert psychologists, psychiatrists, social scientists; a statistician to check validity; refugee lawyers; and all levels of RSD decision-makers, including one or more judges. The process would be ‘owned’ by all parties and ensure that the evidence generated would address relevant questions and meet the needs of decision-makers. As with the NICE Guidelines, a process would need to be established for regular updates.

Using either of these methods, or a combination ensuring the best features of both, an RSD system could provide its decision-makers with relevant scientific findings that are methodologically sound, impartial, independent, balanced and up-to-date, and that are presented in a fully transparent form, with reference to original studies or reviews.

Requiring decision-makers to engage with this evidence would not compromise their independence. Asylum systems internationally have recognized that the obligation to source and refer to COI does not interfere with a decision-maker’s independence. On the contrary, a decision-maker's failure to consider relevant COI may be a reviewable legal error.⁷

There would also be a number of advantages to requiring that decision-makers consider and refer to this evidence when their credibility conclusions hinge on findings about psychological processes. Doing so may allow them to demonstrate that they are making decisions in line with the latest science. If it makes them more conscious of the assumptions they are making, it may also lead them to acknowledge cognitive biases—the first step in addressing them (Burke 2006). And it may reveal gaps in their understanding and the need for further training.

Conclusion

State medical institutions and state-registered medical practitioners are required to provide services based on scientific evidence. If you go to a doctor, you expect them to know the latest thinking about how to treat you. Medical ethics requires no less.

Although RSD decision-makers are required to keep abreast of current COI, aside from bespoke training sessions, the authors know of no mechanism currently used by any RSD system anywhere in the world for ensuring that they engage with social scientific evidence. And, according to multiple studies of credibility assessment (e.g. Millbank 2009; Herlihy et al. 2010; Kagan 2015; Dowd et al. 2018; Evans Cameron 2018; Smith-Khan 2019), decisions about refugee status are made and upheld, despite flying in the face of established scientific evidence. For some people this will be a matter of life or death. For others it may mean a long appeal or review process, with costs for all involved.

A few years after the development of the Judicial Criteria for Assessing Country of Origin Information (IARMJ 2006), Storey (2013) was able to say that,

[t]hanks to a number of major studies and initiatives taken by governments, UNHCR, major non-governmental organizations and the IARLJ in combination with significant advances in the jurisprudence, there are now well-established substantive criteria that decision makers, including judges, need to apply when evaluating COI: e.g., that it is relevant, reliable, accurate, up-to date information which has been gathered in a transparent and impartial manner, making use of all relevant sources of information. (Storey 2013: 120–121)

Storey goes on to conclude that ‘there is increasing access to and use of reliable COI, so that one of the two main dimensions of refugee decision-making—the country dimension—is becoming more transparent and rational’ (Storey 2013: 125).

This article has argued that RSD systems badly need processes that enable decision-makers to access and use reliable psychological evidence in their credibility assessments—the other mainstay of RSD. By enabling decision-makers’ access to and use of reliable social science evidence, the credibility dimension might also be made more ‘transparent and rational’, leading to fairer outcomes.

Accordingly, this article proposes that the efforts that have been put into ensuring that country information is properly sourced, examined and presented to decision-makers also be applied to psychological evidence. The article has shown that there is a body of evidence regarding the psychological processes at play in asylum claims and has argued that this evidence should be made available to decision-makers in the same way that COI is. Drawing on the models developed for COI and those in the world of medicine, this article proposes the formation of working groups comprising representatives from all of the disciplines involved (decision-makers, lawyers, researchers) who would be responsible for setting up and maintaining procedures to collate, assess and distribute relevant psychological evidence to the RSD process. Although beyond the scope of this paper, the authors expect that similar methods could also be used to bring to bear evidence from other fields of social scientific study. As noted at the outset, and as UNHCR has affirmed, other fields of social science may also offer important insights, e.g. linguistics—see Smith-Khan (2019); behavioural economics—see Evans Cameron (2018: 10–15) and anthropology—see Good (2007).

For far too long, refugee law scholarship has focused narrowly on doctrinal questions, ignoring the real-world contexts in which RSD operates. This has hindered the development both of critical legal theory and of evidence-based decision-making. Whether the interface of psychological science and refugee law supports reform⁸ or the improvement of current processes as we propose in this paper, the contributions to this special issue on transdisciplinary approaches demonstrate that drawing together the diversity of vision of multiple disciplines can bring integrity to the field. All will benefit, from decision-makers looking to make good decisions, to academic researchers seeking to understand RSD systems in their broader context, to those seeking refugee protection.

Footnotes

Data Availability

No new data were generated or analysed in support of this research.

Acknowledgements

The authors would like to thank Thomas Gammeltoft Hansen and the editors and anonymous reviewers at JRS for their contributions to improving the quality of this article.

References