Change and continuity over time

Longitudinal research involves collecting data over time, usually prospectively although sometimes retrospectively. But, as we noted in the themed section, the time period that such research covers can range from relatively short time periods to many years. The appropriate time period depends on the purpose of the research. We also discussed how interpretation of the data evolves, perhaps even changes, over time, both for participants and researchers.

Our study of lone mothers and their children in effect fell into two parts. The first three rounds (with interviews in 2002/3, 2005 and 2007) were focused on the transitions to work, on whether and how the women could sustain work, and on what this meant for the lives of the women and the children. As noted above, those were years of policy expansion, especially for working families, as the Labour governments aimed (with varying degrees of success) to increase employment rates, to make work pay, and to eliminate child poverty. The second part involved just one further round of interviews, in 2016, with a selection of the families we had interviewed previously. In fact we had not intended to continue the research after the first three rounds. But the banking and financial crisis of 2007/08, and the austerity environment after 2010, prompted us to want to explore what this had meant for our participants. Finding all the families again would have been challenging and resource intensive, so we focused on fifteen cases, chosen to reflect a range of circumstances when we last spoke to them in 2007. We thus had a purposive sample, a longer time-period of data, and a changed policy context. This gave us a wider lens and broader perspective to understand how continuity and change differed over the short and longer-term.

In our analysis of the first three rounds we had seen lots of change in the lives of the participants (Ridge and Millar, Reference Ridge and Millar2008, Reference Ridge and Millar2011). Getting into work was not a simple linear process: rather, for many of the women this was extended over time, and involved lots of changes of jobs, or of working arrangements (for example, in hours, days or work places). Sometimes it was the work itself that was unstable, and sometimes it was the women who were trying to find the best fit between their work and their family lives. The children also experienced a lot of change, in their daily lives and arrangements for childcare, and in the time spent with other family members (Ridge, Reference Ridge2009; Millar and Ridge, Reference Millar and Ridge2013).

But by the time of the final round, in 2016, for the mothers it was continuity rather than change that was most striking. The women had mainly stayed in work, and some had increased their earnings and moved into more senior jobs. But many were still on wages at, or not much above, the national minimum wage ^{Footnote 2} . And this meant that their incomes had stayed much at the same level, or fallen when for example child benefit and child tax credit ended as children grew up and left home. This had major implications for their futures, as retirement was coming closer, and for their capacity to help and support their children into adulthood.

Indeed for the young people, who were mostly in their early to mid-twenties in 2016, we again saw lots of change in their lives, including in living arrangements, in work/study, and in partnerships and parenthood. These were often challenging to manage and while the mothers and other family members helped, their resources were often limited. The impact of the reductions in state support for families and young people, under the austerity agenda, were very apparent.

Thus the longer-time frame highlighted that, while there may be a lot of change and churn over the short term, there is often limited capacity to effect a significant and lasting improvement in income and material circumstances over longer periods of time. This longer-time frame also reinforced our conclusions about the importance of income security in people’s lives, in that people could only plan, over short and long term, if there was a degree of certainty in their financial circumstances. This is a challenge to the Universal Credit design where the amount received changes monthly if earnings and circumstances change.

Presenting longitudinal qualitative evidence, engaging in policy debates

Over the years of the research we have been involved in seeking to put evidence from this research to a wide range of individuals and groups engaged in the policy process, including civil servants, politicians and third sector organisations. The research was funded by different external bodies at the various rounds which opened up a range of audiences and channels of access. The first two rounds were funded by the Economic and Social Research Council (RES-000-23-1079), which helped to establish the academic profile of the research, and to start the process of making contact with a wide range of people and producing various types of publications. The third round of interviews was funded by the DWP, which meant that we were able to work closely with the researchers within government on developing the topics to explore in the interviews, and were able to present analysis and findings directly into the DWP and other relevant government departments. The final round was funded by the Joseph Rowntree Foundation (JRF), who are committed to the dissemination of research into policy. Our Advisory Group for the JRF-funded round included representatives from government and welfare rights organisations, and the format of the report and especially the summary ‘Findings’ were designed to reach a wide audience (Millar and Ridge, Reference Millar and Ridge2017).

The use of research evidence for policy is complex and dependent on a range of factors, including context, institutional culture, individual competence and perceptions of political feasibility (Stevens, Reference Stevens2011). Monaghan and Ingold (Reference Monaghan and Ingold2019: 361) suggest that civil servants may feel that qualitative data are important but are not always confident in putting this forward, that they are ‘aware of the potentially rich picture that could be provided by a broad range of evidence but have concerns over their capability to organise this evidence for Ministers’. In our experience, as the longitudinal analysis proceeded we found that the use of case studies – accounts drawn from the individual participants – was a strong and effective way to summarise and present the complexities of experiences, circumstances, and choices. This was true in both analytical and presentational terms. Analytically case studies allowed the space for an in-depth analysis that can focus on how processes and relationships build, change, consolidate or break-up over time, as the context and so opportunities and constraints change ^{Footnote 3} . And the presentation through case studies gave a narrative structure that could capture the imagination and enable connections to be made. As we presented the research we often had follow-up conversations with audience members who spoke to us about being a lone parent or growing up in a lone-parent family and recognising the accounts – ‘that was me’. This sort of ‘authenticity’ is an important source of validation of researcher interpretations of qualitative data (Tierney and Clemens, Reference Tierney and Clemens2011).

However, the selection of case studies requires careful and clear justification, in order to convince that these are a valid representation of the research. The rationale for the case selection needs to be both transparent and purposive. For example, we were interested in whether and how the families were supported by the policy provisions in place intended to support work. This included the system of tax credits, which formed the main in-work benefit at the time (now increasingly superseded by Universal Credit). We selected just one case for a ‘close-up view’, which was a lone mother who worked in accounts: ‘I’ve always done accounts since I left college myself and I’ve got my ‘O’s and my ‘A’ Levels and book keeping and accounting and everything’ … [with] access to the internet at home and at work’ (Millar, Reference Millar2011: 39). Thus we argued this woman was ‘not someone likely to be intimidated by form filling or by complex calculations’ and the challenges that she faced were therefore a good test of the problems inherent in the system.

This is an example of selecting a case that can test the interpretation of the data, not by being typical but by having characteristics that provide a focused lens on key issues. However, contrasting cases with some shared characteristics is also an effective way into the analysis. For example, when we wanted to look at the factors that helped and hindered access to higher education, we concentrated on two of the young people with similar family circumstances but who had made different decisions and followed different paths. This enabled us to explore and to highlight the interactions between family, school experience, motivations and opportunities (Millar and Ridge, Reference Millar and Ridge2017).

Case studies also need to be placed in context: this can mean in the context of the research itself, where the analysis of the case should go alongside thematic analysis, each reinforcing the other. It can also mean, in the context of other qualitative research, other studies where similar issues have been explored and analysed. At the minimum this involves referencing other published research, but can also go beyond this, to meta-analysis and to combining data sets. For example, Ellis-Soan (Reference Ellis-Soan2019: 98) synthesises data from six studies of teenage mothers in later life. This, she argues, not only provides a ‘complex and nuanced picture’ but also gives ‘voice to young mothers so they can narrate their own lives and make clear what is important to them. Patrick and Wright (Reference Patrick and Wright2019: 599) make the case for new analysis across different qualitative data sets: ‘Combined Study Qualitative Longitudinal Research as a new methodological approach’ which seeks out the ‘shared typical’ and is thus able to ‘extend inference beyond the usual study-specific confines of qualitative generalisation’. This is a promising methodological route which, they argue, ‘reaches beyond the uniqueness of the individual and the particularities of their circumstances to reveal broader tendencies of major consequence… it is possible to aggregate findings across time and from multiple studies to explore whether there is evidence of an underlying essence of broadly-shared lived experience that could constitute a coherent big picture’ (Patrick and Wright, Reference Patrick and Wright2019: 609–10).

However, it is placing qualitative longitudinal cases in the context of quantitative data – from surveys and administrative records, for example – that is more common and perhaps also most convincing to policy-makers. Discussing family policy research in the US, Bogenschneider and Corbett (Reference Bogenschneider and Corbett2010: 207) make a pithy argument for this: ‘We think of research illustrated with representative case studies or personal stories as a ‘one-two-punch’. With a convincing story you get a left! With rigorous research, you get a right! With a convincing story and rigorous research, you get a knock-out punch!’. This sort of approach – the survey data to set out the trends and the qualitative data to get behind the statistics – can provide compelling evidence. But we must be careful not to reduce qualitative data to a form of illustration, servicing the cold hard ‘facts’ with some lived experience ‘colour’. This can serve to reinforce a view that qualitative research is, by itself, of limited value in tackling policy questions and issues. We need also to make the case for qualitative data for policy analysis in its own right.

Reflections

Our experience with almost fifteen years of data collection highlighted the diverse ways in which longitudinal data were able to enhance and deepen our understanding of the relationships between social and economic conditions, policy and everyday lives – between historical time and personal time. Revisiting and re-interviewing the families over time meant that we were able to build up confidence and trust. The revealing of personal history meant that the factors affecting their lives and choices became more apparent. The experience of domestic violence was one example of this, as it was only in later interviews that we began to get a sense of how much this had affected the women when they were first lone mothers and trying to establish their financial independence.

We also saw how family relationships, in particular between the mothers and their children, reflected complex and subtle dynamics. The mothers were very committed to providing for their children and giving them a good start in life. But the mothers often lacked the resources to protect or support their children into young adulthood. For the children the impact of financial insecurity in childhood could cast long shadows. The close and supportive relationships between mothers and their children was not without tensions and strains that played out in sometimes close and sometimes more distant relationships over time.

We could also see how these families were so deeply affected by the changing policy environment since the turn of the century. The ways in which politics and policy shape, support or constrain opportunity and choice can only be fully revealed by following people over time. Longitudinal qualitative approaches can thus provide rich insight into these different perspectives on, and experiences of, time, continuity, and change.