Introduction
Nearly ten years ago the British public was told that fiscal consolidation was necessary to secure Britain’s future economic stability, and that public spending cuts of £83 billion – ostensibly affecting all sectors of society – were necessary to deliver a ‘new vision for a fairer Britain’ (HM Treasury, 2010: 7). A decade later, research has compellingly demonstrated that the economic elite has been largely untouched by austerity measures, while the most disadvantaged and marginalised individuals, groups and places have been pushed deeper into poverty (Hastings et al., Reference Hastings, Bailey, Bramley, Gannon and Watkins2015; Gray and Barford, Reference Gray and Barford2018), and in extreme cases destitution (Fitzpatrick et al., Reference Fitzpatrick, Bramley, Sosenko, Blenkinsopp, Wood, Johnsen, Littlewood and Watts2018).
Austerity has been attributed to increases in rates of homelessness in some parts of the UK (Fitzpatrick et al., Reference Fitzpatrick, Bramley, Pawson, Wilcox and Watts2016) given its effects on social security entitlements, housing affordability, and household vulnerability to eviction and repossession amongst other factors (Loopstra et al., Reference Loopstra, Reeves, Barr, Taylor-Robinson, McKee and Stuckler2016; Kleynhans and Weekes, Reference Kleynhans and Weekes2019). Local government functions have also been ‘scaled back’ in response to budgetary pressures, with cuts to funding in homelessness services hitting the media headlines in a number of cities (see, for example, Scottish Housing News, 2015; Goodwin, Reference Goodwin2019). These changes have had a profound impact on investment in and the nature of responses to homelessness at the local level (Fitzpatrick et al., Reference Fitzpatrick, Pawson and Watts2020).
This article draws on two qualitative longitudinal research (QLR) studies, both of which sought to investigate the everyday experiences of homeless people with ‘complex needs’ – such as co-occurring substance misuse and mental health problems or other manifestations of extreme disadvantage (Bramley et al., Reference Bramley, Fitzpatrick, Edwards, Ford, Johnsen, Sosenko and Watkins2015, Reference Bramley, Fitzpatrick, Wood, Sosenko, Blenkinsopp, Littlewood, Frew, Bashar, McIntyre and Johnsen2019). The evaluations were commissioned by third sector homeless services operating in two Scottish urban centres and had similar aims: to understand how individuals navigated local homelessness and housing services; and to examine the impact of these services on individual outcomes and experiences.
Below we document some of the evaluations’ key substantive findings. These add to the evidence base informing the development of interventions targeting homeless people with complex needs (Mackie et al., Reference Mackie, Johnsen and Wood2017), and in particular the value of a flexible, long-term and person-centred approach to delivery. The primary purpose of this article is, however, to establish the methodological ‘added value’ of QLR in these studies, and to reflect on the practical and ethical challenges it presents when employed in research involving vulnerable populations.
The article begins by outlining the ways in which QLR has been used in homelessness research to date, before describing the specific data collection and analytic methods used in the two QLR evaluations drawn upon here. We discuss briefly the substantive findings, before moving to a critical reflection regarding our methods and the difference that QLR made to the insights obtained. It concludes that researchers working with homeless people must recognise the particular ethical and practical challenges of QLR, and in particular the responsibilities associated with the researcher-participant relationship when employing the approach.
QLR and homelessness research
As discussed elsewhere in this themed section, QLR is a relatively new methodological development, distinctive because of its contribution to understandings of ‘time and texture’ (Neale and Flowerdew, Reference Neale and Flowerdew2003: 189). While all research takes place ‘in time’, the growing interest in longitudinal research design has focused on its ability to gain more nuanced understanding of the changes in circumstances, perspectives and outlook of people on each side of transitional or ‘critical’ events. Rather than being a retrospective ‘snapshot’, QLR can be seen as more akin to a time lapse, permitting the linking of contexts, mechanisms and outcomes at the individual level (Cashmore and Paxman, Reference Cashmore and Paxman2007; Thomson, Reference Thomson2007; Siennick and Osgood, Reference Siennick, Osgood and Liberman2008).
The dynamic analysis at the heart of QLR has been recognised as having particular relevance to policy research where there is potential ‘to generate unique insights into the ways that social policies and interventions are ‘lived’ and moreover ‘survived’ by individuals, families, communities and organisations’ (Thomson et al., Reference Thomson, Hadfield, Holland, Henwood, Moore, Stanley and Taylor2014: 2; see also Corden and Millar, Reference Corden and Millar2007). Lewis (Reference Lewis2007) emphasises its capacity to consider different ‘domains of change’ – individual, service, policy and structural – and the connections between. A growing body of work has demonstrated the value of qualitative longitudinal design in this context, producing nuanced and contextualised understandings of the everyday experiences of poverty (Patrick, Reference Patrick2017; Treanor, Reference Treanor2020), unemployment (Lewis, Reference Lewis2007; Shildrick et al., Reference Shildrick, MacDonald, Webster and Garthwaite2012), and families and relationships (Tarrant and Hughes, Reference Tarrant and Hughes2019).
The development of QLR as a methodological tool has particular relevance to homelessness research which, over time, has given primacy to the perspective that homelessness should be viewed as a process, rather than merely a situation at a given moment in time (Fitzpatrick et al., Reference Fitzpatrick, Quilgars and Pleace2009: 10; see also Mayock and Corr, Reference Mayock and Corr2013). Longitudinal approaches nevertheless remain relatively rare in homelessness research. Early examples focussed on generating statistical models to predict the likelihood of tenancy sustainment or returns to homelessness (Sosin et al., Reference Sosin, Piliavin and Westerfelt1990; Dworsky and Piliavin, Reference Dworsky and Piliavin2000). More recent studies have privileged the ability of QLR to explore process and meaning, and focused on developing a stronger understanding of the dynamics of homelessness (Mayock and Corr, Reference Mayock and Corr2013; Fitzpatrick, Reference Fitzpatrick2000), the differing priorities and perspectives of key stakeholders within the homelessness sector including service users (Johnsen et al., Reference Johnsen, Fitzpatrick and Watts2018, Reference Johnsen, Watts and Fitzpatrick2020), and the policies and resources required to enable homeless people to achieve positive outcomes (Warnes et al., Reference Warnes, Crane and Coward2013). The empirical studies examined here build on this literature and inspired critical reflection regarding the benefits, challenges and risks associated with QLR in homelessness research.
Data collection and analysis
As noted, we draw on two separate qualitative longitudinal studies on complex needs homelessness. Emma Davidson was involved in both projects, Briege Nugent in Project B and Sarah Johnsen Project A. ‘Project A’ was a two-year evaluation, conducted between 2014 and 2016, of a drop-in service for rough sleepers and those at risk of rough sleeping in the future. The project is best described as a ‘triage’ service, with its main role being to link users into appropriate services. Facilities included basic amenities (showers, access to computers, needle exchange, store for belongings), general support, advice and referrals, as well as help with form filling, making and attending appointments, and accessing housing and other support. The study included in-depth interviews with thirty-eight individuals, with follow up interviews conducted approximately six months after.
‘Project B’ is a three-year evaluation of a service offering tailored support for seventy rough sleepers with complex needs. Delivered by third sector organisations, the service provides intensive and sustained support to help individuals meet their immediate needs, obtain housing, make a home, and through building a relationship with a key worker, connect users to specialist support to prevent further homelessness. The longitudinal data reported, which is part of a larger mixed-method study, involved tracking twelve individuals at six-monthly intervals. At the time of writing, the study had collected its third wave of interview data.
Although the projects operated in different parts of Scotland, there were key similarities, especially in relation to the ethos underpinning the work. Their values emphasised relational practices that supported service users at their own pace, and on their own terms. Both studies were synchronic, having been conducted in contemporaneous historical, political and social contexts, with the consequences of local government cuts and welfare reform framing both service delivery and evaluation focus. The research designs were also broadly similar. By using a qualitative longitudinal approach, the aim for both was to gain a holistic insight into individuals’ ‘journeys’ through and out of homelessness, as well as their perspectives on their homeless and housing situations over time. Each involved in-depth interviews to understand how individuals were navigating the homelessness system and the contribution of the respective interventions.
In each study, space and time was given to allow respondents to narrate their experiences on their terms, with interviews often accompanied by coffee, lunch and other acts of support, such as driving participants to appointments or helping with laundry or shopping. Between waves, contact with respondents was maintained by a combination of informal visits to the service (project A), letters and verbal reminders by support workers delivering the service (project A and B).
This article came about after Davidson and Nugent met to discuss the emergent findings from the second wave of interviews for Project B. It was noted by Davidson that there were distinct parallels between the core themes and narratives identified in Project A (which at that point had been completed). A common theme related to the means through which QLR methods enabled individuals to narrate, shift and re-tell their stories and identities across time. We therefore wished to develop an approach which would examine common themes by bringing a selection of case histories from both studies into an ‘analytic conversation’. Examples of such analytic conversations include the biographical case study approach employed by Thomson (Reference Thomson2007); Tarrant’s (Reference Tarrant2016) common analytical framework for combining secondary QLR data; and Wright and Patrick’s (Reference Wright and Patrick2019) combined study qualitative longitudinal research. Our initial rationale was that such a conversation would add richness to our understandings of how homelessness is experienced at the project level and provide a means for interrogating the wider social, spatial and structural context of these experiences.
While the projects together could not constitute ‘big qual’ data, we nonetheless wished to build on the methodological work of Davidson et al. (Reference Davidson, Edwards, Jamieson and Weller2019) aimed at expanding the analytical potential of small-scale qualitative studies. Here, the intent is not to ‘scale up’ or make small-scale studies ‘bigger’, but rather to ‘pool’ qualitative data as a means of exploring cross-contextual generalisations. By drawing together these two studies, we add to the current innovations taking place within qualitative research which encourage researchers to consider how single datasets might relate to other data exploring the same topic, but through different contexts, localities or population cohorts.
Given the studies were commissioned separately, we were not able to share the datasets, or bring them together to create a new, composite dataset (as in Davidson et al., Reference Davidson, Edwards, Jamieson and Weller2019). Instead, we followed a process broadly similar to that developed by Wright and Patrick (Reference Wright and Patrick2019). Two initial sessions were held in which the project meta data was compared and discussed, including the research aims and objectives, contexts, the number of interviews, timescales and number of waves, and the interview questions. The authors re-visited the datasets they had been involved in to identify cases which could demonstrate the unfolding of stories and shaping of identities across time. Like Thomson (Reference Thomson2007) our focus was not on comparison, but rather on bringing these cases into an ‘analytical conversation’ as a means of interrogating the historical, social, spatial and structural context of the lived experience of homelessness services. It was on these criteria that four cases – Charles (61) and Bobby (39) from project A, and Dave and Barbara (a couple, both mid-40s) from project B – were selected.
The cases were analysed by a first reading to gain a descriptive account of individual ‘journeys’ through services. A second reading focused on the subjective accounts of these experiences by examining storied accounts, sense making and meanings. A third and final reading was then conducted to locate and observe these experiences in the wider social, administrative and political context in which these individual accounts took place (see Thomson, Reference Thomson2007: 575). Summaries of these biographical accounts were then written up in analytical memos and discussed as a team.
Substantive findings: key themes
Our initial comparison of the findings from the two evaluations revealed commonalities in terms of participants’ experiences of homelessness. The majority of service users reported poor physical health, poor mental health and substance abuse. Unsettled housing histories and long-term episodes of repeat homelessness were also common, as was experience of rough sleeping, begging and reliance on free food services. Accounts of experience of physical violence, intimidation, theft and/or exploitation whilst sleeping rough or staying in hostel accommodation were common in participants’ narratives, echoing the findings of a wider body of research documenting the harms associated with street homelessness and temporary accommodation (see, for example, McNaughton, Reference McNaughton2008; Cloke et al., Reference Cloke, May and Johnsen2010; Parsell, Reference Parsell2018; Watts et al., Reference Watts, Littlewood, Blenkinsopp and Jackson2018; McMordie, Reference McMordie2020).
Participants in both projects – including our four cases – frequently narrated their experiences as a consequence, at least in part, of ‘poor choices’ or bad luck. Yet these individualised accounts also made reference to the consequences of institutional and structural processes such as childhood poverty, inconsistent or inadequate support received in care, long-term unemployment, poor access to mental health services, or the experiences of punitive measures such as eviction and benefit sanctions. The ability of both services to engender positive outcomes for service users were also inevitably influenced by external factors such as (amongst others): the adequacy and short-term nature of programme funding; administrative efficiency of the welfare benefit system; application of statutory homelessness legislation; and accessibility and effectiveness of specialist housing, healthcare and other services. The interaction between such individual and structural features was of central concern to our analysis. As such, we have given value to the testimony of participants, whilst applying an analytical lens that acknowledges the power that structures, institutions and policies wield over choice, agency and sense of self.
We now explore in depth how these issues were interpreted and narrated by Charles, Bobby, Dave and Barbara over time. Their stories are discussed through four central themes identified in both evaluations. These were: accounts of life on the street, experiences of navigating homelessness services, subsequent shifts in their engagement with support, and the ambiguous notion of ‘home’.
Life on the street
Initial interviews with participants involved enquiry into their everyday experiences of homelessness. In each case, this had involved extensive periods of rough sleeping and engagement in ‘street-life’. Drug or alcohol abuse were reported to have escalated during time spent on the street. Poor mental and physical health, feelings of unsafety and unstable relationships were also cited. This was coupled with deep rooted feelings of marginalisation and disempowerment. A consequence of these experiences was a complex and ambiguous relationship to life on the street. Dave and Barbara, for example, explained in the first wave that despite having received offers of temporary accommodation, they preferred to sleep on the street given their assessment of the relative risks associated with hostel or shelter residence and street homelessness. Even once allocated a tenancy, Dave continued to go into town to see his homeless ‘pals’. Everyone on the street was, he explained, ‘like him’. It was a space where, in the context of multiple forms of marginalisation and stigmatisation, he felt recognised and ‘accepted’ by other members of the street population:
‘Everyone in the street is like us…Circumstances have brought us all onto the street, so many damaged people there.’ (Dave, wave 1)
Throughout the early interviews ‘the street’ was discussed by Dave and Barbara as, on one hand, a source of collective identity, freedom and (short-term and limited) excitement. On the other, it was a space fraught with physical danger, societal stigma, and one that they could not envisage an exit from. As Dave put it:
‘I call it [the street] the devil’s playground – the folk living rough, you become part of it, and you don’t get a chance to change. Everyone on the street, like us, and they have no chance to change their lives. It’s like a black hole’ (Dave, wave 1)
Charles expressed a similar, and no less problematic, sense of ‘belonging’ to the street. After becoming homeless he was offered a tenancy which he refused due to its location; unbeknown to him this resulted in his removal from the housing waiting list and a long civil court case against the Council. Over a period of almost three years rough sleeping, Charles established a regular place to ‘kip’ (sleep) and a strict everyday routine which included visiting Project A daily to store his belongings and use the showers. Charles was fiercely independent in the first interview, repeatedly emphasising his desire to do things for himself. Street-life, he insisted, was something that he had adapted to. It had become his life and he could not imagine – at that point – an alternative way of living:
‘I’m trying to keep myself together, totally on my own […] You get acclimatised, and not just to the weather. This is my life’ (Charles, wave 1)
Similar to Charles, Bobby had (in his words) ‘ditched’ his old life. He left his home and family after losing his job and subsequent heavy drinking. He started walking from city to city, experiencing different homelessness services as he went. Over time he had become accustomed to his new life, having learnt the best places to kip, how to keep clean, how to beg, who to trust and so on. Bobby had also established friendships on the street with individuals he met regularly and with whom he shared local knowledge and experiences. Yet Bobby stressed the importance of ‘keeping your distance’, and refused offers of beds in shelters and hostels so as to ‘avoid trouble’.
While the experiences of each case are unique, and narratives included accounts of the stigmatisation and physical harms associated with rough sleeping, each discussed the street as a site where they also created and practiced friendship, relationships, camaraderie, laughter and intimacy to at least some degree. Given the long-term nature of rough sleeping in each case, it had become bound up with their own history, being the place they had lived for a significant period of time. In the face of exclusion and marginalisation from services, all four had sought to re-negotiate their own sense of who they were and where they belonged as a means of adapting to, coping with, and ultimately surviving life on the street.
‘Totally on my own’ – navigating everyday homelessness services
The identities acquired on the street evolved over time. To use Charles’s words, spending so long on the street he had become ‘acclimatised’ to rough sleeping. Barbara felt the same, noting during her first interview that she had ‘forgotten about normal life’. The loss of a past life sat alongside a stoicism about their situation and a focus on survival:
I know there are people in here that are desperately clinging to the trouser edges of anyone on the desk, ‘Oh, please help me’. I think I don’t need that so […] I don’t ask for help. I don’t cry for help. (Bobby, wave 1).
This espoused desire to ‘get through it alone’ was, in part, reflective of a wish to retain dignity and autonomy insofar as possible in incredibly difficult circumstances. Nonetheless, daily life required a schedule of visits to soup kitchens, health centres and welfare appointments which, when missed, affected eligibility for benefits. A further challenge in this schedule was the need to find and access a computer on which to make and check welfare benefit claims, apply for jobs and bid for social housing.
Charles, whose physical and mental health had been worsening, found this daily ‘grind’ ‘exhausting’. He frequently missed appointments or behaved aggressively in meetings, which resulted in benefit sanctions and removal from a hospital waiting list. As a consequence, Charles frequently drew on the support of Project A to make telephone calls and write letters to challenge service providers’ decisions. Similarly, Bobby voiced disdain for the statutory homelessness service, which he saw as purely perfunctory, cementing the impression that he had to get by alone:
‘It was ‘Here’s a piece of paper, fill that in and then naff off’ was pretty much what it felt like’ (Bobby, wave 1)
Barbara and Dave were similarly disaffected and the longer they spent on the street the less they felt able to access and engage effectively with available support:
Barbara: Nah, I wasn’t using any services [before Project B].
Dave: I used to go to the drop-in, but I was just dropping in. I wasn’t engaging with anyone before. I was nervous you know, not about the workers, I was just so used to being part of the street. (Joint interview, wave 1)
Turning points and support
A key benefit of QLR was that it enabled us to trace how participants began, over time, to engage with support and navigate their way toward what they described as a ‘normal’ life. In the first interview, for instance, Charles used project A daily but his engagement was limited. His mental health was deteriorating, he was behaving aggressively to project workers and had presented at hospital for self-harm and suicidal thoughts. The ‘turning point’ was described in the second interview, after he was attacked with eggs while he slept on the street:
‘So I get up and I put it all [belongings] in the skip. It took me a while but I got there […] I’m hardly nineteen anymore. I said, ‘I’ve had enough’ (Charles, wave 2)
Project A was able to work quickly to move Charles off the street, first into supported temporary accommodation and then into his own housing association tenancy. By wave two, it was possible to see both change and continuity. Routine was still important to Charles, yet expectations of ‘his life’, his sense of self, and future possibilities had changed dramatically:
‘I lay there [on new bed] and I said ‘Jesus, uh huh!’. I change my bed on a Friday, once a week. I make sure all my washing’s done and all of that. A routine’ (Charles, wave 2)
Bobby made a similar transition. At wave two he had successfully moved into a private rented flat with a deposit provided by the local authority. Unlike in the first interview, Bobby had in the intervening period become less reluctant to accept help from Project A. While still independent, his relationship with the staff had developed and they had become a crucial safety net:
‘Best thing about Project A is that it is always there […] it’s like a really nice comfort blanket. You can go there and talk to people and be involved. It feels great’ (Bobby, wave 2)
The importance of ‘being there’ was also evident during the waves of interviews with Barbara and Dave. They had initially refused to engage with supportive interventions, describing with some humour hiding around corners to avoid project workers. However, the service continued to visit the street, gradually building up trust and rapport until such a point that the couple felt ready to engage:
I was hiding from them…He just kept on at us, he never gave up on us. I was going round corners and going, ‘Oh nooo! Not him again.’ But he just kept at it, and eventually I decided to let him help. (Dave, wave 1)
In the first wave, Dave spoke repeatedly about the need to ‘sort himself out’, attributing blame for his addiction and homelessness squarely on his shoulders. Notably, by the final wave Dave begins to acknowledge the wider factors involved in his homelessness. In particular, he talked – albeit briefly – about the abuse experienced as a child in care, and the growing realisation that these experiences had contributed to his addiction, homelessness and reluctance to accept support. This is one, of several examples, of where through QLR we were able to gain a visceral snapshot into individual’s relationship to support, and shifting expressions of their past and future selves.
The rough journey to ‘home’
Somerville (Reference Somerville1992) notes that whilst a key signifier of ‘home’ is shelter, the concept has wider symbolic dimensions, including roots, abode, privacy, hearth and heart. These featured, in part, in participants’ accounts of rough sleeping and street life. All four cases had come to regard the street and homelessness as a source of identity and meaning (‘roots’), albeit in ambiguous and problematic terms given the deleterious effects on their health and wellbeing. They also had regular places to sleep (‘abode’), and were making attempts to gain autonomy and control over their situation (to establish ‘privacy’). Absent from their early accounts was ‘hearth’ (emotional and physical well-being) and ‘heart’ (loving and caring social relations), both of which emerged as the four engaged in support and gained their own tenancy.
This transition is illustrated by Barbara and Dave who in the first wave reminisced with stories about ‘enjoying’ life on the street. As their relationship with the researcher developed, the couple began to speak more honestly about the extent of their drug use, the associated challenges and the multiple health problems accrued after many years living on the street. The couple also opened up about their ‘real’ views of the accommodation they had been allocated at different stages of their journey. In wave one, they were keen to stress how grateful they were to be in temporary accommodation, and expressed only a few concerns about the poor quality of the housing, which was observed by the researcher. However, by wave two the couple had moved to a permanent tenancy, where they had been able to choose the furniture and decor, and the quality was markedly higher than their previous tenancy. Only at this point did they talk openly about the previous property, describing it as ‘bogging’ (horrible). In the third wave, they reaffirmed these views. This emergent familiarity and comfort with the research process was accompanied by a new lexicon for describing everyday life. Words such as ‘contented’, ‘settled’ and ‘happy’ feature in their third interview, and for the first time in several years Barbara and Dave were managing their drug and alcohol use, organising their finances and eating well:
‘When I get my Jobseekers and when Barbara gets her PIP [social security benefits] we just go and get the freezer stocked up and then we know what we can spend. We put in £20 every two weeks into the electricity and that does us’. (Dave, wave 3)
Similarly, by wave 2 Bobby had started to re-build his social relations including making contact with his children, volunteering at the local church, and working part-time as a cleaner. Charles had no family, but was continuing to engage with his new housing association and, significantly, had been attending a course. This was a remarkable shift, and something that Charles, in wave 1, could not have envisaged himself doing.
Yet for all four individuals the rough journey to ‘home’ did not end with the provision of shelter. For Bobby the move into his own tenancy meant losing his sense of purpose, something previously driven by the need to survive on the streets. At the final wave, he admitted that despite loving his flat he ‘felt lost’:
‘I’ve got a TV, I’ve got a kitchen but it’s when you sit there alone and you have those darkest thoughts. I do anything I can to be out of the house to be honest’ (Bobby, wave 2)
At this point Bobby was able to share with the researcher that he was still regularly using Project A for practical support (e.g. free showers and phone charging). He had chosen not to inform workers about his tenancy for fear of being denied access. A deeper issue was that the project represented a source of support and a connection to an identity that Bobby concomitantly hated, yet felt afraid to lose.
Charles expressed a similar sense of unease. Despite ongoing mental health issues he was only receiving basic tenancy support (for example, support organising white goods and setting up electricity). At the second interview, he described the adjustment to the tenancy as difficult, noting that he found it challenging to leave the house:
‘Well, I feel like when you’ve been on the street and you’ve suddenly got this place, it’s a bit of a, is this really for me? I had all this before but it was my stuff. Anyway, everything’s pretty and beautiful shelves and that, and, cheers.’ (Charles, wave 2).
Dave and Barbara also described feeling isolated and lonely as a result of the loss of social connections from the street as they tried to stay clean from drugs. They fretted about whether they ‘deserved’ their flat:
‘I would like to have two, two seaters so we can have more room for people to sit, and I think we should get two leather sofas … I feel like it is my home, but you will never understand, that took a lot of time for me.’ (Barbara, wave 3)
The difference in the transitions experienced by Charles, Bobby, Barbara and Dave relates to the sustainability of the support they received. Project B continued to offer Barbara and Dave intensive support throughout their journey and by the final interview was gradually ‘pulling back’ with the aim of enabling the couple to be more autonomous and self-reliant. Nonetheless, Barbara and Dave continued to feel supported and aware of the ‘comfort blanket’ that Bobby described as so important. Charles and Bobby, conversely, had to move into new tenancies without (formally at least in Bobby’s case) ongoing support from Project A and the team of workers with whom they had developed strong and trusting relationships. The evaluations thus add to a growing body of evidence that intensive flexible support that ‘sticks with’ homeless people in the long-term, even if their behaviour is ‘difficult’ and/or engagement intermittent, is effective (Mackie et al., Reference Mackie, Johnsen and Wood2017).
Benefits, challenges and risks of QLR
In the final part of this article we reflect on the benefits, challenges and risks of QLR. Scholarship in QLR is represented by a wide range of disciplinary and methodological traditions, bound by its interest in, and engagement with, time. Given this diversity, the proceeding discussion focuses on those issues which have emerged from our own engagement with prospective qualitative longitudinal interviewing in homelessness evaluation. These reflections are used to consider QLR methods more broadly, and their contribution to social policy research.
Methodological benefits
Writing in the 2007 Social Policy and Society themed section, Lewis (Reference Lewis2007) notes that a longitudinal approach can help identify how, when and why changes occur over time, making it especially useful in policy evaluation. Not only can QLR explore individuals’ changing relationship with services and outcomes over time, but it can highlight the dynamic relationship between individual, service, policy and structure. The temporal interaction described by Lewis (Reference Lewis2007) fits the conceptual shift that has taken place in understanding homelessness. Far from being a static event or linear journey, there is now an understanding that ‘[h]omelessness, mobility and spatiality shifts over time, often charting complicated pathways into and out of different accommodation, different ‘resting places’ (Cloke et al., Reference Cloke, Milbourne and Widdowfield2003: 32).
Our studies drew similar conclusions. ‘Exiting’ homelessness was a complicated pathway and for those with extended experiences of rough sleeping and complex needs a tenancy does not constitute ‘the end’. In both evaluations, the central benefit of qualitative longitudinal interviewing was in capturing the dynamic relationship between the individual, services, policies and structures described by Lewis (Reference Lewis2007). By ‘walking alongside’ participants, albeit for a short time, we were able to examine different stages of engagement with homelessness services, consider perspectives before and after events, and explore shifting identities and expectations. For example, at the first interview, Charles sees the street as ‘his life’. In the period between interviews, the interplay between individual experiences and support services provoked a critical change in his living circumstances and sense of self. At the same time, his home and what Charles anticipates for himself remains fragile. A qualitative longitudinal lens is especially well placed to reveal these sorts of transitions.
In this way, QLR ‘provides access to the ‘interior logic’ of lives, discerning how change is created, negotiated, lived and experienced’ (Neale, Reference Neale2019: 9). QLR, in other words, recognises not only the important role of human subjectivity in shaping how individuals navigate and experience services and policies, but it sees this relationship as temporal, dynamic and shifting. For example, in both projects we identified dynamic expressions of agency and independence. In the early interviews, participants sought to emphasise their capacity to act, to make decisions, shape their own lives and defend their decision making. Examples from across the cases, and projects more widely, included refusals of temporary accommodation, rejections of support and sustained expressions of stoicism, self-sufficiency and freedom to choose. Micro-expressions of agency were, however, set against and interplayed with a backdrop of socio-structural factors, in which institutions, policies and regulations interacted to prevent, delay or support progress. As individuals developed their relationships with the projects, the later interviews were critical in illuminating the processes, ‘turning points’ and/or individuals that facilitated progress. In particular, the interviews revealed the shifting role of support over time, as individuals built trust and respect with workers, overcame their reluctance to accept help, and recognised their rights and entitlements. Latterly, intensive support assisted individuals to overcome inadequacies in provision further ‘downstream’ (such as the lack of affordable housing).
The temporal frame of the projects was relatively short; indeed, in project A participants were only interviewed twice, with some informal contact between. To use the case of Bobby, in spite of the time frame for the evaluation, QLR made it possible to examine the ways in which his values, perceptions and strategies for navigating homelessness interplayed with the structural factors that first kept him on the street, and then enabled his movement into a tenancy. Understanding human subjectivities might seem a primarily academic endeavour, yet they are pertinent to evaluation and policy-focused research. Recognising the richness of subjective accounts can help those without lived experience of homelessness understand life on the street. They lend insight into why someone living on the street might choose to reject offers of support, avoid temporary accommodation or hide from support workers. These accounts bring to the fore the lived realities and intended and unintended consequences of policy (Thomson, Reference Thomson2007). Thomson (Reference Thomson2007) refers to this as the ‘long view’ offered by QLR, an aspect illustrated well by Barbara and Dave. Their accounts reveal life on the street as simultaneously a source of misery, yet also a locus of (fragile) support networks, meaning and identity. By meeting Barbara and Dave on multiple occasions, it became possible to illustrate more clearly the challenges service users face in moving away from street-life, such as the loss of connections, loneliness and social isolation. These, in turn, provide evidence with which to lobby for more appropriate service provision.
A further aspect of QLR that can benefit social policy and evaluative research is that changes and continuities in understandings can be examined from different points in time. An example of this, discussed above, is Dave and Barbara’s changing description of their temporary accommodation across interview waves. In the case of Bobby, we can also observe continuity in terms of his future-orientation. He remains, across waves, committed to self-improvement; to proving to himself and his children (who he had not seen for several months) that he could get his ‘life back together’.
A final benefit identified from our use of QLR was the ability to give those considered ‘vulnerable’ the time and space to share their narratives. Welfare and social services (for example, access to disability allowances or universal credit) too often expect individuals to narrate their ‘story’ as an entire or finished entity. QLR, conversely, was used in these evaluations in a person-centred and flexible manner, allowing participants to tell their story at their own pace. While retrospective interviews can, of course, adopt a similar ethos, our experience was that multiple interviews afford a richer time lapse of a period in someone’s life; something particularly useful in bringing to light shifts in identity. This helps people to look back, as well as forward, at different points, promoting reflection, and supporting individuals to arrive at or come to terms with what could be regarded new insights about themselves, or to mature perspectives. For example, as discussed it was only in the final wave that Dave was in the position, perhaps because he was stable and now had adequate support, to look back on his difficult past and confront historical experiences of abuse. Likewise, in the second interview Bobby felt comfortable revealing his need for a ‘comfort blanket’, and shared his anxiety about the potential loss of support from Project A, despite having moving into a private rented property.
Challenges and risks
While both projects revealed the value of QLR to homelessness research, the work was challenging, most significantly in terms of practices of care. As a methodology, QLR can amplify the ethical concerns and challenges expected in one-off interviews with participants. As Neale (Reference Neale2019: 85) notes, it is within QLR that the usual protocols of qualitative research can ‘quickly unravel’. This unravelling requires researchers to be vigilant to balancing ‘reciprocity and professional boundary maintenance’ (Neale, Reference Neale2019: 78). We have emphasised the benefits that building up trust and familiarity with participants can bring, and the ways in which it can generate rich, nuanced data. Yet this same intimacy can introduce the potential for harm through the unnecessary disclosure of information to researchers, emotional involvement between parties, dependency and problems with ending contact. While these issues are likely more significant in complex or long term QLR, they were encountered in both projects discussed here. Researchers would, for example, enact everyday acts of kindness, such as buying a coffee or sandwich, providing help understanding an official form, supporting access to a computer, or folding clothes. On one occasion the researcher when asked for help drove a participant to the methadone clinic and purchased basic shopping. While procedural ethics were followed, ethics in practice requires a deeper reflection on the extent to which these acts of kindness might interfere with professional boundaries, and more significantly the voluntariness of participation.
It was keenly felt by researchers that the conversational nature of interviews resulted in participants sharing extensive and intimate aspects of their lives. Often this included events and details unnecessary to the specific aims of the projects. These interactions undoubtedly allowed emotional connections to be formed with participants, their stories, and their experiences ‘churning’ within the homelessness system. Emotional labour can, of course, be present in all forms of qualitative research. However, repeat visits coupled with ongoing contact with the project on the progress of individual cases afforded a richer understanding of participants, and in turn, a deeper familiarity with their lives and experiences. This highlights concerns about the ethics of care and respect in QLR. As QLR researchers we become involved in the lives of those we study using methods lauded for their relationality and nurturing qualities. Yet these same processes have the potential to expose participants to exploitation and harm. Lessons from this study include the importance of ongoing ethical reflexivity.
A particular ethical consideration with QLR is how to end contact sensitively. Despite the relatively short time frame of the studies, for some interviewees saying ‘goodbye’ seemed abrupt. With Charles, for example, the interviewer began to close the final interview by asking him if there is anything else they should have asked, and Charles replied ‘Has this been a good year for you?’, attempting to keep the conversation going and engage the researcher to find out more about them too. This also highlights the inevitable power disparity that exits between the participant and researcher. A short meaningful conversation ensued between Charles and the researcher, and when the interview was finally closing Charles asked ‘I won’t be seeing you again, will I? That’s it?’. Similarly, in Project A, the researcher ‘bumped into’ Bobby several times after the research. Despite such encounters being welcomed, interactions felt awkward. While no longer governed within the parameters of the objectives of the evaluation, the researcher described feeling unnecessarily weighed down by the intimate knowledge of Bobby’s personal life, and his ongoing (unauthorised) use of Project A’s services.
In terms of the challenges for the researcher, involving homeless people with ‘complex needs’ in QLR meant that recruitment and organising interviews was extremely time consuming and involved emotional labour (Bergman Blix and Wettergren, Reference Bergman Blix and Wettergren2014). In project A, recruitment involved spending significant periods ‘hanging out’ at the drop-in, chatting informally about the study and what it would involve. In project B, recruitment was less protracted, since all participants had already been allocated keyworkers who supported the research. However, service users were dispersed which meant fieldwork required intensive time commitment, visiting service users wherever they were – including prison, hospital, support services and their own accommodation. We adopted a minimal but carefully targeted approach, texting people directly and, in particular, using the support of project staff to set up follow-up interviews. On one occasion, the researcher spent the day with a project worker, searching for previous interviewees in various ‘kipping spots’ across the city (without success). There was an element of attrition across both studies, although this related in part to service users having moved from the area and no longer being contactable. Two had very sadly died. Significant fieldwork time was accumulated due to cancelled interviews or ‘no shows’, an experience which emphasises the need for QLR involving homeless people to be adequately resourced and flexibly designed to accommodate the circumstances of those with complex needs.
Of course, while time can create ethical challenges, it can also provide researchers with solutions. Unlike evaluations that involve one-off interviews, in these studies the longitudinal element provided researchers with more time to develop sensitive, ethical frameworks that were situated at the heart of the research. This approach was supported throughout, by the projects and their staff. Project workers, for example, were able to support researchers in scheduling interviews at times and places that met the mental and physical health needs of participants. As project workers were frequently in touch with participants this also meant that they could update us (with consent) on any critical moments (for example, imprisonment, relationship breakdown, hospitalisation). With this information, we were able to cancel interviews – even at the last minute – thus ensuring participants were never interviewed at an emotionally challenging time.
Further limitations are that the two projects discussed were commissioned evaluations. The commissioned outputs are not in the wider public domain, which limits their influence. QLR is undoubtedly time consuming and expensive, and it has been said that when working with vulnerable individuals it is unethical ‘to divert any energy from surviving into doing anything that does not have the potential to be useful to them’ (Martin, Reference Martin1998: 3). Moving forward we face the continuation of austerity and the growing likelihood of a recession post-Covid 19. It is critical for social policy research to have a means not only of understanding how policies in this epoch are being navigated, but also the outcomes they are delivering for whom in what circumstances (Pawson and Tilley, Reference Pawson and Tilley1997). In giving time to individuals to tell their stories as they unfold, something that they may otherwise find difficult, this method allows for those who are often not heard from to have their stories heard by, and where necessary confront the assumptions of, policy makers.
Conclusion
The purpose of this article was to reflect on the potential methodological ‘added value’ of QLR, and the practical and ethical challenges it presents in the study of ‘vulnerable’ populations. It has brought into analytic conversation case studies from two longitudinal qualitative studies of homelessness. The approach taken was not an attempt to make qualitative data ‘big’, but rather a means of making comparisons, introducing cross-contextual generalisations and finding common themes, experiences and challenges. The cases presented provide a time lapse, charting the rough journey ‘to home’, from being on the street, navigating everyday homelessness services to engaging with meaningful help and obtaining a tenancy. Empirically, they highlighted the highly complex, often ambiguous and evolving relationships that participants had with street-life and street-based identifies, as well the range of factors impinging upon or fostering the establishment of a sense of home after settled housing was obtained.
The methodological benefits of QLR included capturing the dynamic relationship between the individual, services, policies and structure. QLR proved to be well placed to consider perspectives before and after events, to explore shifting identities, human subjectivity, changing expectations and to reveal transitions. This method also gave vulnerable individuals more time and space to share their narratives than is typical in one-off retrospective interviews. These rich subjective accounts can help those who have no experience of homelessness, including policy makers and other stakeholders in positions of influence, to understand lived realities.
QLR also presents challenges, and the greatest concern is related to the ethics of care for participants, with the blurring of boundaries reinforcing the need to give due emphasis to ethical reflexivity. Ending contact sensitively can be particularly difficult, and for some, and those especially vulnerable, the emotional labour and time invested can make any ending hard to endure. For the researcher, balancing recruitment and retention with the need to maintain and reaffirm boundaries is a challenge, and the emotional labour and time invested through repeat visits and ongoing contact substantial.
It is imperative that social policy research has a means of understanding ‘what works’ and how interventions are being experienced, especially as we face potential continuation of austerity given the predicted post-Covid 19 recession. Our analysis not only re-emphasises the valuable contribution that QLR can play in this regard, but highlights the wider opportunities created by bringing QLR datasets into an analytical conversation. Now, more than ever, a case can be made for exploiting the benefits of QLR, to close the distance between policy makers and those affected, whilst remaining mindful of the challenges and risks when employing the approach with vulnerable populations.
Open access
