Utilizing data from a lost sociological project: experiences, insights, promises

In 2000 data from a little known sociological study was ‘re-discovered’, stored in an attic office. The archived data comprised original interview schedules that documented the early work experiences of Leicester’s youth in the 1960s. Forty years on, the original respondents have been traced and re-interviewed as they make the transition from work to retirement. This article examines the complex methodological issues associated with reanalysing, tracing and reinterviewing respondents after such a considerable time lapse. We examine our methodological approach using the concept of qualitative longitudinal research (QLR) as a framework for understanding this process. We also reflect on the value of such longitudinal qualitative research. We conclude by drawing out some of the issues surrounding QLR and the implication of our experiences and insights for those who are now building such longitudinal datasets and the promise such data hold.


Introduction
Recent interest in the importance of archiving qualitative data and the associated methodological approaches to revisiting such data has created significant debate among social scientists as to the value to reusing qualitative data (Hammersley, 1997;Laub and Sampson, 2003;Fielding, 2004;Heaton, 2004;Savage, 2005). At the same time associated approaches to qualitative research such as the role of qualitative longitudinal research (QLR) have also attracted increasing attention (Neale and Flowerdew, 2003;Farrall, 2006). As a contribution to these debates this article focuses on our recent foray in to the field of qualitative data reuse and QLR. The discussion presented here draws upon our discovery, in 2000, of data from a large-scale yet little known sociological study carried out in the 1960s. The dataset, stored in cardboard boxes in an attic office, originated from the project 'Adjustment of Young Workers to Work Situations and Adult Roles' led by Norbert Elias at the University of Leicester between 1962 and 1964 and funded by the DSIR. Never fully analysed, the dataset comprised of 894 original interview schedules that document the work and life experiences of Leicester's youth in the 1960s plus original supporting documentation relating to the project such as letters, memorandum and notes. This dataset has subsequently been used to form part of a wider ESRC project 'From Young Workers to Older Workers: Reflections on Work in the Life Course'. This new study, carried out some 40 years later, involved the reanalysis of the 1960s data and the tracing and re-interview of respondents who were now making the transition from work to retirement.
Our 'new' dataset therefore provides individual narratives of the life course spanning complete work histories. However, despite the quality and breadth of the data we have continuously grappled with the methodological issues involved in reusing and reanalysing data and then tracing and re-interviewing respondents from a study which took place almost five decades earlier. As such, this is a project that has proved hard to define in methodological terms. Our difficulty stems from the lack of clarity within existing literature as to how best to classify and define research which uses data both retrospectively and prospectively. Definitions of the various approaches to such research are unclear and there is an ongoing debate not only of the value of revisiting existing data but also of the methodological framework underpinning such studies.
To explore these debates further the article is structured in the following way. First, we explore the methodological approaches outlined above and move towards providing an explanation of our own approach. We follow this discussion by outlining the methodological framework used in this study and by describing the process of tracing and recontacting a cohort of respondents 40 years after initial contact. We then reflect upon the process of reinterviewing respondents and explore the value of such research for examining and understanding change over time in the context of employment and work histories. We begin by exploring in more detail the issues surrounding qualitative longitudinal research.

Qualitative longitudinal research
Recent interest in the reuse of qualitative data or the secondary analysis of qualitative data has generated substantial debate in the methodological literature. This is evident in the publication of journal special issues on the subject of data reuse (see, for example, Sociological Research Online 12(3); International Journal of Social Research Methodology 7(1); Forum Qualitative Sozialforschung/Forum: Qualitative Social Research 1(3) and 6(1). Access to archived data in the Qualidata Archive has enabled researchers to revisit and reuse data from classic studies of the past, such as: Thompson's study of the Edwardians (1975) and Goldthorpe et al.'s (1968) seminal study 'The Affluent Worker', in order to make use of wholly pre-existing data by interrogating it with new research questions. Alongside the relatively recent interest in data reuse, there is a long tradition in social science research of carrying out restudies and replications, particularly of classic sociological community studies (see Crow, 2002 for a detailed discussion). By contrast to the approach of reusing existing qualitative data, restudies and replication are based on the creation of new data generated by new respondents but, crucially, using the original research questions and revisiting the same geographical location. Davies and Charles (2002: 1.1) define this approach clearly, adding the caveat that a restudy should not have been part of the original research design: By a re-study we mean a deliberate intent to repeat insofar as possible a previous research study using the same research design and methods to investigate similar theoretical concerns usually with the goal of better understanding social change. Re-studies differ from longitudinal studies in that they are determined upon post hoc, rather than being part of the original research design.
A limitation then of restudies is that while broad patterns of social change may be identifiable it is not possible to understand the impact of these changes on the individual life course because, for the vast majority of restudies, it is the place that is revisited, not the individuals. This is where the new Young Worker Project takes on a very different and valuable perspective to other restudies and begins to move us away from the field of restudies and in to the field of qualitative longitudinal research.
QLR is, to some extent, a combination of the methods outlined above. In common with data reuse, it involves reanalysis of pre-existing data from an earlier study using new research questions. However, in addition, QLR, unlike simple data reuse, involves the generation of new data and the use of the original research questions. The most significant way that QLR differs from both reuse and restudies is that the research is focused on the same respondents as the original study.
Despite the lack of a clear definition, how can we best define our own approach? On the one hand our own approach can be classified as simply a form of data reuse. However, it can also be defined as a restudy or replication. Perhaps most persuasively we can argue that our methodological approach, or at least our approach in the follow-up study, best fits the field of qualitative longitudinal research. QLR, has, like data reuse, attracted increasing attention in the literature in recent years (Holland et al., 2006;Farrall, 2006;Corden and Millar, 2007). Farrall (2006: 2) defines QLR as: 'a range of mainly in-depth interview-based studies which involve returning to interviewees to measure and explore changes which occur over time and the processes associated with these changes'. QLR, he argues, is valuable when 'studying a process which has a notion of a "career" of some sort or which involves a developmental process'. This approach, as Thomson et al. (2003: 186) note, appears to have been used most frequently in studies of youth transitions, 'suggesting a particular relevance … when the process of change is continuous over an ill-defined period of time'. Therefore, we can argue that this methodological framework suited our purpose very well as we were interested in exploring the process of the life course, with an emphasis on work and employment spanning a 40 year period. Yet again, however, definitions remain unclear.  for example, suggest that a study such as ours would not be considered as QLR because the longitudinal dimension to the research was not part of the original research design. They argue that 'what distinguishes longitudinal qualitative research is the deliberate way in which temporality is designed into the research process making change a central focus of analytic attention' (p. 185). Using their rather restrictive definition our approach would be considered simply as a follow-up study. Yet even for Thomson et al. (2003: 185) who initially appear to be very clear in their definition of QLR some confusion remains as they go on to suggest that 'longitudinal research designs can be prospective cohort or follow up studies' and these 'can emerge by both "design and default"'. Neale and Flowerdew (2003: 189) go some way to explaining why there is still confusion and significant debate around the concept of QLR, suggesting that 'the contours of this developing paradigm have yet to be articulated and clarified in any detail'.
Our approach mirrors that of Corden and Millar (2007: 586) who propose a broader and looser definition of QLR, arguing that definitions such as those by  'imply that the longitudinal focus must be part of the study from the outset. In practice, however, longitudinal qualitative studies may come about in more diverse ways'. For them the important defining feature of such research is not that the longitudinal element was included in the initial design but that the research includes 'the capacity to analyse change over time' (Corden and Millar, 2007: 586). This, we argue, represents a fundamental strength of our data. Although the longitudinal element was not part of the original design we now have a dataset which allows us to examine change over time, to examine process rather than outcomes (Thomson, 2007).

The problems of qualitative data reuse
However, despite the obvious benefits and attractions of this approach it also raises a number of practical and ethical concerns -confidentiality, anonymity, the researcher-researched relationship and the possible problem of 'auditing'that need to be considered. First, in qualitative research it is usually those who collect the data who are analysing the data (Dale et al., 1988). During the research process it is usually those in the field who provide the guarantees as to how the data will be used and offer assurances relating to anonymity and representation. Indeed, as Corti et al. (1995) suggests, a main concern with reusing qualitative data relates to the promises made regarding confidentiality. A second concern is that in qualitative approaches the researcher often becomes the research instrument (Dale et al., 1988) and the data produced is the end product of interview, reflection and interpretation. Here the researcher cannot be separated from the data and as such one must question the ability of anyone else to re-analyse and reuse the data. Indeed, the collection of primary qualitative data is based upon the relationship between the researcher and the research but no relationship exists with later researchers who wish to reuse the data. It is also rare for those reusing data to have access to contextual information from the original study and therefore it is difficult, if not impossible for the 'new' researchers to test the arguments of the original researchers. Savage (2005: 3), for example, argues that given the limited archival materials which still exist relating to Bott's work the researcher has 'absolutely no prospect of "testing" Bott's argument'. This, he argues, is because a 'great deal of her reported analysis was based on her intensive and tacit knowledge of the respondents and is not now amenable to be "checked" against the existing data'. Indeed, few researchers would suggest that their aims in reusing data are to test the validity of earlier conclusions. Finally, Hammersley (1997) argues that the secondary analysis process may lead to an 'auditing' of social research, raising ethical problems for the researcher as well as the researched. For Hammersley 'the audit model could be taken to imply that the efficiency and competence of researchers can be assessed on the basis of archived material' (p. 136). Corti et al. (1995) also report that researchers are concerned about secondary analysis due to the possible methodological criticisms that could be made of the original research.
However, we feel that these concerns should not limit our usage or restrict QLR that builds upon data already in the public domain. For example, we have anonymized the data and removed personal identifiers once the tracing process was completed. We have given assurances to the original research team and interviewers that their anonymity will be maintained as far as possible. They in turn have given permission for the analysis to take place and both understand and accept that it is 'not our intention to produce an audit of the efficiency and competence of researchers, nor is it our intention to imply in any way that the original research design was incorrect or invalid' (Goodwin and O'Connor, 2007: 378). We also agree with Heaton's (1998) assertion that, while one of the limitations with secondary analysis of qualitative data may be the inter-subjective relationship between the interviewer and the interviewee, it is often the case that more than one researcher was involved in the generation of the data and, therefore reuse may not be as problematic as first thought.

The Young Worker Project
Where restudies have been undertaken the vast majority have been based upon classic sociological studies. Perhaps the best example of this is the ' Affluent Worker' study (Goldthorpe et al., 1968) currently the subject of a restudy by Savage (2005). It may be that these classic studies have become the most reused and restudied because of the importance they have assumed over time. This scenario seems likely as it is natural that such high profile research would attract the most attention from future generations of sociologists because these studies form the bedrock of modern British sociology. However, it cannot be ignored that another reason for the popularity of these studies lies in the accessibility of the data and its reuse is simply pragmatic; it has been archived and is easily accessible therefore it should be the first port of call for any researcher interested in restudies. Certainly, had other studies been archived in the same way it is likely that these would also have achieved prominence in the field of data reuse. As Corti and Backhouse (2005) lament, too many classic study datasets were lost before the Qualidata data archive was set up.
The Elias project is, however, an anomaly when compared to classic studies of the same time period. While it clearly had the makings of a classic sociological study -it was a large-scale study led by a prominent sociologist of the period; it received considerable funding from DSIR, the predecessor the ESRC; and the project personnel all went on to become successful sociologists -the Young Worker Project differs from the classic studies simply because it never came to fruition as originally intended. There are a number of reasons for this; for example, throughout the research Elias and the researchers disagreed about the theoretical framework underpinning the research. For Elias the project aimed to capture initial work experiences as he felt that the young people's experiences prior to work would not prepare them for employment. However, according to Elias, the researchers had interpreted the research to be a study of 'work' per se and that their approach was to explore causal relationships between the different aspects of the sample (what Elias referred to as 'billiard ball' causality). Elias argued this approach ignored the relationships and processes that constituted the young workers experience and that the researchers were actually undertaking a different project to that originally proposed by him. Elias suggested that the researchers lacked a basic understanding of his approach which, in turn, led the researchers to perceive that Elias had no confidence in their ability to complete the research. As one letter from the research team indicated 'every meeting with you has ended in your saying that we cannot understand your ideas, that we are too concerned with numbers and statistics, too pedestrian, too inhibited by our training' (Keil, 1964: 1). Second, there were methodological disputes. Those collecting the data had previously worked on other large-scale quantitative research projects elsewhere, whereas Elias, profoundly at odds with 'empiricist' approaches, preferred to concentrate his analysis on more detailed cases studies. In an interview for Platt's (1976) Realities of Social Research, Elias suggested the researchers didn't believe his assertion that it was possible to develop a precise and straightforward research instrument that could access the 'experiential side' documenting 'actual' experiences of the young workers. For Elias the value of the study was not the production of statistically significant quantitative data but in the ability to explore the transition to work as the young people experienced and accounted for it themselves. However, the research team were very critical of this 'case studies' approach, arguing that the only valuable data would emerge from a factually based survey with a large and properly constructed sample. As Keil (1964: 1) suggests 'why else is there a need to interview 1,000+ young people if one is only concerned, in fact, with the experiences of 20 or 30 young people'. Finally, there was an ongoing argument between Elias and the research team over the researchers' right to publish. Those who had collected the data wanted to publish articles and to disseminate their early results. This Elias vetoed, preferring instead a final report in which he would acknowledge the 'contributions' of the research team.
These disputes have a number of implications for our research and analytical approach. First, the arguments over publication meant that the bulk of the data was never published nor was a final report ever produced. Indeed, it was some 10 years after the conclusion of the 'failed' project that a publication emerged. The book 'Young Workers' (Ashton and Field, 1976) was based on a small sample of interview transcripts from the project and it became a seminal text on youth transitions. It is fascinating, however, in the context of recent debates, that the Ashton and Field book was, in fact, the product of data reuse as neither author was involved in the original research programme but the data all came from the original project. However, the bulk of the data was never used and, in effect, provides us unused data from the 1960s that we used as the basis for the new QLR project. Although not originally designed with that purpose in mind, this is an effective reuse of the data and means that the original research was not undertaken in vain. Second, the methodological and theoretical arguments led to a vast array of data being collected ranging from rudimentary event history data to more exploratory questions focusing on money, relationships, work, education and family. The compromise reached between Elias and the researchers over the design of the research instrument has ultimately led to the individual young workers' lives being recorded in far more detail than a single survey or number of case studies could provide. We have in effect 894 in-depth interviews that have, in the main, remained untouched since the mid 1960s.

The Young Worker Project revisited
The overall aim of this study was to examine the process of adjustment to working life and retirement of a single cohort of male and female workers over the four decades 1962/3-2002/3. This project comprised two key stages: first the data from the original project was analysed and second reinterviews took place with 100 of the original young workers, now facing retirement. This approach enabled us to explore (a) the adjustment these workers made on first entering the labour market in Leicester in the 1960s, (b) the subsequent adjustments they made to changes in the local labour market and associated structure of opportunities in mid-life, and (c) the ways in which they are currently tackling the approach of retirement in the 21st century.
In order for the reanalysis to take place, it was necessary to invest considerable time and resources to digitally archiving the data where possible. The first stage of this process involved entering all 894 interview transcripts into a database. This was an immense task given that each interview included some 85 mostly open questions alongside often detailed, extensive interviewer notes. Often, as Gillies and Edwards (2005) discovered in their exploration of historical data, the handwriting was difficult to decipher and this made the task considerably more laborious. We know, from the interviewer notes and from paperwork relating to the project, that the interviews were originally tape-recorded. However, as far as we are aware these tape reels were destroyed at some point and we have not been able to access these.
We also collated files of paperwork relating to the original design of the project, the grant application, original minutes from meetings and written correspondence between team members. Some of these materials had been archived along with the interview transcripts, members of the original research team donated other material. We were also extremely fortunate that Elias donated his life's work to an archive in Germany where it was placed after his death. The archive held additional information on the project, much of it embedded in letters (and copies of replies) exchanged between Elias and the Head of Department (Ilya Neustadt) and Elias and the research team. Such detailed records have enabled us to reconstruct, in part, the history of the project. This has led us to consider researchers of the future as it is likely that given the advent of media such as email it is unlikely that they will have access to such a wealth of resources. Perhaps, as Fielding (2004) suggests, qualitative researchers should be mindful of the archiving process from the outset of any research and keep all records relating to projects. This would, of course, now include e-resources such as emails. However, given the controversy of the archive one has to question how many depositors would be prepared to include such information for future researchers to access (see Moore, 2007 for a more detailed discussion).
Having (re) analysed the original data we went on to trace and reinterview a sample of respondents from the original project. These respondents, who were school leavers in the 1960s, were reinterviewed as they approached retirement. The value of the original data increased dramatically as the study transformed from being simply a data reuse project and became a rich piece of qualitative longitudinal research. Undoubtedly this find represented a research 'treasure chest' and few researchers will ever have such an opportunity to revisit a large-scale sociological research project. However, the existence of the Qualidata Archive means that others do have access, or at least part access, to data from classical sociological research from the past.
Few researchers have attempted to carry out a restudy of the same scale as the Young Worker Project. This is, in many ways, hardly surprising. Access to archived qualitative datasets has only recently become possible and such data is completely anonymized making the follow up of original respondents impossible. Nevertheless, Thompson (2000: para 41) argues that: the most valuable qualitative datasets for future re-analysis are likely to have three qualities: firstly the interviewees have been chosen on a convincing sample bases; secondly, the interviews are free-flowing but follow a life-story form … and thirdly, when practicable recontact is not ruled out.
In principle we would certainly agree with Thompson and, indeed, we would argue that our follow up study fulfilled all three of Thompson's suggested requirements for reanalysis. However, it is important to consider the argument of Gillies and Edwards (2005: 29) who suggest that: it would be extremely time-consuming, and in many cases impossible, to trace those who were interviewed in the 1960s in order to seek their explicit informed consent. Further, this practice could be viewed as unethical in itself, given that interviewees were often told that no further contact would be made.
We now move on to consider these issues in more detail.

T R AC I N G A N D R E -I N T E RV I E W
Tracing, or attempting to trace research respondents many years after the initial research encounter does, without doubt, raise a number of ethical issues. This was a serious concern for our own research which, to be successful, required us to trace and reinterview respondents from the earlier study. Fortuitously for us, although as far as we are aware this project was not designed as a longitudinal study, the original interview schedule included a question which asked respondents to indicate if they would be happy to be reinterviewed at a future date. In order to comply with current ethical guidance as far as possible, we decided to attempt to trace only those individuals who had indicated that they were happy to be contacted again. Although it is unlikely that any respondents at the time of the original interview would have anticipated contact being re-established some 40 years later, we felt that we had, through this question, obtained a degree of informed consent. However, it is important to point out that when re-establishing contact and reinterviewing respondents we followed standard contemporary ethical procedures such as giving respondents multiple opportunities to withdraw from the research at any point and providing assurances regarding data protection and data use.
We originally aimed to trace and re-interview 200 of the original 894 respondents. Approximately 500 of the target sample indicated in the original interview that they would be willing to participate in further studies and, in line with ethical considerations considered above, it was this group who were the main focus of our tracing attempts. We adopted a multi-strand approach to tracing. First, for those respondents who originally gave consent, we wrote to the last known address. Second, we made use of publicly available sources such as the telephone directory, the electoral register and the website 'www.192.com'. We also made use of the website 'www.friendsreunited.co.uk'. This site aims to bring together people who were at school together and includes names of those who register with the site organized according to school(s) attended and date of leaving school. As we had details of the schools attended by the respondents we were able to identify those individuals in our sample who had registered on this site. With the exception of Power et al. (2005) little attention has been given to the potential of online resources such as Friends Reunited and 192.com in enabling researchers to make contact with individuals for the purposes of carrying out research. The use of such websites raises ethical issues which must be considered (Power et al., 2005) and we were mindful of this, for example, before using this method of contact we requested permission from the site providers. However, we would agree with Power et al. (2005) that in terms of ethics this method of approaching individuals differs little from the more traditional 'cold-calling' approach. This method of tracing proved fruitful as the examples from email responses below illustrate: Thanks for your e-mail. I really cannot remember taking part in the survey and would be interested to know what I said. Yes, of course I would be pleased to talk to you.
Sorry it's taken so long to reply to your e-mail. I have a vague memory of the interview in 1963 but it was almost 40 years ago. Please let me know how I can be of help. I will in any way I can.
Nice to hear from you. No, I don't remember but delighted to help. Ask what ever you want.
Alongside our mail-based campaign we used local print and broadcast media to appeal for those who remembered being interviewed to contact us directly. Although this method of tracing respondents did generate a large amount of interest in the project it was not a successful means of tracing those who took part. Direct approaches to 'known' participants proved far more effective.
The final tracing method was the use of personal contacts. We found that invariably those we interviewed began talking about their old school friends at some stage of the interview. We made a decision to capitalize on this and follow up on leads when interviewees mentioned individuals from our original sample. These proved very important not only in tracing certain individuals but also it enabled us to update our database; for example, we were often told about people who had died in the intervening period and others who had moved away from Leicester.
Using these methods we traced 157 of the original respondents of whom 97 were re-interviewed. The tracing methods varied in their success, yet we would suggest that tracing around 18 per cent of the original sample represents a positive outcome for the study. However, while tracing worked well we became aware early on that our methods were not capturing many women. Although the fact that fewer women than men were included in the original sample explains this in part, it does not fully explain this situation. The main problem that we faced was that clearly many of the women had married since the original interviews and were highly likely to have adopted their married names. This meant that using techniques such as searching the electoral register and telephone directory, which worked when tracing men, were ineffective when it came to tracing women. A trawl of local marriage registers also proved not only extremely time-consuming and expensive but also ineffective. The difficulty in tracing women is clearly reflected in the final numbers of respondents traced.
Our final sample consisted of 97 individuals: 87 men from an original total of 634 and 10 women from an original total of 260 female school leavers.

R E -I N T E RV I E W
The re-interview was semi-structured, covering topics outlined in the proposal and a worklife history diary. The research instrument included qualitative questions to allow the respondents to elaborate on aspects of their lives. Despite our research instrument having some structure, the interviews tended to be more open and qualitative in nature. The respondents often began the interviews by talking about significant life events which meant the interview schedule had to be adapted during the interview process. We did not prevent any of the respondents from discussing issues that they felt were significant. The time lapse of some 40 years did cause difficulties with recall; some respondents for example, brought complete CVs to the interview so that they could guide us through their work histories. We were asking respondents to reflect on a lifetime of work and while this proved relatively easy for those who had held only one or two jobs, others had more chequered work histories and recounted careers with 40 or more different employers. However, as interviewers we found ourselves in the unusual position of having access to data about each individual 40 years earlier. This sometimes meant that we were able to prompt interviewees or to help them date particular events such as the date they left school or to furnish them with contextual information such as how they secured their first job, who they worked with and what time they had to be at work. Such data reflects the minutia and mundane nature of everyday life which is likely to be forgotten after such a large time lapse yet such information served as a valuable prompt for some interviewees. As Farrall (2006: 7) suggests 'QLR allows for respondents to reflect on the changes or lack of them which they have experienced since the previous interview' indeed 'reporting to respondents what they said in a previous interview elicits better data'.
The majority of the interviews were taped with consent. Although not specified in the proposal, we decided to photograph all of the respondents and to collect photographs of them from the 1960s. These were used as prompts during the interviews. The photographs have subsequently been collated to form the basis of a photography exhibition, combining the photographs with text from the research. Once the interviews were complete we invited all respondents and their families to an evening event at the University of Leicester where we disseminated the findings of our work. We displayed the photography exhibition and gave brief talks reflecting on key themes arising from the study.

Reflections on reinterviewing respondents from a lost project
It is evident then that the Young Worker Project has a long and complex history both in terms of the evolution of the project and the various methodological issues which have arisen over the past five decades. The project has impacted on many lives, not least the respondents. We believe that the account of the project is intrinsically valuable to our understanding of the development of British Sociology because of the methodological insights it provides and the fact that it highlights the quality of previous studies and the potential promise that they hold for current researchers. The accompanying interviewer notes from the original project provide unrivalled insight in to the experience of being a researcher during the expansion of Sociology during the 1960s. However, this is not our concern in this article. In this final section of our discussion we reflect on the value of data reuse in a project such as this and we draw out the issues which we think gives this project significant value to contemporary sociologists, in particular the potential for understanding change over time.
We have highlighted above various approaches to data collection and use. Each of these approaches: data reuse, restudies and replication, and QLR are characterized by the significance of the concepts of time and change and retrospection and reflection. Data reuse for example, is only possible because of the lapse of time between projects and the opportunity this provides to examine change through a temporal lens. Restudies and replication are possible only because time has elapsed and the opportunity to research change over time becomes a reality. Finally QLR is characterized entirely by research carried out over an extended period of time. It is, therefore the concept of time that links these approaches together.
Nevertheless, there are issues relating to the definition of each of these approaches and while in broad terms there is agreement over the meaning of each approach, there is little agreement in terms of process at the micro-level. To some extent, we have fallen in to the same trap as many others, by using different methodological terms such as restudy and revisiting interchangeably. This, we argue, is because our project is so multi-faceted methodologically that is has proved impossible to define in simple terms. However, we argue here that our own confusion has helped us to clarify the concepts used throughout the literature and to identify areas lacking precise definitions.
Our purposes in carrying out this piece of research were manifold. We wanted initially to simply (re)use the data in order to understand more about the transition from school to work in Leicester in the 1960s. The rich dataset that we discovered enabled us to do this. However, unearthing such a vast dataset led to the discovery of much more than simple data. We were also able to access extensive documentation relating to the project, giving us important contextual information. Added to these primary data sources was the key publication by Ashton and Field (1976). This publication, based on a sample of the original interviews, led the authors to devise a model of school to work transition which continues to be used as a framework for understanding school to work transitions. We discovered that Ashton and Field (1976) had identified each case by making a note on the interview schedule. This system of noting the category for each respondent has enabled us to some extent, to 'test' the predictions of Ashton and Field for the future careers of these young people with the luxury of both hindsight and foresight . This is not to say, however, that the reuse of qualitative data should be concerned with 'testing' earlier arguments. The strength of such restudies lies not in the potential the revisiting gives us to be 'critical both of what was said in the previous research and of what was overlooked' but more so to 'ground the analysis … in a way that attempts to discover the interconnected nature of the various social forces at work' (Crow, 2002: 4.3). Indeed therein lies the strength of a restudy; the unparalleled opportunity such an approach provides to begin to theorize social change over time. The desire to explore change over time was, in turn, influenced by other aspects of Elias work -especially his writings on the 'retreat of sociologists into the present' (Elias, 1987) in which he argues: The narrowing of the sociologists' focus of attention and interest to the immediate present, in some respects, undoubtedly represents progress in the development of the discipline. Sociologists are now much better able than before to study and in some cases solve short-term problems of their own society…. The immediate present into which sociologists are retreating, however, constitutes just one momentary phase within the vast stream of humanity's development, which, coming from the past, debouches into the present and thrusts ahead possible futures. It is not surprising, therefore, that the recent abundance of empirical sociological enquiries went hand in hand with an impoverishment in other respects. (Elias, 1987: 223-4) Or as Laub and Sampson (2003: 284-5) have more recently suggested: Critics whose gaze is limited to the present thus suffer in the end from a lack of scientific sense. Because our focus is on within-individual patterns of stability and change, we must rely on longitudinal data that other investigators began collecting many years ago in order to empirically study various life adaptations over the long term. There is no other way to proceed.
Indeed, these data have provided the opportunity to reflect on stabilities and changes over time and allowed us to reconsider the past. Such opportunities represent the true value of this type of material. While other writers have only been able to interrogate the past by asking respondents to reflect and recall what has gone before, we were able to reconstruct the past through access to data contemporary to that time. This has enabled us to question the accepted and orthodox views of the past and re-consider the implications of past experiences for the present and possibly the future. For example, we have questioned the idea that the 1960s was a 'golden age' for school leavers where jobs were plentiful and the prospect of unemployment unimaginable. However, examining these data reveals that unemployment was a reality for many school leavers of this time and that it was not always easy to find jobs (Goodwin and O'Connor, 2005a). This in turn has had consequences for how their subsequent careers developed and how they currently experience work and employment.
Likewise, by approaching the data reflecting longer-term historical processes, rather that treating the past and the present as static and unconnected periods of time, has allowed us to reuse this 'historical' data with contemporary concerns at the forefront of our minds. So, for example, at the time that the original data were collected the aspects now central to understanding society, such as issues of gender and ethnicity, were not considered overly important to the analysis. We have re-examined the data, so far, by looking in particular at what we can learn about gendered transitions from school to work in the 1960s, an aspect of the transition experience neglected at the time of the original project (O' Connor and Goodwin, 2004;Goodwin and O'Connor, 2005b). Indeed, we have found ourselves in a very privileged position in having access to a cohort of individuals and data about those individuals both at the start of their working lives and as they approach retirement. Such data allows us to reconstruct individual lifecourses based on historical data and therefore not as subject to the well-documented problems of recall in interviews. The lifecourse data we have also provides a snapshot of the impact of macro-changes on individual lives and the way that fateful moments or epiphanies (Giddens, 1991) can be seen to shape the future of each individual. So, for example, we have data on individuals who left school with a seemingly linear and secure career path ahead in well-established local industries such as hosiery and boot and shoe factories losing their jobs due to the relocation of manufacturing. Such personal crises, or 'fateful moments' forced these individuals to 'seek advice, undertake research, and develop new skills' (Plumridge and Thomson, 2003: 216) in order to secure new and very different types of employment.

Conclusion
The original Young Worker project was a large undertaking and immense amounts of data on many aspects of young people's lives were generated by the original team. The re-interviews have also generated a large volume of data and, when these datasets are viewed as a whole, the complexity and sheer size of the project is somewhat overwhelming. We would certainly agree with Thomson and Holland (2003: 242) that the analysis of qualitative longitudinal datasets is 'highly labour intensive' and presents a whole set of new problems such as how best to 'do justice to the data while preserving participants' anonymity'. As such we have hardly scraped the surface of the data available. We plan to continue our visit to the past by looking in more depth at other aspects of the interviews, for example, we have detailed data relating to the transition to adulthood and independence which may provide us with a new perspective on this important life stage.
We therefore believe it is important to share our experiences and the insights we have gained in carrying out this study. Indeed the Young Worker Project offers an unparalleled opportunity for exploring many key methodological issues as it encompasses the full gamut of data reuse: restudies, qualitative data reuse and qualitative longitudinal research. This experience has led us to question existing use of the terms restudy, replication, data reuse and qualitative longitudinal research as current use of these terms in the methodology literature is problematic and muddled. There is no doubt that such research is both valid and valuable and, given the increasing importance attached to the archiving of research data, it is imperative that researchers share their own experiences and add to our understanding of the methodological issues which arise when making use of data from pre-existing studies.