Multiperspectival designs and processes in interpretative phenomenological analysis research

ABSTRACT

Researchers using interpretative phenomenological analysis (IPA) within applied research typically use homogenous samples exploring shared perspectives on a single phenomenon of interest. This article explores the challenges and opportunities involved with developing rigorous and epistemologically coherent research designs for capturing more complex and systemic experiential phenomena, through the use of multiple perspectives to explore the same phenomenon. We outline a series of multiple perspective designs and analytic procedures that can be adapted and used across many diverse settings and populations. Whilst building upon existing approaches within qualitative methods and IPA, these designs and procedures are intended to scaffold clear routes to practical application, psychological intervention, the design of behaviour change interventions, and other recommendations for policy and practice. We discuss a variety of conceptual antecedents which situate these designs within phenomenology, pluralistic idiography, qualitative psychology, and wider debates within psychology and other social and behavioural sciences.

KEYWORDS:

• Impact
• IPA
• methodology
• multiple-perspectives
• systemic

Researchers using interpretative phenomenological analysis (IPA) within applied research typically seek out a single and reasonably homogenous sample of participants. Depending upon the study and its context, this may or may not be captured via a sampling strategy, which begins by defining a uniform set of demographic characteristics. In IPA, “homogeneity” refers to a probable shared perspective upon the phenomenon of interest. For example, we might explore a research question such as, “how do people make sense of the experience of living with young onset dementia (YOD)?” by conducting research interviews with people who have a diagnosis of YOD. However, many IPA researchers are working within conceptual frameworks which recognise that an experience or process such as “living with dementia” is not solely located within the accounts of those with the diagnosis (e.g., see La Fontaine et al. 2015). The phenomenon is also located within the accounts of other people who belong to the “lived world” of the person with the diagnosis, such as his or her partner, children, friends, and colleagues. A number of studies (on topics as diverse as experiencing psychiatric hospitalisation, interpreting genetic tests, understanding foster placement breakdown, coping with Parkinson’s disease, and participating in nonmonogamous relationships) have used IPA to explore complex experiences from more than one perspective (e.g., see Dancyger et al. 2010; Rostill-Brookes et al. 2011; Smith & Shaw 2016; de Visser & McDonald 2007). Our aim here is to show how an overarching view of these innovative studies can open up new ways of thinking about the potential of IPA research. Thus, we explore the challenges and opportunities involved with developing rigorous and epistemologically coherent research designs which seek to capture more complex and systemic experiential phenomena through the use of multiple perspectives.

For experiential qualitative researchers, it is becoming increasingly important to understand the impact of our work. Qualitative work can have an effect upon the world at many levels. For example, it can adopt an advocacy role, where the voices of participants raise our awareness of an experience. It can highlight processes of marginalisation or identify contexts in which people are misunderstood. Some participatory projects may generate assets and capacity within local communities. Other projects may have an impact through their effects on theory or the ways in which policy is discussed. In many respects, however, the “inferential range” of experiential studies is often limited by their very commitment to depth of focus. The aim of illuminating particular experiential perspectives through idiographic data collection and analysis (e.g., the patient’s perspective) can sometimes appear to present us with a compromise, that is, with powerful insights but limited reach.

In this article, we outline some strategies for designing larger, programmatic studies using experiential qualitative methods which maintain a commitment to depth but also augment it with a systemic and multiperspectival dimension. Such an approach extends the potential reach and impact of experiential research in the “real world.” In our experience, it does this because audiences often respond positively when they recognise an experiential insight, but they may then express doubts regarding its representativeness. When such insights can be evidenced from more than one point of view, such doubts are often assuaged. Multiperspectival IPA retains a commitment to idiography in data collection and analysis but extends this by combining two or more focal perspectives, permitting us to consider the relational, intersubjective, and microsocial dimensions of a given phenomenon. These analytic designs are more complex. They begin with a traditional idiographic approach (i.e., the unit of analysis is still how we understand our participant understanding their experiences within their lifeworld), but they synthesise these analyses not only within a sample but also between samples. They may require an additional analytic focus upon how participants’ accounts are grouped according to certain criteria (e.g., patient and health care provider, as in Borg Xuereb, Shaw & Lane 2015), or according to their roles within clearly relational phenomena (e.g., spouses within a dyad, as in Loaring et al. 2015; or members of a family group, as in Burton, Shaw & Gibson 2015).

Increasing the inferential leverage of idiographic and phenomenological inquiry

Typically, phenomenological work in psychology focuses on personal meaning, and so the relationship between person-and-world is operationalised at the individual level. Thus in IPA projects, the most common research designs involve collecting qualitative data from a reasonably homogenous group who share a certain contextual perspective on a given experience (e.g., people living with chronic fatigue syndrome, as in Dickson, Knussen & Flowers 2007; or new mothers, as in Smith 1999). Thus, “we ask questions about people’s understandings, experiences and sense-making activities, and we situate these questions within specific contexts, rather than between them” (Smith, Flowers & Larkin 2009, p. 47). These designs give us an in-depth view of a specific experience and do so with a recognisably personal scale to that view. They highlight the role of the case study and the value of the idiographic perspective in illuminating people’s relationship to the lifeworld. However, they also give us a one-dimensional perspective on the meaning of events and processes. Whilst this in itself can be of tremendous value, it can also have limitations. The traditional approach can become self-limiting when the broad object of inquiry, or overall research question, also has a strong relational or systemic dimension (such as understanding a caring relationship, making sense of both parties’ experiences of a dyadic therapeutic intervention, or understanding problems in the implementation and translation of effective interventions in specific social or cultural contexts).

For traditional kinds of experiential research, caution is required when it comes to the generation of explanatory or process accounts:

The logic behind sample-specificity is related to the inductive logic of IPA and has consequences for the applicability of findings. Cases and accounts are held to be local, and so analyses are cautious and are built cumulatively. They must therefore be dealt with in detail, and in context. The logic is similar to that employed by anthropologists conducting ethnographic research in one particular community. The anthropologist reports in detail about the view from within a particular cultural frame, but does not claim to be able to say something about all cultures. Subsequent studies may add to this, so that very gradually more general claims can be made, with each founded on the detailed examination of a set of case studies. (Smith, Flowers & Larkin 2009, p. 51)

Often a given group’s perspective is missing from the literature, or else it is present but misrepresented. For example, there is an extensive qualitative literature on the experiences of carers and family members of people with learning disabilities. By comparison, the literature exploring the experiences of people with learning disabilities themselves is sparse. In such situations, the value of standard designs in “giving voice” to a particular perspective far outstrips the caution required by the requisite sampling strategy. At other times, however, we may wish to treat people’s experience as a lens for illuminating the broader meaning or consequences of an event or process to understand its wider constitution, dynamics, or mechanisms. This may occur in applied settings when we are working within an action research paradigm; when we have aspirations to provide evaluation, theory-development, or social critique; or when we simply aim to produce the kind of psychology espoused by the phenomenologist Merleau Ponty, who argued that instead of simple cause and effect models, psychologists need to understand the reasons and construals of people to understand their actions (Moran 2000, p. 420).

In these situations, it can sometimes be helpful to adopt more complex designs. For example, in some cases the phenomenon may be especially relational or social (as in the example of nonmonogamous sexuality). In other cases, we might anticipate some conflict between the different perspectives (as in the example of psychiatric hospitalisation), which needs to be understood if the research is intended to inform and change practice. In further cases, there may be some constraints upon the capacity of the participants to verbalise their experience, and additional perspectives may supplement this.

Multiperspectival designs retain IPA’s strong connection to phenomenological and hermeneutical concepts, but they also build upon links to concepts from systemic psychology, which have been noted elsewhere (Rostill et al. 2011; Glasscoe & Smith 2011). Through the analysis and synthesis of multiple perspectives, it is possible strong and persuasive analytic accounts can be developed. In the remaining sections, we describe and illustrate some of the key features of these designs, outline several key ideas which can underpin them, and discuss some of their advantages and challenges.

Features and types of multiple perspective design

There are many ways in which multiperspectival designs can address relational, systemic, or other socially nuanced research questions, and these can be further extended by the addition of cross-sectional or longitudinal components. It is beyond the scope of this article to detail all potential designs, but here we briefly focus upon the shared elements of multiple perspectival designs. These use the building blocks of well-delivered traditional IPA designs. So depending on the key components within a system or the number of actors within the relationship of interest, multiple traditional IPA studies are conducted and combined. Multiperspectival designs then focus upon identifying the synthesis, integration, or resonance between the findings of the contributing IPA studies.

The taxonomy in Table 1 is not intended to prescriptive, and the distinctions among “types” of designs are intentionally fuzzy. The taxonomy is intended as a tool for thinking about different ways in which samples and cases may be constructed and for exploring the logic underpinning them. Indeed, “caseness” itself is a multilayered concept here. Any given participant within a multiple perspective study constitutes a case at the personal level. Layered above that are other more complex forms of case (e.g., dyads, families, systems), each of which may be conceptualised differently, according to the researcher’s curiosity and interest and on the basis of what arises from the data. We can anticipate some, but not all, of the parameters when we articulate our research question, and we plan our research design. Underpinning these commitments will be a conceptualisation of “what lies between” our participants and our interest in them. For example, in a study of counsellors and their clients, we might be interested in understanding the therapeutic relationship; in a community study of nurse specialists, patients, and general practitioners, we might be interested in understanding how signs of remission and relapse are interpreted within the triad; and so on. There will, however, be times when the developing research offers us other, less anticipated forms of caseness, which can invite us to add a further layer to our analysis.

Table 1. Emerging Taxonomy of Multiple Perspectives of IPA.

Design type	Distinctive feature	Example	Metaphor
Directly related groups	When subsamples are immersed in the same experience, but are likely to have different views of it.	Rostill-Brookes et al. (2011) — process of foster placement breakdown is the topic, and interviews are conducted with young people in foster care, with social workers, and with foster carers. What they have in common is a shared experience of “fragmentation” — of things falling apart without anyone being entirely sure who is responsible for holding them together.	“All surfing the same wave”
Indirectly related groups	When people are linked by some underlying quality of two which bridges between otherwise distinctive phenomena.	In Larkin and Griffiths (2004), participants are interviewed who are involved in recreational drug use or in dangerous sports. Both activities share some features, particularly the juxtaposition of short-term rewards with uncertain levels of risk.	“Sharing the breeze”
Families, teams and other cohorts	When people are involved in a system or group, and have shared an experience.	In Dancyger et al. (2010), the perspectives of different family members are brought together to illustrate the complexities of decision-making about genetic testing for hereditary cancers. In this study, the effect of the multiperspectival design is cross-cutting — it shows how the family itself can constitute a coherent unit of study. Thus, some families adopt a shared narrative which supports testing, and others do the reverse.	“Tangled in the same web”
Dyads	When the research is concerned with shared and distinctive features of an experience which is important to two people (e.g., doctor and patient; service-user and carer; siblings; couples)	Dyadic work can be conducted and written up with the focus on the pair as a pair (see, e.g., de Visser & McDonald’s 2007 study of jealousy in heterosexual swinging couples). It can also be written up at group level, if there are ethical issues arising from the presentation of the data at dyadic level (e.g., see Larkin, Clifton & De Visser’s 2009 study of the meaning of consent for patients and psychiatrists).	“Two sides to every coin”

We might, for example, plan a study of multidisciplinary teams’ experiences of dealing with the suicide of someone under their care. We might anticipate that there will be caseness at the level of each team, and beneath that at the level of the individual team members. When we begin to analyse our data, we might discover other important forms of caseness that invite our attention and merit some structure. There may, for example, be units of analysis within the team (professional or personal alliances, each offering distinctive and competing accounts of how the team copes). There may be caseness too, according to time.¹ Organisational cultures, resources and practices change: how things were, and how things are, might turn out to be important ways of making sense of our participants’ perspectives on their experiences. Teams who have coped with more than one suicide in recent times will also have more than one case to discuss, and the circumstances and aftermath of each one are likely to be distinct. As researchers, we would need to decide how to work with these additional opportunities. In such situations, it is helpful to return to the research question and to reflect upon the planned focus of the study. A clear research question puts the “aboutness” of our interest in plain sight and allows us to make informed decisions about which kinds of case are relevant to our interest and commitments.

Studies pertaining to directly related groups involve samples that are immersed in the same environment or involved with the same phenomenon, but that are likely to have distinct perspectives on it. If a research question can be addressed most effectively with a multiperspectival design, then conceptualising the multiple perspectives represents an important first step in fleshing out the overall design. This means identifying the key groups involved in a given event or process and collecting data from each group. Although other qualitative methods (grounded theory, in particular) examine social processes and relationships through people’s experiences, the focus here is distinct. In multiperspectival IPA designs, the unit of study is the case (e.g., the person, dyad or system), so the decisions made in relation to different dimensions of caseness (as above) are important. IPA is focused on idiographic analyses of patterns in people’s meaning-making rather than on producing a model or theory of an underlying process. Rostill et al. (2011), for example, examined the experience of foster placement breakdown. Interviews were conducted with young people in foster care, with social workers, and with foster carers. The three groups all had direct experience of placement breakdown, and the authors present themes from each set of interviews, organised around a central systemic narrative describing each group’s failure to understand the position of the other two. What they have in common, as the title of the article suggests, is a shared experience of “fragmentation,” of things falling apart without anyone being entirely sure who is responsible for holding them together.

A variant on this design involves indirectly related groups. This is where two or more samples are analysed together because a theoretical or conceptual argument can be made that while they may not share a direct connection or context, they may still share a perspective on an implicit or underlying feature of the phenomenon. For example, one could examine the experience of solitude and conduct one set of research interviews with nuns and another with women prisoners. There are important differences between them (e.g., coercion, spirituality), but there are also interesting commonalities (e.g., gender, institutional life, separation from the “wider world”).

Families and other naturally occurring groups (e.g., teams) can also provide a logical and insightful perspectives on shared psychosocial phenomena. In these studies, individual analyses are often compared and synthesised at the within-group level and then at the between-group level. In Dancyger et al. (2010), the perspectives of different family members are brought together to illustrate the complexities of decision making about genetic testing for hereditary cancers. In this study, the effect of the multiperspectival design is cross-cutting, showing how the family itself can constitute a coherent unit of study. Thus, some families adopt a shared narrative which supports testing, and others do the reverse. In contrast, Penny, Newton and Larkin’s (2009) study of British Pakistani families caring for a young person with psychosis demonstrates how experiences and understandings can shift within the family as well as across families. In Penny et al.’s study, some generational aspects of the meaning of psychosis are illuminated, with siblings describing some concerns which are at odds with those of parents and grandparents.

The dyad also offers a commonly encountered social context where research questions and research designs can require multiple perspectives. There are many psychological phenomena which can be understood more fully if considered from the point of view of a dyad. Dyadic designs can maintain a particularly strong idiographic focus alongside their relational analyses, because couples and partnerships of one kind or another present us with a very coherent and recognisable unit of analysis. The shared experiences of live liver transplantation, for example, are captured in McGregor et al.’s (2014) study that explores both donor and recipient perspectives on live liver transplants. Another example is Loaring et al. (2015). In this study, partners within eight heterosexual couples were interviewed. The final analysis focussed primarily on the shared dyadic perspectives of couples concerning the same relational phenomenon (sexual intimacy following breast surgery) by interviewing both members of each couple.

The core idea in each of these variants is that the researcher aims to take up a series of perspectives around a given phenomenon. The challenge for the analyst is to retain IPA’s commitment to understanding participants’ claims and concerns (when, across the sample as whole, there may be more variation than in a traditional samples) whilst also illuminating those insights gained through inclusion of additional perspectives. These additional insights can arise from matters of congruence, contrast, or both.

There are practical challenges to consider. Such designs give rise to additional ethical issues, especially around threats to internal anonymity. Audiences from outside of the study are generally no more likely to recognise participants in a multiple perspective study than they are in a traditional design. But audiences from inside the study (i.e., other participants) are likely to recognise their own quotes, and from the link between their pseudonym and the pseudonyms of other participants in their system they may be able to identify the source of some quoted data. This issue is discussed in some length in Ummel and Achille (2106). To resolve it, the researcher must respond on several levels. First, multiperspective researchers must be especially careful to offer clarity and negotiation during the recruitment and consent-taking phases of the study, so that participants understand this risk. Second, the risk can also be mitigated by careful decision making about how the data are presented. For example, in Loaring et al. (2016), some sensitive extracts are presented without attribution (rather than with pseudonyms) to prevent partners from identifying each other. Similarly, in Haskayne et al., dyadic data are presented at the general, group level (i.e., without identifying which therapist was linked to which client) for similar reasons. Multiple perspective designs tend to highlight the dynamic nature of good ethical practice and to require the researcher to take responsibility for responding in an ethically appropriate manner to issues that arise not only during the planning phase but also during data collection, analysis, and dissemination. In many of these designs, power will have to be considered while the research is designed and conducted: it is likely that one sample will have more recourse to socio-cultural capital than the other(s), and this may need to be discussed and monitored during supervision. Ummel and Achille explore many of these issues in their article on dyadic designs, and their work provides excellent prompts which researchers can use to reflect upon, and evaluate, their plans and practice.

A further practical challenge is presented by the matter of dataset size: these designs tend to involve more data points and thus require more time and resources. The final stage of analysis is particularly challenging because it is effectively a mini-meta-synthesis. There are difficulties in terms of what to foreground, what to omit, and how to deal with areas of concern that do not overlap, as we discuss below.

Analysis issues

Analysis of these designs can be complex. It makes sense, as with other IPA designs, to begin with the each personal case and then from the personal level to move “outwards.” The direction of one’s movement outwards might vary depending on the design and the data. To begin, many of the key elements of traditional IPA are employed in that for person A in sample 1, participant-led experiential data are collected and analysed idiographically. This is repeated for each person and each sample group. At this point it is possible to present a case study for each participant. Further analysis then focuses on thematic development within the next “unit of analysis” (e.g., within the dyad, family, or subsample) and then, finally, between and across those larger units.

For example, in a dyadic design, the next move after individual analysis might be to look at the other half of the dyad. In a grouped design, it might make sense to stay with people who share the first participant’s perspective and to work with one group at a time. In a more dynamic, systemic design, it could be helpful to work within one “micro-system” at a time before moving on to look at the next.

As an example, the way this worked in Borg Xuereb et al. (2015) was as follows. The study investigated consultations for the diagnosis of atrial fibrillation and decision making about oral anticoagulation treatment, focusing particularly on decisions about whether to take warfarin. In this study, there were three groups of patients: those who accepted warfarin, those who discontinued warfarin, and those who refused warfarin. There were also groups of physicians with different roles: general practitioners in primary care, general hospital physicians, cardiologists, and registrars. Each individual was treated as an individual case and analysed at an idiographic level. Analysis then proceeded with the traditional process for cross-case analysis within the groups. Resonance and disparity among patients who accepted warfarin were noted in the emerging development of themes. The other patient groups were then analysed in the same way, all the time adding to instances of resonance and disparity. A set of themes representing the multiple perspectives of these patients was developed. The physicians were analysed in the same way.

To make sense of our analyses at the multiperspectival level and to present them in a coherent way to readers, we need to develop a narrative about how the experiences relate to one another. Identifying patterns and connections, or indeed conflicts and differences, between and across groups, dyads, or systems, can be less straightforward than drawing out the superordinate themes identified in a standard IPA design. In some studies, there may be a shared experience which enables us to understand an event or process (e.g., the sense of fragmentation in Rostill et al. 2011). In others, there may be characterisable stances or positionalities which allow us to describe how the different perspectives relate to one another (there are some examples of this in Wane et al. 2007 and in de Visser & McDonald 2009). Conversely, conflicts and differences can sometimes illuminate an underlying structure of experience (or understanding of it), as in Larkin, Clifton and de Visser (2009). In Table 2, we list some strategies that may be helpful in thinking about thematic development when working with more complex data.

Table 2. Analytic strategies for thematic development from complex data.

Identify consensus overlap or conceptual overlap	In some cases, participants with different perspectives will still explicitly express the same concerns, and make similar experiential claims. This sort of explicit consensus is the easiest kind of convergence to identify, but it is worth considering it carefully are there conditions or contexts in which participants take up other stances? Sometimes the overlap between the meanings made by participants speaking from different perspectives will be implicit, and the analyst may be able to suggest a concept which underpins the participants’ concerns, such as “fragmentation” in Rostill et al.
Identify conflict of perspectives	Sometimes there will be a clear disagreement between the different perspectival samples. In these cases, theme titles can capture this conflict, and the analysis can explore the variations in experience, and the contexts from which they emerge.
Identify reciprocity of concepts	In other cases, participants may express concerns which complement each other very well. For example, the nurses in Thompson et al. (in press) express their frustration at not having the time to deliver more one-to-one work with their patients; the patients in the other part of this study (Fenton et al., 2014) talk about how important their fleeting experiences of one-to-one contact are for their recovery.
Identify paths of meaning	Sometimes we notice shared experiences with divergent meanings, or shared meanings attributed to different experiences, as in the different interpretations of the consultation offered by patients and physicians in Borg Xuereb et al. (2015). This can be a useful and important way of exploring some of the analytic work developed under the above points.
Identify “lines of argument” (after Noblit & Hare 1988).	This involves “storying” the most important dimensions drawn from the preceding strategies to provide an analytic narrative to the audience which captures the structural and/or procedural aspects of the system which has been studied.

The overarching aim here is to produce an account that capitalises on multiplicity and offers a plausible interpretative perspective on how the participants’ lifeworlds interact and overlap. This need not involve the loss of idiography, the artificial construction of consensus, or the denial of difference, but the analyst should be sensitive to these risks. Practical lessons can be learned here from techniques involved in framework analysis and meta-synthesis (e.g., Shaw et al. 2016; Bennion et al. 2012). A matrix can be developed with cases on one axis and themes on the other. This extends the “lines of argument” process advocated by Noblit and Hare (1988). A matrix provides us with a visual representation of themes gathered from the multiple voices within the sample, and this helps us identify similarities and differences between and within them. The important thing to remember is that the objective of this exercise is not to simply pool accounts and generate a consensus but to demonstrate the ways in which accounts from multiple perspectives relate to one another and to reflect upon how those differences can co-exist.

Conceptual and theoretical underpinnings of a multiple perspective phenomenology

A number of approaches to research incorporate some degree of multiplicity in their conceptualisation, design, and sampling (e.g., see Ribbens McCarthy et al. 2003 or Rose, Thornicroft & Slade 2006). Grounded theory (Charmaz 2006), for example, does this via theoretical sampling, and its underpinning logic for multiplicity is driven by the need for theoretical completion (via sample saturation). The emphasis here, then, is on theoretical power. In facet methodology (Mason 2011), on the other hand, multiplicity is informed by a creative approach to knowledge construction, where multiple aspects of a phenomenon are explored via a range of imaginative means and then integrated (via analysis) in a form of bricolage. Facet methodology’s emphasis is on different ways of seeing. In IPA, the emphasis underpinning multiplicity overlaps to some extent with these qualitative cousins, but the key component is the sense that important meanings are often located “in between” persons.

In this section, we explore the conceptual material from phenomenological philosophy and related fields which can underpin and justify the use of multiplicity in IPA research.

Profiles

The perspectival nature of our relationship to reality is foreshadowed in Husserl’s work (e.g., see synopses by Giorgi 1995 & Moran 2000). Husserl describes how the “outer world” of things and events is perceived via a series of adumbrations, profiles, or aspects. Thus, perception has a partial and perspectival quality, and this is a function of our spatial or relational place in the world. So our sense of “what happens” in the world is a consequence of our position and perspective in relationship to the object we observe. This is in contrast to the inner world, which Husserl ultimately concludes to be encountered through a series of temporal profiles. This sense of the person as a being always “in-relation-to the world” was somewhat marginal to Husserl’s work because he was more concerned with the “inner world.” The concept proved prescient, however, and it was ultimately given a much stronger emphasis — and a rather different language — in the later work of Heidegger (e.g., see Larkin, Clifton & Watts 2006 for discussion about IPA) and Merleau Ponty.² These two successors to Husserl were keen to emphasise the mutuality of person and world (dissolving the distinction between “inner” and “outer” perceiving). Both of them emphasised the inevitably situated nature of human experience, but each with a distinctive focus on either our worldliness (Heidegger) or our embodiment (Merleau Ponty).

Intersubjectivity

Intersubjectivity is, in one sense, a concern with the location of experience. Conventionally, in Cartesian and individualistic models of personhood, experience (and with it, emotion and cognition) is contained “in the head.” Bradley (2005) views mainstream psychology’s commitment to individualism as a deliberate act of “estrangement,” however. When “everything one is and feels and thinks is steeped in sociality” (p. 82), it may seem wilfully obtuse to insist on the individual mind-brain as the basic unit of psychology. Hermeneutic phenomenology is one approach that contests this view. It is concerned with the directedness and relatedness of Being — its fundamentally intersubjective qualities — and, as such, much of its analytic attention is concerned with what happens “in between” (in between persons and other persons, in between persons and objects, and in between persons and cultures).

Thus, when Bradley draws upon Habermas to describe how “we live in a complex multi-dimensional space of concurrent inter-relations” (p. 88), he describes a view of reality and of our relationship to it which lends itself to a multiperspectival view. Events and processes are not simple and discrete. They are complex and dynamic and are subject to processes of mutual meaning-making.

Systems

One key idea underpinning many approaches to family therapy and human systems theories is that events and processes are best understood by exploring what happens in between the individuals involved, whether in terms of their interactions with one another or the stories and language which give meaning to their realities:

It is the observer who generates the distinctions we call “reality.” One’s picture of or knowledge about the world will be the basis for one’s attitude to it. Because persons experiencing the same world “out there” make different pictures of it, problems will arise when they debate which picture is right: either mine or yours. […] One should think of the picture and its explanation more in terms of both-and or neither-nor, and leave out the either/or. (Andersen 1987, p. 415)

Systemic therapy’s engagement with the social constructionist ideas which drive these concerns has often been positive but cautious (e.g., see Boston 2000). While the importance of patterns of relating and construing are central to most forms of systemic practice, so too is the idea of acknowledging the lived realities behind the family’s “multiverse.” Ultimately the therapist is present because of some commitment to “making a difference” in situations when the people in the system are in difficulty. These latter positions violate the relativist stance of full-blown constructionism. Thus, in many ways systemic theory’s relationship to social constructionism is similar to that of IPA and of other “postconstructionist” approaches to qualitative research (Larkin, Eatough & Osborn 2011). That is, both systemic theory and IPA draw upon interest in the functions of language, but neither is primarily defined or constrained by that interest. Both may also be implemented with the view that differing personal perspectives on the world can be understood by a third party focusing on patterns of meaning-making, provided that one begins from the position that each perspective illuminates an important aspect of a shared experience.

Summary

We have argued that, phenomenologically speaking, events and processes in the world are perceived from somewhere and thus are encountered in “profile.” This means our experiences of events and processes are intersubjective and relational. Meaning is “in between” us, but is rarely studied that way in phenomenological inquiry. The meanings of events and processes are often contested and can sometimes be understood in a more complex manner when viewed from the multiple perspectives involved in the system which constitutes them. Multiple perspective designs can be a useful way for IPA researchers to address research questions which engage with these phenomena. We have described some strategies for dealing with the data that can arise from this kind of enquiry. These strategies seek to draw out patterns of meaning (themes) which can be about convergence or divergence. One potential advantage to these sorts of designs is their capacity for greater impact. The convergence and triangulation of viewpoints can be more persuasive than an analysis drawn from a single sample.

Reflections

We are not proposing that this greater persuasiveness sits upon an implicit model of causality. Greater inferential range does not mean we shift out of phenomenological analysis and into a more empiricist frame of reference. Analysis of data from these designs is focused upon meaning, not causality. Instead, persuasiveness is enhanced via triangulated consensus (when consensus is present) and via transparency (especially when perspectives are in conflict), and both of these are consequences of involving more than one stakeholder group. The effect is to increase the potential contextual range of the analysis; if it is not just “these people in this context” who have these concerns but also “those people, in this context” (or “those people in this other context”), then both author and audience begin to feel more confident the analysis is telling them something substantive about how the world can look when a particular phenomenon (e.g., illness) is foregrounded.

Interestingly, in the transition from Husserl to Heidegger, the idea of phenomenological knowledge which transcends context is largely discarded, primarily because the phenomenological enquirer is assumed to occupy a single position (i.e., that of the phenomenologist). Heidegger argues persuasively that the phenomenologist cannot step out of this position. But what if, as in IPA, each participant is acting as one among many phenomenologists? Collectively, they offer us a range of positions and perspectives. The trap of the situated observer is a little less biting. Thus, while the social science researcher drawing upon phenomenology is, in some respects, compromised by the “third person” nature of his or her inquiry, in other respects (i.e., recourse to multiplicity), the researcher has an advantage over the purely philosophical investigator.

There are still limitations. The researcher is a sort of “meta-phenomenologist,” re-interpreting all of the participants’ individual interpretations and still caught in the trap. But the researcher can make use of the multiplicity of evidence available to her and of the requirement for rigour and transparency in ways which are more difficult for us if we only examine our own experience. As a consequence, these designs may have the potential to bring a moderated version of “generalisability” or “abstraction” back online for phenomenology, albeit a complex, cautious, and contextually sensitive one.

Additional information

Funding

ML was supported by a European Research Council Consolidator Grant (‘Pragmatic and Epistemic Role of Factually Erroneous Cognitions and Thoughts’; Grant Agreement 616358) for Project PERFECT (PI: Professor Lisa Bortolotti).