Researching Displaced Persons During COVID-19: Catering for Users’ Specific Needsin Media Accessibility Projects

Coronavirus disease 2019 (COVID-19) has created additional barriers to human interactions. For those conducting research with displaced persons, the barriers posed by COVID-19 add up to the existing linguistic, cultural, geographical and ethical obstacles that this type of research involves. In most cases, researchers have resorted to technological solutions to bridge the communication gap caused by the pandemic. However, the heterogeneous profiles and disadvantaged circumstances of displaced persons require further considerations and planning. This paper examines the experiences of researchers conducting research with displaced persons during COVID. It outlines the special considerations taken and provides recommendations for those conducting research in similar contexts. The communities that engaged in this research were based in Greece, Poland, Italy, Lebanon and Spain. While the focus of the study is displaced persons, the insights presented can be of benefit to those conducting research with other vulnerable groups.


Introduction
The restrictions imposed to control the spread of coronavirus disease 2019  have created additional barriers to human interactions in most countries. For those conducting research with displaced persons, the barriers posed by COVID-19 add up to the existing linguistic, cultural, geographical and ethical obstacles that this type of research involves.
In this paper, we present the considerations taken and the implications that arose from conducting research with displaced persons during COVID-19 (April to December 2020). The key purpose of this research is to uncover limitations and offer recommendations for empirical research with displaced persons during the COVID-19 pandemic, including those regarding accessibility, and to map the practices of those conducting research with these persons during these challenging times. This model could be applied beyond the pandemic as it is adaptable to other research contexts, testing environments, diverse audiences, and disciplines.
In this article, the term displaced persons refers to all people subject to non-voluntary migration, regardless of their current administrative status, i.e., those currently seeking asylum, those under subsidiary protection, undocumented migrants, refugees, and any person in need of international protection. This term is used by the World Health Organisation (WHO) to encompass refugees, internally displaced people, asylum-seekers and those under temporary protection (WHO, 2021).
This study is framed within the field of media accessibility studies. Media accessibility studies were born to cater for the access needs of people with disabilities or impairments, through the provision and research of subtitling for the deaf and the hard of hearing and audio description (AD) for the blind and visually impaired (Greco, 2016;Orero, 2004). More recently in the field of media accessibility, attention has been drawn to secondary or extended audiences that would benefit from accessibility services (Greco, 2018;Jankowska, 2019). For example, the elderly may benefit from AD and dubbing (Ellis, 2016;Jankowska, 2019), and learners of foreign languages or those watching screens in a noisy environment from subtitles (Romero-Fresco, 2013). We posit that displaced persons can also benefit from accessibility services when they arrive in a new country and may not speak or read the host country's language. Research has demonstrated that linguistic, digital and literacy barriers prevent newly arrived migrants and refugees to access that information (Federici & O'Brien, 2020;Translators Without Borders, 2017). Indeed, the universalist definition of media accessibility works towards providing access to media products, services and environments to those who would not be able to access them otherwise (Greco, 2016;Szarkowska et al., 2013). Access to information related to regulations on asylum and migration, and access to any support services in the host country is key for displaced persons. This condition is, however, not always met. Applying accessibility principles to communications with culturally and linguistically diverse communities (CALD) would facilitate their access to information and aid their participation in new societies.
The study presented herein is framed within a larger study, REBUILD 1 , on the creation of Information Communication Technology (ICT) tools for displaced communities. The next section provides more information on this international project. Among other responsibilities, the role of TransMedia Catalonia in REBUILD is ensuring that the ICT tool is accessible for as wide an audience as possible. The project was ongoing during the COVID-19 outbreak and a number of research activities were carried out following mobility and social restrictions that impacted the way the data was gathered. Like REBUILD, other European projects involve research with migrant populations. Because of the challenges that the restrictions posed to the research of REBUILD, we set to examine how similar projects faced those challenges and what accessibility considerations were adopted when conducting research with displaced populations. The results of that investigation are presented in this article. This article is structured by first outlining the background and the research context. This is followed by an overview of the methodology. Finally, the findings are reviewed and discussed along with the provision of recommendations for conducting further research under similar circumstances. The insights gained and the reflections of the authors along with those of other researchers and NGO personnel involved in this type of research are also discussed.

Background
This paper reports on the considerations taken and the experiences of researchers and NGO personnel when conducting research with displaced persons during the COVID-19 crisis. Specifically, we report on the research conducted in three contexts: two H2020 projects, namely, REBUILD and SO-CLOSE, based in Greece, Italy and Spain, and on research carried out on the UN World Food Programme (WFP) in Lebanon. The following paragraphs provide an overview of these three research projects. While none of these projects focus specifically on accessibility, it is a requirement in all three initiatives given the diverse needs of their target groups.
The REBUILD project is a H2020 project, the aim of which is to create ICT tools to support asylumseekers, refugees and migrant communities. These ICT tools are designed following co-creation and user-centred methodologies to ensure that the tools meet the requirements and needs of its endusers. Another requisite is that the tools and the information contained therein are accessible for as wide a population as possible. This project is currently ongoing (January 2019 to December 2021). During the COVID pandemic, two research actions were conducted: in-depth, online interviews with NGOs staff; and online and face-to-face questionnaires with asylum-seekers, migrants and refugees. The REBUILD project targets displaced communities and service providers in Greece, Italy and Spain. Due to their geographical location, these three countries are often the entry-point for migrants and asylum-seekers from Africa and the Middle East into Europe.
SO-CLOSE 2 is another H2020 project of three years (January 2020 -December 2022), created to contribute to social cohesion and fight refugee marginalization and exclusion through innovative digital and artistic tools. The first two research activities of the project were conducted during 2020. The first activity aimed to find a methodology and tool to gather information from the target group in order to inform the subsequent phases of the project. The second activity aimed at developing the co-creative methodology and tools to gather information from the target group, users, cultural institutions and academics. For these two research activities, the project conducted semi-structured interviews and focus groups during the COVID-19 pandemic (September to December 2020). SO-CLOSE follows a universal design approach and adds social accessibility to the accessibility agenda.
The World Food Programme 3 (WFP) is a UN initiative whose objective is the end of hunger and the attainment of food security and nutrition by 2030. Though not exclusively, WFP supports refugee families in Lebanon by providing them with cash assistance to buy food. The programme also helps vulnerable families to meet their essential needs through long term initiatives. Lebanon has the world's highest refugee presence per capita; it is estimated that 25% of the population is made up of refugees (World Food Programme, 2021). In 2020, WFP conducted a series of multi-purpose cash assistance studies with Syrian refugees in Lebanon. The accessibility of the methods and tools used for data collection in this project was a condition to ensure equal participation by all within the target community in the study.
Research for the three studies presented above involved engaging with heterogeneous groups under special circumstances propitiated by COVID-19. Thus, a number of considerations had to be taken prior to developing the methodology and collecting data. Such considerations include the accessibility of the research tools, materials, and activities. The next section presents specialised literature on the implications that research with displaced persons entails, and the recommendations for such research during COVID-19, including accessibility recommendations, by two international organisations conducting research in social and humanitarian projects and experts in the field of accessibility.

Methodological and Accessibility Considerations When Researching Displaced Persons
Research with vulnerable groups involves a series of special considerations. Displaced persons are considered vulnerable because they have typically experienced life-changing upheaval, danger, loss and fear, and face enormous adjustments to life in a new environment (UNHCR & IDC, 2016). In the host country, the majority also face linguistic, cultural and legal barriers. Some are also survivors of trauma and torture, due to experiences in their home country or the migration journey itself, which can be traumatic and involve torture and violence, including gender-based violence. Other vulnerabilities factor being stateless, a victim of trafficking, disabled, an unaccompanied or separated child, a pregnant or nursing mother, addicted to substances and/or elderly (UNHCR & IDC, 2016, p. 22). In some cases, one person can experience more than one of these circumstances, which translates to double or multiple vulnerabilities. Mental health issues are also more prevalent among displaced persons. Indeed, a recent study has shown that Syrian refugees can be over ten times more likely than the general population to develop post-traumatic stress and other disorders (Peconga & Høgh Thøgersen, 2020). Furthermore, while some individuals resettle with ease, others risk prolonged and indefinite detention (UNHCR & IDC, 2016). Lack of accommodation and other essential necessities and overcrowding in refugee camps or other asylum centres are recurring issues for newly arrived asylum-seekers (Peconga & Høgh Thøgersen, 2020).
As a result, refugees, asylum seekers and migrants need specific safeguards in terms of research ethics (European Commission, 2020). Some of these special considerations include data minimisation (the principle of limiting data collection to only what is required to fulfil a specific purpose as per the EU GDPR Directive 2018), careful consideration of the data that is collected, the justification of all interactions with participants and the assurance that participants fully understand the implications of participating in the study (European Commission, 2020). Regarding informed consent, the European Commission suggests adopting less formalised procedures than in traditional research, such as including a cultural insider in the process, working with a recognised NGO or using oral consent (European Commission, 2020). The EC (2020) states that for oral consent, the knowledge and approval of a competent Research Ethics Committee is required (p. 3). Written consent involves complex and legal information that is difficult to translate into other languages. Signing consent forms could also jeopardise the participants' anonymity (European Commission, 2020). The research should also add value to the lives of the people that are being researched, who ought to be recognised as subjects in the process and not simply sources of data (Hugman, 2005(Hugman, , 2010. In this regard, Pittaway, Bartolomei and Hugman (2010) claim that in the case of research with displaced persons, the notion that research data belong to the researcher and the institution should be contested.
Other recommendations include involving individuals from migrant or refugee backgrounds in the research to mitigate potential risks and power differences (European Commission, 2020; Pittaway et al., 2010) or, more preferably, using refugee interviewers (Temple & Moran, 2006). To preserve the anonymity of participants, it is recommended that participants be invited to use a pseudonym and to also avoid recording the sessions unless it is necessary (Temple & Moran, 2006). Indeed, voice has become a focal point for supervisory authorities, and it is considered one of the most sensitive and dangerous aspects of personal data to be exploited in our current data era (Mediartis, 2020).
The many regions and countries of origin of refugees require that research with refugees be multilingual. In this regard, it is necessary to involve translators and interpreters to overcome the linguistic barriers. Translation and interpreting are indeed fundamental in creating relationships and opportunities for communication and exchange. In addition to facilitating effective communication, linguistic mediation, ideally through professional interpreters, could neutralise negative impacts and strengthen relationships among researchers and participants (Hunt et al., 2019;Pittaway et al., 2010). However, the provision of translation and interpreting is still deficient in NGOs and public service providers supporting migrants and refugees (Jiménez-Andrés, 2021).
Finally, taking into account disability-related considerations alongside linguistic and cultural barriers to participation in research enables the "most invisible," and often under-researched demographic of displaced persons, to engage in research (Harris & Roberts, 2006, p. 157). Harris and Roberts's study provides practical measures on research with disabled refugees, the challenges faced, and how these were overcome in their qualitative study. These included producing materials in a variety of languages and formats, choosing venues that do not pose social and physical barriers to participation, engaging with disabled refugee organisations and, above all, flexibility and inventive approaches to overcome the barriers and issues that are likely to arise during the research process (Harris & Roberts, 2006). For example, consent forms should be translated into the languages of the participants given the complexity of legal jargon that these usually employ. In some cases, the consent forms may have to be read out loud to illiterate participants, or to persons who cannot read different alphabets. Easy to read consent forms and other research materials could also benefit the engagement of certain participants as they facilitate comprehension especially for persons with reading and learning difficulties (Arias-Badia & Fernandez-Torné, 2020; Bernabé Caro & Orero, 2019). As recommended by experts on accessibility studies (Romero-Fresco, 2013;Orero et al., 2018), taking accessibility into account from the beginning of the design process, in this case, the research design process, allows for the requirements to be identified as research is being designed.
Due to various restrictions and other health-related constraints posed by the COVID-19 pandemic, additional considerations are required when conducting research with displaced persons. The next section outlines the recommendations provided by two international organisations for research conducted with displaced persons during the COVID-19 pandemic.

Considerations When Researching Displaced Persons During the COVID-19 Pandemic
The previous section details the considerations that research with displaced persons deserves. The COVID-19 pandemic has posed additional challenges, driving researchers to modify their approaches to data collection. Because of the vulnerability of the target community, research with displaced people during the COVID 19 pandemic requires careful thought. This section will look at two recently produced guidance notes that outline a series of recommendations for research during the COVID-19 crisis with displaced persons, as well as the opportunities that the new context can afford. The guidance notes were produced by the Humanitarian Advisory Group (HAG) and Glow Consultants (2020), and CartONG (2020).
The Humanitarian Advisory Group (HAG) and Glow Consultants produced a guidance note as a part of the Humanitarian Horizons research program, aimed at adding unique value to humanitarian action by generating evidence and creating a conversation for change. The note is intended for researchers and operational actors as a guide on how to adjust to remote data collection. The directions given in the guide are technically drawn and provide key takeaways from their past collaborations with a range of humanitarian agencies (HAG & Glow, 2020). The second guidance note discussed below was published with the support of the French Development Agency and a Non-Governmental Organization CartONG, specializing in Information Management. The document has a parallel focus, and its key purpose is to answer their partners' questions regarding collecting data during the COVID-19 crisis (CartONG, 2020).

Recommendations
HAG and Glow (2020) provide a comprehensive guide on data gathering and remote monitoring methods and tools during the COVID-19 crisis. Their approach is underpinned by the following criteria: reducing risk, simplicity and rigour (p. 4). Their key recommendations are: (1) limiting the amount of data that actors are willing to collect by prioritizing sites and targeting communities; (2) ensuring that the remote tools used are appropriate for targeted communities (certain displaced communities may lack internet access, and using online platforms would not be advised in such cases); (3) collaborating with others who work in the research field to share the benefits of the already collected data and prevent the recollection of data that have already been collected by other trusted researchers. In this way, actors can overcome the challenge of travelling restrictions (HAG & Glow, 2020). Indeed, sharing the experiences of other researchers is the essence of the present paper.
CartONG's guidance note (2020) provides a summary of generic recommendations applicable to data collection management in the context of the COVID-19 crisis. The list includes (1) avoiding "risky" data collection for teams and participants, (2) sharing your data as much as possible to limit unnecessary data collection and (3) integrating a data protection component into our new tools (p. 2). Additionally, the note provides a list of technological alternatives to data collection and the exchange of information with displaced persons. The alternatives to survey-type data collection were divided into two groups. The first one is the interviewer administering a questionnaire via phone call or using CATI software. The interviewer asks questions over the phone and records them using a computer interface such as Microsoft Excel or Google sheet. Alternatively, the interviewer can use CATI software, which takes care of the call administration. Unlike the first group, the second group specifies alternative ways in which the interviewee can self-administer the questionnaire, including SMS, USSD, mobile web, or SVI/IVR. The interviewee receives a simple structured SMS and sends back the answers as the survey proceeds. USSD (Unstructured Supplementary Service Data) is also used to send text messages and it is considered a faster alternative to the regular SMS since the answers are recorded right away on a web platform. Thirdly, an organisation may share a link with the interviewee so that they may complete the questionnaire directly via their mobile. Lastly, the final remote alternative to survey-type data collection was using an Interactive voice response (IVR) in which the respondent uses keys on their smartphone to answer pre-recorded survey questions (CartONG, 2020).
The choice of one technology over others should be based on the needs of the participants. For example, if a participant is blind or illiterate, phone interviewing will be the most appropriate. However, if the interviewer and interviewee do not speak the same language, or a participant does not have a high level of the language of the study, sending either a written or a translated version of the questionnaire would allow this person to participate in the study. While these guidelines have been drafted for collecting data from displaced persons, they also present useful tools for research in other contexts.

Challenges and Opportunities
Among the drawbacks, HAG & Glow (2020) report that while data collection was challenging, some opportunities arose throughout this process. First, it bolstered researchers' skills to remotely collect data and access communities during lockdowns. Secondly, it entailed coordination with partners who had access to vulnerable groups and displaced communities, resulting in an increase in data implementation and in the expansion of networks. Travel restrictions lessened access to destinations considered high-risk. This limitation reduced face-to-face interactions with interviewees, which resulted in a lack of in-depth interviews. Finally, the continuous need to respond to the emergencies posed by COVID-19 increased the workload for the organizations. As a result, these organisations had less time and resources to collect data safely and with the necessary rigour.
Researchers also faced additional challenges when conducting their studies resulting from the necessity to employ online data collection methods (CartONG, 2020). First, a large number of participants did not have access to the planned alternatives, such as a telephone or internet access. Even though some of the targeted groups owned telephones, these owners were proportionally more likely to be young men. Therefore, accessing certain ages and genders was difficult. The fact that some communities may not trust technology, considering it a spying tool, forced the researchers to work on these negative perceptions before collecting data in order to overcome the challenge of inherent bias. Furthermore, the bias of illiteracy appeared when using technology for data gathering because it required written exchanges. Analysing the data collected using alternative methods required a certain degree of reservation and an acknowledgement of the identified biases (CartONG,2020).

The study
This study involves data gathered from three larger projects: REBUILD, SO-CLOSE and a third-party independent study on the WFP. These three studies conducted research with displaced persons in 2020. To gather the experiences and insights of the researchers involved in these activities, a qualitative questionnaire was designed and administered in order to form an overall picture of the topic under study. Table 1 provides an overview of the respondents and the research activities that they conducted in 2020, amidst the COVID-19 pandemic. The following subsections outline the specific research activities conducted in each project.
The qualitative questionnaire was designed by the authors and administered to researchers from REBUILD, SO-CLOSE and WFP in December 2020. The questionnaire aimed to gather their experiences, insights and reflections on conducting research with displaced persons during COVID. The questionnaire consisted of five open-ended questions enquiring about the ways in which they had adapted their research in light of COVID-19 and their insights on the research process. Their methodological recommendations for future research with displaced persons under similar circumstances were also collected. The questionnaire can be found in Annex 1.
The qualitative data collected from the questionnaires were analysed using thematic analysis, following the guidelines provided by Braun and Clarke (2006). This process involved reading the questionnaire responses iteratively and identifying repeating themes. Then, the themes were coded or tagged. Once coding was completed, the codes were checked against each other and back to the original data set to ensure they were coherent and comprehensive. This process generated a thematic "map" of the analysis. The findings of the thematic analysis are presented in Section 4. Table 1.

Research for REBUILD
Five qualitative interviews were conducted with NGO staff during May and June 2020 in the framework of the REBUILD project. The objective of the interviews was to map the digital and multilingual practices of third sector organisations working with refugee communities. More information on the study and its results can be found in Jiménez-Andrés (2021). Because the interviews were conducted either during or shortly after the first lockdown (March -May 2020), the participants shared information relevant to this study on the ways in which they had adapted their work to the new situation and their new modes of interaction with their beneficiaries.
In addition to the qualitative interviews, comprehensibility tests were administered to migrants and refugees in May 2020. These comprehensibility tests were part of the evaluation of the visual communication of the REBUILD app. In total, 31 participants completed the tests. The participants resided in Greece, Italy and Spain and their countries of origin were Syria, Afghanistan, Iraq, Iran, Nigeria, Pakistan, Senegal, Morocco, Ukraine and Venezuela. The tests were conducted online using Google Forms, and in print (face-to-face). Lastly, online and phone questionnaires were administered to potential end-users to gather data that would train the artificial intelligence in the technological tool.

Research for SO-CLOSE
To gather information during the initial stages of the project, fieldwork was conducted with all major stakeholders: refugee and asylum seeker communities, policymakers, cultural institutions, NGOs and academics. In total, 197 semi-structured individual interviews were conducted across the following four EU member states: Italy, Spain, Poland, and Greece. The individual interviews were conducted online and face-to-face, depending on the availability, location and needs/profile of the participants. Following the interviews, a series of focus group discussions were conducted with stakeholders to delve deeper into the data gathered during the interviews. These were also conducted online and face-to-face, following the same criteria as the focus groups. The focus groups involved between four and six participants.

Research for the UN's World Food Programme
The WFP commissioned independent research on WPF's multi-purpose cash assistance to Syrian refugees in Lebanon. To that end, a number of research studies were conducted with displaced persons during the COVID-19 pandemic. These were (1) multi-sectoral impact studies of the multipurpose cash (MPC), (2) accountability studies of affected populations, (3) value for money studies, and (4) user journey studies. The overall purpose of such studies was to propose recommendations to improve the WFP's multi-purpose cash programme.

Researching Displaced Persons During COVID-19: Insights and Recommendations
The thematic analysis of the qualitative questionnaires reveals three key areas that deserve consideration when researching displaced persons during a pandemic or in circumstances similar to those created by COVID-19. An overview of these can be found in Table 2. They are presented and discussed in the following subsections in relation to extant research. Table 2.

Recommendations for data collection during the COVID-19 crisis
Categories Recommendations

Methodological recommendations
Limit data collection to essential data: consider using data from secondary sources and contemplate involving fewer participants than initially anticipated if possible Adopt innovative and creative approaches For face-to-face data collection, prepare for unexpected circumstances (change in restrictions, participants being infected or having to isolate, among others) Make data collection activities brief Avoid group data collection methods (such as focus groups) Acknowledge the limitations of the sampling strategy Translation, interpreting, and accessibility For remote data collection, ensure the forms are as accessible as possible (easy to fill in, using easy to read words, and available in various languages and formats) Engage interpreters and cultural mediators in the data collection exercises Ensure that interpreters can conduct phone interpretation with the means available

Other considerations
Invest time building trust with participants remotely Examine whether or not your participants have a negative perception of technology Take the data collection interaction as an opportunity to support the participants and raise awareness of the pandemic Source: Authors' own elaboration.

Methodological Considerations
Above all, planning and executing research during COVID-19 required innovation and creativity (Respondents 1 and 2). For the REBUILD project, the comprehensibility tests were originally planned face-to-face with groups of ten to fifteen participants. Due to the restrictions to mobility and group gathering, they were subsequently conducted in different formats, face-to-face both individually and in small groups, and remotely with participants who could take them autonomously. Overall, more time was invested in planning and designing the data collection activities (Respondents 1 and 4).
Following the guidelines provided by HAG & Glow (2020) and CartONG (2020), prior to the start of data collection, researchers in the REBUILD team pondered whether the data collection was necessary. This recommendation is relevant not only during a pandemic -it is a general recommendation by the European Commission for conducting research with displaced persons (European Commission, 2020). The recommendation from the EC to limit data collection to essential data had been observed when the study was designed, but the team reassessed what data was critical for the project to continue and what data could be collected from secondary sources. Hence, the team limited data collection to essential and critical data for the implementation of the project and halved the number of participants per study to reduce risks for staff, partners and participants.
In line with the recommendations by HAG & Glow (2020), Respondent 1 suggested that using data from secondary sources may be the most responsible and effective way to ensure the study goes forward. Restrictions can be imposed from one day to the next and valuable time may have been misspent on designing a data collection activity that could not take place due to a change in the restrictions. Besides, other researchers may have already collected similar data or be planning to do so in the near future. Connecting and sharing data with other researchers and projects carrying out similar studies could allow researchers to use the data that others have already collected or plan to collect.
An important consideration is to keep data collection activities brief in the cases where they are necessary (Respondent 2). Asylum-seekers and refugees in some contexts receive multiple calls from organisations that support them-UNHCR, NGOs, and associations-and spend a lot of time taking part in research, which can lead to research fatigue (Respondent 2). Because of this, it is recommended to keep the interactions with the participants to a minimum. It is important to take into account that, in some cases, the team in charge of the data collection is the same one that is supporting the displaced community. During the first months of lockdown and the restrictions that followed, the workload for these teams multiplied, especially during the first weeks of lockdown when they had to move all their services to online platforms (Jiménez-Andrés, 2021). Researchers must be mindful of their increased efforts and workload when designing the methodology and selecting data collection tools for their study (Respondent 1).
In some cases, researchers were unable to travel to the field to collect the data. Those in the field (i.e., social workers, enumerators) were trained remotely to collect data on the premises (Respondents 2 and 4). This involved having regular meetings and providing a step-by-step guide to those gathering the data on how to do so adequately and safely. Due to the current circumstances, the possibility of having various data collection phases or activities was not recommended or, in some cases, feasible. Hence, anticipating potential complications was necessary so that data could be collected regardless of any unforeseen circumstances (Respondent 1). For example, the participants invited to the study might not be able to attend because they tested positive or had been in contact with somebody who tested positive.
In other cases, data collection was carried out remotely, with no research team in the field. In such cases, data collection tools needed to be self-explanatory and accessible (Respondent 1). In the REBUILD project, the comprehensibility tests were conducted both face-to-face and online. The test was translated into the languages spoken by the participants and sent as a link to those who had an internet connection and a smartphone or computer. For those without devices or non-literate persons, one-to-one or small group sessions were planned in open spaces and the tests were printed and conducted face-to-face. In this regard, the team ensured that completing the form online would not have undesirable effects on participants, such as consuming a lot of data or having to deal with a form which would be overly difficult to fill in (Respondent 1). Online data collection activities allowed for the fact that some participants have little data allowance on their phones, which they might be using for other, more pressing, purposes (Respondent 3). For example, the questionnaire was conducted using a free, easy to fill online survey tool. The questionnaire contained text only (i.e., background pictures were avoided) to ensure that it consumed as little data as possible.
Regarding data collection tools, group interviews and focus groups were replaced by individual or small-group tools (Respondents 1 and 2). For example, in the WFP, focus groups were substituted by individual phone interviews. In the REBUILD project, for the face-to-face comprehensibility tests with small groups, the research team was mindful of whom to recruit, avoiding involving participants who were more vulnerable to COVID such as the elderly, immunocompromised participants, or those with pre-existing medical conditions (Respondent 3). The recruitment of participants was also affected by the restrictions to movement imposed by COVID-19. As a result, in the three projects detailed in this article, the participants who were involved in face-to-face data collection activities resided in the same municipality as the researchers or in the areas where the researchers could travel to. Both limitations to recruitment were acknowledged in the reporting of the results, as the sampling method prevented researchers from having the diverse sample that was initially planned.

Translation, Interpreting and Accessibility
As introduced in the previous section, the accessibility of the data-gathering tool enabled participants to take part in the studies autonomously (Respondent 1 and 3). Accessibility is understood here as making the data gathering tool accessible to as many participants as possible: be they people with disabilities, speakers of other languages, non-literates or low literates, those using mobile devices or those with slow networks. The accessibility of the tool should be taken into consideration at the design stage (Web Accessibility Initiative, 2021). Some accessibility recommendations include delivering content in various formats, making content easy to understand, minimising distractions and using elements that are familiar to the participants (MDN Web Docs, 2021;Web Accessibility Initiative, 2021). For research purposes, a recommendation is to make questionnaires as easy to complete as possible.
For data gathered over the phone, phone interpretation was required with participants that speak other languages (Respondent 3). In the REBUILD project, part of the data was gathered via phone from participants who could not visit the organisations' premises nor fill the questionnaire independently. For participants who spoke other languages, interpreters were involved in the data gathering. This type of interpreting requires a three-way telephone conversation; a practice that had not been adopted in this organisation before. This required additional planning and training of the interpreters involved. In this organisation, the interpreters are refugees that have received training in translation and interpreting from another NGO. Furthermore, the interpreting was not supported by any specialised equipment or software. As a result, the team experienced some initial complications when conducting this data gathering exercise (Respondent 3). Because conducting a pilot study with potential users may not be a possibility during the COVID-19 crisis, a recommendation is to pilot a collection activity internally to anticipate any complications before data collection starts.
It is also recommended to engage with interpreters or cultural mediators and to inform participants about the research and the implications of taking part in the study before the data-gathering exercise, especially for face-to-face interactions (Respondent 5). In other circumstances, this can be done on the same day as the data gathering activity. However, due to the health risks associated with COVID-19, informing potential participants about the purpose of the research, its implications and the conditions under which the research will take place before the activity is vital so that participants can make an informed decision about their participation in the study. In the REBUILD project, translators, interpreters and/or bilingual interviewers and researchers supported all data collection activities. In the research activities concerning this article, the research team talked to the participants taking part in face-to-face data collection by phone prior to the data collection exercise (Respondents 1, 3 and 4).

Other Considerations: Trust, Technology and Support to the Participants
Other considerations raised by participants concerned the relationships between researchers and participants, the use of technology, and the provision of non-research related support to participants.
In terms of the relationships between researchers and participants, two respondents raised their concerns about the difficulties of building trust with participants remotely (Respondents 2 and 5). Respondents explained that they found it more challenging to create a safe space for participants to speak openly given that the discussions were held over the phone. Similarly, it is important to note that, in the case of online data gathering, some asylum-seekers and refugees have developed mistrust towards the Internet. As stated by CartONG (2020), it is recommended that attention be paid to the negative view of these tools prior to data collection. Engaging interpreters and cultural mediators in the study can also build more trust with participants (Hunt et al., 2019;Pittaway et al., 2010).
With regards to technology, an ICT tool that is widely used is WhatsApp. WhatsApp is popular because it is easy to use, affordable and allows for multimodal messaging (text, voice, video, image). For organisations supporting displaced persons in Europe, it is often the default ICT tool for communication with its beneficiaries (Arias-Badia & Jiménez-Andrés, 2021; Jiménez-Andrés, 2021). However, its suitability is contested as sharing confidential information via WhatsApp does not comply with the European GDPR and US HIPAA Act of data protection (Bernadette, 2018;Masoni & Guelfi, 2020). While these organisations do not favour its use, they often do not have a viable alternative for communication with beneficiaries (Jiménez-Andrés, 2021). Respondent 2 recognised that they had no other way to send the consent forms. Researchers are advised to find alternative tools to interact with their participants, and share and collect data with each other, especially if the participants are considered vulnerable.
Finally, given the context in which the data collection is taking place, researchers must anticipate that the data collection exchange can be an opportunity for displaced communities to discuss any issues or questions that they may have due to the COVID-19 pandemic. Respondent 2 recommended having a service map at hand in case participants need to be directed towards a service provider to receive support. The data collection exercise is also an opportunity to inform and raise awareness on the pandemic; the researchers and teams involved in data collection are therefore encouraged to share important messages regarding hygiene practices and transmission prevention during data collection (CartONG, 2020). Research activities ought to be responsive to needs of the researched participants, and ultimately empower the community, more so when working with vulnerable groups.

Conclusions
The COVID-19 pandemic has affected field research across the world, hindering data collection activities for many scientific projects. As a result of COVID-19 related restrictions, researchers have had to reformulate their research methodologies, shifting to alternative methods of data collection. This paper has examined researchers' experiences and insights during 2020 and offers recommendations for conducting studies with displaced persons during 2020, the year that the COVID-19 pandemic hit the world. This model has already been successfully replicated in the H2020 funded project TRACTION 4 which had to test inmates at a prison in Portugal.
Drawing on the results of an online qualitative questionnaire, this study has gathered the considerations and alternative methods of data collection employed by researchers in three projects: REBUILD, SO-CLOSE and WFP. Underpinning these considerations and recommendations is an awareness of the vulnerability of the target community. The key recommendations include limiting activities and data collection to critical data, adopting creative approaches and favouring individual versus group data collection methods. Alternative methods of data collection frequently require the use of technology, which may not be suitable for certain communities. In this regard, researchers shed light on the importance of making online tools for data collection as accessible as possible, by including easy-to-understand content in multiple languages and formats. Engaging interpreters and cultural mediators throughout the process has also proven essential to establish effective communication with participants and build trust in the study. Finally, researchers are encouraged to consider the data collection activities as an opportunity to spread awareness of the pandemic among the participants of the study.
Due to the uncertainty of the duration of COVID-19, innovation and creativity will play a vital role in the research conducted during the pandemic. It is hoped that this study can support other researchers in conducting safe and rigorous research with displaced persons for the duration of the health crisis.