Societies everywhere run on social networks. Even in the Netherlands or in the United States, connections make social science research much easier. But in North West Europe and in Anglo-Saxon societies, networks may be a little less central than in most other societies, and it is sometimes possible, for instance, to get an elite interview based purely on a professionally worded e-mail and some follow-up phone calls. In non-western societies, networks of trust are typically much more important to everyday interactions, and getting meaningful information from strangers is much less likely. In authoritarian contexts, where the cost of having an even slightly politically loaded conversation can be high, trust and willingness to help come from who recommends you, and from how much effort you invest in assuring your respondent of your trustworthiness and respect for them.

In this chapter, we discuss how we have built our networks in the field and reflect on the centrality of personal connections and the currency of trust. We consider the relations we have with local collaborators, the responsibility we have towards them, and the (sometimes mutual) considerations of risk in such relations. Then, we hone in on one of the most important relations in authoritarian fieldwork: relations with potential interviewees, or as we will consistently call them in this chapter, respondents. We will discuss why interviews are often not refused outright, but evaded, in authoritarian contexts. We describe the ways in which we approach potential respondents to try and maximize our chances of building trust and discuss our experiences with encountering—and sometimes overcoming—three frequent obstacles to trust in authoritarian contexts: suspicion, ideological hostility, and fear. Our ascriptive characteristics (i.e. nationality, ethnic background, age, and gender) partly determine how we are perceived by respondents, but we describe how we ‘work with what we have’, presenting a version of ourselves that helps us build trust and get information. We explain how we navigate interviews with officials and activists, respectively, and reflect on a few cases where we have been subject to manipulation by local contacts and respondents. Finally, we consider the debt we owe collaborators and respondents in the field, and the limited ways in which we can do something in return.

Building Connections

We all come by our respondents via the ‘snowball method’, but a snowball is really the wrong metaphor, since it is the warmth of the connections that matters. People meet you, not necessarily because they are very interested in your research, but more because they trust you or the friends who introduced you to them. How warm these connections need to be depends on the kind of respondents, the topic, and the authoritarian context. In the least repressive context we work in, our Malaysia researcher found it relatively easy to meet activists who were public figures. The main concern was getting them to make time in their busy schedules. Talking to ordinary people who had occasionally attended demonstrations about their experiences was even easier. But even here, speaking to people who sympathized with the protests, but were for various reasons fearful of attending them, did really require personal introductions from their friends.

Practically, getting an introduction may just mean getting a phone number, but symbolically, it means much more. A certain capital of trust comes with the introduction. They take responsibility for us, providing a kind of safety check on behalf of their contact. In the experience of our China and our Kazakhstan researcher, a conversation with a potential respondent will almost always start with them getting a sense of the warmth of our connection with the liaison person: how do you know this person? How did you meet? For how long have you known each other? They may also try to discover whether anyone else in their circle is a friend of ours. According to a colleague of North African descent, someone with local roots may need even more time before the interview can start. Much more than in the case of a foreign researcher, the respondent needs to make a detailed assessment who is in front of him or her, and to what networks they belong. Are they linked to anyone in the regime or the opposition? These things will influence the extent to which the respondent will trust and help us.

Sometimes, the centrality of personal recommendations makes a respondent impatient with our habit of explaining our project and getting informed consent: a trusted person has vouched for us, the respondent feels, so let’s skip the formalities and jump to the questions you have. They may express it in so many words: so and so is a friend, so tell me what I can do for you. Occasionally, an institutional contact may be a functional equivalent of a personal connection. Our Mexico researcher got various journalist contacts via a human rights organization, and found its name to be a magic word among potential respondents: it was the one organization they completely trusted. As always, there are exceptions to our rules, also to our ‘networks are everything’ axiom. One of our colleagues working in the Arab world has great experiences with going through secretaries or personal assistants to make appointments with officials. More than their bosses, they may be under the impression that we are ‘important’ because we come from abroad, and if they have a return favor to ask, it may be something small related to our knowledge of the western world.

In circumstances where being introduced by someone really means coming with their stamp of approval, contacts may not always be willing or able to make the introductions we desire. Our Iran researcher has experienced that people mention names of potential interview partners, but they may say ‘maybe it is not so helpful that I introduce you because there are differences, we have a different political orientation’. Or they will just say ‘it would be good for you to talk to this state secretary and I can give you his mobile number, but if you call him just like that it will not help you to get an interview’. Our China researcher has also come across such reluctance. Sometimes she is given contact details, but without the all-important personal introduction, sometimes they introduce her to another connection who could link her to the interviewee, and sometimes even after a potential interviewee has agreed to talk, they behave reluctantly: they reply to SMS very slowly or not at all, or remain evasive about a concrete meeting. It is not always possible to glean why our existing contacts are reluctant or unable to make forward introductions. It may just be that they are cultivating their own social capital carefully, and see a potential risk without advantage in making the connection for us, or it may be because the contact is not warm enough to successfully establish the connection. Or, as was most likely the case in Iran a few years ago, it is a sign of a contracting political climate and shifting alliances. Finally, there may be a gender dimension to a reluctance: some men may be considered inappropriate for a female researcher to meet, or vice versa.

When a local contact does make a connection for us, we may not fully know what goes on between them and the respondent. It may be that we unknowingly become an asset for the contact, who is trying to impress someone with their international network. Or the other way around, the respondent may extract something from the contact for having been exposed to us. It may not always be necessary to fully understand what exchange is going on, but we should at least be aware of the possibility that either we or our contact are being put in an awkward position of owing someone a favor, especially if the interview does not in fact go well.

There is a tension, in authoritarian contexts, between the need to rely on local contacts and snowball sampling and the need to be discrete about who our interviewees are (see also Ahram and Goode 2016, 843). When we rely on local contacts and respondents to provide us with further contacts, it is quite natural for them to be asking who else we are speaking to or have interviewed. In some contexts, giving this information may be entirely innocuous, but in other cases who we speak to may actually be sensitive information that might get respondents into trouble. It may not always be possible or appropriate to be entirely silent about interviewees while soliciting further contacts, but discretion can be exercised, especially about names we think might be sensitive, unless our local contacts are trusted friends. In general, a policy of openness should not extend to telling respondents details about each other.

Local Collaborators

Sometimes, ontacts may do much more for us than just act as go-betweens. Both our Morocco and our Iran researcher nowadays rely on a few people they know well, who will not only share contact details but help assess who will be useful to talk to, whether they are easy or difficult to approach, and whether approaching them might have consequences for the rest of our research, making it better to meet them towards the end of their stay. They also give their advice on security issues, on wording, or on how to interpret some initial impressions.

Some of these people come to work with us on a more sustained basis, because they are interested in learning social science methods, because they want to put their work for us on their CV, because they are simply curious or willing to help, or because they have become friends. Others are paid by us, as research assistants, translators, survey experts, or local supervisors. Mostly, the relationship is beneficial to both sides. Our Kazakhstan and Malaysia researchers have both had great experiences working with local survey companies, the staff of which were very skilled and professional as well as genuinely interested in the research projects in question. The Kazakhstan researcher also worked successfully with paid research assistants. Apart from simply saving time, she finds that they can help with wording things appropriately (see Chap. 3) and have their own networks and knowledge that facilitates the accessing of information. Apart from payment, she has helped them by discussing and commenting on their PhD theses or job applications. The seniority relation was the other way around when our India researcher did his first fieldwork: in accordance with the customary procedure in his master’s program, he had a paid local supervisor. This payment allowed him to make use of the local academic’s contacts, and use him as a sounding board, without feeling that the relationship was exploitative (see Carapico 2014, 27–28 on the sometimes shameless imposition by foreign scholars on local researchers). Since then, he has had occasional help with translating, both in India and Mexico, from people his own age who offered their help, and for whom he has paid travel expenses but no fees. He has been reluctant to pay these people, whom he considers as friends and stays in touch with. In China, students typically get paid for research assistance to foreign scholars, but also see it as an important opportunity for research training and experience. Our China researcher has used an assistant once, to find and interview a few people in Cantonese. Payment would have been considered offensive since this was a friend of a friend who volunteered to help, but our researcher wrote her a reference letter for study abroad in return.

The relationship with local collaborators can raise ethical concerns, however, when local collaborators are either inexperienced, or dependent on us, or both. Inexperienced collaborators, especially people with no social science or advocacy background, may not be fully aware of the possible risks connected with the collaboration. Even collaborators who are aware of the risks might agree to do more than it is safe for them because they feel a sense of obligation, either in the name of friendship and collaboration, or because they are being paid. Our Kazakhstan researcher felt greatly responsible, for instance, for volunteers who undertook a door-to-door survey for her and a colleague, something that is not very usual in Kazakhstan, and can make the police nervous. They took all the measures they could think of to make sure the volunteers were safe, gave them security training, passed on advice from professional survey-takers, and provided them with letters on university letterhead, stating that the researchers took sole responsibility for the survey in case they were stopped by the police, and so on. The survey was taken without incident.

Our general view is that we bear responsibility for research tasks undertaken for us, whether paid or not. While relying to some extent on our local collaborators’ judgment, we also need to make our own assessment of whether what we are asking is safe for them to undertake, being aware that their position is different from ours because they have to continue to live and work in the country, while we can leave. Most of us feel that we can undertake such risk assessment, and working with local collaborators is a valuable way to save time, get the benefit of a local interpretation, and get access to additional material and networks.

While the greater concern should be whether our research causes risk to local collaborators, we should also be aware of the opposite possibility, that a local collaborator’s other activities and local status compromise our research. As Hilhorst et al. (2016, 19) write ‘researchers also need to consider the social position of their local colleagues, such as co-researchers, translators and drivers. If these people are controversial, the researcher will become controversial as a consequence’. Collaborators can get into trouble, not because of their research assistance to us but for other reasons. Our Morocco researcher recently paid a local collaborator she had known for a decade, for what was meant to be joint research. She did not hesitate about hiring him because he was a journalist known to be critical of the regime, and the work he would do for her was innocuous by comparison to his own publications. After showing her some initial work, he dropped out of contact for almost a year. She never discovered exactly why, but given the authoritarian context there is always the possibility that he may have received some sort of warning that caused him to cease all political activities. Having also had a complicated experience with undertaking dangerous research herself at the beginning of her career (see next chapter), she is now quite reluctant to work with research assistants, because of the risks on both sides of the relation, and would only consider the possibility if she were to start afresh in a new country with no contacts. Even those of us who do sometimes rely on research assistance still undertake the bulk of interviews ourselves. Below we describe how we go about our interviews in the authoritarian field.


Even with introductions, chasing respondents for appointments is probably the most arduous aspect of fieldwork, and failure is not unusual. It happens that a potential respondent simply says no. Our India researcher once had a former rebel leader initially agree to an interview, but then later sending a text message simply stating ‘I don’t think this interview is in my benefit’. And our Iran researcher once found that a journalist and media scholar categorically refused to speak to him about a specific topic, despite the best possible introductions, because he found the topic too sensitive. But such frank, point-blank refusals are extremely rare.

For cultural as well as political reasons, the much more common experience is that respondents will agree to meet in principle but invent a series of excuses to make it practically impossible. Our Kazakhstan researcher experienced this while trying to speak to Kazakhstani students in the United Kingdom. She introduced herself to local university Kazakh societies and met apparently willing respondents there, but then when she tried to follow up, they were sick, they were busy, they had class, until she had left town. This may have been because she lacked the right introductions, or because she did not have enough time to follow up the group meetings with more informal personal chats to gradually build trust, or possibly because being abroad actually made the students feel more under the radar of the government than they would have been at home.

Our Chinese researcher has had quite a few such experiences, especially with policy-makers: they are busy, not in town, not in the office. One eventually agreed, after being pursued for a month and a half, to speak by phone, but then went silent completely, no longer replying to SMS messages. Another, after many e-mail reminders, asked to see a list of questions in advance, and then never responded. In some cases, this may be to do with the sensitivity of the topic. But in certain contexts, in our experience especially in China and Iran, meeting someone from a foreign university may in itself already be considered sensitive, whatever the subject, especially for officials. Such meetings may not pose an immediate risk to them, but they are on a career ladder, and meeting researchers may be held against them at some point. They are just trying to avoid doing anything that could be construed as wrong.

Testing the Waters

When meeting our respondents, we do not get straight to the point. Instead, we invest time in building the relationship, discussing health, families, traffic, and mutual acquaintances if possible. Sometimes this just takes the first few minutes of a meeting, especially when people are busy and used to being interviewed; at other times the entire first meeting is devoted to testing each other out. Sometimes, interviewees will not engage in a first meeting without the presence of the people who introduced us to each other, and only afterwards make an appointment for the second time. Thomson et al. (2013, 6), who work in the Great Lakes region of Africa, explain why so much effort needs to be invested in building trust: ‘(a)s researchers we cannot expect people to respond to us with openness, nor expect that they will tell us their real opinions and experiences when they have just met us. This is equally true for someone in a high-ranking government or rebel position as it is for someone in a remote rural area or someone meeting you in the centre of town. Why would anyone divulge sensitive information, that if known beyond the confines of your interview could get them into trouble with neighbours and local authorities alike?’

Knowing the local language is usually not an absolute necessity, but it can be an important asset, since most people are simply more comfortable in their mother tongue. Our Iran researcher believes that respondents are more open with him speaking Farsi, in part because they appreciate his knowledge of the language and in part because it makes interviews feel less official. But for others, the choice of language can have complicated political connotations, especially for researchers of local or dual heritage. Like our own Morocco researcher, a colleague of Moroccan descent who has grown up in Europe conducts interviews in French or English. This is not because he has difficulty speaking the local dialect of Arabic, but because his accent would lead a respondent to make immediate assumptions regarding his family’s regional background and social class. His fluent English and French and prestigious academic affiliations can offset this, but in turn cause him to be seen as more of an outsider. By contrast, a colleague of Turkish descent who has spent most of his adult life in Europe finds that his accent gives him easy access to elite circles, but can cause him to be distrusted by ordinary Turks, and even more, Kurds. He prefers to use English in his engagement with Kurds in Europe.

Even when we get started with the real interview, there is always an initial part where you try to measure each other. As researchers, we try to assess how much we can push, not starting right out with the big questions. If we feel that there is a difficulty with a particular question, we may skip it and return to it later. Not all of this is about political sensitivities: not everyone is used to giving concise, to-the-point answers even to straightforward questions, and multiple approaches may be needed to get a question understood and answered. But when sensitive topics are being broached, it is all the more important to engage in a careful ritual dance. Showing up like a newshound, pen in hand and firing off questions will alienate even well-disposed respondents.

We have all experienced situations in which establishing trust was unusually difficult. We will illustrate three of the four most common reasons, as we have encountered them: suspicion, ideological hostility, fear, and personality. An example of suspicion was experienced recently by our China researcher, who was exposed to a kind of reverse interview by a policy-maker. For half an hour, she was asked directly and indirectly what she wanted, why she does this kind of research, why foreigners are interested in the details of Chinese policy-making, why our project got a big grant, what the EU’s interest in providing such grants might be, and so on. Testing her patience, she must have given the desired answers, for she eventually got her interview.

In the survey of Middle East scholars undertaken by Clark (2006, 418), ‘(t)wenty-seven percent of researchers specifically identified anti-westernism (usually in the form of anti-Americanism) and the general suspicion and distrust of U.S. policies and perceived agendas as impeding their efforts to undertake field research’ (see also Carapico 2014, 27; Jourdan 2013). We have not been as pervasively exposed to ideological hostility to westerners as an obstacle in field research as Clark reports, perhaps because we are not US nationals, and have mostly worked in European universities, but we have also come across it. Our Morocco researcher got used to always having to sit through a lecture on how Islam is not against women before being able to ask Islamist political actors the questions she was actually interested in. She once got a more than usually hostile reaction from a female Islamist activist, who attributed a question she did not like to the researcher’s ‘Judeo-Christian culture’, and ended the meeting soon afterward, despite the researcher’s protestations that her interest in the reform of Morocco’s Family Law had nothing to do with her personal background. Our Kazakhstan researcher experienced ideological hostility in a spontaneous encounter with young party officials. After being briefly introduced by her contact, an intern, she was left to introduce herself. As soon as she started explaining that she was a PhD student in political science from Italy, a young man started talking about western scholars who go around criticizing other countries while neglecting the study of their political problems at home. He kept standing and declaiming these ‘truths’ in a loud voice and accusatory tone. While it was an unpleasant experience, and it made any actual interviews impossible, the incident could be considered as ‘relevant data’ in her research on the party’s role in the authoritarian governance of Kazakhstan. Other encounters she and local researchers had with young party cadres confirmed that this aggressive way of arguing is something at least some youth branch leaders use against pro-democracy organizations or activists, to delegitimize their claims as something foreign, alien, and potentially destabilizing harmony in society.

Fear of reprisals nearly caused an intriguing interview for our Mexico researcher to be aborted: he had an initial chat via a mutual friend with someone who had an incredible story of state corruption—the building she had worked in was partly destroyed supposedly due to a gas explosion, but as an engineer who worked in the building she knew there to be no gas pipes in or under the building, and suspected self-sabotage. The respondent’s indignation may have sparked her to initially tell the story, but she went on to cancel the planned second meeting in which she would discuss the incident in more detail, because she feared repercussions. Eventually, on a second fieldwork trip, and after another informal meeting, she opened up and told the full story. Fear was the main obstacle to getting interviews with civil society organizations in Burma under the military junta, undertaken by Matelski. She found that organizations that had no regular contact with westerners were reluctant to meet her, and even those that did ‘remained reluctant to share information’, as ‘it could raise suspicion of secret dealings with journalists or activists’ (Matelski 2014, 68–69). When we come across fear, we may try to put a potential respondent at ease, but it behooves us not to push too hard but to respect their judgment as to whether meeting us may pose risks for them.

Our experiences of respondents’ suspicion, ideological hostility and fear all relate directly to the authoritarian context. But respondents in authoritarian circumstances are also just people, who can be stand-offish, overbearing, offensive, or dishonest for no particular reason. Sometimes, we just do not succeed in establishing a connection at a human level, and our interview experience remains frustrating. And sometimes, the ‘click’ occurs precisely when we give up hope and stop trying to fish for information. Our China researcher pursued a well-connected local scholar for over a month, only to be harangued for fifteen minutes about how ‘out of date' and 'pointless' her research was. Just when she was ready to give up and leave, the scholar unexpectedly said that she could try to arrange a meeting with a significant insider. She followed through, and the subsequent interview turned out to be quite important. Our India researcher experienced something similar when interviewing a supporter of the Maoist movement in India: not getting substantive answers to questions, he let go and allowed the interview to turn into a free-floating conversation about historic revolutionary trajectories, to which he contributed his own reflections. Somehow, these ideas enthused the respondent, and he started sharing precisely the personal information on his own motivations that he had previously been holding back. These are just two examples of a common experience, that when we relax because the interview, whether frustrating or fruitful, is over, our respondent also relaxes and shares something with us that turns out to be more meaningful than anything we had heard from them before. Markowitz (2016, 904) actually recommends ending an interview with a ‘concluding ritual’, such as ‘putting a pen cap on the pen and putting the pen on my notebook’, precisely because it ‘proved to be remarkably useful in relaxing informants and often they began a “side point” or “one more thing”’. Formally ending an interview but then asking ‘just one minor question’ is also advised by Art, who refers to Inspector Columbo in the classic TV series as the master of this stratagem (Art 2016, 981).

Work with What You Have

We think of relations with respondents as a kind of role-play: we do not pretend to be anything other than ourselves, but we do present particular versions of ourselves that we think will help us build the relation, and establish the best possible connection with our respondents. There are widely different strategies for this, which have a little bit to do with one’s personality, but actually more with how we are seen by others on the basis of ascriptive categories like nationality, age, and gender. Below, we set out some opposites to illustrate how different characteristics propel us towards different roles—but these are stylized foils. In reality, our interviewer persona is not so entirely fixed: it may evolve over time, it partly depends on the type of respondent, and sometimes we may even change gear in the midst of an interview to get more traction.

Our researcher in India and Mexico, a young man doing research on forms of repression, typically tries to make the relation with respondents as informal as possible, often meeting people multiple times, looking for ways to break the ice, trying to make it ‘click’. This approach seems quite natural when meeting journalists or human rights activists, who often have values similar to our own. But he has also applied it with ‘agents of repression’: police or security officials. One of his most revealing interviews has been with a security agent in a bar, who was detailing how he harasses opposition politicians for a living, while drinking and using cocaine. For our Morocco researcher, a young woman operating in the Arab world, such an approach would be a recipe for disaster: diving into informality with a relative stranger, seeking repeat meetings, being in places where alcohol or drugs are consumed could all lead to misunderstandings. Instead, she stresses her professional persona, makes it clear that interviewing people is ‘work’, and while engaging in the necessary small talk as described above, she deflects questions about her private life. Our Kazakhstan researcher, operating in a context less marked by stereotypical views about the sexual morality of western women, steers an in-between course. She would not meet respondents at a bar or in the evening, but her approach is a little more informal. She uses her nationality, making assumptions about respondent impressions of Italian culture, and emphasizing similarities like the importance of hospitality, late dinners, and big weddings.

Akin to the informal/professional divide, but a little different, is the choice to present ourselves as well-informed or naïve. Naivety is a commonly used interview strategy (Goode 2011; Solinger 2006; Henrion-Douncy 2013), typically more available to young women and foreigners. Our China researcher, while not actually foreign, uses both elements, especially with older men in senior positions (whether in government, research institutions, or companies) who often possess both stereotypical views of young women and valuable information. When it comes to gender, the experiences of other China scholars we know include one that illustrates our sense that women are considered less threatening, and may sometimes have greater access to officials precisely in authoritarian circumstances: two foreign researchers both did research regarding the top leadership of the party (i.e. the most sensitive kind there is, see our previous chapter) at roughly the same time. Both were given permission to do fieldwork, but when the male academic approached government officials, many people refused to speak to him. When the female researcher did, it was all green lights.

Apart from her gender, our China researcher also employs her partial outsider status, playing up the fact that she has been abroad for very long, which makes it possible for her to ask some relatively more sensitive questions that foreigners can usually ask, but that would not be available to a scholar who has remained in China. But she can switch from ‘naïve’ to ‘professional’, showing her familiarity with relevant details, when she feels that a respondent is spinning her a line. As we already illustrated in relation to the use of language, there is great variation in the positionality, and strategies available to, researchers with local or non-western roots (see also Malekzadeh 2016, 867). A colleague of Moroccan origin finds that he can only feign naivety to a certain extent. It would seem disingenuous to act as if he does not understand how things work in Morocco, and sometimes even when he is really baffled, locals still expect him to understand. Yom on the other hand, an American Middle East scholar of Korean descent, often found himself inexplicably cast as more of an outsider to the field than white western academics, and hence considered either more ignorant or more objective than they are (Yom 2014, 18).

In contrast to the naïve stance, showing that you have ‘done your homework’ has been the strategy of our Malaysia researcher when approaching well-known activists, who are typically very busy, sometimes a little self-important, and in regular demand for interviews with researchers and journalists. Display of expertise worked for him, for instance, with an activist who had been elusive and eventually told him to come and meet outside a bar where he was drinking with friends. The respondent initially took a rather abusive tone in front of his friends. Our researcher immediately launched into a detailed question about events twenty years ago, which eventually led to an in-depth one-to-one conversation and a follow-up at the respondent’s home. In line with her professional persona, our Morocco researcher, who interviews mainly officials, similarly prefers to set out her ‘case knowledge’ of a topic early on so as to be taken seriously. She also builds trust by showing that she knows what she is doing in procedural terms, for instance, discussing the status of the conversation as off-the-record, anonymous, and so on. But she has also experienced how, working as a pair with a researcher less steeped in local knowledge, they managed to have it both ways, using expertise and naivety in tandem. Her colleague, who had a more basic command of French, could at some point break into the interview, posing quasi-ingenuous questions that she herself did not feel able to ask, but the boldness of which could be attributed to her inability to express herself more subtly in French.

A final consideration is whether to present ourselves as an ‘important’ or a very junior person. Again, our choices are very much constrained by age, gender, and position, but each of us can work with what they have. A male middle-aged full professor may well find it easier to get an appointment with an official than a young female PhD student. But by being considered more consequential, he may also be more threatening and may have less fruitful interviews. While our Morocco researcher regularly uses her doctoral title to get in the door, our Iran researcher by contrast still sometimes introduces himself as ‘a student’ to diminish his importance, years after finishing his PhD.

Where to Meet

We generally like to leave it up to the respondent to suggest a venue for meeting, but again there is a considerable gender divide, almost regardless of the context. Our Iran, Malaysia, and India-Mexico researchers are all male and interview primarily activists and journalists. For them, cafés, bars, and shopping malls are obvious meeting places. All of them have also occasionally met respondents at their homes. The women among us by contrast will avoid meeting strangers at their homes, at a hotel, or late at night. Beyond the gender issue, meeting a respondent in a public place implies that neither they nor we are uncomfortable about being seen together. This fits with our general commitment to being open about what we do, and not behaving like spies. Both our China and our Iran researcher have experienced that whether one meets in a café or restaurant or in someone’s office may affect the flavor of an interview. Going to people’s offices implies that the interview becomes more official, they will consider you more as a professional, and they use more official language; at a dinner people will be more relaxed and talkative. Going out to dinner does mean being seen together, possibly also by government agents, but both in China and Iran, people who are concerned about being associated with us would probably refuse a meeting in any case. Dinners with our liaison or others present are in a way also less political. Reny (2016, 918) describes, also in the Chinese context, how she was once refused an interview by an official, and given the message that the contact person ‘should have invited him out to dinner’ and then have the researcher join, so the meeting could have been presented as spontaneous. Dinner meetings do pose challenges for note-taking or recording however: neither are very practical or appropriate while eating food, so the researcher must somehow rely on her memory and write it all up immediately afterward.

Triangulation, Not Confrontation

While we have very different approaches to how we present ourselves and where we meet, we have considerable consensus on what to avoid during an interview: confrontation. Confrontation is a frequently employed style of interviewing in journalism. We do not know to what extent academic researchers in democratic circumstances ever employ confrontation as an interviewing strategy, but we do know that for researchers of authoritarianism, it is not a helpful approach to improving our understanding of how things work. It is not our job to influence the views of our interviewees, let alone change them, but to try and establish how they see things, or at least how they choose to present them. Our approach is markedly different in this respect from the one recommended by Markowitz (2016, 905), who writes that in authoritarian circumstances, interviewers ‘need to be prepared to raise hard questions that bring tension into the room, to challenge informants when they are giving the official line and not their own viewpoint, and to identify discrepancies even to the point of calling him/her out on a false statement’.

There are two elements to our stance. First, while we are obviously critical in many ways of the political systems we study, we do not arrogate to ourselves the right to prescribe to the locals how they could be improved. The western societies we belong to or have chosen to live in are also flawed. Moreover, they may be directly or indirectly implicated, today and through a shared global history of exploitation and domination, in the authoritarian rule of the societies we study. The second reason is more pragmatic: confrontation is simply not productive to our line of work. It can lead to respondents shutting us out and spouting well-worn ideological positions rather than answering questions.

That is not to say that we swallow whole everything respondents tell us. Regardless who we speak to, we always assume it is their perspective, informed by their worldview, and often their interest, that we get. Our job is not to push back during the interview—at least not aggressively—but rather to place it in context based on triangulation with other sources. While there is great variation in how much interesting information respondents have, how much they want to share, and how well they can remember and tell us, there is no such thing as a bad interview. There is always something worthwhile about every encounter, even if it was not at all what we were looking for, and never makes it into our written work, it still adds to our overall understanding of things.

We are undoubtedly more at ease speaking to like-minded people, that is, people who in some way have a critical stance vis-à-vis the regimes they live under, than speaking to zealous or bureaucratic government officials. We quickly establish a friendly connection based on shared understandings, and they may be happy to encounter someone they can have an open conversation with. There is a risk here though, perhaps more than when we speak to officials, of bias. Spending a lot of time with people who may already be closer to our worldview than many of their compatriots, we are likely to be influenced by their ideas, views, and discourse (see also Ahram and Goode 2016, 843). At the same time, being more similar to us than others does not always make them the most interesting respondents. In fact, some of us have the experience that people who see themselves as intellectuals can be difficult interviewees, because they like to speak in abstract terms and share their entire worldview, when we are looking to hear factual details, or personal experiences. We have also found activist views to be sometimes marked by vanity and bravado, at other times by excessive cynicism about their role, or by a predilection for conspiracy theories (to which we find authoritarian settings generally conducive).

In interviews with officials, we approach people with respect, knowing that just as we are doing our job, they are doing theirs. Even if somebody is literally telling us lies, we will not openly contradict them. This would only humiliate and alienate them. Instead, we treat their version of the truth as interesting. To give an example, our Kazakhstan reporter interviewed the director of an NGO. She knew from other sources that his was not a ‘real’ NGO but an organization reliant on state funding, without significant grassroots support. Instead of challenging him about the absence of a popular constituency, she asked him more neutrally about the organization’s collaboration with state agencies in its activities, more gently exploring with him why and how he chose to present his organization as an NGO, what its actual status was, and getting answers.

Sensitive Information

Having made successful efforts to gain the trust of respondents, we will often be entrusted with sensitive information. While we take every possible care to keep our transcripts safe (see Chaps. 2 and 6), we cannot give an absolute guarantee that it will never fall into the wrong hands. In this context, we typically trust our respondents to make their own judgments about the information they entrust to us. We assume that they consider what they want to share with us, and that they are aware that there may be some risk in speaking to us, however minimal. We do this in the knowledge that their judgment is not infallible, but ours would not necessarily be better. Some will err on the side of caution. Indeed, our Morocco researcher once had a respondent spell it out in so many words when she asked him about the religious legitimacy of the King: ‘if you can assure me that you can protect me I will give you my answer, I will tell you what I think—but since you cannot I will not answer your question’. Others may be less forthright but make the same call.

Our general experience with officials is that they are very much aware of exactly what they can and cannot say, they are used to weighing their words, and they will rarely be reckless. The experience with activists is a little more varied. On the one hand, they move in a complicated political setting, and we know they think about this a lot. On the other hand, they would not be activists in an authoritarian setting if they were entirely risk averse. Sometimes they can be too nonchalant, or too much affected by a stranger’s interest in hearing their views and experiences. Our India and Mexico researcher has had the experience that a respondent instantly began sharing some quite inflammatory information without making much effort to properly understand who was interviewing him, for what purpose, and with what safeguards. In such a case, we might reconsider whether it is actually responsible to use the information we have been given, whereas when we feel that the respondent has carefully considered what they have told us on the record, we will have fewer qualms (see also Chap. 6 on how we write up and reference our sources).

Being Manipulated

As described above, we assume that respondents have a particular version of the truth that they want to share with us, perhaps even an agenda, and while it can be challenging to figure out how to value and triangulate different accounts, we consider this part and parcel of our craft. But occasionally, we have encountered more blatant forms of manipulation. It is easy to suggest in the abstract that one should not fall for such attempts, but in practice it can, for different reasons, be difficult to avoid being manipulated. We describe a few of our experiences and try to draw some lessons.

One such pitfall is the ‘resentful respondent’, who may have valuable information to share, but has his own reasons (resentment against the system in general, or against a particular person) to speak to us. On one level, such a person can give us access to information that we would otherwise not be able to get. However, precisely because such data cannot easily be triangulated, it is more than usually difficult to assess its veracity. A second risk is that, unless we can keep the contact entirely confidential, it may interfere with the willingness of others to trust and speak to us. Both our India and our Kazakhstan researcher have encountered a ‘resentful respondent’. Whereas the India researcher has cultivated the contact while being aware of the risk of manipulation, the Kazakhstan researcher decided to forego the opportunity. The difference lay not so much in their ability to judge the veracity of the respondent’s account, as in their research topic, and the associated opportunity cost: the Kazakhstan researcher feared it would interfere with the relationship she was building with other party cadres, whereas the India researcher was already focusing on regime critics rather than insiders, so had little to lose.

Our Iran researcher has experienced the opposite situation: being manipulated by a former dissident turned regime informer. He met this young journalist, who wrote about films and had reformist leanings, on his first visit to Iran, spent time with him regularly, and developed a friendship. On a return visit, the journalist told our researcher he had been arrested before and feared another arrest. He instructed our researcher that if he should drop out of contact, the researcher was to deliver a prewritten press release reporting the journalist’s renewed arrest to Reporters Without Borders and other advocacy organizations. When the journalist did in fact disappear, our Iran researcher, concerned for the fate of his friend, did indeed send the press release. It later transpired that this second arrest was fake. During the first arrest, the journalist had been put under pressure and had become an informant for the intelligence agency. The pretend second arrest was intended to dispel suspicion among the fellow government critics on whom he was informing. With hindsight, our researcher believes he should have discussed the situation with other contacts before sending out the press release. If he had done so, he might have picked up on rumors already circulating that this person was an informer, and he would have reconsidered sending the press release. But it was precisely the authoritarian context that caused him to be secretive and trust no one.

A third instance of manipulation, experienced again by our Kazakhstan researcher, involved her being invited to write (a very minor piece of) propaganda for the regime. At the time, Almaty was bidding to host the 2022 winter Olympics, a bid eventually won by Beijing. A warm contact, a young official she had known for many years, and had seen rise in various government positions, called with a small request. She was asked to write an English-language article (in fact she was sent a suggested text but told she was free to write whatever she liked) for a website about what a nice city Almaty was, so that the local media could in turn quote this foreigner’s piece supporting the bid. Despite the fact that the article did not need to be overtly political, and might have been published on an obscure website, she decided that writing it would compromise her integrity as a scholar. But there was a price to pay. She attempted to limit the offense to her contact by citing her team leader’s prohibition as the reason, but her refusal nonetheless did irreparable damage to her relationship with a very helpful contact.

Since every situation of manipulation is different, it is difficult to draw general lessons. The last of the three incidents sketched here is perhaps the easiest to adjudicate: what we do for a living is write, and one of the reasons our writings are valued by society is their independence. We do not necessarily take a ‘neutral’ stance, but straightforward advocacy either for or against the regime interferes fundamentally with our ability to form and disseminate autonomous views. For the other situations, an important suggestion is not to act precipitously, and to recognize that you do not have to be all alone in making these judgment calls. A decision taken after reflection and consultation both with trusted local contacts and academic colleagues at home (without unnecessarily divulging sensitive details) may be a better decision, even when we do not have as much information as we would like on which to base it.

Doing Things in Return

There are different reasons why respondents are willing to talk to us. At the simplest level, it is because we ask them, and most people like granting a polite request better than refusing it. Some respondents feel under an obligation to the people who introduced us. Many have an agenda of some sort, which may be their own or that of the organization they represent. Government officials may be flattered that their views and experiences are deemed relevant to scientific research, considering it a boon to their personal prestige to be interviewed by a foreign researcher. Activists are often just happy that there is one person more who knows what they are doing. Without the willingness of all sorts of people to speak to us, share their time, their views and their knowledge with us, we would not be able to do our work. The same is not true for them. We may occasionally be in a position to do a respondent a favor, but overall, we cannot be as crucially useful to them as they are to us. This puts us in their debt. We try to handle that debt as best we can.

This begins with recognizing the debt, showing appreciation, and signaling our willingness to do something in return. As Loyle (2016, 933) puts it, ‘acknowledging a basic responsibility to contribute in some small way goes far in demonstrating respect for the individuals who give of their own time for our research’. At the simplest level, after a good conversation, we can send a respondent a message to thank them. If we promise to share results with respondents, in the form of a published article, then we would normally need to keep that promise. But making such a promise and keeping it is not always possible, either simply because respondents may not be able to read English, or, more specific to the authoritarian context, it might be problematic for them to receive the text. One of us conducted interviews with activists in Cairo during the Morsi period, and intended to send them her findings, but by the time the research was published, the situation in Egypt had changed so dramatically that she did not think it safe to send them the report or subsequent articles.

We try to stay in touch with respondents after the interview but it does not always happen; we are busy and perhaps they are too. While we will acknowledge our sources, they often need to be anonymized, and sometimes their contribution to our understanding needs to be downplayed for their safety (see also Chap. 6), making our acknowledgment appear inadequate. Our Iran researcher, for instance, ended up not acknowledging his Iranian tutor in his PhD thesis, published during the most restrictive period after the Green Movement in 2012. Likewise, the research assistant and translator on the Cairo interviews mentioned above was never acknowledged by name.

When it actually comes to returning favors, there are often—very contextual—cultural as well as ethical constraints on what is appropriate. Our local academic contacts we can sometimes repay very appropriately, by offering to teach a class to their students. Even if we are wary about teaching political science topics because they might be sensitive, we might teach on our home country’s politics, and there is always methods teaching. Some of us may be able to teach quantitative methods, but most likely, we have expertise in qualitative methods, such as interviewing and document analysis, which is a topic under-taught at most western universities, let alone non-western ones. At other times, we may be helpful to respondents by helping them navigate forms or websites or write in English. But we also need to be clear about what is not possible: we can help respondents fill out applications for visa or scholarships, but we cannot help them get the visa or scholarship.

We often meet respondents in cafés or restaurants. From a research ethics perspective, we all consider it entirely appropriate that we should offer to pay the bill. We see it as a token of appreciation, not as any kind of remuneration. However, we have found that cultural norms are widely divergent on this matter, and unintended offense may be given if we get it wrong. In China, paying is completely acceptable: after all you have invited the person, however senior they might be. Our Morocco researcher also frequently pays for a meal, and in fact finds the offer of lunch or dinner to be a useful response to officials who say they are too busy to meet during office hours. But our Kazakhstan and Iranian researchers’ experience is that most respondents would be deeply offended by a paying researcher, taking it as a refusal of hospitality by a guest in their country. Students might be more willing to be treated, but even they might at least make a show of attempting to pick up the bill. Our Malaysia researcher found that, while he paid for drinks for ordinary young people as a matter of course, activists were adamant about paying for themselves for a different reason. Possibly because of government allegations regarding foreign funding, they had a heightened sensitivity towards accepting even the smallest thing that might be construed as a bribe. When it comes to small gifts, our practices also diverge. Most of us might bring presents for personal friends to our fieldwork, but not for people we meet for an interview; but in China, a bookmark or some tea, worth less than 5 euros, is quite acceptable.

None of us ever pay for interviews. We can imagine circumstances in which it might be appropriate to do so: with respondents who are particularly poor and vulnerable, to the point that recording their material hardship without doing anything to alleviate it would be unethical. This might make sense, for instance, when interviewing undocumented migrants, people in IDP camps, or slum dwellers. But our own research has not typically focused on the most marginalized. In the authoritarian field, we believe payment poses various problems. First of all, any significant sum would cause a power shift: the respondent goes from being a ‘creditor’, giving us their time and insights, to being in our debt. This might cause them to feel pressured to answer even when they are not comfortable to do so. Second, it can lead to ‘desirable’ answers, with respondents telling us what they think we want to hear, causing validity problems. Finally, in some contexts, payment would be considered suspicious by the authorities. This could pose risk to the respondents as well as delegitimize critical findings (see also Loyle 2016, 933, who lists the same ethical and validity problems with paying respondents in authoritarian circumstances).

More of a conundrum is posed by interviews with respondents from non-governmental organizations. Since they represent causes beyond their own self-interest, they can be quite forward about requesting a ‘voluntary’ donation in exchange for their time. The ethical implications of complying with such requests depend a great deal on the circumstances: it may be less problematic when we interview staff than when the organization acts as a gatekeeper vis-à-vis members or beneficiaries we would like to interview, and less problematic when solicited after an interview than before it. We have sporadically made such donations but emphasized that these were personal contributions to a cause we valued, not a form of payment, and indeed we have not expensed them. More often, we try to fend off such requests without giving offense, saying we will think it over.

Chapter Conclusion: Patience, Trust, and Recognition

Building relations with respondents for primarily interview-based research is not fundamentally different in authoritarian contexts than in politically open societies. In all cases, it requires social intelligence: fine-tuned antennae for the words and body language of people we interact with, and the ability to interpret and respond to these signals. It is just that, in authoritarian contexts, the need for these skills is further accentuated. Building trust is both more difficult and more crucial. Getting to the nub of what we really want to know requires more patience. As we have shown in this chapter, it is useful to reflect explicitly on our relation with our respondents, in both directions. This is even more so with local collaborators, where we must think through consequences of their possible dependence on us, especially in relation to risk. In interviewing, we have to find a balance between openness and presenting a persona that contributes to a constructive conversation. We also have to be accepting of the versions of the truth, some more plausible than others, that respondents share with us, and treat them as interesting and worthy of respect as well as in need of triangulation and critical analysis. While we are occasionally at risk of blatant manipulation, the more common situation is that we are indebted to our respondents, a debt that we acknowledge, but cannot fully repay.