Doing research in peoples’ homes: fieldwork,ethics and safety – on the practical challenges of researching and representing life on the margins

The challenges of interviewing vulnerable research participants’ in their homes

The health and personal circumstances of social research participants often prevent interviews from taking place in public venues, making university researcher guidance difficult to observe, particularly if people express a preference for interviews to take place in their own homes, where privacy and confidentiality can be maintained. Other common reasons for vulnerable people requesting that interviews take place in their homes, include: caring commitments to others; difficulty accessing transport and/or venues due to disability; and health conditions such as anxiety/phobia which prevent them from leaving their ‘safe’ home environment.

The additional knowledge and contextual information gained when interviewing people in their homes can prove valuable for research projects. Field visits allow researchers to better understand individual’s financial circumstances and place factors such as local amenities, access to services, and their social networks, which can be indicators of poverty and social isolation. Such opportunities for gathering additional information also present risks, by virtue of entering unfamiliar environments.

Unfamiliar environments can be risky sites for fieldwork. Researchers may feel ‘vulnerable’ when working in deprived communities ¹ and visibly standing out as ‘outsiders’. But the internal environment of participants’ homes also presents risks and has required researchers, including the author, to negotiate various challenges. Several interviews highlight a series of risks encountered by the author (referred to as ‘researcher’ herein). The following example illustrates just how little researchers can know about the disposition of their participants and their environments prior to the interview.

Kieron is alcohol dependent, but when his contact details were provided by his social landlord to participate in the research project, apart from his basic details, no further information was available. The researcher was therefore unaware of his dependency. The discussion during the initial phone call to arrange the interview failed to illuminate any information of cause for concern, which might have given the researcher better foresight and reason for more suitable preparation such as working in pairs for safety. After imparting essential information about the research project, a few informal comments were exchanged about taxi firms operating in his local area and the nearest bus/train station to his home. This conversation put him at ease quickly and built some rapport, which was demonstrated during the second telephone call to remind him of his interview the following day. Kieron helpfully provided directions and offered to put his untrained dog outside during the interview to put the researcher at ease. Again, there was no reason to suspect any problems. However, when the researcher arrived at his home early next morning, he had already been drinking heavily. While consuming alcohol he explained that he drank considerably every day, and his addiction was mentioned repeatedly, during the interview:

Unfortunately, we always have to budget in that there’s enough money for me to have at least four to six cans a day just to be able to take away the shakes in the morning

The poor condition of Kieron’s home was visible, which included dog urine on the floor and on the sofa. Bluntly, he instructed, ‘mind the piss on the floor and on the settee’ (data drawn from research diary extracts), and whilst his change in tone since the two telephone conversations was of concern, more worryingly, a very agitated Pit bull terrier could be seen barking and scratching her paws frantically outside the patio window to get inside to her pups, sleeping behind a flimsy makeshift barrier of cardboard.

Kieron was aware of the dog’s agitation but continued talking about the highly sensitive subject which had influenced his involvement in the research, and clearly caused him anger and frustration: cuts in benefits and getting into rent arrears. Although he voiced his main views before the interview commenced, his expectation was to share his experiences in an interview. So the possibility of not proceeding with the interview, in the researcher’s judgement, was likely to perturb him. Therefore the interview went ahead against the background of barking and snarling. Whilst the researcher lacked confidence in Kieron’s ability to control the large dog outside, had he given in to the dog’s apparent distress and let him in, his partner instilled some confidence in that she appeared stable and she clearly moderated Kieron’s behaviour. For this reason, she was invited to sit in on the interview and contribute, having sought permission from him first. In light of his inability to recollect information increasingly, as he continued to drink, he welcomed her presence. Luckily, she agreed, and immediately helped to calm him down. Consequently, the pace of the interview increased, and the researcher was able to finish and leave quickly and safely.

Upon reflection, the researcher could not have foreseen and therefore prepared for the situation that unfolded, but instead had to think of a solution immediately to reduce an immediate risk. Using an intermediary in a challenging situation, who appeared stable, acted as a calming influence, and helped diffuse anger and frustration, allayed researcher concerns. In weighing up the situation, terminating the interview might antagonise a disconcerted participant even more. Although guidance for researchers suggests that, ‘if the encounter feels unsafe, they should bring the encounter to an early conclusion’ (Ritchie et al., 2014: 105), this is not always feasible, particularly when the participant cannot be reasoned with due to their alcoholism and their strong motivations for doing the interview. In this case, the insistence on being interviewed does bring to attention the ethical issue of the individual’s right to participate in research and being listened to, but also highlights the tension between this right and ensuring researcher safety.

In a second example, Gordon had expressed an interest in the research and was contacted to arrange an interview. This was for longitudinal research over eighteen months, involving three face-to-face interviews and short telephone catch-ups in-between the interviews. When arranging the interview Gordon struggled to listen and communicate smoothly, but his responses were coherent and didn’t set off any alarm bells about his alcoholism. Initially, finding his home proved difficult, but eventually he came out and led the way to his house.

Diary extracts reveal that Gordon was living in grim conditions: rubbish was everywhere; empty alcohol bottles and cigarette butts littered his living space; and excrement was on the walls. He explained that he would shut his dog away because he perceived him to be a risk. The setting was unnerving, but his condition more so, as he began to answer the interview questions. Gordon was polite and cooperative, but he appeared uncomfortable; sweating and stammering, and struggling to understand some of the questions. The interview was cut short so as not to prolong his discomfort, as well as the researcher’s, who increasingly became concerned about his health, demeanour, and her own safety due to his quick change in mood, from calm to anger. After the interview he gladly received his incentive payment, explaining that this was his motivation for participating.

Gordon’s life could be described as ‘chaotic’, but this is even more reason to understand his vulnerability and the barriers that he faced as changes occurred, which directly affected him. Vital information was gleaned about his deteriorating personal circumstances in light of his addiction and welfare reforms. The research team learned a great deal about his changing financial situation, interactions with his landlord and benefits agency, and his support networks. Researcher safety was not compromised in doing this, and flexibility and persistence allowed his experiences to be heard. Nevertheless, this case illustrates the limitations of research guidance in failing to convey the complexities experienced in the field when researching vulnerable people. Ritchie et al.’s (2014: 105) advice for example, that researchers ‘ensure public transport is running’ upon completion of fieldwork is simplistic and assumes an ideal situation - far from Gordon’s case where the researcher found that he had been blacklisted by his local taxi firm which resulted in their refusal to pick anyone up from his address, including the researcher, who then had to negotiate with them for transport.

In a third case, practical arrangements were a cause for concern during the interview and were reflected upon afterwards in a diary entry. Brian was contacted through his landlord inviting him to participate in the research, which he agreed to. The research team had very limited information about Brian’s health or financial circumstances. Only when the researcher arrived at his bedsit did Brian’s vulnerability and dire financial situation become evident. Brian explained that he was being rehabilitated for alcohol dependency, and that he was suffering from cancer. He was frail and didn’t appear to present any risk. However, the room within which the interview took place was far from ideal from a safety perspective. It was dark, with only one small sofa, which was initially shared by both Brian and the researcher; and the sofa being so small that physical contact was impossible to avoid. So, the researcher moved onto the floor, explaining that it would be easier to layout and organise the research paperwork that way.

Whilst the researcher didn’t feel any threat, the practical issues were noteworthy, and this experience underlined how elements of health and safety policies and guidelines are at odds with real-life circumstances, and if most aspects were observed, it would prove near impossible to consult with those people at the very periphery of society. Brian’s home was bare, and he explained how he was struggling to make ends meet due to the welfare reforms. When he was interviewed, both the heating and lighting were off, and he sat in a bitterly cold room wearing layers of clothing. As darkness fell, it suddenly changed the state of balance that had been established during the interview. The researcher began to feel uneasy as the text on the topic guide was no longer visible, so she indicated that he should turn on the light:

The challenges of interviewing in unpractical settings and in conditions largely out of researchers’ control are highlighted by this case. For example, adequate lighting in participants’ homes, maintains a sense of safety, but cannot be guaranteed in an interview situation, particularly in cases where participants are living in severe poverty. Whilst every effort is made for interviews to take place during daylight hours, this is not always possible, especially during the shorter autumn and winter days. However, for safe research practice, spatial-temporal considerations are paramount as different spaces present different challenges, and more so, as day becomes night.

Clearly, guidance is provided for researchers to limit risk, but some of the practical advice, such as sitting facing the exit door, or not interviewing in a bedroom are not always realistic (SRA, 2006) Brian’s case raised more than just practical safety concerns, but ethical ones too about the degree of his vulnerability. These are discussed later in the article.

Unintentionally unearthing ‘sensitive’ data while exploring housing issues

Housing research projects delve into diverse subjects ranging from, sustaining tenancies, housing and worklessness, poverty, and so on. All are subjects which have the potential to cause distress. Potential issues are identified beforehand by the research teams and discussed in terms of how to best manage situations should they arise. However, on occasion participants shared highly sensitive information to the unprepared researcher particularly as they became more trusting and open. This is more common with longitudinal work as good rapport is built over time.

The following example demonstrates how delving into past experiences, can accidentally unearth sensitive data which can be cause for disillusionment and a sense of failure for the participant. When Gloria was contacted to arrange her first interview, she was talkative and helpful, even offering to meet the researcher outside her block of flats, to ensure she found her way easily. Gloria engaged fully with the research questions and provided useful insights into the issues explored. She revealed that she was alcohol dependent, and suffered from anxiety, which sometimes confined her to her home, and that she relied heavily on her family for support.

As the researcher delved into questions about money management she did not anticipate one response:

A fragile and vulnerable woman emerged during the interview, about whom the researcher had limited information prior to the encounter. Whilst in-depth interviews with vulnerable people are vital for understanding the full context of their lives, they can also be deeply problematic from an ethical perspective, in terms of evoking memories which may cause immediate, or delayed, distress after the interview. Some of the potential issues will be discussed later in this article.

The occurrence of distressing life events during the course of the research

Longitudinal housing research captures changes in circumstances over a certain period of time, and significant changes did occur during the course of a few participants’ involvement in the research. Whilst illuminating for the research findings, they are ethically deeply problematic, for both participant and researcher.

In one such example, the participant, Marion, was personable and welcomed the researcher into her home. She provided directions to her home, introduced the researcher to all of her family, made her tea, and made the researcher feel comfortable immediately. The first interview revealed that Marion was struggling financially, and a short catch-up by telephone a few months later, indicated that her situation had deteriorated considerably with bailiffs taking goods from her home. A debt support worker was present during the second interview, which generated some discussion of the family’s changing household composition, which was to the detriment of their finances. Up to this point, Marion had remained in good spirits regardless of her hardship, and she continued to show commitment to the research. The researcher interacted with other members of the family too and learned a great deal about them. Her young son was present during the second interview, and he exchanged some banter with his mother and the researcher.

It was during a telephone catch up two weeks later, that Marion dropped the bombshell: she broke down and explained that her son had suddenly died. The researcher was shocked and saddened and after expressing her condolences, she suggested that Marion take time out from the research or consider withdrawing. However, Marion was adamant that she wanted to continue. The final interview visit was extremely difficult to prepare for, and the topic guide was pored over, to approach subjects with utmost sensitivity. Having spoken to the family who were in good spirits in the past, despite their challenges, the researcher found it extremely difficult to speak to them in such somber circumstances. The impact of this situation is explored later.

Aside from longitudinal research, in one-off interviews there is still the possibility of catching a participant on a day when they have received some bad news, and worse still, news which the participant has then used to implicate the researcher in their misfortune. The next example demonstrates how one participant’s suspicion, accusations, and misunderstandings based on receiving bad news on the actual day of the interview, put the researcher at significant immediate risk. When the researcher initially telephoned the participant, Sarah, to arrange the interview, she requested further information as she had difficulty understanding the information sheet that she had received. Sarah’s feisty attitude was noted during the discussion, and she seemed very suspicious of the motives for the research. She was very blunt and asked many questions, but was appeased by the information provided over the phone, and especially when it was reiterated that the researchers were entirely independent from Sarah’s landlord and the Benefits Agency.

However, unbeknown to the researcher, on the morning of Sarah’s interview, she had received news about the loss of her housing benefit, and as the researcher walked into the small two- bed flat, she was confronted by a very aggressive and agitated woman, and worse still, by two very aggressive dogs. While an irate Sarah demanded to know if the researcher had reported her to the council, the dogs set about gnawing and tugging at the hem of the researcher’s trousers. Sarah did nothing to stop her dogs due to her own anger, and the dogs increasingly became uncontrollable. At that point, the two dogs presented the most immediate risk to the researcher, and luckily Sarah’s son had just emerged from the balcony having heard the noise and commotion, so the researcher took the chance to firmly instruct him to take the dogs into another room, so that she could clear up any misunderstanding, uninterrupted. It took almost twenty minutes to clarify the research aims, and the independent position of the researcher, to alleviate suspicion and to diffuse the aggressive situation that had unfolded. The participant insisted on doing the interview, during which she repeated:

I got a call from the council office this morning saying that somebody said that I’d moved out the area and I’ve moved to London. So at the minute they’ve stopped my benefit.

Her anger and anxiety made sense in light of the information she went on to provide. Sarah was a lone parent of a teenager, long-term sick, reliant on benefits, and struggling to cope financially. After the interview, she revealed her drug dependency, which was not picked up by the researcher before her visit to conduct the interview. Again, although researchers can gain some information about a participant e.g. their nature, during the initial stages of arranging an interview, there is no way of knowing what the individual’s real disposition or circumstances are, even if their contact details have been provided by a statutory agency, such as a housing provider. Researchers can do prior research into any issues in the fieldwork area, but again, know little about how participants live in their home, who visits, their pets and their unpredictable nature.

Notably, university personal safety and lone working guidelines are usually generic and only scratch the surface of the complex situations encountered in the field. For example, pieces of advice relevant to Sarah’s case are incompatible with the reality faced, it states: when experiencing hostility or aggression, ‘if already inside, leave immediately’; ‘keep yourself between the angry person and the escape route; and ‘be aware of your possible escape routes and position yourself accordingly’ (Sheffield Hallam University, 2013). But, in this case, Sarah and her dogs were blocking the only exit, and she demanded answers. To attempt to leave would probably have further aggravated the situation. Only guidance on dealing with aggressive behaviour is appropriate here. This includes, remaining calm, listening carefully, and using the correct body language to diffuse the situation. However, nowhere in this guidance is the recognition that in rare situations staying and resolving misconceptions may be the safest option.

Impact on the participant

All six examples used in this article emphasise the vulnerability of each participant by contextualising their financial circumstances and/or health conditions. Attention is also drawn to the fact that the level of vulnerability of the participants was not known by the researcher before going into the interview, and that this presented a number of ethical dilemmas, which the researcher tried to address sensitively at the time of interview, and reflected on some of the more complex issues afterwards. Firstly, she considered whether the vulnerability of her participants’ had increased further, having imparted such personal information and feelings that they might have otherwise locked away. The researcher reflected on the impact of asking Gloria questions which inadvertently reminded her of her painful past. A lot of Gloria’s inability to cope was bound up with the loss of her husband, who provided her security, particularly in financial matters. By focusing on her inability to manage, the interview might also have potentially caused Gloria to reflect on her perceived ‘failures’ and led to disillusionment.

Unintentionally increasing the interdependency of participants involved in the longitudinal research was also of concern. A few of them clearly valued having someone willing to listen and to share their problems with, interviews were not always confined to the scope of the research subject and spilled over into other personal aspects of their lives. With reassurances of confidentiality, the repeat visits and telephone catch-ups, a very good rapport was built which encouraged participants to open up. The researcher feared she might have fostered a false sense of dependency amongst her socially isolated participants.

The researcher was particularly concerned that she had asked several participants extremely personal questions for research purposes during a very difficult time in their lives. For example, in Brian’s case, dependent entirely on his carer, he was struggling with two serious health conditions and his financial circumstances. The researcher feared that the interview had unavoidably ventured into potentially distressing subjects such as Brian’s illness, family support (or lack of it), difficulties with his landlord, including threat of eviction etc. His responses seemed quite emotional throughout his interview and significant feelings of failure and guilt were shown, which the researcher feared might have caused him more anxiety, even after the interview when he was alone. The researcher learned that severe loneliness and isolation characterised Brian’s life, so, after the interview, they talked at length about how he might be able to access various local support groups and opportunities for social interaction. Just how effective this discussion was in terms of encouraging Brian to find out more about the local support available, is unknown, but weighed on the researcher’s mind.

Impact on the researcher

Very rarely is the impact on researchers, of doing challenging research, explored and/or reported. Yet, in some situations, researchers may find it difficult to distance themselves from the distressing circumstances of their participants. For example, in Marion’s case, the researcher had tracked her circumstances, which deteriorated as the research progressed. Rapport had been built up, which improved with each communication, and the researcher had spoken to several members of Marion’s family, who were all spirited and welcoming, even in light of their financial struggles. She was shocked and saddened to hear of the son’s death, only two weeks prior she had been sitting talking to him on the same sofa. In this situation, the researcher did share this experience with the research manager, who listened with sensitivity, but nevertheless, the researcher was left feeling concern about how the family would deal with the significant challenges in their lives. Their financial situation was so dire that when Marion revealed that she still owed ‘money for the funeral’, the gravity of this statement was deliberated over for some time by the researcher and she was left with feelings of sadness and guilt.

The researcher was also left wondering if Brian had taken steps to overcome his isolation, but a follow-up telephone catch-up might have been a step too far in terms of researcher involvement, especially if Brian’s dependency on her increased. Having considered his illness and stressful life events, such ethical issues were difficult for the researcher to reconcile.