Complexity and flexibility: interviews with three-generation families in their homes

Growth of multi-generational living

Some 64 million Americans (in 2016; Cohn and Passel, 2018) and 4.3 million Australians live in multi-generational households (in 2011; Easthope et al., 2015). It is estimated that 1.8 million households in Britain contain at least two adult generations, an increase of 38% between 2009 and 2014 (NHBC Foundation, 2017). In-migration from cultures where multi-generational living is the norm is a global phenomenon and is most obvious in large internationally connected cities such as Sydney, Australia (25% of all households in 2011; Easthope et al., 2015) and in increasingly multicultural cities such as Auckland, New Zealand.

In New Zealand, where this study took place, since 2001, the number of people living in extended family households has increased by 57% (Lynsar and du Puis, 2015). Multi-generational living is usually associated with Māori, Pacific Island and immigrant communities. Māori comprise 16.5% of the total NZ population: Asians 15% and Pacific populations, 8.1%: all these population groups are increasing as a proportion of the total population (Statistics New Zealand, 2018). Distribution is also concentrated as Auckland (population 1.5 million) is home to two-thirds of the Pacific and Asian populations, half the Middle East, Latin American and African populations, and 24% of Māori live in Auckland. Multi-generational living is generally associated with these population groups. Further, in Auckland, in 2013 two-thirds of those living in extended families lived in an extended family of three or more generations (Lysnar and Dupuis, 2015). It is notable, however, that the rise in multi-generational living is also occurring in cultures and locations not traditionally associated with multi-generational living. Between 1996 and 2013, New Zealand experienced a rise of 142% in multi-generational living across all population groups, including large increases amongst New Zealand/European families. In addition to international in-migration, very high house prices continue to add to the rise in multi-generational households. Sydney (third) and Auckland (seventh) feature in various lists of the 10 most expensive cities in the world for housing. Auckland’s income to house price ratio is 1:9 and Sydney’s 1:11.7 (Bertaud, 2019). Unaffordable housing is a key driver for the rise in multi-generational living.

We argue that given the trend towards rising numbers of multi-generational households in many countries, a phenomenon that seems to have attracted little attention in terms of research methodology, this is an area of significance for qualitative researchers. A better understanding of this trend is especially necessary for housing, planning, design and social policy-related professions. Given the importance of and rising numbers living in multi-generational households, the need for robust qualitative research methods directed at this significant group is imperative due to the current lack of research guidance. Although our research did not address issues around multi-generational living, it highlights issues relevant to multi-generational research techniques generally.

The study

The three-generation study examined in this article was part of a much larger study of urban greenspace being undertaken (People, Cities and Nature: Restoring indigenous nature in urban environments) exploring engagement with green spaces and nature. In previous work, we have explored relationships between children and green space, the changes in older adults’ relationships with nature as they age and people’s backyard relationships (Freeman et al., 2017; van Heezik, Dickinson and Freeman 2012; Freeman et al., 2012; van Heezik et al., 2018; Freeman et al., 2019). The last four all used home-based interviews. These studies alerted us to the fact that different generations use urban green spaces quite differently. Understanding this difference is of fundamental importance for the lead author’s profession of planning, as it became evident that different life stages are associated with different greenspace use. The ‘one size fits all’ construct of greenspace provision is clearly inappropriate.

As part of the wider study, we undertook a focused piece of research on a subset of three-generation families. The 57 participants (age range from 6 to 80 years) included 20 children (9 females, 10 males), 16 parents (15 females and one male) and 21 grandparents (11 females and 10 males), living in a range of suburbs and socio-economic situations. The study rationale was that by recruiting three-generation families where all generations lived together or in close proximity, all family members theoretically had access to exactly the same range of neighbourhood biodiversity and greenspaces. In this way, differences relating to life-stage could be separated and identified. The primary research question for the three-generation study was: ‘do different generations living in the same neighbourhood value and use the same green spaces and if not why not’? We decided to use three-generation families rather than interviews with three generations from separate families in a neighbourhood, as the close residential proximity of the families in this study reduced the influence of additional locational variables. It also enabled the research to better identify family characteristics of greenspace use.

Interviews required the completion of a questionnaire, either by the participant (depending on age and comfort in doing so) or by the researcher who completed it in conversation with the participant. Interviews were with individuals in the family home. On occasion, other family members could listen in, but the interview responses were those of the participant being interviewed. One activity included selecting known and preferred bird species from a selection of photos laid out on a flat surface, ideally a table. Another activity included selecting preferred landscapes from a series of photos. The interviews were not recorded or transcribed. The mix of activities meant a more fluid interview process in different parts of the home (and garden) was required, not altogether conducive to recording so we used written questionnaires and reflective notes instead. Many participants self-completed the questionnaire and where it was completed by the interviewer, any unclear responses and/or wording were checked with the participant. Two generations, the grandparents and parents, completed the same interview, and all three generations completed the photo activities. A child-appropriate version, modelled on the adult interview, was used for the children. Given the experience of the team in interviewing across age ranges, across different ethnic groups, locations and ages, the differences and level of the challenges experienced with multi-generational family-based interviews were unanticipated.

After each family interview, a written interview-reflection was compiled that recorded time, place, people in or passing through, evidence of comfort or discomfort, other concurrent activities and general incidences that took place during the interviews. These reflections form the primary data for this article. Reflections ‘provide non-textual or auditory information about interviews and focus groups, useful in understanding participant meaning . . . allow for transmission of the full depth of the study context’, as described by Phillippi and Lauderdale (2018: 382). The reflections were written after the interviews whilst it was still fresh, allowing the researcher to follow Maharaj’s advice for interviews of making ‘more transparent the process of entering a research context [the family home], establishing a role as a participant observer, and transforming observations into a written account’ (2016: 123). Related to reflection is recognition of ‘emotion’ in research, especially the working practices around the researcher. The researcher’s personal identifications and experiences when set in an intensely social context such as multi-generational interviews in a family home will influence and be influenced by the wider emotional context. As Loughran and Mannay explain, the researcher becomes part of the social events and processes they observe and help to narrate a setting that calls for active reflexivity (2018). Our experience with this study brought home strongly the necessity for different methodological and contextual considerations and the practice of active reflexivity for us as researchers when working in this, for us, new, but demographically increasingly important group: multi-generational families.

Who are family?

Jamieson et al. state that: ‘one of the challenges . . . to studying families and relationships lies in identifying who or what is the unit of analysis’ (2011: 4). There are several different and changing ways of defining families. The definition that most closely describes the families in this research is: ‘a network of related kin’ and ‘the extended family comprise a nuclear family plus relatives – such as grandparents, aunts, uncles, grandchildren, in-laws and so on – living in the same household or nearby’ (Steel et al., 2012: 19). Our families included families living in the same home, adjacent homes and close by homes. They included divorcees, widows, re-partnered parents, couples, biological children and other people living in the household such as international students. Our only requirement was that the family had three generations where one of each generation would agree to an interview, and the families included much of the diversity and complexity of modern family structures.

In defining families, we let our families self-define but also followed the advice of Baker, who asserts for researchers the definition of family should encompass those with ‘caring and enduring intimate relationships regardless of their legal or blood ties’ (2014: 3). In addressing the question of, ‘who constitutes my family?’ Finch points to the fact that the researcher cannot adopt the default position of pointing to the household they currently inhabit. Rather family is represented though ‘embeddedness in a particular set of relationships’ where families are constituted by ‘doing family things’ (2007: 68, 79) and gives the example of a restaurant meal that included current and ex-partners and children, where the meal displayed family connections between the different households and partners. This concept is vitally important to us in understanding ‘family practices’ around greenspace use. Defining a family is necessary as it has ramifications for three key elements of the research process: organising who participates and when, the interview setting and issues of privacy and confidentiality.

Recruitment and informed consent

It was not possible to follow standard channels of recruitment as three-generation families have a few common points of contact where all three generations can simultaneously be aware of and agree to join in the research. A small number of families were recruited through a local newspaper advert but most were recruited by being recommended to or known to the researchers, or where an interviewed family recommended another family. Informed consent is a difficult issue when recruiting within families, as degrees of interest within the family vary and family power structures and dynamics can play a role. In some cases, one family member acted as the gatekeeper, defined by Jensen and Laurie as: ‘key people who let us in, give us permission, or grant access. . .individuals who are known and trusted by potential participants’ (2016: 121). Sometimes, we made contact with one of the adult generations who then spoke to the other generations. Indirect communication to participants through the family intermediary did have the potential for miscommunication or non-transfer of information but it is important that communication preferences, which mostly but not always worked, were guided by participant families and not externally imposed by the researchers. Other times, the family preference was that we spoke to and explained the study to the other family participants directly, rather than the ‘gatekeeper’ taking on potential responsibility for recruitment. Similarly, participants selected the time and place for the interviews.

Recruitment across families varied from just one from each generation (three participants) to all eight members of one family. All participants, including children, prior to the interview were informed of the voluntary nature of the participation and they signed an informed consent form. In cross-generational family interviews, a particular issue is child consent where: ‘Parental insistence [of child participation] can be tricky for a researcher to ignore’ (Coad et al., 2015: 9). Children’s participation was communicated by the parent, but it was made very clear to children by the researcher at the interview that they had the right to not participate and the decision was theirs. Kirk (2007) alludes to the fact that the adult centredness and unequal power relations between children and adults are duplicated in the research process. So too Mannay in her interviews with mothers and daughters discusses the ‘landscape of power’ in families, one that was evident in her study’s mother-daughter relations, and one that is illustrated below (Mannay, 2013).

In the reflection extracts drawn from the researcher’s interview notes, all names are pseudonyms and identifying information has been omitted. A power differential was present in our study in at least one family, as indicated in the following case extract:

In one family the researcher was informed all 4 children would participate. However, on completing one parent, Aidan’s interview, Carter relayed that his brother Jayden did not want to do the interview. The interviewer explained this was fine, he didn’t need to do it. However, the father Aidan insisted this was not ok and he should do it. Aidan retired to speak to Jayden and then relayed that Jayden did not mind the subject matter but was too shy to talk to the interviewer. The interviewer suggested if Jayden was willing he could do the interview by himself, or with his father, which happened. Later on both came through to the living room table where they did the photo exercise. The two were observed to be comfortable but at no stage did Jayden make eye contact with the interviewer and when being thanked carried on focusing on feeding the family pet.

In the above example, family dynamics clearly prevailed where children were expected to participate in a ‘family exercise’, but the mode of their participation could be adjusted to minimise their discomfort. In another interview, a younger boy who was expected to be interviewed also opted out as his friends had come over to play. In this case, again the interviewer indicated this was fine and no interview took place. Of the five teenage boys interviewed, only one was comfortable with making eye contact and talking to the interviewer. The others were described in the interview reflections as follows: ‘a man of few words’ or ‘did not elaborate or interact with me’ and ‘preferred to do the interview independently filling in the questions on their own’. The dictum that a neutral, comfortable, even empathetic relationship be established (Jordan, 2006) was unrealistic with four boys. The lines of consent in these four cases are blurred, being recruited by their parents. At the interview they were fully informed of their right to not participate by the interviewer; however, they indicated that they were willing to complete the questionnaire and tasks, but usually completed these independently of the researcher. It was recognised that these tasks were unlikely to be their preferred activity choice. As a group, their voices are seldom heard when it comes to greenspace planning and the data they supplied were vital to the research and in some ways significantly different to that supplied by other family members.

Timing and time commitment

The timing of interviews with families is important given the range of other competing family commitments (Clendon, 2007). For the interviewer, the first visit is to an unknown setting. Wengraf (2001: 191) advises: ‘you need to count for yourself on a minimum of 30 minutes setting up time’ . Gray (2009) counsels that preparing for the interviews cannot be rushed and gives the following advice: ‘Before commencing the interview it is important to get your interviewees as relaxed as possible. . .’ (2009: 380). But this excellent advice is not usually practical in a family setting, especially where the participants, not the researcher, ‘control’ the interview environment. In a home, the researcher is a guest and as researchers: ‘We need to be mindful of our obligation and responsibility to be not just good researchers but ‘good guests’ (Coad et al., 2015: 15). In this scenario, the researcher does not have the luxury of a preparatory 30 minutes. Yes, introductions and information exchanges need to take place prior to the interview but the researcher is ‘intruding’ on family time and should not exceed their welcome. Some families feel a sense of obligation to treat the ‘guest’ well (Jordan, 2006) and don’t feel comfortable until the researcher has accepted the minimal offer of a cup of tea or coffee as in the following scenario:

I arrived at the participant’s home at 11.00 as planned. I was greeted by Jules and taken through to the lounge. She told me she was making scones for us to have with morning coffee. She continued baking the scones and making coffee whilst we talked about things in general. We sat down and had scones and coffee before starting the interview.

While the above was an ‘easy scenario’, that is a quiet home with one grandparent participant, a cup of coffee and scones was appropriate. However, interviewers need to develop the skill of negotiating and politely sidestepping all offers of hot drinks, without causing offense, especially in family home scenarios where the family may be putting on hold other family activities and commitments. As a guest in the family home, the researcher becomes part, by virtue of their very presence, of the ‘everyday encounters’ in the home from discussing meals, to being part of a scenario where a child may resist going to bed as they want to watch the researcher. Nonetheless, this participation in itself is important as it generates researcher insights into family practices (Mannay and Morgan, 2015).

As the interviews took place in the family home, this was a place where interviewees were more relaxed. It was apparent to the researcher that this was the participant’s space as many normal family activities carried on during interviews: watching TV, reading newspapers, children playing, arguing, asking questions of parents, feeding pets and preparing meals. One significant issue was the time of day for interviews. For grandparents, generally, interviews took place at a time and in a place convenient to them, so were generally calm and unhurried. However, for parents, this was seldom the case, given child, work and various caring commitments, and their homes had more people. The interviews were all timed around families’ schedules taking into account day, time and how many interviews they wanted per visit. Schedules were necessarily flexible to accommodate changes as in the next extract.

In her research with children in their homes, McLean advises: ‘Willingness to work flexibly and to fit the research around the many demands in the parent’s and children’s lives was the key in building strong research relationships’ (2011: 58). Flexibility was required for many interviews. For example, illness is a part of family life, and two interview times were changed due to illness. In one case a child being off school meant an interview could be conducted at a more leisurely pace during the day, rather than in the post-school period which was often an active time with sports, friends, meals and general fetching and carrying. Most parent interviews and child interviews had to be after school or work. This has implications for the interviewer as their work is primarily, thus also outside normal working hours, taking place in late afternoons or early evenings. Issues arose through participant’s work-induced delays or lack of communication, or a change in the interview process as in the following example:

The arrangement was to interview Cody and his grandmother Anna, at 4 pm. When the interviewer arrived Cody opened the door, Anna was in the house, knew about the research but didn’t realise the interview was then (she hadn’t received Cody’s mother’s email about the timing). She said she would go across the road to her own house and come back in half an hour for her interview. As Anna left a young person (later revealed to be a sibling) arrived clearly discomforted to see a ‘stranger’ in the house, she asked if the researcher was there to teach Cody and left. While the grandparent Anna was interviewed Cody watched TV and his mother Michelle came home from work. Michelle’s interview was then completed. By then, it was 6:30 and the family needed to get on with dinner and family activities. Michelle needed to talk to her daughter about a medical event that day. After this it was decided to limit interviews to a maximum of 2 consecutively to reduce interference in family time and routines as the interviewer had been present for 2:30 hours during what is usually full-on family time.

In several instances, communication between family members as above had failed, or other factors intervened such that the interviewee was not present when the interviewer arrived and the family member opening the door did not know of the interview. In several cases, the teenagers needed additional communication to return home for the interviews. It is not just the timing but also the time commitment that matters. It is not appropriate to take more family time than is reasonable. The total time for all 15 family interviews was 3015 min (50 h and 15 min) with the time per family varying from 1 h 30 min in one visit to 5 h 45 min over three visits. This means that while an individual interview may last on average 30–45 min, the family will find that they have to suspend or partially suspend most family activities for the total duration of the interviewer’s visit

Interview location

The 15 families were interviewed in 28 homes, including one interview in a café at the interviewee’s request due to a sick partner. Grandparents were usually interviewed in their own home although one child was also interviewed at the grandparent’s home. The place for the interview presents its own challenges. Interviews at home can be expensive (primarily researcher time) and overall more time consuming (Fargas-Malet et al., 2010) but in our case absolutely necessary. All interviews except the café interview took place in communal areas, at the dining, kitchen table or breakfast bar or on the sofa. In their research, Coad et al. (2015) talk of the importance of ‘location, location, location’ and asserts that the ‘where’ of the method has received minimal attention. Within the home, practical limitations were encountered: lack of physical space, the interview space being communal, and for us poor light and lack of a table or a clear space. As part of the interview, bird photos had to be laid out, but in two homes the light was poor so they could not be clearly seen. Where no table was available floor space was used, but this is not ideal for older adults. However, there are also many positives to interviewing in home; for example, a grandfather observed their partner’s interview and then asked if they could also take part, so an interview was scheduled for a later visit. MacDonald and Greggans described their home-based research interview experience as follows:

Our expectations as researchers in terms of the process of collecting data, through a question and answer technique in a calm quiet environment that was time limited for 30 minutes were very far from the reality of what happened in practice (2008: 3124).

For us too, the assumption that there will be a quiet, clear space available was somewhat unrealistic. However, our participants were comfortable in the setting and seemed able to function effectively in places that would in no way meet the advised ‘quiet secluded location’.

In our study, family members all seemed to freely and comfortably move between the different homes, as while interviews were going on in one home, family from the other home often called in. When new adults enter their home, there is often a need to break the interview to do introductions and explanations of who the new person and the interviewer is. Some were family who were not taking part in the study, or who were interviewed at a different time but were, nonetheless, present during at least part of interviews. In family and home-based research, we concur with the observation: ‘that multiple interruptions are part of the complex social reality of other people’s lives’ (MacDonald and Greggans, 2008: 3124). This fluidity immediately raises issues of confidentiality, comfort and disclosure as the participant may not wish to disclose information to other family members or visitors, or may adjust their responses as a consequence. It is also possible that if a child is being interviewed they may adjust their responses to those that are most appropriate for the parent to hear and vice versa. However, as reported by Coad et al. (2015), we found that the parent could also act as a ‘valuable proxy’, helping the younger children especially, for example, reminding them of places they go and things they do, then letting the child continue to develop that recollection. Our adult participants showed genuine interest in their children and grandchildren’s responses. Gray advises ‘Confidentiality should be offered to respondents when completing questionnaires so clearly the same respect should be afforded those participating in an interview’ (2009: 391). We could not guarantee confidentiality without imposing inappropriate interview conditions on a family home, for example conducting the interview behind closed doors. As our research was not especially sensitive, we do not believe confidentiality was as major an issue as it might be in other types of research.

Busy family life

In their thought-provoking article on the counter narrative of a ‘failed’ interview, Nairn et al. assert that: ‘the complexity and messiness of the research process is often concealed within articles that present a coherent narrative’ (2005: 223). Interviews within busy family contexts repeatedly exhibit the complexity and messiness that Nairn et al. reference and difficulties would be experienced in trying to identify or construct a coherent narrative. However, it is this complexity and ‘messiness’ that is at the heart of family life, especially where family life intersects across generations and households. It also contains within it the adaptability and responsiveness to changes that enable family life to succeed. In order to succeed, the interviewer needs to accommodate the fluidity of family living and be flexible. Flexibility also revolves around family schedules.

One family’s interview with the oldest child Aria began at 7:30 so the other 3 children could meet the researcher but then go to bed leaving a quieter environment for Aria’s interview. She in turn went to bed so the mother Hazel could be interviewed in a quiet post-child environment. Even so one sibling did come in to see what was happening with her sister’s interview but got told by Aria: “what are you doing? You should be in bed”, enabling the older sister Aria to reassert her difference in terms of family expectations around bedtime.

Two interviews with mother Maia and daughter Leilani were scheduled to take 90 minutes, starting at 4 pm. However, Leilani was absent on arrival and the interviewer was informed that the child needed to leave at 5.10 pm. Then Maia texted to try and locate Leilani. About five minutes before completing Maia’s interview Leilani, her brother – (who talked loudly to his mother about his day), and two men came in. This meant that there was insufficient time for the interview but Maia preferred to press on. Maia’s interview was completed and ‘there was now lots of noise and excitement in the home’. Leilani overcame her initial surprise, settled well and despite the limited time for introductions, completed her interview in the time available when all three left the house.

Family life is fluid, busy and often full of competing demands. Thus, although the interviewing process may not match the ‘expected standard’, families do seem comfortable with adjustments and changes as family life necessitates. The discomfort seems to be of greater concern to the researcher and their notions of a ‘proper’ interview environment than the family.

A key aspect of busy family life is interruptions, parallel and competing activities. We empathise with MacDonald and Greggans (2008) who report: ‘Capturing the chaos and complexity is imperative in qualitative enquiry but in our experience dealing with this perceived chaos is harder in reality than is detailed in the literature’ (2008: 3129). It was the level of this complexity and fluidity of movement that was most striking and most different to any other home-based interviews we had done. With hindsight, the fluidity makes sense as these are large families who live in close proximity with shared lives, so why should such shared normality cease because a researcher is present? The following descriptions indicate how this shared normality played out in two interviews:

The grandmother Zoe, had muddled the time and expected the interviewer earlier than agreed so the interviewer hurried over. The great-grandmother Annie who also lives there put the kettle on. A friend of Zoe’s was present and needed to finish their conversation before baking biscuits for an evening event. During the interview it was decided to fetch the photo album of the grandmother’s childhood to illustrate some answers so this had to be located and family photos discussed. The living space was small and dark and the table already full so there was no room for the interviewer’s bird photos to be laid out. During the interview the grandfather Michael came home and was introduced. Then the grandmother’s daughter Isobel and granddaughter Jessica (both also participants and the youngest child in the study) came into the house. All looked at the family photos and the interviewer talked to the child to set her at ease as she would be interviewed on another date. Jessica agreed to the interview as long as she could give the interviewer a ‘makeover’ with her new [beauty] set post-interview (this happened). The great grandmother Annie was frail and wanted to see the bird photos that were on the floor. The grandmother Zoe started to get tense as she needed to get ready for an evening engagement. The grandfather Michael then asked if he could take part and an interview time was agreed. At one time there were, 7 people (including the interviewer) in the small kitchen-living space, aged 6 to 80+, baking, on the computer, chatting, looking at photos and making tea. The grandfather Michael’s interview on a different day was easier with just 5 people though the TV and radio were both on.

The next example included a grandparent who lived next door and a young family with three children.

In the first visit and interview with 10 year old Janey, her younger brother kept coming over and enthusiastically answering the questions when Janey was trying to think carefully about the answers. The brother also disagreed with some of his sister’s answers and challenged her. At some stage the grandmother Millie came in from next door. When not at the table the brother was outside with his younger sister, with arguing and noisy playing being audible. Beth, Janey’s mother’s interview experienced a number of interruptions due to the need to mediate between the children outside, and became pressured as Beth needed to leave after the interview. The next visit was easier. An adult friend was visiting as were two small children, playing with the youngest sibling. Beth finished off her interview. The younger boy then came home with four friends and asked if he could play outside rather than do the interview, this was agreed. In all, 12 children and adults in addition to the researcher were present but the interview ran smoothly.

This free flow was common in the family homes where the children lived. In all, a total of around 64 adults and children passed through or were present in the interviewing space in addition to the person being interviewed. In two-generation research with children in their home, Bushin (2009) found during interviews only eight of the 39 took place without any interruptions. Not all disruptions are problematic, they vary greatly in scale and intensity from the more minor ones of the noisy turning of newspaper pages, to Janey’s brother vociferously challenging her responses. We did notice that in most cases, the adults seemed relatively unfazed by interruptions from children or people, but more discomforted by time pressures, the need to undertake family activities and deal with work or other pressing commitments. In three-generation families, even more than two generations, especially where the family is in the same house or physically close, it would be the quiet and uninterrupted that is unusual with interruptions being the norm.

As well as providing potential distractions, people-related activity meant the interviews were not confidential, and given the presence of some non-family members were not confidential to the family either. University ethics processes are generally predicated on the maintenance of anonymity and privacy, clearly not possible in a family scenario where for reasons of ease, comfort and practicality interviews need to be undertaken at home and cannot be conducted behind closed doors. This concern has been identified by several family researchers (Bashir, 2018; Coad et al., 2015; MacDonald and Greggans, 2008). In Bashir’s study of housing for vulnerable people, her participants expressed a preference for home interviews as they felt this could afford greater confidentiality (Bashir, 2018: 642). In our case, the reverse was true and confidentiality can be impossible to maintain. However, our study was not considered sensitive and personal disclosures were not expected as part of the study. However, as the 10-year-old girl Janey demonstrated, she was clearly discomforted by her brother’s listening to her responses. On a more positive note, parents’ presence did make some children feel more comfortable and the younger children did sometimes ask parents for clarification. Some parents, grandparents and children showed great interest in their different answers and elected to listen to each other’s interviews. The interviewer let the family dynamics dictate the presence of others at the interviews, and mostly these additional presences were positive.

The interviewer and the unexpected

In her analysis on ethics and safety when interviewing people in their homes, Bashir (2018) refers to researcher vulnerability, in her case from alcohol-dependent participants but also from potentially aggressive dogs and identified issues that cannot be anticipated in advance. The unexpected is an inevitable component of family interviewing. Our researcher was a highly skilled researcher, very experienced in home-based interviews and as an ex-medical professional was very experienced in working with a range of people, including older and vulnerable people. Yet, even so, not all eventualities can be prepared for. Like Bashir, our interviewer also encountered a barking dog that felt menacing. She encountered family members who didn’t know why this stranger was arriving at or was in their home, a vulnerable older person with dementia who was confused by the repeat visit of the researcher and needed new introductions each time, and other people living in or passing through the home. Interviews can be interrupted as life events carry on, and can need immediate discussion; outcomes from doctor’s visits, preparations for a birthday party later that day, organising child pick-up from school, children’s friends coming to play, a child unexpectedly home due to illness and teenage children absent when they should be home and needing to be located. All these and others happened during interviews.

McGrath et al. (2019) identified 12 tips for qualitative researchers. Tip nine is ‘Be prepared to handle unanticipated emotions’. While McGrath’s advice was intended primarily for medical researchers, unanticipated emotional disclosures were part of our research experience. Possibly, the mature age and experience of the interviewer and her previous experiences as a nurse meant the interviewer was someone who several of the older participants shared experiences and confidences with. Sometimes these were ongoing situations, such as discussions with two different family members about coping with a relative with dementia, information about sick children overseas and some shared confidences with the interviewer about living with multiple generations in shared households. The interviewer almost takes on the role of caring listener in these cases, with these more intimate discussions usually playing out over a shared cup of tea or coffee. Spending time with families and in the intimacy of family life means such ‘sharings’ are increasingly likely. These illustrate how the researcher becomes part of the social world, facilitators of shared accounts (Atkinson and Coffey, 2001) as they enter what Mannay and Morgan (2015) call the ‘waiting field’, defined as a place where real lives carry on. As families become comfortable around the researcher, the chances of the researcher’s inclusion, intentional or not, increases. As has been mentioned previously, there is little guidance on researcher’s responses to and ways of coping with intimate information. On the positive side, these ‘sharings’ are indicative of the acceptance and trust families had of the interviewer.

An unexpected affirmative element was the interest different family members showed in each other’s responses, between siblings (although they may disagree) and grandparents, parents and children. Part of the interview involved choosing photos of birds they recognised and would like to see more of, and preferred landscapes. Even the ‘shy’ teenagers showed a keen interest in the photos and came to see what other family members selected. We found the interviewee would sometimes informally interact with others, as in the case of Zoe’s family, where all present clustered around the family photo album. Sometimes parents and children sat together on the couch so they could listen to the other’s responses and were usually mutually respectful and did not interrupt. This was also a way of family members learning about each other and showed the intimacy of concern for understanding between family members. One parent commented on how revealing it had been to see how her daughter interacted with a stranger, as she had only previously seen her interact with family and known people.

Discussion

We wrote this article initially in response to what we found was a lack of guidance on the complex realities of one-on-one multi-generational family interviews in home environments. When starting this research, we found we had underestimated the complexities associated not just with interviewing families, but the added complexities of interviewing across three generations. We successfully completed some 57 interviews with family members aged between 6 and 80 years plus. Here we identify the limitations we encountered, the lack of guidance on interviewing multi-generation families and offer our perspective on the process.

Family-focused research – current focus: Published studies on family research invariably focus on two sets of participants: adult carers and the cared for (usually medically oriented research). In medical research, the aim is usually to solicit the views of both the adults and children impacted by medical conditions (Borra et al., 2003; Gannoni and Shute, 2010; Poston et al., 2003; Young et al., 2003). One study (Gannoni and Shute, 2010) interviewed parents and children about serious medical conditions: cancer, chronic renal failure and type 1 diabetes. The interviewers noted a surprising number of commonalities between responses, including emotional reactions on the part of the children and adults. Although that study was on a very different and far more sensitive subject, in our study we also observed some commonalities in family responses and shared understandings, evident in the different characters of our 15 families.

Another area where two-generation interviews are common is childhood studies (Bushin, 2009). In both medical-based and childhood studies, considerations around the interview location, confidentiality, interruptions, wider family participation or presence are all identified. The upsurge in research with children, especially from childhood studies, children’s geographies and related disciplines, has contributed to a huge expansion in research involving interviews with children and often also with parents. Van Blerk and Kesby reveal the complexities around the child–parent research relationship and children’s own agency in research (Van Blerk and Kesby, 2013). In her research on changing geographies of use of the neighbourhood, Karsten’s interviews with parents and children enable identification of generational changes (2005). Growth in including children and multiple family members in research has led to the recognition of a range of ethical and practical issues. Home-based research has brought into focus many of the issues we identified, perhaps most notably the fact that as Bushin recognised, no home is the same and every interview setting is likely to be physically different (2009). This is unlike interviews in schools, offices or other more controllable settings. To this, we would add that every family is also socially and emotionally different, depending on the makeup of the family. If the family comprises three generations, these differences are exacerbated. If issues are encountered such as disruptions, competing activities and organising interview times, these may have to be discussed and negotiated across three generations, rather than with one interviewee. The key is flexibility in timing, adaptation to the physical space and the wider physical context (e.g. noisy children), but also mental and emotional flexibility in moving seamlessly between a young child’s interview and one with an older person.

There are no guidelines as to how multi-generational research should be conducted, although they are desperately needed. In addition, the form they should take in view of the intense individuality and range of family scenarios will be a challenging question to address. University ethics assume a standard one-on-one interview scenario that takes place in a public or professional space. The issues experienced here highlight the need for ethical behaviour, especially in the absence of clear guidance, but also the ability of the interviewer to control the situation. In particular, like Bushin (2009) and Mannay (2013), we became very aware of the need to carefully negotiate the power differential embedded in family practices, especially between children and parents, but also in one instance between grandparent and great grandparent. We were faced with an issue where a fourth-generation great grandparent was interviewed, but due to memory lapses it was uncertain whether they really understood the research process and were able to fully give informed consent. However, the family was a very close one and the great grandparent was present throughout, and to exclude them would have also been unethical. Given the uncertainty around informed consent, though, the data for this participant were not used.

Our research showed that in multi-generational interviewing when multiple members are interviewed from each generation, the ethical challenges become magnified and similarly the challenges of one-on-one interviews in family homes. In some ways, the challenges reflect those of any family interviews but on a potentially larger scale. Some researchers believe: ‘The strength of the interviewer-participant relationship is perhaps the single most important aspect of a qualitative research project’ (Knox and Burkard, 2009: 569); we would concur with this statement. Our interviewer was a highly skilled, medically trained professional who had worked for many years as a research interviewer. Her experience as a nurse and in geriatrics was, we believe, critical to, not just the success of the research, but to gaining access to the families and their homes. This was not research that could be negotiated by an inexperienced researcher. Neither would it be fair to expect a new researcher to be able to cope with the sometimes sensitive disclosures that participants make, or with the tensions that arise where there has been a lack of in-family communication about the ‘stranger’ who appears at or in their home. Personal or sensitive disclosures can be anticipated for topics such as refugee experiences (Bögner et al., 2010), or post-disaster research (Freeman, Nairn and Gollop, 2015), but we would argue also that disclosures are possible and likely when dealing with complex multi-generational families in settings where families become comfortable with the researcher. These disclosures can be emotionally challenging but are countered by the immense privilege of being enabled to enter into the family home and observe the strength of caring and sheer organisational capacity of everyday family life.

A new research direction

We call for multi-generational research like ours to be the focus of greater research attention. Developing a better understanding of the changing patterns of urban living will increasingly necessitate engaging with multi-generational families, as their growth looks set to continue. At a general societal level, multi-generational living is associated with both positive and negative characteristics. It provides a positive affirmation of the social benefits accruing from family support, cultural understandings of family life and global migration taking traditional extended forms to new countries. Negatively, multi-generation living is reflective of social poverty, as families cannot afford separate homes and are forced into extended and multi-family living as a survival mechanism. In the research context, we see in-depth qualitative research using in-home family interviews as becoming a very significant methodological direction in understanding this growing phenomenon. Just as there was little thought given by researchers to interviewing children prior to the 1990s, now there is a substantial base of research and methodological information around child interviews (Kirk, 2007; Skelton, 2018; Tisdall et al., 2009; van Blerk and Kesby, 2013). The same reversal needs to become more evident in multi-generational research. The central notion must be to give everyone, regardless of age or position in the family, the opportunity to have their voice heard. Doing so will enable researchers to enjoy the depth of family contribution that we experienced in our relatively small study, and ideally multi-generation research that can be embarked on with greater guidance from the research community.

We finish with the words of three generations of one of our families:

Grandmother Millie: Father was an organic gardener. . .Grandparents lived in the Catlins [a beach and forest area] we had holidays with them. . .I try to help the grandchildren see the beauty around them

Mother Beth: Grandparents were incredible gardeners and I spent a lot of time in their place, precious memories. . . all my happiest memories as a child are out in nature.

Granddaughter Janey [Interviewer- How much time do you spend in your garden?] from 10-3.00 or so whenever it’s sunny [What does nature mean to you?] It means that animals can survive. Nature’s practically everything, without nature we wouldn’t be here. . .

This quote shows the connections with nature passed through and integrated into this extended family. Much of their story shows a strong continuous environmental ethic referenced across five generations from 10-year-old Janey to her grandmother Millie who in turn referenced her own grandparents. It is one that would have been missed had we only interviewed one generation. Janey’s responses suggest Millie’s hopes for her grandchildren are being realised. Families have stories to tell and only by putting together the stories across all generations can we, as researchers, gain a real understanding of what it is that is special and different about multi-generational family living.