Challenging ethical issues of online ethnography: reflections from researching in an online translator community

ABSTRACT

This article explores challenging ethical issues involved in the conduct of ethnographic fieldwork in an online translator community. How online ethnography is different from its offline counterparts is increasingly being discussed. Many discussions have revolved around the research ethics potentially changed by such differences. It is not uncommon to see conflicting views and suggestions of ethical conduct in these discussions, and resolving these ethical conflicts during the design and conduct of a research project can take time and effort. This article focuses on our conduct of ethnographic observation, surveys, and interviews in an online LGBT+ subtitling group. It highlights four ethical issues involved in this online ethnographic fieldwork that required particular attention and effort: access to a site of enquiry, participants’ informed consent, data management, and relationship management. It provides a reflective account of our decision-making when dealing with these ethical issues. We conclude that these ethnographic issues are not necessarily new nor exclusive to fieldwork in an online translator community. However, resolving them requires more nuanced reading and creative application of existing ethical guidelines.

KEYWORDS:

• Ethics
• online ethnography
• internet research
• online translator communities
• fieldwork

Introduction

This paper was motivated by the authors’ experiences of conducting an in-depth ethnographic project to research translator-initiated online activities, specifically the everyday experiences and motivations of subtitlers in an online Chinese LGBT+ subtitling group. At the outset, we were aware that there would be significant ethical issues to consider in the conduct of this research: representation of sexual minorities in media in China is sensitive politically and socially, and online subtitling groups in China can be legally ambiguous depending on the media content involved and its copyright. We were prepared to deal with our research participants as people exposed to several layers of social, political, and legal vulnerability. However, we were unprepared for the time and effort that would be required to explain the context of translator-initiated online activities in China to academic colleagues from diverse disciplinary backgrounds as part of institutional ethical review. We also underestimated the ethical issues that would be presented by proposing to conduct an ethnographic study of these activities exclusively online. We provide these reflections on our experiences to help others who may be considering similar projects, taking into account that the study of translator-initiated online activities is increasing in the discipline of Translation and Interpreting Studies (TIS) and they are not a minor phenomenon (Li 2017; Pérez-González 2017; Yu 2020a). In short, we wish that we had read such a paper prior to the start of our own project.

We begin the paper by exploring the problem space of ethical issues involved in ethnography and its conduct online. We then briefly describe the study of an online Chinese LGBT+ subtitling group that we used to reflect on these issues. We focus our discussion on four main ethical issues – access to a site of enquiry, participants’ informed consent, data management, and relationship management – and propose tentative conclusions based on our reflections.

The ethics of ethnography and its conduct online

Ethnography is a form of research used to interpret and explain the life experiences and behaviours of a social or cultural group (Madden 2017; Coffey 2018). It has developed and been used in different forms across a number of disciplines over the last century. This includes increasing use among TIS scholars in recent years.

Some researchers in TIS see a natural affinity between ethnography, translation, and interpreting through their shared concerns with processes of interpretation, writing, dialogue with external research sources, and an aim to translate one culture in terms intelligible for another (Wolf 2002; Buzelin 2009). Scholars in TIS have also drawn on ethnographic research as ways to deal with issues of partiality, historical contingency, representation, and understanding of the Self and Other (Agorni 2007; Bachmann-Medick 2009; Wolf 2002). More significantly, for other TIS scholars, ethnography has provided them with a repository of methodological tools and approaches to collect and analyse data about translation, interpreting, translators, and interpreters in new ways. They have used large-scale ethnographic projects to examine translational phenomena at independent publishing houses, government institutions, non-governmental organisations, hospitals, and more (see, e.g. Davidson 2002; Angelelli 2004; Mullamaa 2006; Koskinen 2008; LeBlanc 2013; Tesseur 2014; Duflou 2016). TIS scholars have also started to apply ethnographic methods to the study of online translator communities (e.g. Dombek 2014; Li 2015; Wongseree 2017; Yu 2019, 2020b; Lu and Lu 2021).

Certain characteristics are fundamental to ethnography in whatever form it is conducted and whatever the disciplinary background. These characteristics relate in particular to the authenticity and validity of the sites of enquiry involved, the nature of participation of researcher and researched, their relationships and positions, and the variety of data sets used to generate meaningful interpretations and explanations.

Ethnography involves describing social and cultural phenomena in great detail, or ‘thick description’ in Geertz’s (1973) term, in authentic settings (Brewer 2000) where the researchers immerse themselves in the phenomena and use a variety of methods, including observation of participants, interviews, documentary analysis, survey questions, artefact analysis, and various forms of recording (see e.g. Coffey 2018). A setting is authentic and valid when it allows the researcher to become immersed in a particular social context at different levels – physical, social, mental, emotional, etc.–and generates a rich experience of the research context (Jones 2010). This implies that the researcher must negotiate emic-etic tensions in their work and acknowledge their insider and outsider perspectives on the social group (Wolcott 2008). From this richly described immersive experience of a social group in a site of enquiry, a researcher attempts to find systematic interconnectedness and relationships between large numbers of small data points (Agar 1980) to achieve phenomenological understanding and create meaning (Jones 2010).

These fundamental characteristics of ethnography raise numerous general ethical questions. For instance, if access to and description of a social group in an authentic site of enquiry is required, what right does the researcher have to access that site and write about someone else’s world and worldview and should the researched participants be made aware of and able to read and contest the researcher’s account (Angrosino 2005)? In other words, what authority does the researcher hold to represent the social group being studied (Stanley and Wise 1983)? If the researcher is already part of the social group prior to the beginning of the research and feels they can ensure access and representative authority, will the findings of the research have consequences that may even change the character of the situations that were studied (Hammersley and Atkinson 2007)? How will the researcher reflect on and critically engage with their own participation in the ethnographic frame (Tedlock 2005)? In addition, should only the researcher get to identify the questions to ask, the problems to attempt to solve, or what to report about the social group (Fontana 2004)? In contrast, should ethnographic projects place participants’ voices at the centre of the process (Sundar 2006)?

In addition to general ethical questions arising from ethnographic practice, there have been many discussions on the ethical issues of applying ethnography online (e.g. Reid 1996; Hine 2000; Buchanan 2003; Driscoll and Gregg 2010; Boellstorff et al. 2012; Kozinets 2015; Sugiura, Wiles, and Pope 2017; Lavorgna and Sugiura 2020). Moreover, many academic associations such as The Association of Internet Researchers (AoIR), The British Educational Research Association (BERA), The Council of American Survey Research Organizations (CASRO), The British Psychological Society (BPS), The British Sociological Association (BSA), and The Market Research Association (MRA) have proposed and revised guidance on how to conduct online research ethically.¹

Guidance from different sources such as the above can conflict, and there are four main issues on which sources disagree. First, there is disagreement on establishing the boundaries of an online site of ethnographic enquiry, especially if the research involves the observation of people’s interactions online. Some scholars suggest that online interactions are text-like and these texts are publicly published online (e.g. Bassett and O’Riordan 2002; Walther 2002), while others contend that online interactions such as direct messages occur in private spaces (e.g. Kozinets 2015). This disagreement might be due to the fact that some online interactions happen in more private spaces where (the archives of) the interactions are only accessible to those who participate in them, while other online interactions happen in more public spaces that are publicly accessible (Roy and Beckman 2005). Because there is no consensus on whether online interactions are public or private among researchers and participants, there are then no definitive ethical answers as to whether permission to access such interactions should be obtained or not. Second, some researchers argue that all online observations involve human subjects (e.g. King 1996; Sharf 1999), whereas others observe that some online interactive information has no identifiable creator and are public archives (e.g. Frankel and Siang 1999; Bassett and O’Riordan 2002; Kozinets 2015). Therefore, whether consent should be gained for using a piece of online information also depends on the nature of the information. Third, there is little consensus on the anonymity that an online identity affords. Some argue that online identities are mostly fabricated and online user names are automatically pseudonymised and semi-anonymous (e.g. Reid 1995; Frankel and Siang 1999; van Hout 2014). In contrast, there is also the opinion that there is consistency between people’s online and offline identities, whereby online anonymity is an illusion and can be deciphered (e.g. Correll 1995; Katz 1997; Zimmer 2010). Such disagreements result in confusion about the copyright/ownership of online information and whether and how collected online information should be credited to the online identities or fully anonymised. Finally, although scholars seem to agree that intimacy is heightened online, they do not agree on the level of intimacy that ethnographers should maintain with participants. Some argue that intimacy is the key to immersive participation (e.g. Driscoll and Gregg 2010), while others strongly object to excessively intimate relations between researchers and those being researched (e.g. Boellstorff et al. 2012; Kozinets 2015). There is further disagreement on the extent to which power is (im)balanced between researchers and their participants, particularly that self-declared researchers are not necessarily in their traditional position of authority when interacting with their online anonymised participants, whereby the research might even be steered by the participants (Sade-Beck 2004; Boellstorff et al. 2012; Kozinets 2015; Lavorgna and Sugiura 2020).

In sum, it can be seen that guidance and academic arguments in the area of ethnography online and offline are not entirely clear. The ethical conundrums that we encountered as we designed and conducted our own ethnographic project led us to the question that we are trying to answer in this paper: What are the challenges of dealing with ethical issues when applying ethnography online rather than offline, particularly when researching translator-initiated online activities? To answer this question, we used our own experiences of conducting an in-depth online ethnographic project.

The study

The study aimed to explore the everyday experiences of subtitlers in an online Chinese LGBT+ subtitling group (hereafter referred to as the Group). This group of subtitlers were working together online, subtitling LGBT+ related media content such as films, TV series, music videos from outside China into Chinese and making subtitled versions available online to a community of Chinese audiences (hereafter referred to as the Community). The Group is active on a number of online media platforms such as CHAT (a chatting platform), BLOG (a blogging platform), VSHARE (a video-sharing platform), and QWEB (the Group’s own website).²

The overarching research questions of the study were to enquire on the subtitlers’ motivations to engage with the Group and how they vary in different possible ways (e.g. initial and developed motivations, and individual and collective motivations). In order to uncover the complexity of their perceived motivations, an online ethnographic research approach was adopted. This approach allowed the lead author of this article to conduct a lengthy period (12 months) of fieldwork on the Group’s multiple online sites, enabling close contact with the subtitlers and an immersive experience of their everyday practices in the Group.

The research followed an inductive process. The data collection and analysis were designed to be adaptable to the field and to respond to the fortuitous situations of fieldwork. There were three phases to the process, where later steps were informed by and developed upon earlier findings.

Phase one was to understand the general organisational and technical structures of the Group. Such general information could be collected through online ethnographic observation, a form of digital archiving of online information as left by interactive internet users. The lead author of this article, who is fluent in Chinese and was an existing member of the Group prior to the research, conducted netnographic observation continually for 12 months in the Group, by taking screenshots of platform/web-based information as they were archived online. The screenshots were stored with the researcher’s descriptions and became the ‘archival data’ (Kozinets 2015, 165) of the first phase of fieldwork.

Phase two was to understand the demographic and motivational constitutions of the Group. Such information had to be collected from individual subtitlers. Given that we only needed a general understanding at this phase and that we needed a large variety of responses to know what possible demographic and motivational elements were present in the group, an online survey in simplified Chinese and English was used to gather data. The survey asked individual subtitlers for certain demographic information, about their motivations, and their feelings towards certain online activities observed in the first phase. The survey was carried out via SurveyMonkey, sent to respondents via a web-based link around the middle of the observation period, and stayed active for 45 days to collect the ‘elicited data’ (Kozinets 2015, 165) of the second phase of the fieldwork.

Phase three was to understand in more depth how the motivational factors found through the survey were negotiated and fulfilled individually by the demographically diverse group members. This phase mainly involved online interviews, which were conducted towards the end of the 12-month fieldwork. Participants had the freedom to choose among emails, written-text messages, audio messages, and audio(visual) calls as the medium of the interview. The content of the interviews was transcribed and translated and became the elicited data of phase three.

In sum, our reflections are based on a three-phase ethnographic project conducted online involving online ethnographic observations through digital archiving, an online survey, and online interviews. An additional method, reflexive journaling, was also adopted throughout to maintain the reflexivity of the research and provided ‘produced data’ (Kozinets 2015, 165). In the next section, we discuss the unique ethical issues involved in these online data collection steps.

Ethical issues of online ethnographic fieldwork in online translator communities

Based on our experience, there were four ethical issues involved in conducting online ethnographic fieldwork in an online translator community that required particular attention and effort: access to a site of enquiry, participants’ informed consent, data management, and relationship management. This list is by no means exhaustive. However, it encompasses four particularly contentious areas of debate about the ethics of online ethnography among academics and research institutions.

Access to a site of enquiry

The first ethical issue we encountered during fieldwork was related to gaining access to the ethnographic field that would constitute the site of our enquiry. To conduct online participant observations or other online ethnographic fieldwork, researchers often need to be in the same online location or site with participants. This may require researchers to negotiate access with site gatekeepers so as to respect participants’ privacy and autonomy (Reeves 2010). However, as mentioned before, there was no consensus in ethical guidelines on the boundaries between public and private online sites and hence on whether access to a site of enquiry needs to be negotiated. It is also worth mentioning that gatekeepers of online sites are not easy to identify because access to online sites is often automated. We had to look at specific situations in our research and made nuanced arguments on this matter.

In our research, the Group was active on multiple sites (CHAT, BLOG, VSHARE, and QWEB) in different manners. The purpose of BLOG and VSHARE for the Group was to make public announcements to the platform users who do not have to be followers of the accounts to see the content thereof. In other words, these two sites were already publicly accessible. In contrast, the Group’s sites on CHAT and QWEB were semi-closed and required gatekeeper negotiation. CHAT is a chatting platform with a chat-group function that allows users to create cohort-like groups. People who are part of the group can access the content shared within the group including messages and also documents, whereas outsiders cannot. To join, one had to contact the Group’s official BLOG account. Similarly, outsiders had to register on the welcome page of the Group’s website (QWEB) to access the content inside; and for the registration, one also had to contact the official BLOG account of the Group and ask for an invitation code. The researcher observed that these four sites should be treated differently in terms of access negotiation. On the one hand, the Group’s BLOG and VSHARE accounts were already publicly accessible, to which we decided it was not necessary to negotiate access (Bassett and O’Riordan 2002). This decision was also made with regard to the nature of the two sites – BLOG and VSHARE – that were relatively mainstream online media which practised self-censorship under regulatory authorities’ guidelines. This means the content on the Group’s official accounts on these two sites would not be sensitive enough to necessitate an access negotiation in our judgement. On the other hand, the Group’s sites on CHAT and QWEB were semi-closed, which we considered required gatekeeper negotiation, and the registration/joining processes made it obvious that at least one gatekeeper was behind the Group’s BLOG account.

The lead author’s own circumstances as an established member of the Group facilitated the negotiation of access with gatekeepers but did not remove its necessity: group members needed to be made aware of the research and give permission for the research to be conducted on these sites. It was before the online survey was to be launched that the lead author initiated formal gatekeeper negotiation with a group manager. The online survey marked the beginning of the researcher’s direct interactions with participants and was considered the right moment to start the negotiation. Before the access negotiation, the research was ‘covert’ (Davidson 2006, 48) and based on observation on the Group’s public BLOG and VSHARE accounts, which we considered was a form of digital archiving of only publicly available information without the researcher interfering or the participants knowing. After access to the CHAT groups and QWEB was granted, the fieldwork became overt and the observation/archiving extended to these two semi-closed sites of enquiry, where relatively less-public online information was archived with the participants knowing, in addition to the more personal data collections through the survey and interviews.

It should also be noted that there were informal conversations with the manager about the research beforehand. The manager was aware of the start of the lead author’s interest in research and it was at that time that the researcher talked with the manager about the intention to gather information from other Group members. Since then, the manager expressed unquestionable interest in and support for the research. However, they also suggested changes to the research, and discussed gatekeeping matters with other managers. Although the formal request of a ‘go-ahead’ was eventually made in a rather straightforward manner, it was mounted through a series of informal negotiations with the gatekeeper(s) beforehand. Moreover, different levels of (in)formal negotiations were necessary because there might be cohorts of population managed under different gatekeepers within a single organisation (Mulhall 2003). In the Community in question, there were CHAT groups branched out from the main CHAT groups, which were managed by different people or gatekeepers. Some of these branch groups have autonomy from the Group, which we considered was part of the Group but required separate gatekeeper negotiation. This is also why gatekeeper negotiation is a constant process that starts with access to sites but has to be maintained throughout the research (Duke 2002). For this ethical issue of negotiating access to a site of enquiry, we argue that the necessity of gatekeeper negotiation may depend on the statuses of the site (closed or public) and should be reviewed at different phases of an online study (human subject exempt or not). Indeed, even after formal gatekeeper negotiation, permission to collect input from individual participants may also need to be obtained from individuals, which relates to the next ethical issue in our discussion: informed consent.

Participants’ informed consent

Informed consent involves informing potential participants of the research as much as possible and then obtaining their (written or oral) consent to have their outputs used for the purpose of the research (Boellstorff et al. 2012, 133). This is an ethical step to respect participants’ privacy and autonomy. According to most of the ethical guidelines we reviewed, it is necessary to obtain informed consent when observation-based research involves human subjects. However, when it comes to online observation, there has been no consensus on whether online observation is the observation of human subjects or collection of digital archives or documents. In addition, when certain online spaces assume some level of anonymisation and pseudonymisation, obtaining informed consent in the form of a participant signature or a selected button on a graphic user interface, as institutional ethical approval norms prescribe, may in fact collect participants’ personal information and bring chances of extra exposure to them. In this case, it is debatable whether informed consent offers more protection or harm, and whether it offers more protection to the participants or the researchers. The necessity and means of obtaining informed consent from individual participants again depend on specific situations.

In our fieldwork, three methods were used to collect participants’ outputs. Most outputs collected through the three methods could be associated with an online profile (pseudonymised or not). According to general guidance, informed consent should be obtained from participants whose outputs are to be cited or presented in the research (Buchanan, Markham, and Ess 2010). However, we argue that different measures are required online depending on the type of data involved.

Our data from the covert observation/archiving were already publicly available, a form of ‘publication’ or published information like books or newspapers (Bassett and O’Riordan 2002, 235). Having considered different views on these types of data, we decided that informed consent was not necessary for this particular case because the information collected was already open access on relatively mainstream online media (VSHARE and BLOG) where sensitive content would have already been excluded by the platforms themselves and free to be referenced. The data from the overt observation/archiving were gathered in access-restricted (registration required) online spaces where information was shared publicly within the space. We deemed it necessary to obtain informed consent for these data. However, the private spaces of the Group (CHAT groups and QWEB) were shared by a large number of people whose unstable participation in the spaces made it very difficult for the researcher to collect informed consent from them individually. Eventually, the researcher asked the manager to collectively inform the community members of the research with a semi-permanent announcement and gave them an opt-out option in which their data would be excluded from observation on request. As an extra ethical step, the overt observation did not collect any personal information, and the researcher took all possible measures to reduce the potential harm from the research done.

The data from the online surveys and interviews were elicited data uncovering people’s thoughts in a direct and personal manner. These data would not take shape without the researcher’s intervention. As such, it is clear that consent has to be obtained from participants. To protect participants’ privacy or at least to preserve informants’ online anonymity, informed consent has to be obtained carefully.

In the case of the online survey, the researcher used a paid/subscribed version of SurveyMonkey which offered contracted data protection, whereby the only party who had access to survey respondents’ digital footprints was SurveyMonkey and they were legally bound to protect such private data. In the beginning of the survey, a statement plainly describing the survey including researcher’s contact information was provided to inform potential respondents. They were asked to tick off statements, thus anonymously indicating whether they understood the information and gave their full consent for their responses to be used for the described research before they could start the survey. In addition to the anonymous consent, they were also reminded of not leaving any identifiable personal information in the blank spaces of the survey, which would be erased if found by the researcher.

In the case of the online interview, where a box-ticking response through the graphic user interface was not available, an informed consent form with a plain language statement was sent to potential participants in the form of a digital document through the CHAT file transfer. They were asked to sign the document with a designated pseudonym and erase their digital footprints on the file (erased by the researcher if found not erased already) before they sent the signed version back to the researcher. The signed pseudonyms were thus marked on both their informed consent and their interview transcription. These pseudonyms were different from their CHAT pseudonymised profile, which thus had to be linked in case that the researcher needed to go back to a specific informant for further information or if the participant wished to withdraw from the study before the research was completed. The links between pseudonyms were saved in an encrypted file at the researcher’s institution that was only accessible to the investigators. This pseudonym file was permanently deleted immediately after the research had been finalised.

For the matter of participants’ informed consent, we observe that the necessity of informed consent depends on what and how data are to be collected from participants. Online ethnographic observation can be based on digital archiving of publicly available information and informed consent may be unnecessary. If the observation collects information archived in a shared but private/access-restricted online space, it may be necessary to collect informed consent from participants but there may be certain logistic difficulties to obtain consent individually from each participant. In such a case, consent can be informed collectively but rejected individually. For surveys and interviews where data are elicited by the researcher in a personal manner, informed consent has to be obtained in a way that does not give informants unwanted exposure. We can already see that whether and how personal information is collected together with data play a decisive role in obtaining informed consent online because it aims to protect participants from potential harm caused by exposure, which is discussed in more detail below.

Data management

Before researchers collect or use the data from their ethnographic site of enquiry, an important ethical question to ask is whether such information can be used in the research. In other words, if used, would the information bring any potential harm or benefits to the information provider, to the researcher and their affiliates, or to any other individual? Furthermore, if the information is deemed to be essential to answer the research questions, are there any means to minimise the potential harm or maximise the potential benefits during and after the data collection? Internet researchers have suggested measures such as pseudonymisation, anonymisation, and fabrication – using metalanguage instead of actual examples when presenting data (Reid 1995; Frankel and Siang 1999; Markham 2012; van Hout 2014; Kozinets 2015). However, there is no agreement on when and how to take such measures because the risk or potential harm of each data collection step varies.

There are many reasons that a piece of data can be potentially harmful or beneficial if used. In our study, the focus was on an LGBT+ subtitler group. Issues of sexuality and identity are sensitive subjects, particularly in a country like China where sexual minorities are not represented in official and/or mainstream media and hold an ambiguous social status (Bao 2019). In addition, while our study did not examine the Group’s subtitled media or subtitling practices, the broader subtitling work of the Group is not always authorised by the media copyright holders and may have a legally ambiguous status in China. These two sensitivities make our participants vulnerable subjects, and similar vulnerability is now widely reported in online ethnography done in TIS, particularly in studies of online translation activities (e.g. Boéri 2012; Li 2015; Baker 2016; Pérez-González 2016; Yu 2017; Lu and Lu 2021). There is a risk that referencing vulnerable subjects or using their information may bring potential danger to them. However, academics should also ‘avoid any unintentional erasure of minority groups in research’ (Bassett and O’Riordan 2002, 244–245) which further excludes already marginalised social entities. One answer to this dilemma is to include such subjects without exposing their identities, anti-sanctioning tactics or any other sensitive information that we know would make them more vulnerable.³

In order to protect participants from potential harm or sanctioning brought by unwanted exposure, it is important that the researcher and their regulatory body rather than the participants make the decisions about the degree of anonymisation and referencing (Reid 1996). In the case of our study, all participants had online pseudonyms and avatars, and the data also included significant information stored online. Powerful search engines and IP hackers nowadays allow people to easily trace sources of online information with, for instance, a short excerpt from a post or an online pseudonym or avatar. We reasoned that, if participants were fully cited, the extra exposure would bring greater risks to them than they already faced in their everyday life. In that regard, we fully anonymised all participants, did not record their tactics of avoiding authorities’ censorship and sanction, and took significant efforts to prevent any tracing back to participants’ own online pseudonyms or real-life identities. Specifically: 1) we anonymised the Group and the Community throughout our research, 2) we ran the online survey in a manner that rendered the respondents non-identifiable, 3) we pseudonymised the Group’s sites and all interview participants, encrypted the pseudonyms, and deleted the encrypted file of pseudonyms as soon as the research was finalised, and 4) we did not observe their detailed anti-sanctioning tactics. As a further measure to protect the data from exposing our participants, all online archival data were ‘fabricated’ (Markham 2012). The lead author rendered (paraphrased and translated) archival data in such a way that did not change its original meaning but that the data could not be traced back to its source.

Based on our experience of data management, we maintain that the vulnerability of participants should be evaluated on a context-informed and case-specific basis. If participants are deemed vulnerable, the researcher should take measures of anonymisation and/or fabrication to avoid exposing participants to a degree that is beyond their daily exposure. We also suggest that these ethical decisions regarding data management should be preferably made by the researcher and their institution instead of by participants, given that participants may not have the best knowledge of the potential harm brought by the research to them. However, this should not stop the researcher from listening to participants’ own concerns. This decision would be more informed if there were mutual trust built between researcher and participants, pointing towards the importance of relationship management in the field.

Relationship management

From the above three issues, we can see that building a good relationship with participants is vital for progressing ethnographic research. Ethnography requires the constant management of relationships with participants to gain their trust, to share their experiences and understandings, and to negotiate the insider-outsider tensions and power balances of research conducted in an ethnographic frame. Scholars in TIS who conduct ethnographies may even be at an advantage in their ability to manage participant relationships effectively. In conventional anthropology, ethnographers have to learn to speak local people’s languages and adapt to their cultures so they can better communicate with the participants. TIS scholars, on the other hand, are very often translators who speak the translator communities’ languages and are familiar with important cultural practices that will help facilitate good relations. But what of the ethics of these relationships?

Managing participant relationships ethically can be based on a number of ethical principles discussed by Murphy and Dingwall (2007): avoiding bringing them harm (non-maleficence), making our engagement with them beneficial (beneficence), and respecting their rights (autonomy and justice). In practice, it is conventionally suggested that an ethnographer should use no deception, avoid engaging in sexual or romantic relationships with participants, and take leave after the research is finished (Boellstorff et al. 2012). However, these three pieces of conventional ethical guidance may require more nuanced understanding in an online context.

No deception can have two levels of meaning. First, the ethnographer in a conventional offline setting is guided not to conduct observation unannounced like ‘a fly on the wall’ (Boellstorff et al. 2012, 142). However, in online settings, ethnographic observation may only involve archiving publicly available online information that do not contain sensitive content. Such observation may not need to be announced if it does not reveal participants’ identities or personal information and if it does not project harm on those being researched (Walther 2002, 212–213). In the case of our study, as explained above, we did not announce the observation in public sites while we did clearly announce the observation in access-restricted domains, the online survey and interviews. Second, no deception can also refer to an ethnographer not assuming a false identity (Boellstorff et al. 2012, 142). In an online environment, the question of identity itself may be more nuanced: identities on the internet, especially in communities of socially vulnerable individuals, may already be fabricated and forged to a certain extent for good reasons. How can a researcher conducting ethnography online balance this reality with conventional guidance that ethnographic research may be compromised if the researchers do not introduce their true selves to the field (see Boellstorff et al. 2012, 143)? In our research, for example, the participants were first slightly frightened by the fact that the researchers used real names and contacts in the informed consent forms. In such a way, following conventional guidance towards no deception as we did may even mean that the researcher’s presence changes the virtual dynamic of the online community and that the online researcher is at a more vulnerable position than their participants, who may be able to retain their pseudonymised identities. While this could not be avoided, it was reflected upon when we interpreted the data.

Online ethnographies are often conducted remotely, and this mutual physical distance may make it less likely for ethnographers to engage in unethical romantic or sexual relationships with participants. However, this should not be taken for granted. Given the online dating is becoming increasingly normal, particularly for sexual minorities, the intimate relationship built between an ethnographer and their participants could potentially elevate into something romantic even in an online setting. Levels of intimacy can be ambiguous and unpredictable (Boellstorff et al. 2012, 144–145) and require careful monitoring to ensure that they are kept at appropriate levels. In the case of our study, the use of reflective journaling proved an effective tool to monitor the trajectory of participant relationships and in fact all the above decision-making. Sharing these reflections among all researchers on the team helped to ensure the maintenance of appropriate relationships between those conducting the research and those being researched at all times.

Finally, ethical guidance to take leave when the research is finished may be complicated for a number of online ethnographies in TIS, especially in studies of online translation activities such as subtitling. As was the case in our study, one or more of the researchers involved may already be an established member of the community under examination and may not want to give up their membership after the completion of the research. This membership may have inspired their initial research idea, facilitated their access to research data, or even become an integral part of their lifestyle. This was true of our study. Rather than leaving the community after the research is completed, the researcher informed participants of the ending of the fieldwork and clearly flagged the change in their status and position within the Group.

Conclusion

This paper was motivated by the fact that guidance and academic arguments in the area of ethnography online and offline are sometimes conflicting, and it can be difficult at times to apply conventional ethical guidance to some of the online communities that are increasingly popular objects of enquiry in TIS. We used our own experiences of conducting an in-depth ethnographic study of an online Chinese LGBT+ subtitling group as the basis for our reflections on the ethical challenges that such a study presented and that might be faced by researchers with similar research interests. By focusing on our conduct of ethnographic observation (digital archiving), a survey, and interviews online, we identified four ethical issues involved in conducting online ethnographic fieldwork in an online translator community that required particular attention and effort: access to a site of enquiry, participants’ informed consent, data management, and relationship management.

We found that ethically accessing a site of enquiry online required some thought because the threshold of public and private can be important to define and, where access to a site is automated, it may be challenging to identify gatekeepers with whom to negotiate access. Furthermore, we found that the method of access will depend on the level of covertness in the design. Avoiding identity deception may be difficult to achieve in online communities based on already (partially) fabricated identities. We encourage researchers interested in conducting online ethnographies of translation communities to be clear from the outset and be able to justify any aspects of the research that they intend to keep covert and to prepare to engage in a (possibly time-consuming) period of informal and formal gatekeeper negotiation for any aspects of the research process that will be overt.

We also found that conducting ethnography online rather than offline makes consideration of participants’ informed consent a unique ethical issue. This is because what we observe online are often people’s outputs (e.g. websites and posts) as publicly available information and may not require consent to be used. This is also because obtaining informed consent in the form of a participant signature or a selected button on a graphic user interface, as institutional ethical approval norms prescribe, may in fact collect participants’ personal information and bring chances of extra exposure to them. Overall, we suggest that a researcher intending to seek informed consent from online participants should focus on whether the participants’ personal information will be recorded or not and whether quoting this data will bring extra risk or harm to the participants than they already face.

Furthermore, our reflections indicated to us that the management of sensitive participant data elicited online may require more than the conventional steps of anonymisation, pseudonymisation, and restricted referencing that are already well-established in academic practice. Powerful search engines and IP hackers nowadays allow people to easily trace sources of online information with, for instance, a short excerpt from a post or an online pseudonym or avatar. Identifying digital traces can easily be left behind by participants in documents that they share, sometimes without their knowledge. Further to standard guidance on consideration of steps to anonymise, pseudonymise or selectively quote participants, we encourage researchers conducting ethnographies online, especially of vulnerable communities or sensitive topics, to consider rendering (paraphrasing and translating) data in such a way that does not change its original meaning but would not allow it to be traced back to its source. We also encourage use of data management protocols to check any documents shared with the researcher for any potentially revealing metadata.

Finally, the reflections made by us in this paper suggested that conventional ethical guidance on the management of relationships between researchers and participants may need a more nuanced reading for online ethnographic settings. In particular, announcing the beginning and the end of the fieldwork period, signalling the concomitant relationship shifts, and even use of real identities online may generate tensions and reactions from participants online that would not be expected offline. We recommend that prospective ethnographic researchers online make extensive use of reflective journaling from the beginning of a research project – even at the proposal or design stage – to help them be aware of their position (which may change as the research progresses) and be conscious of how this position may affect their relationships with participants.

There are certain ethical issues that we have reflected on that we assert are particular to ethnographies conducted in online translator communities. Firstly, we argue that the use of a community’s translations as data pose specific ethical challenges, even though we did not use such data in our research. The translations produced by an online translator community could provide valuable ethnographic insights. However, these translations might be unauthorised, infringe copyright or be of dubious legal status, posing an ethical dilemma that needs to be addressed. We suggest that ethnographers with such concerns seek guidance from their local institutions. Secondly, we argue that TIS scholars conducting ethnography in online translator communities may be better positioned to address some ethical issues than ethnographers from other fields. TIS scholars as translators themselves may already be well versed in a community’s local languages and cultures. This background may give them advantages in not only gatekeeper negotiation and field relationship management, but also in understanding and guarding the nuances of documents and data translated between the languages used in their institution or publication and the languages of the communities researched.

Overall, the issues raised in this paper required nuanced reading and creative application of existing ethical guidelines to be dealt with as part of both ethical conduct and institutional approval. In conclusion, we want to emphasise that the ethical measures we took were based on our decisions made in our particular research (context), which means they are not absolute solutions to these problems in all cases. We hope our experience and reflections could be of help to other colleagues in and beyond TIS who share similar research interests.

Disclosure statement

No potential conflict of interest was reported by the authors.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This research was funded by the DCU Faculty of Humanities and Social Sciences PhD Scholarship Programme.

Notes on contributors

Boyi Huang

Boyi Huang is currently a PhD candidate of Translation Studies at the School of Applied Language and Intercultural Studies. He is a member of the Centre for Translation and Textual Studies (https://ctts.ie/) and a member of the Indirect Translation Network (http://www.indirectrans.com/). Boyi’s current research project is about an online Chinese LGBT+ subtitling group. Generally, his research interests include Audiovisual Translation, Citizen Media, Ethnography, and Queer Activism.

Patrick Cadwell

Patrick Cadwell Patrick (Pat) Cadwell is an assistant professor of Translation Studies at the School of Applied Language and Intercultural Studies. He is a member of the Centre for Translation and Textual Studies (https://ctts.ie/). Pat’s current research interests centre on communities of practice and the impact of translation and interpreting on experiences of crisis and disaster. He is a member of INTERACT: the International Network on Crisis Translation (https://sites.google.com/view/crisistranslation/home) and of the now completed COST Action IS1201 on Disaster Bioethics (https://sites.google.com/view/crisistranslation/home). His current PhD supervision covers research into translation carried out by communities of practice and largely non-professional translators in Japan, Vietnam, and China. These PhD projects are being carried out within an ethnographic framework and combine various digital and in-depth qualitative approaches.

Ryoko Sasamoto

Ryoko Sasamoto, BA, MPhil, PhD, is Associate Professor in the School of Applied Language and Intercultural Studies (SALIS), and a member of Centre for Translation and Textual Studies (CTTS). She has a background in Linguistics and her research expertise is in the interdisciplinary area, working across different disciplines such as Pragmatics, Audiovisual Translation, and Reception Studies. She takes both theoretical and empirical approaches in her research and uses eyetracking and corpus analysis from the perspective of Relevance Theory. She is particularly interested in the communication of emotions, as well as pragmatics of AVT with a focus on telop and fansubbing. She has published extensively on onomatopoeia, telop, and other multimodal communicative phenomena.

Notes

1. AoIR: https://aoir.org/reports/ethics3.pdf. BERA: https://www.bera.ac.uk/publication/ethical-issues-in-online-research. CASRO: https://www.insightsassociation.org/issues-policies/casro-code-standards-and-ethics. BPS: https://www.bps.org.uk/sites/www.bps.org.uk/files/Policy/Policy%20-%20Files/Ethics%20Guidelines%20for%20Internet-mediated%20Research.pdf. BSA: https://www.britsoc.co.uk/about/latest-news/2017/april/ethics-of-digital-research/. MRA: https://www.mrs.org.uk/standards/code-of-conduct.

2. These are pseudonyms of the media platforms that the Group is active on. They are pseudonymised here in an effort to protect the Group’s identity and avoid unwanted exposure.

3. It is not always easy to decide what is sensitive information for an individual and what will cause harm in ethnographic research. In this case, the researcher reflected on their positions as both a researcher and a participant to gauge what could be sensitive for the participants and what potential harm there might be from the research to the participants, and took all possible measures to reduce the potential harm.