Qualitative_Analysis_in_Psychology_and_the_Arts

Jarrod Israelstam Qualitative research is a multifaceted set of tools which has arisen, in part, from the inadequacies of the quantitative paradigm. The data collection and analysis methods it utilizes are aimed at representing the subjective nature of reality. As Guba and Lincoln (1994) argue, the division between ontology and epistemology breaks down in the qualitative paradigm because ontology is no longer about ‘essences’ and epistemology is no longer about capturing them. Thus, in discussing qualitative epistemology, this essay will make reference to the epistemology or knowledge of real people. In other words, because we cannot capture the objective realities of participants in research, we must try to capture their constructions instead (Denzin & Lincoln, 2005). Specifically, this essay will examine the nature of the self, the nature of data produced in interviews, focus groups and naturally occurring data, and how methods such as CA, MCA, narrative analysis and thematic analysis can be useful tools in the continuing development of the qualitative paradigm. According to Denzin and Lincoln (2005), qualitative research is the interpretive study of phenomena in their natural setting. It is a means of representing the world, as well as inevitably transforming it. It is, at its base, an approach which does not seek to avoid issues of bias or present unquestionable facts. Rather, the qualitative researcher admits that she or he is biased in very particular ways and makes an attempt to reflect on and tackle these biases (Guba & Lincoln, 1994). She or he recognizes that ‘facts’ can always be contested and are constantly changing. Qualitative research is able to move past this seeming obstacle of uncertainty. It evaluates claims by how well they fit all available information, not according to some guidelines for ‘absolute truth’ (Denzin & Lincoln, 2005). This is because truth itself is a construction, and not ‘real’ like a physical object (Denzin & Lincoln, 2005). The constructed nature (or ontology) of psychological experience and truth demands research methods and analysis which speak to this nature (Guba & Lincoln, 1994). Qualitative methods employ hermeneutics, a way of interpreting human psychological life through human interpretation. Freeman (1993) argues that in a basic sense, there is no un-languaged world. It has been argued by many that human social life is constituted through language, and thus the study of meaning must proceed through analysis of “linguistic data” (Polkinghorne, 1988, p. 7). Thus, hermeneutics is interpretation through applying a researcher’s particular, linguistic understanding. This is in opposition to the interpretation of impersonal data via numbers i.e. statistics. As Szasz, a fervent supporter of the anti-psychiatry movement put it, “man’s sign-using behaviour does not lend itself to exploration and understanding in the terms [of physics and chemistry]” (Szasz, 1961, p. 3). Qualitative research produces a picture of the world that seeks out and accounts for ‘inconsistencies,’ rather than hopes they don’t exist (Denzin & Lincoln, 2005). Parenthetically, critiques of arrogant quantitative truth claims (in a manner of speaking) can also be found in non-Western literature which is in no way concerned with qualitative methods. Sayeed Qutb’s Islamist critique of modernity is that ‘reason’ is the new god, who we pray to by conducting experiments (Euben, 1997). We must thus keep the limits of reason and sense-based investigation in sight. We need not assume that can get nor even need to have absolute knowledge. However, whilst we might thus see reliability and validity in research in a different way, we need not abandon what it was designed to protect (Guba & Lincoln, 1994). Two important purposes behind these quantitative concepts is that knowledge should not be arbitrarily generated or produced without theoretical justification (Guba & Lincoln, 1994). However, the qualitative approach has replaced these with criteria which do not assume the existence of an objective social reality against which findings can be compared. The notions of trustworthiness, transferability and dependability correspond to internal validity, external validity and reliability, respectively (Guba & Lincoln, 1994). However, Guba and Lincoln (1994) argue that these criteria are suspect exactly because they mirror the quantitative criteria. Thus they propose other criteria such as ontological authenticity (prioritizes participants’ view), educative authenticity (increases understanding of others), and tactical authenticity, which “empowers action” (Guba & Lincoln, 1994, p.114). These criteria can be met by, amongst other ways, triangulating data. For example, one may use several different methods and sources, represent the perspectives of diverse participants and make use of more than one researcher (Denzin & Lincoln, 2005). The above is a general account of the model which qualitative methods are founded on. However, as the social, natural, political and economic climates of the world have changed, qualitative methodology has evolved. Notably, this evolution is put in sharper focus by a corresponding stasis in the quantitative paradigm (Denzin & Lincoln, 2005). For example, the qualitative approach has grown methodologically from participant observation and ‘objective’ accounts thereof, to textual studies and narrative analysis. It has also grown theoretically, with the incorporation of critical, feminist, and hermeneutic ‘sub-paradigms’ (Denzin & Lincoln, 2005). Quantitative research has stayed in what Denzin and Lincoln (2005) call the traditional and modern historical moments, which ostensibly ended in 1970. These periods are characterized by cause-and-effect experimentation on human subjects. Only the degree of the certainty of truth claims changed between the two periods, characterized by positivism and post-positivism (Harré, 1981). That is, whilst some psychologists in the quantitative tradition adopt(ed) a post-positivist epistemology, all still essentially remain in the same mould. This ‘mould’ is scientism. This is the belief that the methods of the natural sciences are supreme and applicable to all disciplines, and is only part of the problem. Sarbin (1986) points out that the quantitative perspective takes as its starting point a totally misunderstood version of the physics model. Bourdieu (1999) argues that scientism ironically leads to unscientific enquiry. Instead of viewing the world as a dynamic thing (in which parts interact with each other) as physicists actually do, quantitative psychology has ‘inherited’ a law-bound model of relationships between static things (Harré, 1981). As Sarbin (1986) argues, the lessons of quantum physics are being taken more and more seriously in the natural sciences in general. These are the lessons of uncertainty. For example, Heisenberg’s uncertainty principle of 1927 essentially argued that the very act of measuring affects the experiment (Sokal, 1996). Ignoring this now classic principle, the quantitative psychology model seeks to describe cause and effect relationships as if they were discoverable in the world ‘out there,’ as if they were the workings of a machine (Sarbin, 1996). Once such “causes” are discovered, they inevitably become concrete – or reified. From that moment, they are publicly viewed as ‘real’ and therefore everything else is evaluated in relation to these ‘real’ things (Denzin & Lincoln, 2005). However, Sokal (1996, p. 218) contends that physical reality, just like social reality, is essentially a “social and linguistic construct.” Therefore, as Denzin and Lincoln (2005) argue, we cannot have a value-free science, and should not aim to have one. Scientific knowledge reflects and maintains the hegemonic ideology of a particular time and culture (Sokal, 1996). Thus, many qualitative researchers respond to this as a challenge, and seek to engage in knowledge production that equalizes the power balance and considers unheard voices (Denzin & Lincoln, 2005). For example, qualitative researchers are generally expected to think constantly during research about their perceptions and biases and record these in both reflexivity journals (during research) and a reflexivity section of their final report (Henwood & Pidgeon, 1992). This injunction towards reflexivity, unfortunately, did not save qualitative researchers of the past from representing non-Western cultures as strange and exotic subjects of study (Denzin & Lincoln, 2005). However, it has the potential to prevent such mistakes in the future. Qualitative research is still, after all, interested in studying all cultures and is concerned to gain an inside perspective on whatever its subject-matter may be (Denzin & Lincoln, 2005). This is known as ‘emic’ research (Guba & Lincoln, 1994). ‘Etic’ research, usually associated with quantitative psychology, imposes theories, research questions and norms (moral norms and test norms) which come from one population on another population (Guba & Lincoln, 1994). The former is typically Western, although not always, whilst the latter is traditionally non-Western. In this way quantitative research seeks to generalize its findings on many different populations and thus add to its own credibility as a paradigm (Denzin & Lincoln, 2005). Qualitative research, on the other hand, first produces local data and then cautiously theorizes about it, applying what is locally relevant (Denzin & Lincoln, 2005). Studying participants in their natural context is indeed a pillar of qualitative research. However, in an important sense, any context will be unnatural to some degree. Whilst one can attempt to do ‘natural’ or field research, and whilst this may have advantages over lab experiments, one must realize that ‘natural’ will always have to be placed in scare quotes (Potter, 2002). As Potter (2002) argues, we need to consider most data naturalistic rather than natural. Only if the data would have existed in exactly the same form if the social scientist was dead, is data actually natural (Potter, 2002). As Speer (2002) notes naturalness can be viewed on a continuum, depending on the methods used to gather data and the involvedness of the researcher. However, Speer (2002) cautions that all data is affected by the researcher, whether through recording participants’ speech or only in analysis (e.g. of texts). Thus, Speer (2002) proposes that we view the naturalness of data through how participants orient to it – whether their actions demonstrate that they are in a natural context or not. Speer (2002) argues that the natural/unnatural distinction could be abandoned as basically meaningless. However, Potter (2002) points out that his ‘naturalistic’ data is still more natural than some interviews for example, and importantly, than experiments. It could be argued that this is a valid point, even if we do take into account the participants’ orientation to the situation as Speer (2002) suggests. The notion of “insider epistemology” has interesting ramifications for the debate on studying people and cultures in their natural context (Fay, 1996a, p.9). This notion is essentially that to understand people of a certain type, one has to be one of those people (Fay, 1996a). Some qualitative researchers seem to adopt this perspective. For example, Bourdieu (1999) conducted research, partially borrowing the method of Labov for conducting narrative interviews. Bourdieu (1999) argues that if one employs an interviewer who is practically the same as the interviewee(s) in ways relevant to the context, one can be more in step with the feelings, thoughts, and explanations of the interviewee. Bourdieu (1999) implies here that this kind of interviewer can have more empathy or even solidarity with the interviewee. Bourdieu (1999) assumes that even if, unavoidably, the interviewer objectifies the interviewee, that participant is aware that she or he shares the aforementioned relevant characteristics with the interviewee. Therefore, she or he will feel at ease, as the two parties share “the risks of exposure” created by the interviewee’s answers being known by others (Bourdieu, 1999, p. 611). Whilst it is possible that participants may feel more comfortable with interviewers who are similar to them, this is not necessarily the case, especially if the participant does not perceive the similarity. Furthermore, even if true, this idea does not imply that the researcher will know the participant better. Fay’s (1996a) argument is that one does not have be one to know one – in fact, sometimes (as in his Hitler example) being one is a hindrance to knowing one. Thus we should conduct and evaluate ‘emic’ research critically, or risk operating on simplistic notions of what it means to know another. This is not to say that researchers should not make every effort to see things from the participants’ points of view as given by the participants. As Bourdieu (1999) argued, it seems inevitable that researchers will objectify participants, whether the research is quantitative or qualitative. Participants will sometimes even accept the way they are thus positioned by the researcher (Bourdieu, 1999). However, by viewing participants not as passive vessels of knowledge but as an agents who make their own orientations during an interaction, researchers can gain greater insight into how knowledge is constructed (Holstein & Gubrium, 2004). In this sense, qualitative research seems to have the upper hand over quantitative. Potter and Hepburn (2005) point out that interviewees are often recruited because of their membership in some kind of social category relevant to a particular study (such as ‘Indians’). However, these authors caution that in analysis, one must be careful to see how the participant has actually oriented. Potter and Hepburn (2005) point out that interviewees may resist answering questions in the way they were designed for example by deeming some of it irrelevant in their answer or by answering it from a different footing than was expected. Participants even construct their talk, at times, as if it were someone else’s opinion, which performs self-protective functions (Wilkinson, 2004). Stokoe (2006) argues that the assumption of gender as a category relevant to participants has been problematic over the years. Stokoe (2006) points out that the problem has often come when researchers have analysed data by simply counting the number of times a concept – such as gender – comes up in a transcript. This suggests the use of quantitative content analysis, although this is not specified. Membership categorization analysis (MCA) has several advantages over such methods as content analysis. For example, it recognizes that categories are conceptually built in collections; thus, the category ‘male’ usually goes together with the category ‘female’ (Schegloff, 2007). Thus, for instance, when such categories are found in unusual configurations, there may be some important meaning for one to examine (Schegloff, 2007). Another useful point about MCA is that one can assume with caution that if Y refers to a behaviour characteristic of a category by X, they are activating that category with respect to X (Schegloff, 2007). We can avoid assuming the use of categories by participants in our analysis. However, it must ultimately be recognized that facts are not discovered from analysis after interviews, but are built up during interviewers by both interviewer and interviewee (Holstein & Gubrium, 2004). Not only are interviews inevitably built around a researcher’s own theoretical concerns, through constructing questions in a closed-ended way participants are influenced to respond in very limited ways (Potter & Hepburn, 2005). This is the sense in which Holstein and Gubrium (2004, p. 142) argue that interviews and interviewers are “unavoidably active.” This activeness is ignored by quantitative interviewers, which compounds the fact that their interviews are closed-ended. However, by making research qualitative and open-ended, one does not avoid this problem. What this does do is make the focus of research interaction (Silverman, 2001). Participants always modify the experiences they have had as they report them, of their own accord or because of something suggested by the interviewer. But this is not necessarily problematic. As Holstein and Gubrium (2004) argue, this begins to take place even before the interview has started. Answers to a particular interview schedule will also be different across interview occasions (whether using the same or a different interviewee). Wilkinson (2004) makes similar points in relation to focus groups. According to Wilkinson (2004) focus group talk does not reveal pre-existing facts about psychology or society, but constitutes these facts in real time. So, as participant-researcher talk is interactive, there must be means to analyse interaction. Quantitative analysis, with its reliance on frequency counts and correlations, and its focus on variables, not people, is unable to supply such means. This is because it is not designed to study interaction; in fact, it is designed to study something which utterly lacks interaction (Guba & Lincoln, 1994). The only potentially relevant tool for ‘linguistically’ analyzing interviews in the quantitative paradigm, it appears, is content analysis (Wilkinson, 2004). This is related to thematic analysis: whereas content analysis counts instances of themes, thematic analysis builds a picture of the relationships between themes [and sub-themes] (Braun & Clarke, 2006). It is only arguable that thematic analysis is a quantitative tool, and researchers often seek to ‘convert’ the qualitative sdata obtained into quantitative data (Boyatzis, 1998). Neither content analysis nor thematic analysis can truly analyse interaction. But in interaction, there is an entire world wholly un-examinable by statistics. By contrast, the qualitative approach has a number of methods for analyzing interaction (Stokoe, 2006). Conversation analysis is concerned with what people do with their talk and how talk is given meaning by context (Heritage, 2004). It is concerned to find out how conversations are constructed such that certain outcomes occur (Silverman, 2001). This is through the fine examination of turns at talk, interruptions, problem-solving (or ‘repair) and many other features of talk such as tone and pauses (Silverman, 2001). All of these details, along with others, are vital for understanding how human beings relate to each other, not in an abstract sense but in ways that are visible (or audible). When a participant says something, we cannot take it at face value, and we cannot know what its meaning is without examining the interpersonal and social context in which it was said (Heritage, 2004). One important kind of context is the institutional context (Silverman, 2001). Institutional talk is structured differently by its very status of being within contexts such as courts of law, television, telephone conversation or university classroom (Heritage, 2004). Of course, people may also ‘just’ engage in conversations, but conversations have their own contexts and constraints. In the same way that talk cannot be taken at face value, textual information – whether written for the purposes of research or ‘naturally’ occurring – cannot be said to speak for itself (Ricoeur, 1981). It is imperative to look at the world behind the word (Ricoeur, 1981). It is argued that power operates through text. Texts are powerful items because the recording of words tends to have a freezing or concretizing effect (Prior, 2004). For example, government records of birth and death tend to be the sole factor determining people’s social benefits (Prior, 2004). The meaning of a text is shaped also by the style of writing, as well as the intended audience (Prior, 2004). Thus, a government agency will expect the use of formal writing, and a historian may write for the audience of Bush’s America. Whilst historians do not simply concoct stories and impose them on a meaningless series of events, the texts of history determine who is portrayed favourably, who ‘won’ and what should be considered as fact - or opinion (Fay, 1996b). Texts are not written with ‘no purpose’ in mind – they are intended to achieve certain actions, and must be read as such (Prior, 2004). These actions may include the construction of experts and non-experts (Prior, 2004). However, more often than not, they include conveying (hidden) moral imperatives and value-judgments (Drew, 2006). It is tempting to simply analyse the semantic meaning of texts, but this is a grave mistake. This approach treats texts in a totally uncritical way which, through not uncovering the power invested in texts, simply re-inscribes it (Prior, 2004). To take Drew’s (2006) example, there were many reports during the ‘70s, ‘80s and ‘90s of violence in Ireland. Such a report could have read: ‘the Protestants were involved in a shooting, as aggressors, who came in ‘hordes’ and seriously injured 100s.’ If a reader simply believes the statistics, the apportionment of blame, and the details of the event then she or he will reproduce the information as fact. What is necessary is that one looks at the source of the information – who is saying it and what their political, economic religious or other commitments are (Prior, 2004). As Drew (2006) points out, texts present accounts (and manipulations) of physical and social reality, both in the writing process and the reading process. Viewing research as ‘only’ getting at constructions and not absolutes has some very important social ramifications. For example, Prior (2004) shows how diagnosis using the Diagnostic and Statistical Manual (DSM) and interviews based on it are implicated in the deceptively convincing ‘reality’ of mental illness. As Eisenberg (2007) argues the makeup of mental disorders such as ADHD and depression has changed so vastly in the past 60 years that it is difficult to believe they exist. Added to this is the huge number of different ways that disorders such as these are defined, diagnosed, and treated between one psychiatrist and another (Eisenberg, 2007). More damning is the evidence that “[drug] company-sponsored studies are more than 4 times more likely” to find the existence of mental disorders than studies with other funding (Eisenberg, 2007, p. 281). Costello and Duncan (2006) show that disorders are constructed in a way that demonizes and marginalizes people who do not exactly fit social norms. Prior (2004) points out that much of this work is done through texts. The use of such means as statistical tables to represent facts about disorders is misleading, as many disorders are diagnosed using essentially arbitrary cut-off points. If the cut-off were to be moved, either fewer or greater numbers of people would have a disorder (Prior, 2004). This, along with the perpetual co-existence of disorders with each other, shows that mental functioning should be viewed on a spectrum, not in categories (Prior, 2004). The DSM represents things in the way it does because it is compiled in a “politically structured space,” by a small elite of interested individuals (Prior, 2004). It might even be argued that it is only because the DSM has been continually edited and re-published in the last 60 years that the concept of mental illness is considered a valid topic (Prior, 2004). In other words, one is left to wonder whether the DSM came first, or ‘mental illness’ did. What is key is to understand the deep meaning behind texts, both semantically and politically. Such an understanding can be produced by conversation analysis and other methods for analyzing talk (Drew, 2006). For example, Drew (2006) shows that in analysing suicide notes, the complex nature of rationality is revealed. It may immediately seem that suicide is irrational, but Drew (2006) shows how suicidal people (try to) convince themselves and others that it is a reasoned decision. For example, it is often constructed as the only available option (Drew, 2006). Conversation analytic studies of suicides can show the way that people who are considered to be contravening one of the most emotively-charged social norms can acknowledge this (Drew, 2006). They may thus hope to be seen, after their death (or failed suicide) as normal, acceptable individuals. Importantly, analysts look at the author’s orientations to norms or relationships (for example) rather than assuming their importance (Drew, 2006). The notion of meaning is another key point of division between quantitative and qualitative methods (Guba & Lincoln, 1994). Because of the sometimes uncritical application of mathematics – numbers – to the outcome of research, it may be argued that the quantitative approach deprives research of its meaning (Guba & Lincoln, 1994). Research is undertaken in order to gain insight, yet the convenience of statistics and its usefulness in some situations has lead quantitative researchers into an unseen trap. This ‘trap’ is essentially the assumption that the number obtained for one participant mean anything the same as the numbers obtained for another participant (Harré, 1981). If one considers just how many numbers are applied to each of, say, 100 participants in almost any statistical research the problem is compounded over and over again. Harré (1981) argues that this problem is added to by conceptual confusion in creating research questions at the ‘input’ end of research. It is suggested by Harré (1981) that this confusion is caused by the same refusal to take linguistic meaning seriously that is implicated in the ‘output’ end of such research. As Harré (1981) argues, it may be that the effect of the experiment was the same for each participant but because of the definitional confusion, the explanation for this is totally uncertain. It could equally be that something different happened to each participant. Of course, quantitative research is sometimes far less confused than the ‘helping behaviour’ example cited by Harré (1981). However, it could be argued that no matter how well-defined its terms are, the original problem of meaning will still remain (Harré, 1981). The social ramifications of this are tremendous, as a great deal of money is spent on research which is potentially useless. Another aspect of meaning and ontology that is ill-understood in the quantitative approach is the nature of the self. The self has traditionally been seen as an objectively-existing, independent thing within individuals which is relatively constant (Freeman, 1993). For example, theorists such as Costa and McCrae have developed theories of ‘traits’ or fairly stable elements of personality (McAdams, 2001). Whilst many theorists today accept that people are generally not determined by these traits but that behaviour depends on particular situations, most are still committed to notions of the single self disconnected from time, other selves or society. This is essentially a simplistic notion of the self. Whilst in body, all people are clearly individual, people do not merely interact with their bodies but rather through sociality (Crites, 1986). Freeman (1993) argues that we construct our own selves – as well as other selves – through narratives. Narratives are nothing more than stories which we tell ourselves. This is not necessarily to say that people delude themselves. However, as Sarbin (1986) argues, all people live according to a plot structure. This plot is, importantly, coherent in the sense that causal links between events are consistently inserted, even where the causal relation is spurious (Gergen & Gergen, 1986). This structure is comparable to a number of different kinds of plot including the tragedy, the development narrative and the moral tale. Importantly, we continuously re-construct our own selves to maintain the kind of identity that we would like to have (Sarbin, 1986). And as Freeman (1993) argues, our narratives are inherently affected by our memory and memory-failures. Thus the idea of an objectively existing self begins to break down. One might still argue, however that the self is relatively constant (Crites, 1986). After all, one might not think about oneself and immediately imagine having different selves. Of course, one may realize that one has changed over time, but then the time periods – past, present and future – are viewed separately (Crites, 1986). The continuity we perceive in ourselves is in large part due to the continuity of memory (Crites, 1986). As Fay (1996b) argues, we cannot conceive of the self – or any other narrative – as set in stone, with clear borders between beginnings, endings and middles. We impose those structures on the narrative, it is argued (Fay, 1996b). The view of the self as sharply defined comes from viewing life histories as simply existing in the world. However, as Crites (1986) argues we are constantly recollecting the past (perhaps the good times of childhood) and projecting the future (our hopes and dreams) in order to create our present. Epistemologically speaking, to know ourselves we must recall, as well as integrate, what we know about our past (Crites, 1986). As McAdams (2001) argues, we need synchronicity or coherence within ourselves in order to function. This coherence must be created, both across time and between our discordant feelings and opinions (McAdams, 2001). From the hindsight of the present, we attribute meanings to past events which could not have been attributed to them at the time (Fay, 1996b). For example, a person may say ‘That was when I started my life’s career’ without even realizing how and when they actually came to know that. Thus, the self is essentially emergent. Is also constructed in different ways for different listeners and in different social situations (Riessman, 2008a). In the example of Sunita’s life, Reissman (2008a) shows how Sunita came into the interview with a set of expectations about what was going to be asked and who and what the interviewer was. Thus Sunita constructed her self, within a narrative, in a way that was comprehensible for a Western interviewer who did not fully understand Indian culture. Thus, Reissman (2008a) argues, there are only intersubjective representations of the self, and not real selves. Thus, in using narrative analysis, researchers try to pick out how interviewees have organized their stories, according to several different criteria of structure. For example, Gergen and Gergen (1986) point out that stories may be structured towards an endpoint or goal which acts as an impetus for moving the plot along. This teleology is made largely with reference to the construction of theory by psychologists, but – as developmental theories may seen as stories – a psychologist’s narrative and an ‘ordinary human being’s’ narrative is essentially no different. Gergen and Gergen (1986) also show that narratives may be designed to include strictly or substantively unnecessary dramatic twists, to gain and hold the audiences attention and to convince the audience of a point. Reissman (2008b) shows that there is, across interviews, a pattern of narrative elements. These are not stages in a story, as the storyteller can move back and forth between them. Rather they are regular ways in which stories are told: abstract, orientation, complicating action, evaluation, resolution and coda. Wengraf (2001) argues for the use of minimalist questions which prompt the interviewee as little as possible. However, Reissman (2008b) argues that there is a strong narrative impulse in people, which does not have be ‘brought out’ by a question phrased in a narrative way. The existence of the narrative elements shows how the self is not objective, but rather a projection of an image for an audience. In conclusion, it appears that the significance of the qualitative-quantitative debate is along three dimensions: ontological, epistemological and social. It has been argued that the role of research in society is important and its power is great. Thus, this paper has advocated adopting the qualitative approach due to the ontological and epistemological advantages it has for producing knowledge. It has argued that there are, at present, several methods of analysis available for understanding the common-sense ways in which people use language. And, without a valid way of understanding selves, how they construct knowledge interactively and how these constructions are co-created by researchers, the fate of psychology may be a regress towards determinism and scientism. References Boyatzis, R. E. (1998). Transforming qualitative information: Thematic analysis and code development: London: Sage. Braun, V. & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3, 77-101. Costello, L. and Duncan, D. (2006). The ‘evidence’ of sex, the ‘truth’ of gender: Shaping children’s bodies. Children’s Geographies, 4 (2),157-172. Crites, S. (1986). Storytime: Recollecting the past and projecting the future, chapter 8 in Narrative psychology: The storied nature of human conduct. New York: Praeger. Denzin, N.K., & Lincoln, Y.S. (2005). Introduction: The discipline and practice of qualitative research, chapter 1 in The Sage handbook of qualitative research. London: Sage. Drew, P. (2006). When documents “speak”: Documents, language and interaction. In P. Drew, G. Raymond & D. Weinberg (Eds.), Talk and interaction in social research methods (pp. 63-80). London: Sage. Eisenberg, L. (2007). Commentary with a historical perspective by a child psychiatrist: When “ADHD” was the “Brain-Damaged Child.” Journal of Child and Adolescent Psychopharmacology, 17 (3), 279-283. Euben, R. L. (1997). Comparative Political Theory: An Islamic Fundamentalist Critique of Rationalism. The Journal of Politics, 59 (1), 28-55. Fay, B. (1996a). Do you have to be one to know one' Chapter 1 in Contemporary philosophy of social science. London: Blackwell. Fay, B. (1996b). Do we live stories or just tell them' Chapter 9 in Contemporary philosophy of social science. London: Blackwell. Freeman, M. (1993). Rewriting the self, chapter 1 in Rewriting the self. London: Routledge. Gergen, K.J. & Gergen, M.M. (1986). chapter 2 in Narrative psychology: The storied nature of human conduct. New York: Praeger. Guba, E.G. & Lincoln, Y.S. (1994). Competing paradigms in qualitative research, chapter 2 in Handbook of qualitative research. London: Sage. Harré, R. (1981). The positivist-empiricist approach and its alternative, chapter 1 in P. Reason & J. Rowan (Eds.). Human Inquiry: A Sourcebook for New Paradigm Research. Chichester: John Wiley & Sons. Henwood, K.L. & Pidgeon, N.F. (1992). Qualitative research and psychological theorizing. British Journal of Psychology, 83, 97-111. Heritage, J. (2004). Conversation analysis and institutional talk. In D. Silverman (Ed.) Qualitative research: Theory, method and practice (2nd ed.). London: Sage. Holstein, J.A. & Gubrium, J.F. (2004). The active interview. In D. Silverman (Ed.). Qualitative research: Theory, method and practice (2nd ed.). London: Sage. McAdams, D.P. (2001). The psychology of life stories. Review of General Psychology, 5 (2), 100-122. Polkinghorne, D.E. (1988). Introduction, chapter 1 in Narrative Knowing and the Human Sciences. New York: State University of New York Press. Potter, J. (2002). Two kinds of natural. Discourse Studies, 4 (4), 539-542. Potter, J. & Hepburn, A. (2005). Qualitative interviews in psychology: problems and possibilities. Qualitative Research in Psychology, 2, 281-307. Prior, L. (2004) Doing things with documents. In D. Silverman (Ed.). Qualitative research: Theory, method and practice (2nd ed.). London: Sage. Reissman, C.K. (2008a). Constructing narratives for enquiry, chapter 2 in Narrative Methods for the Human Sciences. London: Sage. Reissman, C.K. (2008b). Structural analysis, chapter 4 in Narrative Methods for the Human Sciences. London: Sage. Ricoeur, P. (1981). The model of the text: Meaningful action considered as a text, chapter 8 in Hermeneutics and the Human Sciences. Cambridge: Cambridge University Press. Sokal, A.D. (1996). Transgressing the boundaries: Towards a transformative hermeneutics of quantum gravity. Social Text, 46/47, p. 217-252. Sarbin, T.R. (1986). The narrative as the root metaphor for psychology, chapter 1 in Narrative psychology: The storied nature of human conduct. New York: Praeger. Schegloff, E.A. (2007). A tutorial on membership categorization. Journal of Pragmatics, 39, 462-482. Silverman, D. (2001). Naturally occurring talk, chapter 6 in Interpreting Qualitative Data: Methods for Analyzing Talk, Text and Interaction. London: Sage. Speer, S.A. (2002). ‘Natural’ and ‘contrived’ data: a sustainable distinction' Discourse Studies, 4 (4), 511-525. Stokoe, E.H. (2006). On ethnomethodology, feminism, and the analysis of categorical reference to gender in talk-in-interaction. The Sociological Review, 54 (3), 467-494. Szasz, T.S. (1961). The myth of mental illness: Foundations of a theory of personal conduct. New York: HarperCollins. Wengraf, T. (2001). Preparing lightly-structured depth interviews, chapter 6 in Qualitative Research Interviewing. London: Sage. Wilkinson, S. (2004). Focus group research. In D. Silverman (Ed.). Qualitative research: Theory, method and practice (2nd ed.). London: Sage.