All entries for Wednesday 03 August 2005
August 03, 2005
Well, I didn't feel too bad today. Which, considering my general mindset at the moment, is probably worthy of a blog entry or seven. I got up at a reasonable hour this morning and even did some work, though I'm not sure how useful my musings on the theoretical and epistemological status of discourse analysis will be when it actually comes to attempting to conduct some of my own.
Anna and I went to Leam this afternoon to get us hair chopped off. Am concerned that I have now acquired a mullet due to over-enthusiastic hair stylist, but enjoyed the giddiness induced by hairspray asphyxiation nonetheless.
Went to evil Starbucks to caffeinate ourselves with mint-choc-chip-mocha-frappa-whippochinos-with-sprinkly-bits-on-top (well Anna just had a latte, but I always have to get the most revoltingly sickly thing on the menu). Bought stuff with non-existent money. Really exciting stuff, too, like printer cartridges.
Didn't do any exercise, as usual, but got lots of emails from The Friends telling me that they are all running twenty miles a day in preparation for half-marathons left right and centre come September. Am very proud of them all for being so energetic, but think I will stick to eating stuff for the time being.
Meeting Kate and Rhiannon in London for more coffee-related jollities on Friday afternoon, which should be nice as haven't seen either of them since xmas. Kate is currently dividing her time between working in Cheltenham and cat-sitting in Kent, and Rhi is at Reading finishing off her masters. Looking forward to everyone being within spitting distance next year, not that I will actually spit at anyone as tis rude and unladylike.
Staying in London over the weekend avec le chum. Going to Devon with Helen, Thomas and Leonie the weekend after. Naomi is coming to visit the weekend after that to experience the 'delights' of Warwick campus out of term time. Since at the moment she's actually looking forward to starting her MA here, I'm not sure this is a very wise idea.
Hmm I quite like this whole mundane thing (in any case it's not as boring as my 'academic related' category). Can't actually remember what I had for my lunch though.
(more notes / scribblings / thoughts)
(the theoretical and epistemological bit)
Because of the interactive nature of the data, focus group data are often suited to discourse analysis. Discourse analysis is concerned 'with how language is used to create and secure meanings, how
competing accounts are negotiated and how speakers draw on certain interpretive repertoires in making their arguments within a given discursive context' (Tonkiss 2004: 205).
Language is not viewed as neutral or colourless, reflecting an 'objective' reality, but rather as a means of re-presenting it (c.f. Abrams 'The Mirror and the Lamp'):
Language is seen not simply as a neutral medium for communicating information, but as a domain in which our knowledge of the social world is actively shaped… Discourse analysis involves a perspective on language which sees this not as reflecting reality in a transparent or straightforward way, but as constructing and organizing that social reality for us (Tonkiss 1998: 246).
The way that we use language is rarely innocent, and discourse analysis can help to reveal how talk and texts are ordered to produce specific meanings and effects (ibid: 247).
Foucault (1984) argues for a theoretical understanding of discourse as a realm in which institutions, norms, subjectivity and social practices are constituted and naturalised. In this way he links discursive constructions to the shaping of social institutions and practices of social regulation and control.
Rather than garnering accounts so as to access people's views, attitudes and opinions, or to find out what happened (i.e. assuming that there is a tangible social 'reality' that can be accessed and 'objectively' presented), the discourse analyst is interested in how people use language to (re)present their accounts of the social world. 'Language is both active and functional in shaping and reproducing social relations, identities and ideas' (ibid: 248).
As Skeggs (1997) argues, language does not simply speak its ‘truth’ in a straightforward referential way, but rather it is context-dependent and comprised of discourses that are in fact constructed themselves. In adopting a postmodern epistemological position, it is necessary to acknowledge fragmented and multiple subjectivities and reject the existence of an ‘authentic’ self. Language itself is subject to varying interpretations and representations, contains different meanings and adopts a variety of forms, and is put to different uses depending on context and potential for explanatory power (Skeggs 1997: 26).
Language is thus viewed as a social practice which actively orders and shapes people's relation to their social world (Tonkiss 1998: 249).
- Interpretive context
The social setting in which a particular discourse is located. The context is relevant at both a macro- (e.g. gender inequalities in society) and micro- level (e.g. type of interaction, relationship between participants, immediate discursive aims of the speaker)
- Rhetorical organisation
'The argumentative schemes which organise a text which work to establish the authority of particular accounts while countering alternatives' (Tonkiss 1998: 250). How are statements put together? What effects do they seek? What forms of knowledge are privileged? Which speakers will be heard as authoritative? Is discourse persuasive to action?
Doing discourse analysis (the practical bit)
Formulate a research problem – not one that is looking for answers to specific questions but looking at the way meanings are constructed. E.g. how is unemployment constructed as a political issue or 'problem', how are unemployed people represented within public discourses, how do they construct themselves to resist these interpretations, how does unemployment affect discourses of hegemonic masculinity?
Be selective about data – extract sections that provide the richest source of analytic material. Are there contradictions or inconsistencies in the text? Representations that contradict the researcher's assumptions? Are these productive?
Select a number of recurrent themes / sections of data. Categories of analysis that emerge from data may feed back into the research question and cause it to be modified.
Once themes are established – what ideas / representations cluster around them? What associations are established? are paricular meanings being mobilised? what languages are employed (e.g. economic / medical / religious / natural) and bound up into particular discourses?
Are there patterns of variation? How do participants attempt to reconcile conflicting ideas, to cope with contradiction or uncertainty, to counter alternatives? How are seemingly coherent, 'smooth' discourses disrupted? How are discourses brought together and for what purposes? Are discourses founded on a series of oppositions? Are there consistencies within and between texts?
Read for emphasis, detail and silences. Read against the grain of the text and look at gaps – what is absent from the accounts? Alternative accounts are excluded by omission.
Discourse analysis may be concerned with the examination of meaning, but meaning is contestable and specific texts are open to alternative interpretations:
The discourse analyst, like other social actors, aims to provide a persuasive account, which in this case offers an insightful, useful and critical interpretation of a research problem (Tonkiss 1998: 259).
Although internal validity may be achieved by interpreting data closely, external validity is more difficult to claim as it is difficult to contend that the analyst's own discourse is obective, factual, true, not to mention the fact that discourse analysis typically deals with small datasets. Social researchers should question their own assumptions and adopt a reflexive approach to social research (c.f. feminist issues of epistemologies / power).
So basically it's a bit like analysing a piece of literature, is it then?
Foucault, M. 1984. 'The order of discourse', in Shapiro, M. (ed.) Language and Politics. Oxford: Blackwell.
Tonkiss, F. 1998. 'Analysing discourse'. In Seale, C. (ed.). 1998. Researching Society and Culture (1st ed.). London: Sage.
Tonkiss, F. 2004. 'Using Focus Groups'. In Seale, C. (ed.). 2004. Researching Society and Culture (2nd ed.). London: Sage.
Skeggs, B. 1997. Formations of Class and Gender. London: Sage.
(Notes / scribblings / thoughts)
Choice of methods
I chose to use a focus group for my research because the topic was particularly suited to interaction in a group context. It enabled me to gain insight from different perspectives and allowed flexibility to pursue topics which arose through previous discussion. Guiding, stimulating and facilitating the discussion was crucial to the focus group. Providing interesting material for research and ensuring all participants were sufficiently as ease to disclose occasionally quite private info was all crucial. The d. guide for the focus group was devised in a logical order, so that 'warm up' questions were placed at the beginning and built up to the more important and complex questions towards the end (d. guide in appendix).
Before the discussion I outlined the broad research aims, reiterated my position on confidentiality and anonymity, and checked that it was okay to record the discussion. I also encouraged them to express their views as freely as possible by stressing that there are 'no right or wrong answers'.
Focus groups typically involve six to ten people. They need to be small enough to allow everyone to participate, but large enough to capture a variety of perspectives and enable people to bounce ideas off each other. Running focus groups requires the researcher to:
- facilitate interaction
- enable space of different group members to make their views known
- keep the group discussion focused around the core themes (but still allow flexibility)
- deal with dominant voices
- sustain a pace of discussion that means core topics are covered but without constraining or rushing the talk (Tonkiss 2004: 204)
Focus groups may be used to supplement observation methods (as in my case where I used observation first to inform the focus group topic guide) – focus groups allow researchers to elicit info and explore attitudes not easily accessible through observation alone. Not all issues are always and easliy observable 'in the field', and focus groups allow members to define these issues in terms of their own understandings and concerns, producing shared and contested meanings:
Focus groups are an artificial intervention into a 'natural' observation setting, involving the researcher in a directive relation with their research subjects and with the process of data production (Tonkiss 2004: 197).
Methodological assumptions and issues
Key feature of focus group research is its interactive quality. Unit of analysis = group, not individuals. Focus groups are useful to theoretical research that seeks to explore social / cultural meanings, knowledges and discourses.
Underlying methodological assumption is that opinions, attitudes and accounts are socially produced and shaped through interaction with others (i.e. social constructivism?). Group context is important for exploring the way in which social and cultural knowledge and meanings are produced – we can see how people articulate and justify ideas in relation to others – emphasis is on social interaction and collective meanings.
Focus groups in this sense are not simply a means of interviewing several people at the same time; rather, they are concerned to explore the formation and negotiation of accounts within a group context, how people define, discuss and contest issues through social interaction (Tonkiss 2004: 194).
Focus groups capture the inherently interactive and communicative nature of social action and social meanings, in ways that are inaccessible to research methods that take the individual as their basic unit of analysis (ibid: 198).
Focus groups may be empowering, for 'if a group works well, trust develops and the group may explore solutions to a particular problem as a unit, rather than as individuals' (Gibbs 1997). However this benefit may not extend to all members, e.g. shy or inarticulate members. They may also become a 'forum for change' and means of achieving emancipation (ibid).
One methodological disadvantage is that researchers have less control over the data that emerge – a potential gain to participants (can be empowering), but a loss of power on the part of the researcher (issue to flag up in power/ethics section?). Another problem is that focus groups are an insecure basis for generalisation. Any claims I make in term of developing insights into social attitudes cannot be matched by ststematic claims about the representativeness of attitudes. Furthermore, although focus groups aim to reproduce the interactive aspect of naturally occurring social processes, they are not in themselves naturally occuring interactions. They offer no guarantee as to what people say, or how they interact, outside the research context.
Sampling – purposive sampling (participants selected on basis of having a significant relation to the research topic) – but known to each other beforehand as recruited from a particular training group. This may create problems with established relations of power, disagreement or consensus being brought into research setting (Tonkiss 2004: 201–2). Assertive voices are more likely to direct the group discussion (as I found). Furthermore, 'Familiarity can limit self-disclosure and discourage disagreement' (Litosseliti 2003: 54).
Can familiarity be seen as an advantage though? I would argue yes, it is a strength. The men had very varied responses to my questions and their friendship and familiarity allowed their differences in opinions and interpretations to be discussed openly as they knew each other sufficiently to show disagreement and to disclose private information. The fact that they had common experiences also meant that as a group they were able to give specific examples of events or people that they all knew about. I would suggest that the friendship dynmatics of the groups (many of the group saw each other socially after the course had ended), the subject of the research (their experiences of unemployment), a setting for the discussion with which they were all familiar by this stage (a room at the university of Leeds where they had already spent 4 days as part of the course), and the fact that I was known to them and had already built a friendly rapport with them – that all these factors helped the focus group to be successful as the participants felt at ease 'in the familiar' and were keen, lively and open in their interaction with me and the other participants.
Although they shared various commonalities (their age, gender and the fact they were long-term unemployed) their backgrounds were by no means homongenous. They had access to varying amounts of cultural, social and educational capital, and this affected their interpretations of their situations and meant that they often openly challenged other participants' interpretations. They had access to different discourses e.g. one member of the group was familiar and well-versed in pro-feminist discourses, and challenged other participants about their understandings and interpretations of women and work:
If mutliple understandings and meanings are revealed by participants, multiple explanations of their behaviour and attitudes will be more readily articulated (Gibbs 1997).
Gibbs, A. 1997. Social Research Update (19). Department of Sociology: University of Surrey. Available online at: link Accessed 03/06/05
Litosseliti, L. 2003. Using FocusGroups in Research. London and New York: Continuum.
Tonkiss, F. 2004. 'Using Focus Groups'. In Seale, C. (ed.). 2004. Reasearching Society and Culture (2nd ed.). London: Sage.