A Tale of Two Analyses[1]: The Use of Archived Qualitative Data
by Jo Haynes and Demelza Jones[2]
University of Bristol
Sociological Research Online, 17 (2) 1
<http://www.socresonline.org.uk/17/2/1.html>
10.5153/sro.2523
Received: 12 May 2011 Accepted: 14 Nov 2011 Published: 31 May 2012
Abstract
This article provides a unique contribution to the debates about archived qualitative data by drawing on two uses of the same data - British Migrants in Spain: the Extent and Nature of Social Integration, 2003-2005 - by Jones (2009 ) and Oliver and O'Reilly (2010), both of which utilise Bourdieu's concepts analytically and produce broadly similar findings. We argue that whilst the insights and experiences of those researchers directly involved in data collection are important resources for developing contextual knowledge used in data analysis, other kinds of critical distance can also facilitate credible data use. We therefore challenge the assumption that the idiosyncratic relationship between context, reflexivity and interpretation limits the future use of data. Moreover, regardless of the complex genealogy of the data itself, given the number of contingencies shaping the qualitative research process and thus the potential for partial or inaccurate interpretation, contextual familiarity need not be privileged over other aspects of qualitative praxis such as sustained theoretical insight, sociological imagination and methodological rigour.
Keywords: Secondary Analysis; Research Methodology; Class; British Migration; Bourdieu
Introduction
1.1 The debate about the use of archived qualitative data in social research reconfigures some important issues pertaining to the fundamental character of qualitative research (see for example, Bishop 2007; Hammersley 1997; 2010; Heaton 1998; Mason 2007; Mauthner et al 1998; Moore 2007; Parry and Mauthner 2004; 2005; Savage 2005; Silva 2007). Although this debate focuses on the practical and substantive issues surrounding the use of archived qualitative data, it nevertheless provides an avenue for further exploration of the significance of context and reflexivity within the interpretative process and the ontological and epistemological status of qualitative data. Moreover, it refocuses our understanding of ethical issues – principally, the meaning and ongoing management of informed consent beyond the original study.1.2 Recent debate has focused on the issue of fit between archived data and new research questions, whether data can be used in different contexts and ethical concerns about the legitimacy of informed consent for original participants (Bishop 2011; Hammersley 2010; Moore 2007; Parry and Mauthner 2004). The challenges associated with using data in a different research context to the one in which the data was originally collected is of most concern to the discussion here, with the implication being that only those immediately involved in the original research context can adequately interpret the data given their immediate access to and experience of the research context (Mauthner et al 1998). Others have sought to overcome such problems by offering alternative conceptualisations of data and by questioning assumptions about the contextual information and extent to which the original researchers had ‘privileged’ access to this information (Hammersley 2010; Moore 2007).
1.3 Parry and Mauthner (2005: 340-1) suggest that in order to move the wider debate forward, the ‘relationship between the researchers reanalysing archived qualitative data and the original dataset’ should be addressed. This is predicated on their observation that very little is known about instances of ‘reuse’ carried out by researchers that are not part of the original research team. The discussion here addresses this gap by focusing on the use of archived data by a researcher who was not part of the original study, as well as the original researchers. The extraordinary aspect of this example is the similarity between their analyses: Jones’ (2009) analysis of O’Reilly’s archived data on British migrants in Spain used Bourdieu’s ideas to carry out a culturalist class analysis and Oliver and O’Reilly’s (2010) use of the data drew on these theoretical resources for similar purposes. What observations can be made about the use of archived qualitative data when independent researchers not only adopt a similar conceptual framework in their analysis of the same archived qualitative data but also produce similar conclusions?
1.4 By focusing on key interrelated issues of context and reflexivity, this article makes a distinctive contribution to the important methodological debate around the use of archived qualitative data. First, the hitherto unpublished analysis by Jones, along with a revised critical commentary of Oliver and O’Reilly’s article, is presented and discussed. Following this, critical reflections pertaining to the use of archived qualitative data are considered in light of the analytical processes adopted by Jones (2009) and Oliver and O’Reilly (2010). We argue that whilst the insights and experiences of those researchers directly involved in data collection are important resources for developing contextual knowledge used in data analysis, other kinds of critical distance can also facilitate credible data use and challenge the assumption that the idiosyncratic relationship between context, reflexivity and interpretation limits the future use of data. Moreover, regardless of the complex genealogy of the data itself, given the number of contingencies shaping the qualitative research process and thus the potential for partial or inaccurate interpretation, contextual familiarity need not be privileged over other aspects of qualitative praxis such as sustained theoretical insight, sociological imagination and methodological rigour.
The Two Analyses
2.1 The data for Jones’ (2009) analysis was taken from O’Reilly’s (2006) British Migrants in Spain: The Extent and Nature of Social Integration, 2003-2005 archived dataset, and consists of a sample of seventeen interview transcripts made available to postgraduate students (Jones amongst them) for the purpose of studying qualitative research and analysis methods. To be clear, Jones’ (2009) unpublished analysis was carried out independently from and without awareness of Oliver and O’Reilly’s (2010) ‘re-use’ of their data from earlier research. Moreover, the excerpts from Jones’ (2009) analysis included in this article have not been substantively altered in any way in light of Oliver and O’Reilly’s (2010) contribution. This section outlines the choices made by Jones around framing the analysis of O’Reilly’s archived data. It then reproduces a condensed version of the findings of Jones’ analysis, which is compared to Oliver and O’Reilly’s (2010) article.Background to Analysis: Framing Class
2.2 O’Reilly’s (2006) research was carried out with the aim of updating an earlier ethnography of the everyday lives of British residents of the Costa Del Sol, from which a lack of integration with the Spanish host community emerged as a key theme (O’Reilly 2000). The later research involved O’Reilly interviewing Britons in both the coastal areas where she had worked previously, and inland areas of Málaga Province where, it was suggested, integration would be more advanced due to the lower density of migrants and the relative lack of ‘ex-pat’ orientated facilities. The interviews were part of a broader project incorporating a self-completion survey of British migrants, interviews with Spaniards who regularly encountered British migrants, and focus groups with British children resident in Spain and their Spanish peers (O’Reilly 2006). Findings from this and O’Reilly’s earlier data were presented in several publications, with an emerging theme being migrants’ tendency to differentiate themselves, often in terms of superiority, from tourists. Distinctions are framed in several ways: migrants distinguish themselves as having ‘things to do’ while tourists are at leisure, complain about the ‘nuisance’ of the holiday season, and identify the tourist ‘other’ on the basis of how they look, what they like, what they do and where they go (O’Reilly 2000; 2003; 2008).
2.3 Jones’ analysis of the archived data presented below explores how these distinctions are practised by respondents, not just in relation to tourists as described by O’Reilly’s earlier work, but to other expatriate Britons. The choice of class as a framework for the analysis of O’Reilly’s archived data was made following initial exploration of the sample of interview transcripts provided for the purposes of the assignment[4]. The differentiation strategies employed by respondents, and their emphasis on public behaviours, aesthetic and cultural preferences, and social spaces, suggested that an enquiry within these terms of reference would be rewarding. The tendency to identify and position oneself in relation to an inferior ‘other’ on the basis of appearance, behaviour and consumption is significant to Bourdieu’s culturalist approaches to class, and such an understanding is used to frame this analysis.
2.4 The recognition of culture as central to the formation and maintenance of class stratification differs significantly from the Marxist view of class as determined by modes of production. Indeed, it may be seen as having more in common with Weber’s conception of ‘status’, an additional strand of social stratification determined by shared notions of an acceptable lifestyle and the prestige attached to that lifestyle (Coser 1971; Roberts 2001). Although Weber viewed class and status distinctly, in new sociological understandings and popular discourse the notions are commonly conflated, with social standing ‘often indicated by lifestyles and particular patterns of consumption’ (Crompton 2008: 15). The now commonplace recognition of class as a relational concept draws further on Weberian concepts of ‘status’ in arguing that ‘higher’ social groups construct and maintain their claim to superiority by denigrating the cultural practices of other groups (Turner 1988). This relationship has been explored in terms of ‘disgust’, which is identified as a powerful distancing tool: ‘disgust helps define boundaries between us and them, me and you. It helps prevent our way from being subsumed into their way’ (Miller 1998: 50; emphasis in original). The ‘disgusted’ relationship is a trope of media and popular discourse, which frequently presents lower class culture as vulgar, valueless and immoral (Lawler 2005; Skeggs 2004).
2.5 Popular discursive imagining of the ‘Brit abroad’ appeared clearly relevant in O’Reilly’s interviews with British expatriate respondents. O’Reilly’s analysis of her earlier data points to the slew of media coverage following large-scale British migration to Spain during the 1980s, which presented a negative archetype of the lobster-hued ‘Brit’ obsessed with alcohol and sex, interested only in socialising with co-nationals in British-run shops and bars, refusing to learn Spanish, and of dubious character through suggestion that the Costas were a criminal bolt-hole. O’Reilly’s respondents are reported to be acutely aware of this stereotype: ‘[they] felt the media representations of them depicted them in a bad light’ (O’Reilly 2001: 174). Engagement with this stereotype is also explored in Sriskandarajah and Drew’s (2006 59) research amongst British expatriates in several destination countries (including Spain), which reports that ‘sections of expatriate Britons judge and distinguish themselves as being better than other groups’, with integration held as a key measure of status and success. The authors suggest however, that given that many of those who make this distinction are themselves hardly integrated, ‘it becomes apparent that integration is simply a backdrop for a more ingrained set of cultural markers’ (Ibid: 59). As in the media stereotype, the label ‘Brit abroad’ is associated with lack of integration through a vulgar attachment to aspects of Britishness connected with ‘low education and culture’, with a quotation from one Spanish-based expatriate exemplifying the cultural distinctions in operation: ‘It’s like Blackpool and Ilfracombe. There is a difference in the same way as there’s a difference between Torremolinos and Salobreña’ (Ibid: 59). Here spatialisation, identified as central to ‘disgusted’ class relationships (Lawler 2005), forms the basis of distinction: the tacky and touristy is contrasted with the genteel and ‘authentic’, and the cultural baggage this entails.
2.6 The analysis presented by Oliver and O’Reilly (2010) is also framed by Bourdieu’s ideas about class, employing the concepts of the field, habitus, capital, the game and distinction ‘in attempting to understand the (re)production of class in the self-making migration project of British migrants in Spain’ (Ibid: 51). The development of the article stemmed from ‘discomfort’ on the part of both authors around notions of the ‘decline of class’ in lifestyle migration settings (Ibid: 50). The seeming correspondence between lifestyle migration and visions of ‘individualising modernity’ conflicted with evidence gathered from the authors’ own empirical studies amongst British migrants to Spain, ‘which revealed the multiple ways class seems to be re-articulated under new conditions’ (Ibid: 50). Although class has featured in O’Reilly’s previously published work, both within descriptions of the demographic characteristics of British migrants to Spain and the exploration of the denigration of the ‘expat’ as ‘low-class’ in UK media and popular discourse (O’Reilly 2001), class relations among British migrants in Spain and class-based self-representations of migrants have not been a focus of previously published analysis of this data. An exception is a chapter by O’Reilly (2009) which explores the impact of ‘class habitus’ on the dispositions and choices of the children of British migrants to Spain. Access to this chapter came after publication of the Oliver and O’Reilly article in 2010 and thus not prior to Jones’ (2009) analysis of the archived data.
2.7 Although Jones’ analysis is based purely on the selection of O’Reilly’s archived data made available for learning purposes while Oliver and O’Reilly’s article analyses complete data sets collected by the two authors[5], both pieces of work share strong substantive commonalities, foregrounding respondents’ ready engagement with culturalist framings of class stratification, and employing a Bourdieusian theoretical framework. The following section presents extracts from the analysis of archived data conducted by Jones in 2009, with a new critical commentary highlighting similarities (and some differences) with the analysis presented in the Oliver and O’Reilly article (2010).
Comparing Analyses
2.8 Turning first to Jones’ analysis of the archived data, replicating the findings of Sriskandarajah and Drew’s aforementioned study (2006), respondents express disdain for ‘Brits’. They profess to avoid areas they frequent and equate the arrival of large numbers of Britons with an area becoming ‘spoilt’. Whilst one respondent in the archived sample analysed refers to this explicitly in relation to tourists, other respondents either do not distinguish between tourists and residents, or expressly refer to other expatriates: ‘We moved to Mijas Costa first, but I couldn’t stand all the Brits’ (Debbie)[6]. The possible hypocrisy presented by the fact that the respondents are themselves British is negotiated through the construction of these ‘Brits’ as an inferior ‘other’. Several respondents express embarrassment at the possibility that they could be identified with, or mistaken for these ‘others’, and take pains to distinguish themselves, exhibiting resistance to the dominant stereotype of the ‘Brit abroad’, and expressing their difference both in vague and unspecified terms of the ‘wrong type’ or ‘that sort’, and through more corporeal indicators which can be placed firmly within a discourse of class distinction. Oliver and O’Reilly’s (2010: 59-60) article makes the similar observation that some respondents deliberately and self-consciously avoid interaction with other British migrants, and ‘use culture to distinguish themselves from the strong working-class visual presence and associated stereotype of the British in Spain’
2.9 In both analyses, integration emerges as a strong indicator of social standing and class. Jones’ work observes that of the seventeen of O’Reilly’s archived interviews analysed, less than half of the respondents exhibit characteristics indicative of integration, such as proficiency in Spanish and friendships with Spaniards. Nonetheless, the majority characterise ‘other’ Britons by pointing to their lack of integration – their insistence on speaking English to Spaniards and a preference for British-style food and entertainment: ‘they walked into this place [restaurant] and one woman said “ooh, isn’t it Spanish?” like they thought it was a bad thing! They don’t normally go to Spanish places’ (Jane). Most of those who have not integrated themselves still criticise and differentiate themselves from ‘others’. Respondents’ accounts justify their own superiority to these ‘others’ through placing emphasis on their own attempts to integrate, and it is claimed that their own lack of integration is not so bad as they do not ‘flaunt’ it through public displays, and that they ‘appreciate’ Spanish culture even if they are unable to fully partake in it. This replicates Sriskandarajah and Drew’s (2006: 59) finding, that while expatriate Britons may deride compatriots for publicly engaging in ‘embarrassing’ activities such as eating British food and watching British television, they are happy to engage in these activities themselves ‘in a private space where no one can judge them’ - the distinction between privately watching Sky TV at home, and publicly watching it in a bar. Oliver and O’Reilly’s (2010: 60) analysis also identifies a relationship between integration and status: ‘being integrated is thus seen as being cultured and is a means of accruing symbolic capital’.
2.10 As mentioned, spatialisation is central to culturalist discourses of class (Lawler 2005). ‘Disgust’ is reinforced through the imbuing of spaces ‘with a personal aesthetic to render its inhabitants horrific and pathological’. This tendency has historical lineage in the ‘mapping’ of cities in terms of danger and vice, which in turn informed ‘the classification of the inhabitants of such areas’ (Ibid: 433; see also Shields 1991). This form of distinction is practised by respondents in the archived data analysed by Jones, notably through a number of references to British-run bars and clubs, and in particular ‘The London Pub’, which with its theme nights, cheap alcohol promotions and giant-screen Sky TV seems to spatially exemplify the ‘Brit abroad’. ‘The London Pub’ is associated with danger and disorder. When asked how safe from crime he feels in Spain, Peter responds: ‘…outside The London Pub at two o’ clock in the morning, not very safe!’, while Maria recounts forbidding her teenage daughter from attending a party in another, nearby bar: ‘I said it’s not that I don’t trust you… I mean it’s right near The London Pub!’ O’Reilly’s archived annotations to her interviews show that several actually take place in British-run bars or cafes, suggesting that such establishments are not intrinsically associated with danger and disorder nor off-limits per se - only those deemed unsafe, or where activities deemed vulgar and unacceptable occur. Negative attitude towards British-run businesses such as bars and shops selling branded British foodstuffs also features in Oliver and O’Reilly’s (2010: 60) analysis, with frequentation of such spaces identified as a key marker of non-integration and ‘low class’ in respondents’ accounts.
2.11 A further common theme in both analyses is reinvention. Jones’ analysis of the archived data identifies a repeated negative association of the ‘other’ with reinvention. A move to Spain is recognised as a chance to construct a new identity, with respondents seemingly finding this opportunity (as it is taken up by ‘others’) somewhat disquieting. Distrust or hostility is expressed towards those who are reluctant to talk openly about their past, or whom it is suspected are dishonest about their backgrounds in the UK. Katy says, ‘you don’t want to get to know English people here too well cos you don’t know about their background’, and explains that she likes a particular British couple because ‘you can tell they are just normal and there are no lies and no secrets’. Respondents consider reinvention as a chance for the ‘other’ to present themselves as ‘better’ than they really are, through the acquisition of a more lucrative or prestigious trade or profession: ‘They just come out here, and they say “oh yes, I’m a plumber” or “I’m an electrician”. I’m this, I’m that. Really back home they were a delivery driver or a dustman’ (Andrew), or to construct a fictional past, which grants them a higher social standing than suggested by their current circumstances: ‘I mean the things you hear… I was a Concorde pilot at home but now I’m laying bricks’ (Peter). Occupation can function as both an economistic indicator of class through pay-scales and material wealth, as well as a culturalist indicator through understandings of ‘occupational prestige’ and its relationship to social standing (Crompton 2008: 49-50). Here, respondents’ attention to the occupational status and history (or fictional history) of ‘others’ focuses on suspicions of dishonesty surrounding claims of professional or educational attainment. Disapproval is expressed in terms of moral superiority. Andrew says, ‘I don’t particularly want to mix with that type of people’, while several respondents equate reinvention with immorality and criminality, framing their criticism of these ‘chancers’ in terms of threat and mistrust.
2.12 These observations are mirrored in Oliver and O’Reilly’s (2010) analysis. Respondents’ denigration of other expatriates is commonly ‘based on speculations about what economic realities might underpin reinventions’, while ‘Easyjet builders’, who have ‘acquired’ their professed qualifications and experience on the plane to Spain are derided (Ibid: 58). As in Jones’ analysis, Oliver and O’Reilly frame this talk in culturalist terms of moral disapproval and ‘threat’ to the acquisition of ‘social capital’: ‘constant reinventions… result in struggles over how to trust or value others’ (Ibid: 57). For respondents uncertain about the reliability of compatriots’ accounts of themselves and their pasts, and ‘deprived of their usual ways of ensuring they mix with ‘people like us’’, culturalist indicators of class and status are of enhanced importance: ‘where the backgrounds of others are unknown, information is deduced about others through observations of their habitus, seen in their tastes, dress and behaviours’ (Ibid: 57).
2.13 Along with the above similarities, it is important to note that there are also differences between Jones’ and Oliver and O’Reilly’s analyses of the data. In Oliver and O’Reilly’s work, economistic indicators of class assume more stand-alone significance. Differentiation between migrants is described in Marxist terms of the purchase and supply of labour. Migrants who sell their service and labour are distinguished from those who can afford to hire them and their Spanish counterparts, while elderly expatriates who can afford to buy live-in care are contrasted with those who cannot, and so are forced to rely on help from Spanish social services and volunteers (Ibid: 55-6). Perhaps because they were drawn from both O’Reilly and Oliver’s data, these findings were not apparent in the sample of O’Reilly’s archived interviews analysed by Jones.
2.14 Whilst Jones’ analysis makes reference to respondents evoking economistic class indicators such as property ownership and material wealth to differentiate the ‘other’: ‘they’re not rich. They might have a few bob in their pocket when they come out here but that soon runs out’ (Andrew), and ‘most of them are council house tenants’ (Katy), these economistic determinants are framed or accompanied by a reference to the moral or aesthetic. Andrew attributes the failure to make limited money last with ‘going in the bars every night’, while Katy strongly denigrates those she has characterised as council house tenants as ‘the scum of earth’, going on to observe that ‘…the children haven’t got their real Dads with them’. Returning to the literature, within the ‘disgusted’ relationship, low class is frequently characterised by excess. Lawler (2005: 433-4) and Munt (2000: 8) explore the positioning of the lower class in terms of ‘chaos’, ‘weakness’ and ‘waste’, while Skeggs’ (2004: 99-104) gendered reading suggests that an emphasis on lower class women’s bodies, their fatness, nakedness in revealing clothing, and perceived promiscuity or over-fertility, signifies broader characterisations of a lack of self-discipline and control. These stereotypes are clearly a reference point for Andrew and Katy’s responses. Andrew points to a lack of self-control in the drinking habits of ‘others’, echoed by Valerie’s observation that ‘because the booze is cheap they think, Woah! Great! And they down it so fast and they get paralytic’, while Katy’s contempt for step-families points to a moral judgement on the mother’s sexuality.
2.15 Additionally, Oliver and O’Reilly’s (2010: 61-2) analysis highlights an alternative form of ‘othering’ amongst those respondents characterised as lower class, or lacking the social and symbolic capital of their compatriots. These migrants, it is reported, ‘would often complain about the exclusivity, snobbery or pretentiousness of the others’, whom they characterise as ‘the snobs’ or ‘the colonials’. Again, this finding was not apparent from the selection of O’Reilly’s archived data available to Jones for analysis.
2.16 Despite these differences, a clear picture emerges from both analyses of respondents’ engagement with the ‘Brit abroad’ stereotype, and practising of resistance through the relational oppositional construction of their own identities. Integration, or at least an attempt or pretence towards integration, is central to this process, through its broader associations with ‘good taste’ and respectable behaviour. The attributes afforded to ‘other’ Britons through this oppositional characterisation: lack of discipline and self-control symbolised by excessive drinking, promiscuity and rowdy behaviour, the ‘cultural poverty’ implied by a failure to appreciate Spanish culture, and the suggestion of immorality or criminality, posit respondents within a wider contemporary tendency of conformity to Bourdieusian concepts of class. Respondents exhibit a ‘disgusted’ relationship with the ‘other’ through spatialisation, and embarrassment at the thought of being associated with, or mistaken for, one of ‘those kind’ of ‘Brits’, with the ‘Brit abroad’ stereotype providing a doxic reference point against which relational identity constructions occur.
Reflections on the use of archived qualitative data
3.1 Drawing on the critical reflections of the analytical process of Jones (2009), as well as the substantive results of both analyses, this section will discuss two aspects of the debate about using archived qualitative data. First, it considers the significance of contextual knowledge within the use of archived qualitative data and following this – given the similar interpretation of data produced in both analyses – it will consider the role played by reflexivity and its relationship to the status of the data.Context
3.2 A key issue raised by the use of archived qualitative data is the perceived difficulty in analysing data removed from its original context (Moore 2007). Hammersley (1997: 138-9) notes that time spent in the field is a key component of qualitative research, involving the acquisition of a ‘cultural habitus’ and the collection of ‘head-notes’ which allow the primary researcher to interpret data ‘against the background of all that he or she tacitly knows about the setting’. This certainly appears to be the case with O’Reilly’s data, which was collected over an extended period, during which time she was immersed in the social world of the British expatriates. This ‘intuitive component’ (Ibid: 139) is inaccessible to the practitioner of any subsequent analysis, unless it is conducted by the original researchers. However, it is a matter of degree as to how this contextual knowledge configures the analytical process of any research. The impact of this tacit knowledge of research context can vary both within the original research as well as in subsequent analyses (Hammersley 2010). For instance, the typical division of labour when working in a research team means that not all the original researchers can draw on their ‘implicit understandings and memories’ (Ibid: 3) or ‘direct personal involvement’ (Mauthner et al 1998: 738) within this or any subsequent research. Rather than this signalling a fundamental problem that undermines the ontological foundations of all research or suggesting that clear judgments about contextual relations within research can ever be made, the problem with contextual information here is aptly described by Hammersley (2010) as a ‘matter of degree’. In other words, both access to and the reliability of contextual information are variable characteristics of all research and should not be taken for granted.
3.3 Notwithstanding this qualification, whilst reading O’Reilly’s archived interview transcripts, questions such as ‘what does he look like?’, or ‘how does she dress?’ arose; non-verbal signifiers accessible to O’Reilly as the primary researcher, but unavailable here. After reading the transcripts repeatedly, mental images began to develop of some respondents, but fearing this could unduly influence the reading of their words and meanings, attempts were made to suppress them. Similarly, having not visited the Costa del Sol since childhood, Jones developed a mental picture of the places where respondents lived and worked, based on both fragmented memories of the region and media representations. The question arose as to whether the reproduction of possibly stereotypical, inaccurate or outdated images of both places and people are problematic for the analysis process. Again, the non-verbal, observable indicators that contribute, consciously or not, to analysis were unavailable. O’Reilly and Oliver’s revisiting of their data, even with access to their memories and implicit understanding of the earlier research context, is also subject to the mediating – and potentially distorting – influence of public discourses and media representations of British migrants in Spain. Moreover, the passing of time in itself can create some doubt within any subsequent use of data by the original researchers (Mauthner et al 1998). The important point to note here is that all research is potentially shaped by partial or inaccurate contextual information because mediation and interpretation is involved throughout the research process. Whether the subsequent understanding of the research context is based on actual contact or derived from other sources, it can facilitate sound analysis or constitute a source of error. Researchers, after all, are not in unmediated contact with reality (Hammersley 2010), even when establishing the nature of the research context itself. For Oliver and O’Reilly, the question as to whether the particular people and places studied originally were unrepresentative of the wider context of British migrants in Spain would have required some consideration. Therefore, critical reflections about the research context should be an intrinsic aspect of the original and any subsequent analysis – an aspect of what Mason (2007: 2) refers to as ‘investigative epistemology’ – where researchers are ‘purposefully investigative with and about data’.
Reflexivity and the status of data
3.4 Following on closely from the above point, and before concluding this section, we look at what this serendipitous example tells us about the role played by reflexivity within qualitative research (see for example Mason 2007; Mauthner et al 1998; Moore 2007). We then briefly question how to understand data in light of this: is data constituted within the research process itself or does it exist independently? To explore this we will be drawing on Hammersley’s (2010) recent argument that data is best understood as having two phases within the research process.
3.5 Compared to quantitative research, reflexivity is typically made more central and transparent as an intrinsic aspect of knowledge production within qualitative research (Moore 2007). For Mauthner et al (1998: 736), data is the ‘product of the dynamic, dialectical and reflexive nature of a particular research encounter’ and as such, is co-produced by the researcher and the researched. Therefore, data is limited and is shaped by ‘biographical, historical, political, theoretical and epistemological contingencies’, suggesting that any subsequent use is likely to be problematic as sets of data cannot be treated as ‘discrete entities’ (Ibid: 742). Indeed, Mauthner et al’s attempts to reuse their own respective datasets presented a range of difficulties, such as finding they no longer had a privileged insight and discovering that particular data was missing. The conclusion reached by Mauthner et al was that datasets could not ‘elude the conditions of their production’ (Ibid: 742) and that the reuse of qualitative data to generate new substantive findings raises serious epistemological issues.
3.6 The perceived epistemological gaps identified by Jones certainly provided some intellectual discomfort and pause for thought. However, the fact that the findings of Jones’ analysis and that of the original researchers have so broadly overlapped when a similar conceptual framework of culturalist class differentiation has been applied, suggests that the above constraints in the production of data may be overemphasised by Mauthner et al and the methods literature more generally. Despite the ‘biographical, historical, political, theoretical and epistemological contingencies’ (Ibid: 742) associated with the production of data, the two analyses discussed manage to produce new substantive insights that are similar. They do so with different kinds of critical distance. After close examination of a portion of O’Reilly’s data only, and through a theoretical understanding of the cultural mechanisms in the reproduction of class distinctions, Jones (2009) identified processes that reproduced class differences and interpreted these within a framework based on Bourdieu’s concepts as discussed above. Returning to their original data after a hiatus, Oliver and O’Reilly (2010: 62) suggest that ‘there is ample evidence of the reproduction of class among British lifestyle migrants in Spain’. This observation followed from their shared discomfort with the idea that class was assumed to be declining in significance in public and academic discourse. As mentioned above, Oliver and O’Reilly’s analysis places more emphasis on economistic signifiers of class (the provision or purchase of labour), whereas Jones’ interpretation of the data suggests that when economistic indicators were raised it was in conjunction with culturalist indicators. It is possible that originally, this data was generated through observational work rather than gleaned through interviews and thus, is inaccessible to subsequent researchers. Nevertheless, it should be remembered that Jones’ analysis was conducted using just a sample of O’Reilly’s interview transcripts, so this discrepancy between the analyses could as much reflect the limitations of working with an incomplete dataset as any broader, generalisable limitations to the use of archived qualitative data[7].
3.7 Two questions follow from this. First, given the different experiences of both Mauthner et al’s (1998) attempts to use their data again and those discussed here, on what basis can we predict which archived qualitative data to work with? Regardless of how much or how little contextual information is archived alongside the data, there is no definitive way of knowing in advance whether data has ‘potential’ for further analysis. To make a universal recommendation based on topics or method of data collection or a particular judgement based on reflexive relations, especially research where reflexivity is in the form of self-narration (Adkins 2002), would be to narrow the investigative field prematurely.
3.8 Second, what conclusion can be made regarding how data is understood in these examples? If we follow Mauthner et al’s (1998) interpretation of reflexivity – in terms of the centrality of the researcher in the production of data – it is difficult for data to elude the contexts of its production. By privileging this kind of reflexivity, data becomes fixed in the past and in fact, Mauthner et al conclude that the principal value of archived qualitative data is for just that – studying the past (Moore 2007). The two analyses here demonstrate that data is not only able to elude the conditions of its production, but can support similar analytical conclusions through different processes of reflexivity and critical distance from the initial data collection[8]. However, does data need to elude the conditions of its production in order to be reusable? Should it be assumed that the contingency of data on the conditions of production render the data unusable in other circumstances? Moore (2007) argues that all data is recontextualised; in the process of data use, that all researchers, regardless of the relationship to the initial conditions of its production, are engaged in a process of contextualisation. Although both analyses drew on contemporaneous discourses of class and the disciplinary shift in class analysis, only Oliver and O’Reilly (2010) were able to situate these critical insights alongside their experience and head-notes from the original research settings. To conclude that data therefore is a product of the research process, in which it is recontextualised and reconstructed through different forms of reflexivity (see for example Moore 2007) leaves us unable to explain how two independent analyses could correspond in the way that the two discussed in this article do.
3.9 Hammersley (2010) may offer some fresh insight to make sense of this phenomenon with his recent argument that data is constructed as well as given. In doing so, the two analyses – Jones’ (2009) and Oliver and O’Reilly’s (2010) – can be explained without a problematic appeal to the idea that they are simply derived from the data, conceived as existing independently of the research process and as speaking for itself. We are familiar already with the idea of data as constructed through the research process, whereby data is selectively used as evidence to support the claims we wish to make and is therefore contingent upon the question being addressed and the manner in which the analysis unfolds. However, there has to be some limit on how data is constructed, ‘we do not and should not make up our data’ (Hammersley 2010: 5 emphasis in original). This brings us to Hammersley’s second meaning of data, that it must restrict the inferences we make and conclusions we eventually arrive at. In other words, the data must in some capacity ‘exist prior to and independently of the research process’. Furthermore, when we make our interpretations of data we do so from/of ‘the world that is there’ (Ibid: 4).
3.10 Various qualifications are made by Hammersley regarding the precise nature of how the independent existence of data is understood, as well as its constructedness. Hammersley describes the distinction between understandings of data as ‘labelling phases …of a single process’ (Ibid: 4.) where data is either identified as relevant to research or we do the necessary work to produce it. Nevertheless we are limited in what we call data by the intelligible objects available to us, in Hammersely’s (Ibid: 4.) words, ‘by the noumena that are available’. This is a distinction between the ‘what’ and the ‘how’ or ‘why’ of data. Regardless of subjective speculations or valid truths about the data we have either produced or identified it as relevant material to make inferences from. We then use this material as data to provide evidence, reconfiguring it in the process as we are framing it within the context of our arguments in order to address our research questions – the how and why. Returning to the two analyses here, from different vantage points, both identified archived data as intelligible material from which to make inferences relevant to the research work in question. In neither case was the data constructed from nothing. Subsequently, some of this material was used to generate evidence within the context of similar arguments about cultural expressions of class.
3.11 The two analyses are therefore not simply a coincidence – a random event that may titillate or even frustrate those embroiled in the debate about the use of archived qualitative data. They show that the data – whether archived or not – is ‘there’ in some sense, and with methodological rigour it is possible for researchers, who are theoretically informed and up to date, to produce similar, and therefore credible, interpretations.
Conclusion
4.1 The comparison between the two analyses has provided a valuable opportunity to revisit some of the vexed questions associated with using archived data. From a pedagogical point of view, the assignment itself presented a valuable opportunity to work with ‘real’ data prior to embarking on further ethnographic fieldwork. The assignment put into practice the data management and analysis techniques explored in methods classes, including the use of NVivo coding and modelling tools for data organisation purposes. Although somewhat intellectually troubling, the challenges to the analysis outlined above also provided useful training for being attuned to the kinds of non-verbal signifiers researchers need to be aware of when conducting primary observational and interview based research. Ultimately however, in this instance, the missing signifiers did not prevent a credible analysis of the data, a reminder that their status and significance to the research process itself is a matter of interpretation (Hammersley 2010).
4.2 Oliver and O’Reilly’s (2010) ‘re-analysis’ of their data provides an opportune moment to examine the debate about the use of archived qualitative data. We are certain that they too will be surprised to know of Jones’ (2009) analysis, and thus that the data – specifically O’Reilly’s data – was able to evoke a similar portrait of British migration to the Costa del Sol for an independent researcher. The fact that both interpretations of the data – based on distinct processes of reflexivity – produced similar analyses by drawing on Bourdieu’s concepts to illuminate lifestyles and attendant cultural class relations, challenges negative perceptions of the qualitative research process as idiosyncratic and the revoking of any semblance of scientific method. Additionally, it demonstrates that there is, following Mason (2007: 4), ‘epistemological value in allowing for a range of reflexive interpretations of data, some from close range and some from a distance’. In fact, the investigation of class dynamics itself could benefit from changing perspectives on how they are approached (see for example, Savage 2007).
4.3 Although the debate about secondary analysis – like any worthy social scientific debate – is far from over, given the current funding crisis in the social sciences and the co-option of qualitative enquiry within the commercial realm, it is imperative that qualitative researchers are at the forefront of these debates. Through careful consideration of the methodological opportunities as well as the potential hazards associated with the range of uses of qualitative data (Mason 2007), it seems reasonable to be optimistic about the rewards that can come from a range of uses of archived data by researchers not involved in the original study.
Acknowledgements
The authors would like to thank the three anonymous reviewers and the Editors for their very useful comments on an earlier draft. They would also like to acknowledge the access to and use of Karen O’Reilly’s (2006) data from the ESDS QualiData Archive.
Notes
1This could be called ‘A Tale of Three Analyses’ or even ‘A Tale of Four Analyses’ depending on whether you include O’Reilly’s (2006) or both Oliver (2007) and O’Reilly’s earlier analyses of their respective data or not. However, poetic licence aside, we refer to ‘two analyses’ because we focus on Jones (2009) and Oliver and O’Reilly (2010).2Both authors have contributed equally to this article.
3Jones’ original use of the data is contained in an unpublished essay completed as part of her doctoral training in May 2009. The unpublished essay, ‘‘Not our sort’: the reproduction of class distinctions amongst British migrants in Spain’ was produced for assessment purposes for completion of the unit ‘Advanced Qualitative Research’ within the Masters programme in Social Science Research at the University of Bristol. Jo Haynes convened the unit. The assessment required students to conduct data analysis using one of a choice of data samples drawn from archived datasets from UK Qualidata.
4Students are given a list of eight datasets to choose from. The choice typically reflects the range of substantive topics covered elsewhere in the Masters programmes. For each dataset, any contextual information stored in the archive is provided, as well as a sample of the data. Rather than overwhelming students with all of the dataset, a subset of the data is made available enabling them to select a small sample appropriate to their research question or approach. The amount of data provided from each dataset varied depending on how the sample was originally structured. However, students were also allowed to request other data if they believed it was useful for their analysis.
5Oliver’s data was collected during doctoral research into retirement migration to Andalucía (Oliver 2007).
6As in O’Reilly’s, and Oliver and O’Reilly’s presentation of this data, quoted respondents will be identified by their first name.
7Jones’ analysis was restricted to a smaller sample of archived data because of the nature of the assessment that was set for the unit.
8It is probable that Oliver and O’Reilly were carrying out their analysis, writing, or redrafting their article for Sociology at the time that Jones was completing the assignment.
References
ADKINS, L. (2002) ‘Reflexivity and the politics of qualitative research’, in T. May (ed.) Qualitative Research in Action, London: Sage.BISHOP, L. (2007) ‘A Reflexive Account of Reusing Qualitative Data: Beyond Primary/Secondary Dualism’, Sociological Research Online, 12 (3) <http://www.socresonline.org.uk/12/3/2.html> [doi:://dx.doi.org/10.5153/sro.1553]
BISHOP, L. (2011) ‘Using Archived Qualitative Data for Teaching: Practical and Ethical Considerations’ in R. Edwards and G. Crow (eds) International Journal of Social Research Methodology
COSER, L. A. (1971) Masters of Sociological Thought: Ideas in Historical and Social Context. London: Thomson Learning.
CROMPTON, R. (2008) Class and Stratification (3rd Edn.) Cambridge: Polity.
HAMMERSLEY, M. (1997) ‘Qualitative Data Archiving: Some Reflections on its Prospects and Problems’ Sociology, 31 (1): 131- 42. [doi:://dx.doi.org/10.1177/0038038597031001010]
HAMMERSLEY, M. (2010) ‘Can We Re-Use Qualitative Data Via Secondary Analysis? Notes on Some Terminological and Substantive Issues’ Sociological Research Online, 15 (1) <http://www.socresonline.org.uk/15/1/5.html> [doi:://dx.doi.org/10.5153/sro.2076]
HEATON, J. (1998) ‘Secondary Analysis of Qualitative Data’ Social Research Update Issue 22.
JONES (2009) ‘Not our Sort: The reproduction of class distinctions amongst British migrants in Spain’ [Unpublished Masters essay, University of Bristol].
LAWLER, S. (2005) ‘Disgusted Subjects: The Making of Middle-Class Identities’, Sociological Review. 53(3): 429-446. [doi:://dx.doi.org/10.1111/j.1467-954X.2005.00560.x]
MASON, J. (2007) ‘Re-Using Qualitative Data: on the Merits of an Investigative Epistemology’ Sociological Research Online 12 (3). <http://www.socresonline.org.uk/12/3/3.html> [doi:://dx.doi.org/10.5153/sro.1507]
MAUTHNER, N. S., PARRY, O. and BACKETT-MILBURN, K. (1998) ‘The Data are out there, or are they? Implications for Archiving and Revisiting Qualitative Data’ Sociology 32 (4): 733- 45. [doi:://dx.doi.org/10.1177/0038038598032004006]
MILLER, W. I. (1998) The Anatomy of Disgust. Cambridge MA: Harvard University Press.
MOORE, N. (2007) ‘(Re)Using Qualitative Data’ Sociological Research Online 12 (3).
MUNT, S. (2000): Introduction. In Munt, S. (Ed.) Cultural Studies and the Working Class: Subject to Change. London: Cassell.
OLIVER, C, (2007) Retirement Migration: Paradoxes of Ageing. New York: Routledge
OLIVER, C. and O’REILLY, K. (2010) ‘A Bourdieusian Analysis of Class and Migration: Habitus and the Individualizing Process’ Sociology 44 (1): 49- 66. [doi:://dx.doi.org/10.1177/0038038509351627]
O’REILLY, K. (2000) The British on the Costa Del Sol: Transnational Identities and Local Communities. London: Routledge.
O’REILLY, K. (2001) ‘Blackpool in the Sun’: Images of the British on the Costa Del Sol’. In King, R.and Wood, N. (eds.) Media and Migration: Constructions of Mobility and Difference, London: Routledge.
O’REILLY, K. (2003) ‘When is a tourist? The Articulation of Tourism and Migration in Spain’s Costa Del Sol’, Tourist Studies. 3(3): 310-17.
O’REILLY, K. (2006) ‘British Migrants in Spain: The Extent and Nature of Social Integration, 2003-2005’, ESDS Qualidata. <http://www.esds.ac.uk/findingData/snDescription.asp?sn= 527>
O’REILLY, K. (2008): ‘Hosts and Guests, Guests and Hosts: British Residential Tourism in the Costa Del Sol’, Lifestyle Migration Hub. <http://www.lboro.ac.uk/departments/ss/lmhub/lmhubhome.html>
O’REILLY, K. (2009) ‘The Children of the Hunters: Self-Realization Projects and Class Reproduction’. In Benson, M. and O’Reilly, K. (eds.) Lifestyle Migration: Expectations, Aspirations and Experiences. Farnham: Ashgate.
PARRY, O. and MAUTHNER, N. (2004) ‘Whose Data Are they Anyway? Practical, Legal and Ethical Issues in Archiving Qualitative Data’ Sociology 38 (1): 139- 52. [doi:://dx.doi.org/10.1177/0038038504039366]
PARRY, O. and MAUTHNER, N. (2005) ‘Back to Basics: Who Re-uses Qualitative Data and Why?’ Sociology 39 (2): 337- 42. [doi:://dx.doi.org/10.1177/0038038505050543]
ROBERTS, K. (2001) Class in Modern Britain. Basingstoke: Palgrave Macmillan.
SAVAGE, M. (2005) ‘Revisiting Classic Qualitative Studies’ Forum: Qualitative Social Research 6 (1).
SHIELDS, R. (1991) Places on the Margins: Alternative Geographies of Modernity. London: Taylor and Francis.
SILVA, E. B. (2007): ‘What’s [Yet] to be Seen? Re-using Qualitative Data’. Sociological Research Online 12(3) <http://www.socresonline.org.uk/12/3/4.html>
SKEGGS, B. (2004) Class, Self, Culture. London: Routledge.
SRISKANDARAJAH, D. and DREW, C. (2006) Brits Abroad: Mapping the Scale and Nature of British Migration. London: Institute for Public Policy Research.
TURNER, B. S. (1988) Status. Milton Keynes: Open University Press.
