Interpretative phenomenological analysis: a reply to Amedeo Giorgi.
Abstract: From a concern with the scientific standing of phenomenology, Giorgi (2010) criticizes two forms of phenomenological practice: interpretative phenomenological analysis (IPA) and the approach of Collaizzi, along with three doctoral dissertations. In this paper I will respond to the critique of IPA presented by Giorgi.
Article Type: Report
Subject: Existential psychology (Research)
Philosophy of mind (Research)
Phenomenology (Research)
Author: Smith, Jonathan A.
Pub Date: 07/01/2010
Publication: Name: Existential Analysis Publisher: Society for Existential Analysis Audience: Academic Format: Magazine/Journal Subject: Psychology and mental health Copyright: COPYRIGHT 2010 Society for Existential Analysis ISSN: 1752-5616
Issue: Date: July, 2010 Source Volume: 21 Source Issue: 2
Topic: Event Code: 310 Science & research
Geographic: Geographic Scope: United Kingdom Geographic Code: 4EUUK United Kingdom
Accession Number: 288874193
Full Text: Context

I first articulated IPA as one approach to experiential qualitative psychology in the mid 1990s (Smith, 1996). The approach is phenomenological in being concerned with participants' lived experience and hermeneutic because it considers that experience is only accessible through a process of interpretation on the part of both participant and researcher. IPA is also idiographic as it is committed to a detailed analysis of each case. IPA has now become well established as a qualitative methodology in psychology and increasingly in cognate disciplines. Much of the research that has been conducted so far has been in the physical and mental health field. However there is growing work in other areas too- e.g. sports science, education, music. I am currently conducting a review of the IPA corpus (Smith, in prep).This offers both a descriptive overview of the work that has been done and points to ways of evaluating its quality.

Selectivity

I think it is fine for people to critically evaluate my work. I recognise it as part of the scholarly and scientific enterprise. However I am very surprised that Giorgi should choose to base his sweeping critique of IPA on just two book chapters (Smith and Osborn, 2008; Eatough and Smith 2008) and indeed there is only passing reference to the second of these. There is a large corpus of writing on IPA by me and increasingly by others too. This corpus is made up of theoretical pieces, methodolological guides, completed studies, a major book. One cannot say everything in a single chapter or paper and different publications have different purposes. I would have hoped that he might have looked at rather more of this corpus and sampled the range of types of IPA writing before launching his critique. He would then be in a position to make a more complete and less selective assessment.

The book chapter on IPA on which Giorgi bases his critique has a particular aim--to offer to psychology students with little background in qualitative research an accessible overview of conducting an IPA study. Such students are so often daunted and put off by the prospect of doing qualitative research that I thought it valuable to make the chapter accessible, practicable and to offer an encouraging tone. Clearly the chapter doesn't say all there is to say and the interested student can then turn to the book on IPA (Smith, Flowers and Larkin, 2009) and other papers for more detailed treatment of its underlying philosophy and the process of data collection and analysis and for examples of completed studies.

Philosophy

Giorgi contends that IPA does not have a grounding in the philosopy of phenomenology and hermeneutics. He states, for example, that:

The theory and practice he [Jonathan Smith] recommends has little to do with continental philosophical phenomenology.

(Giorgi, 2010, p4)

This is not the case. IPA clearly has theoretical underpinnings in phenomenology and hermeneutics. This relationship is discussed in a number of papers (e.g. Smith, 2004, 2007; Larkin, Watts & Clifton, 2006) and also in one of the two chapters Giorgi himself cites in his paper (Eatough and Smith, 2008). Most significantly however, Giorgi makes no reference in his paper to the book on IPA by Smith et al. (2009) which gives extensive treatment of the key thinkers in phenomenology and hermeneutics (Husserl, Heidegger, Schleiermacher, Gadamer et al.) as well as outlining explicitly the relationship IPA has to them. The book is referred to in one of the two IPA chapters Giorgi cites and it was published before the submission date for his journal article.

Science

Giorgi claims that IPA is unscientific:

I agree with Giorgi that qualitative research should endeavour to be scientific and I believe IPA is. This is because I have a different definition of science from Giorgi.

Two of his key concerns are with prescription and replication. Let us take prescription first. It is not clear from what he writes whether Giorgi believes IPA to be prescriptive or to not be prescriptive. Giorgi appears to criticize IPA for not being prescriptive but then goes on to suggest we are being disingenuous:

Here he seems to be saying that IPA claims to be non-prescriptive but that in reality it aims to be. To support this, Giorgi states that:

Well here we will just have to disagree on the semantics. For me the whole point of suggestions is that they are not prescriptions.

So let me respond to Giorgi's charge. It is the case that in most of my writing on IPA, I argue that it is not a prescriptive methodology. The point I am making is that processes in qualitative research are not the equivalent of the carefully prescribed procedures in quantitative research. And like Allport, I am wary of 'methodolatory'. It is the intellectual and intuitive work that is done at each stage that produces the result of a piece of qualitative research. Doing good IPA requires the development of some complex skills--interviewing, analysis, interpretation, writing, and researchers at different stages will have different degrees of fluency and adeptness at these skills. It is the degree of proficiency in these skills which will influence the quality of the research carried out more than the conscientious following of procedures.

Thus I cannot prescribe exactly how to conduct a good interview and then analyze it. I can offer some guidelines to good practice based on my professional judgement and experience. However, following the guidelines is no guarantee of doing good work. Adhering to the same protocol, one researcher may produce something inspired, another something pedestrian. That is because the result depends on the professional and personal skills of the researcher, as well as the inclination of the participant and the various interactions that occur in the meeting of researcher and participant. Giorgi completes this section of his paper with the claim:

This is not the case. There is not total freedom. Constraints come into play in a number of ways. First one has to achieve a threshold of proficiency in a range of skills in order to do the work. These can be developed during training and monitored in supervision. And in these early days of training in qualitative methods for psychologists, ones expectation of a piece of work should take into account the amount of training received. Second, there are a set of research steps that need to be made and these do need to be made in the appropriate order. But the point is that there is flexibility in how each of these steps is taken- hence we are advocating a balance between stricture and flexibility. Finally there are quality control criteria which help guide the researcher to conduct good work as well as the reader to evaluate it. I will discuss these further below. Now let us turn to the issue of replicability, with regard to which Giorgi states:

Note that Giorgi elides checking and replication here. But let's first discuss replication. I will come back to checking afterwards.

In my experience, the majority of qualitative psychologists do not see replicability as a criterion they aspire to or consider their work should be judged against. This is precisely because the construct derives from paradigmatic assumptions which do not necessarily apply in, or sit easily with, human science research. Indeed there is currently lively debate in the qualitative psychology community over criteria for evaluating the validity of qualitative research. This has largely been prompted by qualitative researchers' frustration at their work being evaluated according to quantitative criteria like replicability and, therefore, one move has been towards establishing alternative criteria for assessing the quality or validity of qualitative research. For an example of a highly regarded paper presenting the debate about quality in qualitative research as well as alternative criteria by which to evaluate qualitative psychology, see Yardley (2000).

Personally I do not see replicability as an appropriate referent for judging most qualitative approaches to psychology (content analysis might be an exception). Qualitative research is a complex, interactive, dynamic process and it is not clear exactly what one would be expecting to replicate. Two highly skilled experiential qualitative researchers might interview the same participant about a topic of existential importance. Both can follow the same interview schedule and the recommended guidelines for good practice. However interviewer B would not be expected, nor able, to replicate the interview of interviewer A. That doesn't mean the interviews were not successful, or that they were not scientific. The evaluation needs to be on different grounds.

Scholars concerned with attempting to define criteria suitable for judging the validity of qualitative research have come up with a wide array of suggestions. Among these are: commitment, transparency, plausibility.. In Smith et al. (2009) we describe in considerable detail how these type of constructs can be employed to help evaluate a piece of IPA research. I am happy for the scientific basis of IPA to be assessed but I think it should be according to criteria which are appropriate to it, rather than those which are not.

Now let us consider the issue of checking. While I don't think IPA should be judged by its ability to produce replication, I agree that a reader should be able to check the results of a study and check how they came about. That is a different matter, though Giorgi presents them as though they are the same thing. Again I have written in detail about how to facilitate checking (e.g. Smith, 2003; Smith et al., 2009).

This checking can be done at a number of levels. For example, a supervisor can check that a postgraduate student's analytic process has been systematically carried out and recorded. It is also possible for the researcher to arrange an independent audit. Here the complete body of material is made available to an 'auditor' whose task is to follow the analytic path made by the researcher from transcripts, codes and themes to final write up and to decide whether the analysis presented is warranted from the analytic trail provided.

As a final example of a type of checking, the reader of a paper can check that the methodological steps have been described and that the analytic report is sustained, coherent, evidenced. Thus when reading an IPA paper, for example, one can check that each theme presented has been supported with sufficient extracts from participants to illustrate both convergence and divergence in how the theme is manifest. Had Giorgi, as part of the preparation for his critique, chosen to look at an example of a completed IPA study in a journal article, he would have been in the position of his notional 'fellow phenomenological researcher'. He could then have checked and evaluated a completed piece of IPA work and he could have done so by the criteria for qualitative research appropriate to it.

Laxity

If one chooses to criticize a piece of work, I think it is incumbent on one to read the work carefully and represent it accurately. Unfortunately, Giorgi has been careless at various points in his article.

He comments on the example of analysis presented in Smith and Osborn (2008) and on p9 reproduces some of the material we present there. However Giorgi presents an incorrect heading for one of the columns which he labels 'Researcher Themes'. As we make clear in the text, the first phase of analysis which Giorgi is describing here involves annotating initial notes from the researcher and these are intended to be wide ranging. These then get transformed, through a subsequent phase of analytic work into 'emergent themes' at the second stage. Thus 'Researcher Themes' is not an accurate heading for the text describing the first phase of analysis and is misleading. As one moves through the stages of analysis, the results of the analysis become more psychological and more abstracted and Giorgi's account does not do justice to this process.

Giorgi assumes the participant we are analysing is a man:

And yet, as we clearly point out in the chapter just before we give the extract, this participant is a woman to whom we have given the psueudonym Helen. And then in the extended extract from our write up provided later in the chapter, we refer to her as Helen again and use the female pronoun to describe her numerous times.

Giorgi claims we do not pick up on the participant's statement 'I just get so hateful'. However, as is made clear in the chapter Giorgi is critiquing, doing IPA is a dynamic and iterative process where prior analytic comments are revisited in the light of the unfolding account. If he had looked carefully at the extract from our write-up on p77-78 of the chapter, Giorgi would have seen that we do indeed directly address, in considerable detail, the issue of the participant feeling hateful, in our consideration of her self-denigration. And this self denigration can, by examining the audit trail, be traced back to our initial comments made on the transcript.

References

Eatough, V. and Smith, J.A. (2008). Interpretative phenomenological analysis. In Willig, C. and Stainton Rogers, W. (eds) Handbook of Qualitatitive Psychology. London: Sage.

Giorgi, A. (2010). Phenomenology and the practice of science. Existential Analysis, 21, 3-22.

Larkin, M., Watts, S. and Clifton, E. (2006). Giving voice and making sense in interpretative phenomenological analysis. Qualitative Research in Psychology, 3, 102-120.

Smith, J.A. (1996). Beyond the divide between cognition and discourse: using interpretative phenomenological analysis in health psychology. Psychology & Health, 11, 261-271.

Smith, J.A. (2003). Validity. In Smith, J.A. (ed) Qualitative Psychology: A Practical Guide to Research Methods. London: Sage.

Smith, J.A. (2004). Reflecting on the development of interpretative phenomenological analysis and its contribution to qualitative research in psychology. Qualitative Research in Psychology, 1, 39-54.

Smith, J.A. (2007). Hermeneutics, human sciences and health: linking theory and practice. International Journal Of Qualitative Studies On Health And Well-Being, 2, 3-11.

Smith, J.A., Flowers, P. and Larkin, M. (2009) Interpretative Phenomenological Analysis: Theory, Method, Research. London: Sage.

Smith, J.A. and Osborn, M. (2008). Interpretative phenomenological analysis. In Smith, J.A. (ed) Qualitative Psychology: A Practical Guide to Research Methods. London: Sage. (2nd Ed.).

Yardley, L. (2000). Dilemmas in qualitative health research. Psychology and Health, 15, 215-228.

Jonathan A Smith is Professor of Psychology at Birkbeck University of London. He has published many journal papers, primarily applying interpretative phenomenological analysis (IPA) in health and social psychology. He is lead author of Interpretative Phenomenological Analysis: Theory, Method and Research. (2009, with Paul Flowers and Michael Larkin, London: Sage). He has also edited a number of books on qualitative psychology and is a former editor of the journal Psychology and Health.
It seems to me that many of the practices being advocated by those
   recommending IPA are not scientifically sound (p6)


There is also the question of whether the promoters of the method
   actually do not have prescriptive intentions as they claim.
   (p7)


To offer "suggestions" ... it seems to me, are prescriptions. (p7)


They simultaneusly want to give any potential follower of their
   strategy total freedom to deviate from what they are doing. (p7)


The ability to check the results of a study or to replicate it is a
   scientific criterion ... Why shouldn't a fellow phenomenological
   researcher want to check the results of a scientific
   phenomenological study whose results she finds intriguing? (p7)


The participant says that he "just gets so hateful" (p9).
Gale Copyright: Copyright 2010 Gale, Cengage Learning. All rights reserved.