Philosophical Assumptions
Philosophical Assumptions
Qualitative interviewing has become a prominent research method in the social sciences. Face-to-face conversation is an everyday occurrence, and this has probably resulted in an assumption that interviewing is a preferred option because the researcher feels, most at ease with this technique. In well-executed research preferences are not the issue, rather the focus is on justification: to what extent can the methodology and methods adopted be justified in relation to the purpose of or rationale for the research? This question brings to the fore a host of issues that need to be carefully worked through, examining our philosophical assumptions about what we can (and cannot) know and associated theoretical perspective(s). With this in mind, this chapter will explore the epistemological and ontological thinking behind qualitative research and qualitative interviewing. We will concentrate on the following philosophical issues as they have direct relevance for qualitative interviewing:
• methodology and methods
°    different approaches to research
• theoretical groundwork and making connections .
• developing a rationale
• epistemology, ontology and qualitative interviewing
©
Methodology and methods
rvised numerous undergraduate and postgraduate student research ing sup ^ ^       c0nfidence that the distinction between methodology Pr°'eCtS'jloCj5 is almost always something that causes confusion. The two are same Methods are easily explained; they are the techniques or proce-S use to collect and analyse data. In qualitative research interviewing <Jures *e^e most frequently used methods when generating data. Other meth-a one     delude observation, diaries, the generation of visual images or other C°of text. In this book we cover the use of qualitative interviewing as a » 'Thod' - a means of collecting and analysing data. Having said this, we will .
m ro make evident how methods are informed by methodology. M thodology, as the word suggests, relates to a process where the design of the research and choice of particular methods (and the justification of these in relation to the research project) are made evident. As such methodology requires more from the researcher than just preference or intuitive appeal to 'usrify the choice of particular techniques of data collection and analysis. It becomes necessary to outline the philosophical and theoretical positions informing the research process. Thus there is a requirement to outline assumptions embedded in the methodology adopted. Often you will find methodology explained as an 'approach' or 'perspective5 that has within it implicit and explicit expectations about how research is undertaken.
Different approaches to research
It is common to refer to qualitative and quantitative 'paradigms' in research (e.g. Holliday, 2002), suggesting that they represent very different ways of thinking about the world. As you will see, this is at best an over-simplification with regard to qualitative research, as it encompasses many different ways of understanding the world and what we can know about it. Portraying qualitative and quantitative research as being in opposition to one another is also perhaps not very helpful or indeed accurate - as the rise in the use of mixed methods (including both qualitative and quantitative elements) clearly shows (Shaw and Frost, 2015). However, there are certainly some broad ways in which most qualitative research differs from quantitative work. In this section we will introduce some essential terms you will need to understand, at least at a basic level, to make and justify choices about how you engage with qualitative interview-based research. In a subsequent section, we will introduce four different approaches to qualitative research and their underlying philosophical assumptions. We recognise that thinking through and explaining your philosophical position can be a challenging aspect of the research process, but it is a crucial one. Remember, though, that you are not expected to become a professional philosopher of social science! You just have to develop a good enough understanding to make and communicate sound and defensible decisions about how you design and execute your research.
©
Interviews in Qualitative Research
Epistemology
A concise definition of epistemology is the philosophical theory of knowledge. Of major importance is the issue of what counts as knowledge, and social scientists are often preoccupied with attempting to.formulate sufficient criteria for evaluating knowledge statements - what it is we can claim to know. Unfortunately, differences between qualitative and quantitative research often become drawn as fervently oppositional rather than merely rooted in different understandings around what we can know, and what we might want to know, as researchers. If we take the idea of knowing as the basis for elaborating on both the differences between qualitative and quantitative research in general, and between the different types of qualitative research, we can hopefully make clear the fundamental methodological issues that underpin the justification for a specific approach. Our rationale for the choice of methods becomes less idiosyncratic (something we will return to later) and more complex when we ask knowledge-based questions about specific issues and phenomena. This then promotes consideration of what might be a reliable route to such knowledge. Therefore epistemology - how we know what we know, a means of establishing what counts as knowledge - is central in any methodological approach. Marshall and Rossman (2006) use the term 'epistemological integrity' when referring to the connections between the nature of the research, overall strategy, research questions, design and methods. Developing such integrity is not always easy and involves thinking through the values and ideals, principles and rules by which the phenomena under investigation can be known.
Ontology
Ontology is likely to be an unfamiliar term to those new to social research and is seldom unproblematic for others who might consider themselves seasoned researchers. It should be said that ontological and epistemological issues often arise together, resulting in a somewhat confusing representation. Blaikie (1993: 6) offers a 'root definition' of ontology: the 'science or study of being'. We are not sure that this takes us much further in our endeavour to unravel methodological approaches. However, he goes on to say that ontology means 'the claims or assumptions that a particular approach to social enquiry makes about the nature of social reality' (p. 6). There are those who would say that, stricdy speaking, we should stay with ontology as the study of 'being' (see, for example, Crotty, 1998), where the emphasis is on the theory of existence. Nevertheless, for our purposes Blaikie's pragmatic view provides us with a clear indication of why we need a philosophical perspective for our methodology. Without a perspective on the nature of social reality, and on how people might exist in the world, it would be impossible to consider what might count as relevant knowledge in the research process. For example, if we assume that people's behaviour, their way of being in the world, is brought about by their interactions in social situations, our view of the nature of social reality would
Philosophical Assumptions
©
ke very different from one based on a belief that genetic inheritance explains behaviour. These are two very different ontological approaches regarding the theory of existence. One approach relates to social practices and people as ocial actors; the other is biological, suggesting that what 'drives' our being in the worid is inherited and located within the individual.
Ontological positions are often described primarily as 'realist' or 'relativist'. Put somewhat simply a realist ontology subscribes to the view that the real world is out there and exists independently from us. For those taking this position, the world is made up of objects and structures that have identifiable cause-and-effect relationships. Indeed the natural sciences (e.g. chemistry, physics, biology) are all broadly founded upon a realist ontology. Quantitative, experimental methods in social research are also based upon the belief that 'real' elements of our existence can be uncovered by using appropriate methods of data collection and analysis. For example, social researchers have used twin studies to investigate criminal behaviour and the impact of genetic inheritance (Walters, 2006). Relativist ontology rejects such direct explanations, maintaining that the world is far more unstructured and diverse. According to this position, our understandings and experiences are relative to our specific cultural and social frames of reference, being open to a range of interpretations. Within relativism, society is not viewed as a pre-existent 'real' entity with objects and structures, but rather as the product of people engaging with one another. Therefore relativism is more consistent with the social practices and interactive explanations of how people exist and live in their world; Existence is therefore explained differendy within these two approaches, and as such the data that would need to be collected to investigate these different versions of reality are not the same.
There are, however, variations that blur the somewhat crude realist-relativist distinction. Offering a version of experience and existence that only takes account of people as social actors (relativism) with no recognition of the constraining impact of social structures has been questioned (see Willig, 1999a). Similarly, conceiving the person as a mere automaton subject to social or biological mechanisms that then deterrnine behaviour (naive realism) has also been treated with scepticism (Bhaskar, 1991). Critical realism is a perspective that retains a core element of ontological realism; behaviour and experience are seen to be 'generated by' underlying structures such as biological, economic or social structures. These structures or mechanisms do not directly determine people's actions; instead such structures have tendencies that may impact on our lives. For example, social structures can create inequalities that.have the potential to influence our existence. Bill Jordan (2004: 3), when exploring the transformation of collective life in modern society, gives a very candid account of how structural inequality with regard to gender and education in the 1950s impacted upon his family life:
My own marriage was to the daughter of close family friends, and the commitment to it made before I went to university. My wife subscribed to the new ethos of autonomy for women, but had none
©
Interviews in Qualitative Research
Philosophical Assumptions
of the advantages of education or opportunities of a profession. We had several children in quick succession, and she justifiably felt excluded from the exciting public life which I enjoyed. I became the inept partner to a resentful, highly competent person trapped in domesticity.
Social structures are seen to be located outside of the individual's control; the sexual division (social structure) of labour inherent at the time had consequences that cannot be disregarded when trying to understand the experience of both Bill and his wife. A critical realist ontology would take account of such structures when attempting to make sense of social reality. Yet, while recognising the 'real' potentiality of mechanisms and structures, critical realism does not propose 'hard' determinism. People can transform their lives having insight into their own contextually located existence.
Generally speaking, quantitative, research subscribes to a realist ontology, with qualitative research having its foundations in more critical realist and relativist approaches. Therefore epistemological questions around what represents knowledge within a particular ontological view expose the connectedness of research. By this we refer to how theory and philosophical understandings impact on what we believe can be known: these beliefs and understandings then influence how we gather and'make sense of information. For example, if we believe that genetic inheritance determines behaviour we would not use qualitative interviews to investigate this explanation. Conversations and words do not provide the kind of data that would be required to explore the genetic transmission of behaviour. However, if we subscribe to a social and interactive explanation for behaviour then speaking with people in order to explore their social experiences would be consistent with our ontological position. Thus research is connected - our ontological beliefs and understandings impact upon what counts as knowledge. Ontology, epistemology and methodology and methods are all connected and cannot be viewed in isolation.
Theoretical groundwork and making connections
Supporting what we have already said, Williams (1998) makes the point that when discussing various methodologies, the differences between quantitative and qualitative approaches are not wholly technical matters. Instead, he suggests that differences result from particular philosophical and theoretical traditions. Theory guides us in research; it can sometimes help to define the problem, offer insight and show us possible solutions. Just imagine as a student that you ask one of your lecturers how to write a good essay. The lecturer could begin by explaining in detail the overarching learning outcomes of the course and how these link to pedagogical issues in higher education. This insight might enable you to understand the broad aims for
course and how these are exemplified in different forms of assessment 1 t require certain strategies. You might argue that a simple set of pointers would have sufficed and taken much less time. Then again you might think that now you know how things work, you have an understanding of how things fit together and can put this into practice. In research this theoretical understanding of how things fit together is fundamental to the research process.
Interpretivism
Qualitative approaches are generally, but not always (see Holliday, 2002), founded upon theoretical perspectives rooted in interpretivism and are variously described as hermeneutics, phenomenology, ethnography, discursive, interactionist - to list only a few. Each of these approaches to qualitative research has distinct features, many of which we will expand upon in later chapters in relation to qualitative interviewing. The point being made here is that within the social sciences the term 'interpretive' is quite broad but can be encapsulated in concerns around how the social world is experienced and understood. Interpretive research is generally idiographic, which literally means describing aspects of the social world by offering a detailed account of specific social settings, processes or relationships. The focus for research might be to uncover how people feel about the world and make sense of their lives from their particular vantage points. Therefore, qualitative interviewing fits; actually conversing with people enables them to share their experiences and understandings. While this might appear a rather obvious comment to make, it is said with a degree of caution and is something to which we will continually return. In research it is all too easy to adopt such simplistic and seemingly rational viewpoints. As we shall see, interpretivism perceives experience and understanding as seldom straightforward; people participate in indeterminate lifeworlds, often attaching different interpretations and meanings to seemingly similar 'facts' and events. Alfred Schutz (1962: 5, cited in Flick, 1998: 31) explains that:
All facts are from the outset facts selected from a universal context by the activities of our mind. They are, therefore, always interpreted facts, either facts looked at as detached from their context by an artificial abstraction or facts considered in their particular setting. In either case, they carry their interpretational inner and outer horizons.
The stance Schutz takes is that what we might see as 'facts' become open to levels of interpretation. The notion of searching for one overarching truth about the reality of how we live our lives is seen to be misplaced; rather we have 'multiple realities' or different interpretations.
0
Interviews in Qualitative Research
Philosophical Assumptions
0
Locating scientific methodology
Positivism
Interpretivism is usually seen as counter to 'scientific' approaches which are more accurately referred to as 'naturalist' approaches to research. The naturalist approach assumes an ontological view that human beings are part of nature and can be studied in the same way as other objects in the physical world. This is a controversial and much contested viewpoint in the social sciences. The idea that human beings and human behaviour are reducible to variables that can be measured and subject to statistical analysis continues to be a major topic for debate. Even so, the traditional theoretical approach within the natural sciences, and dominant also in the social sciences, is positivism (often viewed as akin to naive realism). The positivist approach is nomothetic, which means developing general laws or principles to explain particular phenomena.
The positivist position is situated within the epistemological tradition of objectivism, where objects in the world have meaning that exists indepen-dendy of any subjective consciousness of them. Therefore the underlying aim for research within this tradition is to provide objective knowledge - knowledge that is value-neutral, unbiased by the researcher/research process. Of vital importance for such research is the belief in an objective reality that can be uncovered. This objective reality is more commonly referred to as 'truth' - a belief in the correspondence of knowledge with what can actually be proven to exist. Objectivism and the search for regularities, principles and laws underpin the quantification of scientific research. Aggregate data across large populations, statistical analysis, replication, generalisation and the reduction of intervening social variables are scientific strategies that claim to make known the 'real' aspects of existence.
The positivist ideal of objective knowledge existing independently is distinctly unlike the interpretivist view where meaning arises from our engagement in the world (especially the social world) we inhabit. Take, for example, the issue of climate change. Evidence exists that demonstrates changes in the Earth's atmospheric conditions. Changes in air temperature, solar variation and weather conditions can all be measured and exist independently of our subjective viewpoints. What do we actually know from this evidence? There is continual disagreement among experts about causality and projected consequences. While this may infer that" there is conflicting evidence, there are also the interpretations and investments of the researchers to consider. It may be possible to argue that subjective understandings are different from the scientifically established 'facts' inherent in objects, as with the climate change illustration. On the other hand, it remains hard to comprehend how such facts can effectively be distanced from any interpretation of them.
Banister et al. (1990) refer to a 'gap' between the object of study and the way it is represented. Interpretation is seen as the bridge between representations of particular phenomena - between what we claim is occurring, and the
1 world 'out there'. How we represent data can be. influenced by a host ''i?3 tors, including the context in which the data were generated, moral and
1 "cal concerns and researcher agendas. Pursuing objective knowledge, and ■rla truth, becomes somewhat elusive if it forever pivots on interpretations
d understandings which are open to the vagaries of human relations. It is
ť1- that even those who operate within the confines of scientific method
fair CO sa.y L1ACl * 1
ften acknowledge that preconceptions and suppositions can impact upon ° sanations. The difference is that while positivists would try to ntinirnise the CXP ct these have on the research process, interpretivists would see them as inescapable, and therefore part of what it is we research - especially through the application of reflexivity, which we cover in detail in Chapter 9.
Some research using qualitative methods largely accepts the assumptions underpinning the scientific positivistic model; as we have already noted, this is sometimes referred to rather dismissively as 'naive realism'. We prefer the term •qualitative neo-positivisť (Duberley et al, 2012; King and Brooks, 2017b) as this acknowledges that researchers may have a coherent rationale for taking such a position - for example, they may wish to test the applicability of a well-established theoretical model to a particular case (e.g. McCluskey et al., 2011). Qualitative neo-positivism may also be used in some mixed methods research, especially where the conceptual ideas driving it are rooted in the quantitative tradition (e.g. Maznevski and Chudoba, 2000).
Empiricism
Central to the scientific approach is empiricism, the view that our knowledge of the world is gained from experience. No problem here, you might say - but what this means is that only the systematic collection of sense data via observation gives rise to the development of knowledge. Willig (2001) describes how nowadays few social scientists would subscribe to a pure form of empiricism or indeed positivism. Such forms seem naive when much of what passes as scientific research is founded upon pre-existing theory. What empiricists would claim is that the acquisition of knowledge depends upon the collection of observation data. Therefore on the basis of multiple observations researchers are able to develop general laws based upon the 'fact' that under particular conditions certain effects will always occur. This process and form of reasoning is called induction. David Hume (1711-1776), an eighteenth-century philosopher, did not share such confidence, arguing that there could be no logical justification for such claims. Hume's position is that there can be no certainty, as the seemingly logical connection between cause and effect is based upon expectations created by past experience. How can we claim that something will always have the same effect in the future? Experiences may change and therefore reasoning with regard to cause and effect changes. Karl Popper (1902-1994) proposed the hypothetico-deductive method whereby theory claims are put to the test and are either rejected or retained for the time being. With this scientific method the emphasis is not on conclusive verification, the
0
Interviews in Qualitative Research
Philosophical Assumptions
establishment of a proposition as true for all time, but instead on falsification -facts are not unassailable, rather they are open to constant challenge.
Undoubtedly, the key issue is that science is 'based on facts.... Based on what we can see, hear and touch rather than personal opinions or speculative imaginings' (Chalmers, 1999:1). Interestingly, Kvale (1996) takes exception to the implication that qualitative interviews are necessarily 'unscientific', arguing that this depends on how science itself is defined. He offers the following definition of science: 'the methodological production of new, systematic knowledge' (p. 60). Inherent within Kvale's argument is the view that scientific method based upon hypothesis testing, objective results and generalisability is one way of configuring science. Alternatively, a scientific method could also include systematically produced 'intersubjective reducible data', such as those produced in qualitative interviews.
Verstehen and the issue of causality
While we might question aspects of the scientific process, the overarching principles of data generation, founded upon hypothesis testing, observation and measurement, are sacrosanct in that domain. Alternatively, interpretivism prioritises the interpretation and meaning of human experience over measurement, explanation and prediction. Personal opinions and imaginings are not framed as merely 'speculative', rather they constitute what makes life intelligible. This differentiation is attributed to the thinking of Max Weber (1864-1920), who suggested that the human sciences should be concerned with understanding [verstehen). The following explanation from Strike (1972: 28) is certainly outdated in its gendered assumptions but it appears to exemplify why the scientific process alone may not be sufficient:
the verstehen doctrine will claim that human beings can be understood in a manner that other objects of study cannot. Men have purposes and emotions, they make plans, construct cultures, and hold certain values, and their behavior is influenced by such values, plans, and purposes. In short, a human being lives in a world which has 'meaning' to him, and, because his behavior has meaning, human actions are intelligible in ways that the behavior of nonhuman objects is not.
This search to uncover meaning is then contrasted with an emphasis on explaining [erklären) and establishing causal relationships exemplified by the natural sciences. Demonstrating causality requires the researcher to show that an effect is due to a particular cause/variable. For example, we might undertake research to investigate a causal link between early parenthood and relationship breakdown. Causal explanations are usually in a linear form, stating cause and effect in a straight line; X causes Y. However, if we accept the general import of verstehen, such uncomplicated linear explanations seem incomplete, even unsustainable. The reason(s) for relationship breakdown might be connected
■ h a combination of numerous factors: changing aspirations, differing values in tion to gender roles, factors related to the child, social difficulties. Further, lCjiat counts as 'relationship breakdown' itself may be a matter for differing • terpretations. To prove causality we would need to eliminate the possibility ro3t an effect might be due to something other than the causal variable (e.g. early parenthood). It is the complex meanings that people give to their existence that are of interest in qualitative research, and thus tracking down linear causal relationships often becomes erroneous or unproductive and of little value in developing understanding. Wilhelm Dilthy (1833-1911) similarly contrasted verstehen and erklären, proposing that natural reality and social reality are different kinds of reality and therefore require different methods of investigation. Maybe this is the point at which the researcher has to make some decisions. What kind of reality do we subscribe to and how does this impact upon our rationale for using qualitative interviewing?
It should be noted that not all qualitative research rejects any attempt to explain or to talk in terms of cause and effect. Qualitative neo-positivist research (see below) may well seek explanations of social phenomena, often in terms of existing theory. Work within what we call the 'limited realist' tradition (see below) may also concern itself with causality, especially in realist evaluation research (Maxwell, 2012). However, the goal here is not to establish cause-and-effect relationships between discrete variables, but rather to develop plausible accounts of causality in complex social situations.
Developing a rationale
While much is made of the differences that exist between qualitative and quantitative research, it is undeniable that both paradigms share a purpose. The purpose of research is to enhance knowledge, to in some way enable us to know more. When undertaking research it is standard practice to develop a research proposal outlining a clear rationale for the research: what is it that we want to know, what is the purpose of the research and how might this be achieved? This is- often where those new to the use of qualitative research methods hit problems. Crotty (1998: 13) argues that when discussing the research process 'we need to be concerned about the process we have engaged in; we need to lay that process out for the scrutiny of the observer; we need to defend that process as a form of human inquiry that should be taken seriously'. Nonetheless, the philosophical underpinnings of quantitative research are often not outlined, remaining implicit within the methods used for generating data. For quantitative research the status of 'facts' supported by measurement and observations seems enough to demonstrate that the work has epistemological integrity. However, as we will see in the following sections and in subsequent chapters, a more detailed account for qualitative research is needed. This is not done out of any inherent need to defend qualitative methods. Rather it reflects the fact that there is a variety of philosophical positions within qualitative research so we cannot simply assume an unstated conventional
0
Interviews in Qualitative Research
Philosophical Assumptions
©
position. It is at this point that there is a need to make appropriate connections between the nature of the research, our overall strategy and how we will go about collecting and analysing data.
Box 2.1 outlines two different rationales for research looking at coercive treatment for drug misuse. Using this particular research project, where the second author was part of a research team commissioned by a local agency, we make visible the initial implications of these traditions. Evident are the ways in which theoretical frameworks rooted in specific philosophical positions produce differing rationales for what might need to be known, thus advocating particular research methodologies and methods. In using this example we do not intend to exemplify a full account of any methodological approach. Instead we aim to present the main aspects of interpretivist and positivist traditions in a format that enables a comparative assessment. Also note, as we show in the next section, that there are significant variations within the interpretivist tradition
Box 2.1   Developing a rationale underpinned by philosophical theoretical perspectives
Recent government initiatives around coercive treatment (alternatives to prison) for those convicted of offences linked to substance misuse have resulted in a need to know more about the treatment process and the impact this might have in terms of bringing about behaviour change. The figure below aims to make evident how different philosophical understandings impact upon the rationale we might develop when proposing research in this area. Also evident is how the rationale then impacts upon what kind of data we aim to generate.
interpretivist^—.-
-^^fmethod of data generation:
Interpretivist			Positivist		
	Interpretivist perspective Multiple versions of reality			Positivist perspective t One version of reality	
Need to understand how offenders are engaging with treatment »  How might offenders perceive the treatment providers? • How do offenders engage with treatment services? • What impact do they believe treatment will have on their alcohol misuse and offending?			Need to know the impact of coercive treatment on offending behaviour: • How many offenders are receiving coercive treatment for alcohol misuse? • How compliant are offenders with the treatment regime? • What impact does coercive treatment have on overall rates of offending?		
Semi-structured interviews with offenders aiming to enable participants to present their individual understandings and experiences.
Positivist_
Proposed method of data generation:
Access and collate a range of statistical information that will aim to investigate if any association can be made between coercive treatment and rates of offending before and after the introduction of treatment.
The interpretivist rationale focuses on understanding how individual offenders experience treatment services and what it means for them. The positivist rationale relies on more factual, statistical information where cause and effected be investigated, thus an interest in comparing rates of offending. Both rationales have the potential to enhance our knowledge base around coercive treatment. However, philosophical theoretical underpinnings result in different rationales - different perspectives on what we might need to know.
Epistemology, ontology and qualitative interviewing
Having explained some of the overarching philosophical tensions that exist in the research process, we now aim to situate issues raised within specific philosophical traditions. A useful place to begin is to consider the status of conversation. It is all too easy to view conversation, within the qualitative interview situation, as an uncomplicated exchange of ideas and opinions. Breakwell (1990: 81) states that 'The interview approach relies heavily upon respondents being able and willing to give accurate information'. The assumption here is that accurate information is there to be discovered and thus such knowledge is achievable. We do well to take time to consider how these ideals have been challenged by more critical approaches.
Rorty (1979) emphasised how we constitute knowledge through conversation and social practice. So rather than knowledge being conveyed in conversation, it is brought into being. This has resonance for qualitative interviewing as we become increasingly aware of the constructive nature of social interaction and the part played by active subjects in making sense of their experiences (Gubrium and Holstein, 2003). Indeed, Shotter (1993: vi) describes how 'conversation is not just one of our many activities in the world. On the contrary, we constitute both ourselves and our worlds in our conversational activity.' Here we return to the idea that it is our personal imaginings that make life intelligible. Thus the idea of conversation as no more than observable verbal behaviour, or verbal exchange, where knowledge of an objective reality is described and discussed is continually being extended and challenged.
0
Interviews in Qualitative Research
It becomes clear that what counts as knowledge, and how that knowledge is generated and understood, carries real implications for qualitative interviewing. Methods and methodology do not exist in a vacuum, rather they are subject to new and extended ways of thinking about the world.
In later chapters we will locate qualitative interviewing within the specific philosophical and theoretical perspectives of phenomenology (Chapter 11), discourse analysis (Chapter 12) and narrative research (Chapter 13), endeavouring to reveal how data are intricately associated with beliefs about reality and knowledge. In an attempt to make clear the implications embedded in such beliefs we explore four discrete philosophical positions: qualitative neo-positivism, limited realism, contextualism and radical constructionism. We define these positions in terms of both their ontology and their epistemology, agreeing with Maxwell (2012) that the neglect of the former in many textbook discussions of philosophical foundations is a mistake. Any framework for identifying philosophical positions inevitably involves some simplification, and portrays positions as more distinct and discrete than they actually are. We freely confess this is true for ours! However, we still feel it gives a useful starting point for your thinking about the assumptions you are making in your research, and how they should impact on your design and execution of it.
Of value in the process of exploring these different positions are three questions posited by Willig (2001: 12-13), which she suggests provide a framework for elaborating on assumptions that might underpin particular methodological approaches:
• What kinds of assumptions does the methodology make about the world?
• What kind of knowledge does the methodology aim to produce?
• How does the methodology conceptualise the role of the researcher in the research process?
These questions will be evident as we outline these differing philosophical positions. It might also be the case that these questions prove useful for any researchers who are in the process of interrogating the integrity and coherence of their research. Table 2.1 gives a summary of each position using this framework (based on King and Brooks, 2018).
Qualitative neo-positivism
Qualitative research from this position takes an unambiguously realist view of both ontology and epistemology. We consider the implications of this in relation to Willig's (2001) three questions, below.
Philosophical Assumptions   ( 19
fable 2.1 Summary of four broad philosophical positions for qualitative interview research
philosophical position_
Ontology -assumptions about the world
Epistemology - nature of knowledge produced
Implications for
interviewer role
Neo-positivism
Limited realism
Contextualism
Radical
constructionism
Realist
There is a single real world 'out there'
Realist
There is a single real world 'out there'
Realist
Seeks potentially generalisable, objective knowledge. By following methods correctly, can have degree of certainty about conclusions
Constructivist/ relativist Seeks plausible explanations, sometimes including causal mechanisms. Recognises these can never be free from the 'position and perspective of the researcher(s)
Relativist Multiple, potentially competing versions of reality may exist
Relativist Multiple, potentially competing versions of reality may exist
Constructivist/ relativist
Seeks rich accounts of experience
and/or social phenomena in
specific contexts. Recognises
these can never be free from the
position and perspective of the
researchers}
Strongly relativist Seeks critical understanding of how versions of reality are constructed in specific settings, especially through language. Applies same critical stance to claims about knowledge production and the research process itself
Researcher seeks as much objectivity and avoidance of bias as possible. Higher degree of standardisation likely than in interviews from other positions
Researcher uses reflexivity to understand own influence on research process. Attention to own subjectivity and, where appropriate, perspectives of multiple
researchers, helps produce plausible explanations
Reflexivity essential to producing rich account of data, recognising key role of relationship between interviewer and interviewee
Critical reflection on self and data at all stages of research essential. Attention to range of influences (e.g. theoretical, interactional, political] and recognition of performative nature of language (e.g. how interview questioning may structure interview encounter)
Interviews in Qualitative Research
Philosophical Assumptions
Assumptions about the world
The qualitative neo-positivist position makes the same kind of assumptions about the world as quantitative research in the hypothetico-deductive tradition. Thus it takes the view that there is a 'real' world out there, which exists independently of our attempts to understand it. The goal of qualitative research from this position is therefore to capture an accurate picture of people's real experiences of the world.
There is a tendency for qualitative researchers from more interpretivist traditions to refer to this position as 'naive realism'. We prefer the present term, as used by Duberley et al. (2012), because it recognises that researchers may make a thoughtful choice to adopt a neo-positivist stance, rather than doing so simply out of naivety. Examples of where this may happen include qualitative studies which are strongly tied to existing mainstream theories, and mixed methods studies where researchers are concerned to achieve straightforward philosophical coherence between quantitative and qualitative arms (Hughes et al.* 2010).
Assumptions about knowledge production
We said earlier that epistemology relates directly to a means of estabhshing what counts as knowledge. Those utilising qualitative neo-positivism believe that by adopting particular methods we can describe and (where appropriate) explain aspects of the world out there. Fundamental assumptions about objectivity and reliability, dominant within quantitative positivist research, prevail when using qualitative interviewing underpinned by a realist epistemology and ontology. Attention is likely to be given to representative sampling and gener-aUsability such that even if studying a small number of cases the aim is to be inclusive of those who represent the larger group or population under study. Interview guides in qualitative neo-positivist research will tend to be more structured and less flexible than in other positions, though they should still allow scope for the participants to raise their own issues that the researcher may not have anticipated. With regard to the quality of data analysis, comparisons between independent coders to ensure the clarity and validity of themes generated from the interview data will often be used. This is the only philosophical position of our four where the calculation of inter-rater reliability scores makes sense. (See Chapter 10 for more on quality issues in data analysis.)
Role of the researcher
All qualitative interviewing requires the researcher to consider their role in the research process. Qualitative neo-positivists do not believe that the interviewer can ever be an entirely impersonal and neutral data collector; the impact of interviewer characteristics and interpersonal dynamics between researcher and
rticipartf cannot be denied. However, from this position there would be Pa    attempt to minimise personal impact than from others. This may be bievcd by the type of interview guide used (as noted above), and strict limits 10 personal disclosure by the interviewer. Where there is a team of interview-training is likely to be needed to ensure their interview style is as
to ers
xandardised as possible.
Limited realism
This position shares with qualitative neo-positivism a belief in a reality which is independent of our engagement with it, but it does not hold that we can Icnow that reality with objective certainty by following methods correctly. Its relativist epistemology means that we must seek explanations that are plausible but always to some degree tentative. Approaches that we would consider as fitting within a limited realist category include various forms of critical realism (Archer et al., 1998), subtle realism (Hammersley, 1992), natural realism (Putnam, 1999), and Corbin and Strauss's version of grounded theory (e.g. Corbin and Strauss, 2015).
Assumptions about the world
In terms of ontology, limited realism shares the same view as qualitative neo-positivism: that there is a real world, which exists independent of our perceptions of it and interaction with it. The differences between these two types of realism stem from their contrasting views of epistemology. This is a useful place, though, to emphasise that ontological assumptions do not only exist in the form of the reahsm-relativism dichotomy. For instance, in social scientific research, positions on what constitutes 'human nature' are also a form of ontological assumption. For example, existential phenomenology assumes there is no inborn essence of human nature, while psychoanalysis assumes that we have innate drives that shape our development. Our focus on broad categories based on realism versus relativism is because in terms of qualitative research practice this highlights important broad similarities and differences among approaches.
Assumptions about, knowledge production
Limited realist approaches share a relativist rather than realist epistemology. Their proponents do not believe that through rigorous use of methods we can ever escape from our particular positions in the world. Our interpretations as researchers are always shaped by who we are and the methodological choices we make. In contextualist and especially radical constructionist research, the implication is often drawn that relativism means we cannot 'privilege' any one interpretation over another. Limited realists reject this argument: given that
■ reci views'in Qualitative Research
Philosophical Assumptions
0
there is (in their view) a real world 'out there', we should seek the best possible explanations for processes and phenomena within it, while accepting that our conclusions must always be tentative and partial. Limited realist research often draws on theory and/or seeks to develop theory. Finally, limited realists are often interested in causality - unlike contextualists and radical constructionists -but of a different type than that sought by positivistic research. Rather than looking for general laws, they seek to uncover contextually contingent mechanisms by which the causes of specific social phenomena might plausibly be understood (Maxwell, 2012).
Role of the researcher
The relativist epistemology of limited realism means that understanding the researcher's place in the research is an essential part of the process. This means that reflexivity is of great importance, both personal (what the researcher brings to the research) and methodological (how their methodological choices shape the data and its interpretation - for instance, using focus groups or individual interviews). Limited realists do not seek to use reflexivity to remove 'bias', as a qualitative neo-positivist might; they see the researcher's subjectivity as a vital element in achieving an understanding of the phenomena under investigation. However, their commitment to a realist ontology means they do seek to develop explanations that reflect what is 'really' going on in the setting they are studying, albeit recognising they can only ever be partial and tentative. They may therefore use techniques to make explicit their presuppositions, so that they can better see their impact on the research. Maxwell (2012: 99) advocates the use of 'researcher identity memos' in which at the start of a project researchers write about 'their background, purposes, assumptions, feelings, and values as they relate to the research'. The use of multiple researchers who consciously bring different perspectives to the research can also be valuable in limited realist research (e.g. King et al., 2017a). We cover reflexivity in depth in Chapter 9.
Contextual ism
Assumptions about the world
The basic assumptions of contextualism are that everyday life is set in a particular time, consisting of a myriad of factors, relations and activities, and is in a state of incessant change. From this position 'facts' cannot be commensurate with, or reducible to, a decontextualised view of human nature (Jaeger and Rosnow, 1988). The context of a historical, cultural and social milieu is integral to how we live, understand and experience our lives. The realist, scientific linear process of cause and effect, where directly observable facts are the arbiter of what counts as knowledge, becomes just one of many explanations set within its own historical context (Feyerabend, 1993). This means that all knowledge
. js dependent upon the context, this being inclusive of the perspective ° A oint taken when formulating the research. Hence, contextualism is *ded on the belief that all knowledge is local, provisional and situation f°un , (Madill et al., 2000). Approaches such as interpretive or hermeneu-^^henomenology (Langdridge, 2007; Smith et al., 2009; van Manen, 2014), 0C ^ crivist grounded theory (Charmaz, 2014) and some forms of narrative ^"earch (McAdams, 1993) adhere to a contextualist position.
Assumptions about knowledge production
tne qualitative research interview there is often, but by no means always, a face-to-face verbal interchange where one person (the interviewer) attempts to obtain information or expressions of opinion or belief from another person or persons. Therefore, especially in a contextualist approach, it is important to lenow as much as possible about the context of a particular encounter in order to produce knowledge that acknowledges and understands situated perspectives. Contextualists, unlike limited realists, do not attempt to produce explanations that reflect an objective reality, but rather try to understand phenomena in context from the particular perspective of the researcher(s). Thev may well ask 'What might the world be like for these people in this set-ring?'- However, in contrast to limited realists, they would not seek to reveal causal mechanisms that underlie the phenomena of interest.
Role of the researcher
In many respects, the role of the researcher in contextualism is very similar to that described for limited realism, with a strong emphasis on reflexivity. Within contextualism, researcher influence within the qualitative interview is not treated as a source of 'bias' threatening to undermine the validity and reliability of results, rather the researcher's subjectivity is as much part of the context as that of the participant. In interpretative phenomenological analysis, for example, researchers talk about a 'double hermeneutic', where the participant is providing an interpretation of their own experience, and the researcher is interpreting that from their own unique position (Smith and Osborn, 2008). Contextualist approaches commonly put a very strong emphasis on the importance of the relationship dynamics between interviewer and interviewee, and urge researchers to pay close attention to these (Finlay, 2005).
Radical constructionism Assumptions about the world
Any approach with its roots in constructionism begins by emphasising the role of language. These are by no means immediately accessible ideas as they may
24 )   Interviews in Qualitative Research
Philosophical Assumptions  ( 25
seem counter-intuitive to usual ways of thinking. The belief that language is referential, merely representing reality 'out there', is overwhelmingly brought into question within this strongly relativist approach. Rather than objects having meaning in a world that exists independently of our conscious interpretations of them, our interpretations/representations actively construct objects. Language is then not understood as representing reality but as doing things; for example, Clarke and Cochrane (1998) trace how 'natural' forms of child-care rooted in biological drives can actually be understood as embedded in discourse. Discourse refers to the way that images, stories, statements, ways of talking can produce a particular version of events (see Chapter 13 for much more on this). We can use Hall's (2001: 72) quote to outline further what is meant by the term 'discourse':
Discourse ... constructs the topic. It defines and produces objects of our knowledge. It governs the way a topic can be meaningfully talked about and reasoned about. It also influences how ideas are put into practice and used to regulate the conduct of others.
So, rather than claiming that mothering is a 'natural' biologically-located instinctual drive, Clark and Cochrane make visible the way in which language effectively constructs a host of expectations and obligations that suggest, rather than prove, that forms of mothering are 'natural'. Therefore, language is conceptualised as being productive; this means that language has the potential to construct particular (and different) versions of reality. This contrasts dramatically with positivism, where the one 'true' knowledge of the world is accessible through observation. It is hardly surprising that the rise of social constructionism and related critical approaches (described here collectively as radical constructionism) has to some extent challenged the foundations of existing knowledge, necessitating a radical rethink of what we consider knowledge to be.
Assumptions about knowledge production
Radical constructionism produces knowledge that does not adhere to traditional conventions. Objectivity and value neutrality are seen as discursive devices employed by a positivist science to uphold its powerful grip on knowledge production. The idea that human beings can somehow remove themselves from theprocess of active engagement in knowledge production is viewed with a level of incredulity. As Burr (2015: 172) puts it: 'No human being can step outside of their humanity and view the world from no position at all, which is what the idea of objectivity suggests, and this is just as true for scientists as for everyone else.' Also rejected is the view that there is an objective truth waiting to be discovered. What we have is meaning that comes into existence out of our engagement with the social world. Meaning is not 'out there' waiting to be discovered, rather it is brought into being in the nrrrass nf onrial wrU^™
sequentially, radical constructionism is relativist in epistemology, seeing ^"wledge as historically and culturally located. At different times and places will be different and often contradictory interpretations of the same phe-ena. It is also ontologically relativist, either because it believes there to be n<ultiple and potentially competing versions of reality, or because it in effect 'ubsumes ontological questions into epistemological ones (Maxwell, 2012). S    The philosophical underpinnings of radical constructionism may sound as if they are the same as those of contextualism. However, while contextualists interested in how we can understand experience in specific contexts, though always from our particular perspectives, radical constructionists are not interested in experience as lived at all. Rather they are interested in the position eople take within an interview and what they seek to achieve in so doing, without any attempt to reference internal states or personal motivations.
Role of the researcher
Burr (2015) asserts that radical constructionists call for the democratisation of research relationships, with research being necessarily a 'co-production' between the researcher and the researched. From a radical constructionist position, interviews are not a simple verbal exchange, but provide a social setting (with particular structures) for the negotiating of meaning between the interviewer and interviewee (see Chapter 13 for more on this).
Reflexivity is again important, as we showed for both limited realist and contextuahst research. However, the central focus on language use leads Willig (2001) to argue for the use of what she calls 'critical language awareness'. Radical constructionists should reflect carefully on the language they use in the research process - the categories, labels, forms of description and so on - to develop an awareness of how these shape what they produce from their work.
Conclusion
Qualitative research, including qualitative interviewing, requires a great deal of effort, with the researcher having to explore how they conceive the world. In this chapter we have made evident some of the philosophical and theoretical issues that prevail when engaged in qualitative interviewing. By showing that methods and methodology are distinct aspects of social research we have sought to make accessible some of the more difficult features of qualitative research. We have covered several of the tensions that exist between qualitative and quantitative approaches, outlining theoretical differences and presenting particular philosophical positions that we believe have relevance for qualitative interviewing. In the following chapters, as we examine the process of carrying out qualitative interview research, we will endeavour to show in more detail the ontological and epistemological thinking inherent in particular
apDrOaehp.e Hpfail aimc tn pímin rpcparí-Vifrc with thp triple tr> pncurp that
26 )   Interviews in Qualitative Research
they have the insight and information needed to give credence to their work moving beyond mere description to present theoretically driven and coherent qualitative research.
Recommended reading
Brooks, J. and King, N. (2017) Approaches to qualitative psychology. In J. Brooks and
N. King (eds), Applied Qualitative Research in Psychology. London: Palgrave. The first and third authors provide an overview of the different positions that can be adopted by qualitative researchers and, with examples, explain how commonly used methodologies map on to these positions.
Maxwell, J.A. (2012) A Realist Approach for Qualitative Research. London: Sage. A clear and thoughtful consideration of realist approaches such as those described here as limited realist (note that in this text 'critical realism' is used as a generic term to encompass these types of approach).
Ramazanoglu, C. with Holland, J. (2002) Feminist Methodology: Challenges and Choices. London: Sage.
The first chapter in this book provides an accessible yet detailed account of 'what is methodology in social research'. While this account is clearly situated within feminist methodology, the outline offered is useful for other methodological approaches.
Tuffin, K. (2005) Understanding Critical Social Psychology. London: Sage. The first three chapters of this book examine the ways in which researchers have attempted to understand the social world. Experimentation, science and social constructionism are all covered in some detail.
Ethics in Qualitative Interviewing
The ethical practice of social research with human participants is a complex and demanding responsibility. Throughout the social research process, from initiation to completion, ethical issues will exist and emerge - often raising moral dilemmas that are not easily resolved. Nevertheless, whatever the design, context or structure of the research, we must always be mindful of the ethical implications for all those involved in the process. In light of this imperative most qualitative and quantitative research texts include sections or chapters that outline what are considered to be the main issues that researchers need to consider. This book is no different in this regard, but it is prudent to be alerted to the extensive literature that already exists on this subject, spanning disciplines and ideologies. It would be impossible with such limited space to offer an all-embracing account that accommodates the breadth of historical, philosophical and political thinking underpinning ethics in social research. For this reason, what we offer in this chapter is insight into underpinning debates but also practical guidance and advice that is directly targeted at ethical issues in relation to qualitative interviewing. Being aware of the need to be comprehensive yet targeted and succinct, we will cover the following main areas:
• morality, epistemology and ethical principles
• ethical review processes
• qualitative interviewing and informed consent
(jžsj   Interviews in Qualitative Research
Ethics   ( 29
• confidentiality and anonymity
• physical safety and welfare of the researcher
Morality, epistemology and ethics
We begin by outlining some of the contemporary thinking and debates that surround social research ethics and qualitative research more specifically. Our justification for initially offering a more conceptual view is that it is important to have a sound appreciation of why you are taking a particular course'of action. This is similar to the point we have made in Chapter 2, regarding your choice of philosophical position.
As researchers we all bring to the research process our own individual morality which is an accumulation of understandings, feelings, positions and principles around particular issues. Our moral outlook has been shaped by the different experiences, events, and social and cultural locations that constitute our lives. Morality is therefore not merely a matter of simple universal dichotomies such as good and bad or right and wrong. Rather we each have our own individual moral viewpoints which, although not necessarily consistent and coherent, we nonetheless feel strongly about. This moral compass is there in research, and while there is the possibility of embracing certain general ethical principles, the way in which these are taken up and acted upon is very much reliant upon notions of individual morality. The following quote by Edwards and Mauthner (2002: 16) captures how ethics and morality are intertwined:
Ethics concerns the morality of human conduct. In relation to social research, it refers to the moral deliberation, choice and accountability on the part of researchers throughout the research process.
It is important to note not only the prevalence of morality, choice and accountability here, but also the idea of the whole research process. As qualitative researchers we need to ethically consider, for example the framing of our research question, how this is impacted upon by interested parties and what might be the implications and applications of research framed in this way. Yet interestingly ethical concerns are often only directed at research practice (methods, consent, confidentiality), with knowledge construction seen as an epistemological issue that is not situated within the ethical domain (Doucet and Mauthner, 2002). This is an unsustainable position not least because different approaches in qualitative interviewing are founded upon complex, competing and often political understandings of human experience. Thus there is an obligation to consider the morality of not only research practice but also the various practical, epistemological and ontological assumptions that surround and define the research.
In Chapter 2 we discussed the range of different philosophical positions from which qualitative research may be undertaken. A common asnect of all
but the neo-positivist is the rejection of a realist epistemology. This PoSltl0n1Sications for ethical tlünking; if we do not believe knowledge produces «fflP 1   _ k£ neutral and disconnected from the perspectives of those
|Ca!J in the process, we must consider personal and social context an
ever
rion
"nV°1V^l part °* any etmca* decision-making. Early feminist researchers CSSefi' 1981a- Gilligan, 1982) have shown how an epistemology founded (Oakley' >■ . ______(1_________.______:„;__a„^a „ „^,„^„1
men's experience, as an all encompassing norm, introduced a powerful bias that discriminated against women. There are regrettably many instances 0f knowledge construction where underlying epistemologies have mobilised detrimental cultural, social and gendered beliefs that have until recently remained unchallenged. Embracing qualitative methods with its more contextual located and constructionist roots prompts careful deliberation around knowledge production that is inclusive of inherently complex ethical relationships and responsibilities. We have ethical responsibilities not only to those who participate but also those for whom the knowledge is produced. As qualitative researchers engaged in producing knowledge we are required to act responsibly, being aware of how the research produced will be read, reinterpreted and used. Being attentive and transparent with regard to the personal, theoretical and epistemological assumptions that underpin, and generally inform, the research therefore has an ethical as well as reflexive dimension (we cover reflexivity more comprehensively in Chapter 9).
Ethical theories and principles
Within philosophy in general, and the philosophy of social research in particular, there is a wide range of theories of ethics and ethical behaviour. A detailed consideration of these is beyond the scope of this book (see, for example, Christians, 2017, for further discussion). However, an overview of the main positions is of value, as it sheds fight on the principles underlying ethical review processes (discussed later in this chapter) and the distinctive challenges for qualitative research.
Brinkmann and Kvale (2017) describe three main positions in ethical theory: duty ethics, ethics of consequence and virtue ethics. The first of these is sometimes referred to in philosophy as a deontological position, and it argues for ethics to be seen in terms of general principles that apply in all circumstances. Ethical behaviour comes from adherence to such principles. In a research context, the principle that researchers should always seek to avoid harm to their participants can be seen as reflecting a duty ethics position.
Like duty ethics, the second theoretical position of an ethics of consequences (commonly referred to as utilitarian ethics) also tends to operate through rules and principles, but these are not absolute. Instead, they are based on the notion that research should be judged according to its propensity to produce the greatest happiness for the greatest number. Thus while a deontological ethicist might argue against a risky new cancer drug because of the
likplv V>
arm tr\ rnal rprmiPntc   o nŕi
Ílíŕo
non m
ífrrir
Ethics
Interviews in Qualitative Research
significant and widespread good in the wider population would outweigh p0 sible negative outcomes to a relatively small number of individuals. Utilitaria principles figure very strongly in ethical review processes.
The utilitarian model of ethics is founded upon the ability to predict th consequences of an action, with Tightness or wrongness being dependent 0n the outcomes of an act. Utilitarian ethics has generally been deployed becaus of compatibility with scientific ways of thinking, drawing on rational thought I and the assumption of a single and consistent view of moral action. This is I often the kind of rationality we use in this chapter. For example, when recruit. ' ing participants we consider whether it is appropriate to include those who might be identified as 'vulnerable'; this consideration is generally framed in terms of predicting the possible consequences participation might hold for the participant. Utilitarian ethics emphasises the role of human happiness as a consequence of our actions but this does then translate into a concern with achieving the maximum good. Theoretically the infringement of an individual's rights may be acceptable in order to maximise the happiness of the majority. In research terms this might translate into including the vulnerable participant because, although as a consequence their well-being may be compromised, the well-being of the majority might be served by their inclusion.
There are difficulties with the use of utilitarian ethics in qualitative interviewing, not least the inherent problems associated with predicting the future consequences of actions. Qualitative interviewing seeks to be fluid and flexible, thus having to accurately predict consequences is generally not desirable or realisable. Also, if we return to questions of morality, how might different people, agencies and institutions view the consequences of certain actions? Our individual morality may make it impossible to align ourselves with an ideology that justifies a potentially detrimental act based on maximising benefits for the majority. Indeed, the assumption that one moral code might encompass the complexity of qualitative social research seems hard to sustain both morally and methodologically. In spite of this difficulty, generally speaking there are certain moral principles, drawn from utilitarianism, that are often times used when formally evaluating both qualitative and quantitative social research: respect for persons, beneficence and justice.
• Respect for persons demands that individuals participate voluntarily, having had adequate information about what involvement in the research will entail - including possible consequences. This means dealing with people as free to choose, but also acknowledging more vulnerable people's rights to be protected. There are thus two aspects to respect for persons: autonomy and protection. Being free to choose with regard to participation is generally related to information sharing. The participant should be given comprehensive information such that they are able to give their informed consent. There are particular target groups (e.g. children, older people, those with literacy deficits or mental health disabilities) where the individual's ability to understand and fully appreciate the process and consequences
©
- nation is hard to determine. Indeed there will be persons whose vul-0* ^ |jtjes require that they are protected against involvement in research. "cr3 ^ £or persons is a wide-ranging principle that for social researchers ^'S^well beyond simplistic notions of respecting people's freedoms.
licence relates to the researcher's responsibility to secure the well-•    of participants. The principle is to avoid harm, although whether K^rneans to avoid harm altogether or minimise it is much debated'. As lready stated, underlying this debate is the ideology of maximising benefits or securing the 'greater good', whereby it is understood that to achieve substantial benefits there may be certain risks. Some might think that this
» with researchers being able to jus-
fits or ----~
ubstanrial benefits there may be certain ris».». -„-c rilitarian argument is a slippery slope, with researchers being able to justify unethical research by pointing towards gains in knowledge. While we can historically point towards the catastrophic misuse of this principle (e g 'medical' research undertaken by the Nazi regime), most researchers have the well-being of all participants foremost in their minds. \  Justice is a principle that is often less talked about when looking at generic guidance on the ethical practice of research. Implicit in justice is the notion of 'fairness', and in research terms this translates into the fair distribution of both the benefits and burdens of research (for a review see Porter, 1999). Take, for example, how the benefits of research have not been equally distributed, with men and women's priorities fading to receive the same attention. Such injustice in research also works in the other direction with some groups being overburdened in terms of involvement (e.g. the institutionalised, certain ethnic groups, particular service users). While there are no easy solutions to such matters, having the principle of justice present does offer the likelihood of discussion and possibly accountability.
The third ethical theoretical position discussed by Brinkmann and Kvale (2017) is that of virtue ethics. While it does not dismiss the worth of rules and principles, it treats them as aids to decision-making, which must always respond first and foremost to the particulars of a specific ethical issue. This is in line with the pragmatist philosophy of John Dewey, who emphasised the responsibility of the individual to make ethical judgements in response to specific circumstances, informed by their own values and those of their community. Not surprisingly, this emphasis on ethics in context has attracted interest from many qualitative researchers.
Principles founded upon utilitarianism do without doubt extol notions of individual autonomy, rather than a more relational view. The scientific furtherance of knowledge, being regarded as the overall aim, is also evident. Indeed the existence of explicit principles does point towards a consensus worldview of
32 )   Interviews in Qualitative Research
Ethics   ( 33
what constitutes good ethical practice in social research. However, this consensus is seldom evident in qualitative research practice; as we said previously different approaches to research mobilise their unique understandings of ethical practice that are underpinned by certain theoretical, cultural and moral ideologies. Hence, many researchers have been engaged in working towards a more social ethics (see, for example, Holland, 2014) where a complex and situated view of moral judgements is adopted, in line with the pragmatist values approach described above. Within this broad approach conventions of. distance and impartiality in research are replaced with notions of caring, inter-dependency and collaboration. Carol Gilligan (1982) characterises the female moral voice as an 'ethics of care'; here merely avoiding harm is seen as inferior to embracing an ideology of participation founded upon compassion and nur-turance. Gilligan's ideas, and those of scholars building on her work, have had a strong influence on feminist and wider social ethics thinking and practice (see, for example, Held, 2006, 2014; Bell, 2014).
It is obviously simplistic to typify women as caring/relational and men as being engaged in a more rational approach within social research. Even so, the overall challenge to individualistic utilitarianism is evident, with Denzin (1997,2002) referring to 'feminist cornmunitarianism' as an alternative ethical theory that can take forward qualitative research. He argues that we are now in a period where there is an abiding concern with moral discourse that is inclusive of politics, gender, freedom, nation and community. The idea that we can appeal to an objective, morally neutral viewpoint (e.g. university review boards) is rejected, being replaced with a more localised morality, an ethics of care and 'shared governance'. Research becomes ethically situated in a mutually cooperative domain where the community is served rather than the producers of knowledge and the policy-makers. Within such a model participants have a say in how the research is conducted and may have a part to play in its actual undertaking. Research then becomes far more about social action with the researcher and researched participating together, acting in the best moral interests of both the individual and the community (see, for example, Shaw, 1999, on 'participatory inquiry'). Of course such approaches have their own challenges, not least that of how to achieve the participatory ideal when confronted with differences around methodological know-how (Heron, 1996). Those facilitating the research may find it hard to engender full participation when knowledge and power are so unequally dispersed.
Ethical review processes
When you begin a new qualitative research project you will almost certainly need to go through research governance and/or ethical approval processes. There is a good deal of variation between countries, and between research areas and institutions within countries, so it is always important to check the local requirements. In this section we give an overview of typical expectations, inevitably with a UK focus given our own experience.
Research governance
thical principles concern the rights, dignity and safety of research par-V7lule e esc&[Cil governance focuses on the development of joint standards ric»Pants' es zfaZ permit the proper management and monitoring of research ^unnecessary, allows sanctions to be brought in cases of research miscon-anC^' £j0vernance processes will normally include checking that ethical ^UCt als have been obtained; the actual application for ethical approval may 3'5 rnay not ^e Part °^ ^ overa^ Sovemance process. As part of research gov-0f nee you may be asked for evidence of relevant insurance cover (which will * most cases come from your university), managerial approval to access in nisations, and regular updates on the progress and eventual outcomes of the project. Data protection is an increasingly important part of research governance; approvals will often require a data management plan to ensure you store data securely at all stages, and only share it in line with appropriate regulations and crucially with what participants have consented to.
Ethical codes and review panels
The foundation for ethical codes of practice in research is the Helsinki Declaration of 1964 (World Medical Association, 2001), which formulated a set of core principles for human research in the broadest sense: protection from harm (physical and psychological); respect for individual dignity; right to self-determination; right to privacy; protection of confidentiahty. These feed into a range of ethical codes for specific disciplines, including psychology (e.g. American Psychological Association, 2017; British Psychological Society (BPS), 2014), sociology (e.g. British Sociological Association, 2017), social anthropology (e.g. Association of Social Anthropologists of the UK and Commonwealth, 2011) and education (e.g. Australian Association for Research in Education, 2018; British Educational Research Association, 2011), or sometimes across all human science areas (e.g. Canadian Institutes of Health Research et al., 2014). There are also ethical codes for some specific forms of research across disciplines, including online methods (Association of Internet Researchers, 2012; British Psychological Society, 2013) and visual methods (Cox et al., 2014). Universities may have their own general ethical codes, though these will generally draw on the specific codes as appropriate.
While details of ethical codes vary between types and across countries, there is a good deal of common ground among them. King (2018) discusses nine ethical considerations that might apply to any qualitative interview-based study, as outlined below. We will explore some of these issues in much more depth in the rest of the chapter.
1- Informed consent. The researcher should ensure that participants are fully informed about the research procedure and give their consent to participate in the research before data collection takes place.
®
Interviews in Qualitative Research
Ethics
2. Confidentiality. The researcher should maintain confidentiality regarding any information about participants acquired during the research process, except where this could lead to significant risk of harm for the participant and/or others. Very occasionally a participant (or group of participants) may want to waive anonymity, though even here researchers need to consider possible implications for others before agreeing.
3. Right to withdraw. The researcher should ensure that participants feel free to withdraw from participation in the study without fear of being penalised. This means either withdrawal from the interview process, or withdrawal of data after interview.
4. Assessing risk of harm. The researcher needs to carefully consider potential harm that could arise from the research, to the participant or others (including themself).
5. Deception. Deception of participants should normally be avoided altogether. The only justification for deception is when there is no other way to answer the research question and the potential benefit of the research far exceeds any risk to participants.
6. Debriefing. The researcher should ensure that, after data collection, participants have as good an understanding as possible of the research and its aims, including how data are to be used. Generally in qualitative research most of this is covered prior to data collection, but it is often worthwhile reminding participants of key points at the end of the interview.
7. Use of incentives. Careful thought should be given to any use of incentives for people to take part in your research. While there are circumstances where this is acceptable, as a rule of thumb, incentives should not be of a scale that would be likely to induce people to do anything they would not otherwise do.
8. Limitations to the researcher's role. In some circumstances, you may find that there is the potential for participants to be confused or uncertain about your role. This is particularly likely where you may have an existing relationship of some kind with a participant - perhaps you have interacted professionally with them (e.g. as a health professional, teacher or lecturer, line manager in their organisation), or they are a friend, neighbour or family member. A clear ethical risk in all these cases is that what happens in the interview may have an impact on a relationship outside the research setting. Even if there is no prior acquaintance, aspects of your identity that you have disclosed may create role confusion or uncertainty among some participants. For example, if you were researching experiences of anxiety and you had disclosed your status as a psychologist, an interviewee may think you are able to offer them professional therapeutic advice. For all such situations the crucial message is that researchers must be clear
th
fmcplvpc as tr, thp limifo r>f tW rnlp
participants in the course of obtaining informed consent. As we discuss further below, this will usually mean more than just including a statement explaining role limitations; very often we need to come back to issues of informed consent in the course of the interview.
<j Honesty and integrity in the research process. As a researcher you have an ethical obligation to participants, the research community and the wider society to carry out research in a fair and honest way. Breaches of this would include conscious misrepresentation of findings, failure to declare conflicts of interest, or failure to acknowledge (e.g. through authorship) the contribution of others to your research.
These basic considerations, some of which we cover later in greater depth, offer sound initial guidance for enabling researchers to protect participants from harm, aiming to preserve their well-being and dignity. The important thing to bear in mind is that these are basic ethical considerations that will normally be developed and expanded upon in the process of undertaking qualitative interview research.
Alongside the development of professional codes of ethics, government agencies in many countries now have stringent ethical processes that must be adhered to before any research can commence. For example, in the UK the National Health Service requirement is that any research involving patients or staff must go through the Health Research Authority (HRA) process. This involves completing, and submitting for approval, a lengthy document in which the researcher is required to give explicit details about the aims, methods and outcomes of the proposed research. The researcher is required to identify and comprehensively respond to a range of ethical issues, explaining in detail how these will be addressed. In order to accommodate the inductive fluidity of qualitative research it is possible to gain phased ethical approval. Although this is time-consuming it does mean that researchers can return at a later date with, for example, a more fully developed interview schedule or indeed a change of focus. Whatever the strategy, failure to comply with the strict guidance provided will almost certainly result in the proposed research being denied approval. Universities also have ethical review boards and although, in relation to the HRA process, they may require the information in a less prescribed format they will nonetheless call for similar information and detail. Such processes are not there to hinder research but should be viewed as providing an opportunity to thoroughly engage with the ethical implications that research with human subjects presents. Researchers often ask how approval is secured. There is no secret formula; all ethical review boards have guidance that is readily available. This guidance needs to be meticulously read through with a view to identifying the information required and the level of detail needed. If possible, speak to other researchers who have been through the process and ask if they might be willing to share their 'approved' documentation (providing this, of itself, does not have ethical implications).
Interviews in Qualitative Research
Ethics
0
Researchers do sometimes tell stories of difficult encounters with ethi^ committees. While we do not deny that this may occasionally happen, 0llr experience (having submitted to a range of ethical boards, and having bee^ members of ethics committees) is that the process is demanding but usually fair. Indeed, often the process and advice given can enhance methodologies rigour and pre-empt potential problems. This said, it may be the case that some ethical review boards are hmited in their understanding of qualitative research. This can prove to be testing and it is therefore important that the documentation submitted is written in a style that is accessible and clear. For example, while it will usually be helpful to identify your broad methodological approach, lengthy discussions of philosophy and methodology may alienate reviewers, whose main concern will be to get a clear idea of what ethical issues arise for participants and what the value of your research is.
Finally, when thinking about these processes it is useful to remember that such committees have a demanding remit, often aiming to be supportive of research while still being accountable for its ethical undertaking. It is in everyone's interests to engage in ethical research practice and therefore governance processes that regulate may be time-consuming and demanding but they are also necessary.
Qualitative interviewing and informed consent
Ethical codes of practice emphasise the importance of gaining the informed consent of participants prior to taking part in the research. Participants should be giving their 'knowing consent' (Berg, 2001), making choices free from duress or inducement. The key word here is 'knowing'. Are participants fully aware of what they are consenting to when they agree to participate in a qualitative interview? Do participants know if they are consenting to answer any and all questions that are asked in the qualitative interview? Are participants knowingly giving you the right to use the data no matter what is said? In this section we aim to raise central issues, making it plain that working through the specifics of informed consent is a crucial aspect of ethical research practice. It is fair to say that while it is important to obtain informed consent, participants will only be in possession of fuller knowledge of what participation entails when they are taking part in the interview. While accepting this conundrum, there is a great deal that can be done to ensure that the qualitative interview is a productive and/or enjoyable experience that holds no unwelcome surprises.
Process of negotiation
In order to get to the point of being able to conduct a qualitative interview, the researcher is in our view required to enter into what is in effect a process of negotiation. The researcher should provide as much information as possible so that the participant is able to consider, and potentially negotiate, the terms of their
ent In social research there are of course tensions, as offering explicit information may have implications for data production. If you state, '       je that you are interested in the power dynamics between managers ' "^jkers you may inhibit what certain participants will discuss. Participants ' W°   -ve certain consequences if they reveal too much about the relationship, ^fjs no hard-and-fast rubric here; rather the researcher needs to be conscien-' their efforts to share relevant, understandable and honest information. Of in this information-giving process your concern will be to protect the ■n-endedness and fluidity of the qualitative interview. Even so, every effort uld be made to minimise the risks from participation in social research. The direct personal contact with participants that occurs in qualitative ■erviews arguably enables researchers to obtain information that could help making ethical decisions and to engage in a genuine process of negotiation around ethical concerns (see Rosenblatt, 2000). While such flexibility means that prospective participants can be engaged with and informed in a variety of ways, there are specific issues that need to be covered. In Box 3.1 we suggest a framework for producing a participant information sheet that explains the purpose of the research, the nature of participation and what will happen to the data generated. Of course this will not be the same for all approaches to qualitative interviewing, and we are not suggesting the standardisation of such processes. Rather, the framework aims to highlight key information that researchers should consider providing.
Box 3.1   Information for interview participants
What is the purpose of the research?
• You can give the aims of the research here, but ensure that these are presented in a format that participantscan appreciate and understand.
• If appropriate, describe the potential benefits that the research might have for enhancing practice, informing policy and/or contributing to knowledge. Be cautious though - do not overstate the potential of the research.
Why have I been chosen?
• The research needs to have relevance with regard to the proposed participant's involvement, so be specific.
Do I have to take part?
• Participation is voluntary, and this always needs to be stated.
• If appropriate, make it clear that choosing not to participate will not have any
negative consequences.
(Continued)
Interviews in Qualitative Research
(Continued)
• Make participants aware that they have the right to terminate their involvement in the research whenever they please.
• Participants also need to know that they can decline to respond to questions or prompts at any point in the interview.
What do I have to do?
• Explain exactly what the participant needs to do if they are willing to become involved in the research. This is dependent upon the nature of the research, but generally participants will need the contact details of the researcher in order to initiate their participation.
• Provide some general information about the interview. It is good practice to make participants aware of the kind of questions that will be asked before they consent to participate. This outline need only be general, but it should both alert and prepare the participant for their involvement in the research.
• Give details of the expected duration of the interview and an indication of where and when it might take place.
• People do not like surprises when they are consenting to something that is unfamiliar, so mention that, with their permission, you will need to record the interview.
What happens to the information I give at the interview?
• Explain concisely how confidentiality of the data and anonymity will be handled.
• For some research you may need to explain the limits of these arrangements.
What will happen to the results of the study?
• Give details of dissemination; this should include feedback to participants.
Who is organising the funding of the study?
• If appropriate, provide this information.
Who has approved the study?
• Giving this information can allay people's concerns so, if appropriate, provide this information.
Contact for further information
• This information may have already been given. However, presenting contact details again in a way that encourages the seeking of further information can enable participants to seek clarification and therefore increase participation rates.
Ethics
[ Tpis information can be presented in a range of formats. Possibly the • d format is a letter with the information sheet enclosed. Alternatively, f^1" • be that this information can be verbally explained to participants as 'r rna^ an initial individual or group discussion with the researcher. The P3^ -ation might be presented as text but-with a more engaging graphic T at The way the information is presented will very much be reliant upon h researcher adapting and responding to the needs of the participant up h wou^ be f°ousk' f°r example, to use the information sheet in |ox 3-1 in its current S0*^ witk young schoolchildren. Something more aesthetically appealing and interactive would be needed. However, providing "nformation with this general information structure enables the researcher to outline the specific nature of the research. Participants are given information detailing what involvement will entail and the potential benefits and consequences of participation. "With such information the participant is able to make an informed choice, knowing what to expect if they consent to participate in the research.
Normally, ethics committees and review boards will expect consent to be documented by the signing of a consent form. Box 3.2 gives two examples of consent forms, one that can be informally completed before the interview commences but after information has been shared (Form A), the other (Form B) with a more personal tone being inclusive of specific information that can be sent to participants in advance. The important thing to keep to the fore is that informed consent is about being open, truthful and respectful of people's right to choose. While informed consent in social research is not generally a legalistic process (although there may occasionally be at time when such documentation has a more bureaucratic purpose) the consent form endorses the process, making it visible, memorable and most importantly something that can be returned to if necessary. There may be occasions, though, where it is very difficult or impossible to obtain written consent. This could be because of literacy issues, or because participants have a suspicion or mistrust of formal documents. It may simply be a practical matter, as Lawton et al. (2017) describe in relation to research in emergency medical settings. In such circumstances ethics committees would generally accept alternative arrangements for recording consent; for example, through audio-recording of verbal consent at the start of the interview. In Lawton et al.'s study, verbal consent immediately prior to interview was followed by later written consent.
Thank you for reading this information sheet and considering taking part.
®
Interviews in Qualitative Research
Box 3.2   Consent forms - two different options
Form A
Title of Project:
Name of Researchers:
Please initial box
1. I confirm that I have read and understand the information sheet dated □ (insert date) for the above study. I have had the opportunity to consider the information, ask questions and have had these answered satisfactorily.
2. I understand that my participation is voluntary and that I am free to withdraw □ at any time without giving any reason, and without any consequences for me.
3. I have been informed that the interview will be tape-recorded and □ I give my consent for this recording to be made.
4. I understand that all information I provide will be treated as confidential, □ and will be anonymised.
5. I agree to the use of anonymised direct quotes from my interview in □ publications and presentations arising from this study.
6. I agree to take part in the above study. □
Name of Participant
Signature
Date
Researcher
Signature
Date
Ethics   { 41
Informed consent for master's project
Title: Women's experiences of reproductive choice
rjear (name of participant)
a paster's student in the School of Social and International Studies at the University of Bradford. I would like to invite you to participate in research I am undertaking as art of my studies. The research has been approved by the University's, Departmental Ethics Panel. My research project explores the experiences of women as they consider reproductive choices and the prospect of becoming a parent for the first time.
If you agree to participate this will involve being interviewed once and it is expected that the interviews will last no longerthan one hour. I can undertake the interview at a time and place that is convenient for you and I would want to record and transcribe the interview. All interview data will be treated with the utmost respect and will be stored securely. However, information about the project, including interview data, will be shared with my dissertation supervisor and other appropriate staff at the University.
You may be concerned that other people will be able to know what you've said in the interview. I will do my very best to protect you from this by removing identifying information, for example changing your name and your exact age. You will be able to withdraw from the project at any time until 1 May 2009. After this time I will be at the point of writing up my research and therefore will not be able to remove quotations from the final dissertation. The final dissertation resulting from this project will be publicly available through the University Library.
I appreciate your giving time to this study and if you have any questions please do call
me at_. You can also contact my dissertation supervisor,
Dr_at_.
Thank you
(Name of researcher and signature)
If you are willing to participate in the master's project outlined above please sign below. Signature___
Print name
©
Interviews in Qualitative Research
Ethics
©
The process of negotiation required to obtain consent does not end wjtL 1 the recording of consent in whatever form. Rather, there is an ongoing proces throughout the interview and sometimes beyond. At various points in an intet ] view, you may need to check understanding and consent again - for instance if the interview takes a very unexpected direction that potentially raises ne\J issues, or if the participant suggests uncertainty about confidentiality. This kinH of 'process consent' (Ramos, 1989) can also take place within the interview situation. When interviewing parents who had experienced the death of a child Rosenblatt (2000: 204) gives the following example of how he sensitively used process consent within the actual qualitative interview, giving participants the implicit right to withdraw from aspects, and even all, of the interview:
I don't know if that's an appropriate question to ask
or not
I feel like maybe all these questions are too personal. You can tell me to shut up anytime you want...
Can I ask you ...?
The point of questions like these is to minimise any discomfort a participant might feel in refusing to answer a particular question, or follow a particular direction of exploration.
The right to withdraw and managing emotions
The right to withdraw from an interview, and to request withdrawal of data after the interview, is integral to the giving of informed consent. In reality participants rarely withdraw once they have consented, but if they choose to do so, this should be made as easy and comfortable for them as possible. However, the right to withdraw can be deployed mistakenly when participants become emotional and distressed during an interview. This is often the reaction of a novice researcher who assumes that someone's distress necessitates the termination of the interview, perhaps in relation to the ethical principle of beneficence and the undertaking to do no harm. Understandably, interviewers can often fear the emotional reactions of participants, but automatically invoking a distressed participant's right to withdraw from the process is often not the outcome that the participant wants or needs. Of course the aim of the interview is to gather information and not to elicit a participant's extreme emotional response. On the other hand, interviews aim to provide rich, detailed and in-depth information and it is hard to imagine how this might be achieved without layers of emotional input. Although dependent on the type of interview, in general as an interviewer you should acknowledge, and if necessary try to verbalise, the emotions you observe (e.g. 'Does this make you sad?'). You should not-ignore such emotions; sometimes it is useful to suggest a short break (make a coffee, etc.), after which the participant can be asked if
nt co continue. Having been involved in research that by its very nature U •   to emotions (e.g. drug misuse, intimate partner relationships, care of ~ nSC > we can say with confidence that it is often more useful for partici-■■■iXi° fjnish the interview. Emotional life is an essential part of our ce and therefore will be an integral part of the interview process. : \s well as having the right to withdraw from an interview, participants j in almost all cases be expected to retain the right to ask for their data to "° -moved from the study, after they have completed participation. It is impor-*rCto malce participants aware of the practical limits to this right, though. *rfk    there will come a point where it is no longer realistic to withdraw data; thesis may have been submitted, an article or book chapter published, ecially for student work at any level, it Would often be considered reasonable to not allow withdrawal of data after analysis has finished. These restrictions to the right to withdraw data are in most cases ethically fine, so long as participants are made clear about them from the start, for instance via the information sheet and perhaps a verbal reminder at the end of the interview.
There is a more complicated situation for studies that involve multiple interviews, whether face-to-face or online. A participant may complete some of the sequence of interviews, then withdraw from the study, potentially leaving the question of whether the data collected up to that point should be included in analysis. Again, this is where proper informed consent is essential. In this kind of study, you should make explicit the expectations regarding partial completion. For some studies it may be appropriate to ask participants to state that they wish previous data to be withdrawn, otherwise you will assume it can still be used after they leave the study. In other instances, it may be advisable to automatically remove all data when a participant withdraws. As a rule of thumb, the more sensitive the research topic and the more potentially vulnerable the participants are, the greater the Hkelihood that automatic removal of all data upon withdrawal should be the option.
'Off the record'
In our experience of using qualitative interviews in various research projects there is one thing that seldom varies. Almost always participants breathe a sigh of relief when the recording equipment is switched off but then continue to talk. Often this part of the qualitative interview encounter is crucial, with participants sharing sometimes' highly sensitive information that has direct relevance to the research. This appears to occur regardless of the tenor of the interview. We have known many different researcher responses to this kind of 'off the record' extended conversation - a conversation that at times is about personal disclosure. In our view there is only one ethical response that takes on board the ethical principle of respect for persons, and that is to renegotiate with the participant. As the researcher you can sensitively ask if it might be acceptable to turn the tape-recorder back on or if you might be able to write
{^44^5   Interviews in Qualitative Research
Ethics (45
down what is being said. Whatever course of action is taken in terms 0f recording data, the choice regarding the inclusion of 'off the record' data should always remain with the participant.
As well as comments made after the interview has formally ended, participants may sometimes ask for something that is said during the interview to be kept 'off the record'. Of course, as a researcher you must comply with this and remove it from the transcript, but it is not such a simple matter to delete your own recollection of what was said! You need to be careful that in presenting findings you do not in any way even hint at understandings that were based on what was told you in complete confidence.
Payment for participation
Paying participants to take part in research is controversial and may be seen as an inducement that changes the fundamental nature of the process. With payment, instead of participation being voluntary, the research relationship is seen to be founded upon tangible rewards that may impact not only on consent but also on any data generated. Drawing on an uncomplicated account of power relations, participants who have received payment may feel obliged to respond in a particular fashion, having thus relinquished their free choice regarding participation. An alternative view is one put forward by Hollway and Jefferson (2012) that takes a far more relational view of power. They argue that payment for participation may have an equalising effect, exchanging participants' time for researchers' money, and showing respect for their participation. However, the authors avoid oversimplifying the case, recognising that equalising is set within a structural understanding of power. This understanding suggests that power is unevenly distributed between the researcher and the researched. Such power differentials are often multifaceted, reflecting involved and contextually located fields of inquiry. While conceding that, in some ways, payment may induce participation, this does not negate the relational nature of the research process. Each party in the research process has inputs and investments that facilitate and mediate involvement. Of course there will be any number of factors to consider when making choices regarding payment for participation: when, how much, and to whom is payment made? However, situating payment for participation within a relational dynamic, rather than framing it as inducement, does provide a convincing ethical argument. Even in this context, though, we would reiterate our earlier point that the size of any incentive should not be such that one might reasonably think it could persuade people to participate against their better judgement.
Confidentiality and anonymity
Confidentiality and anonymity are often taken to mean the same thing in research. This is a mistake; while the concepts are related, they have quite
meanings that are critical in relation to qualitative interviewing. In the literature confidentiality is commonly viewed as equivalent to the prin-^of pr'vacy- Therefore to assure someone of complete confidentiality to suggest that what is said in the quahtative interview will remain *?Car and not be repeated. Obviously this cannot be what we want to imply ""^dal research, and more specifically quahtative interviewing, since as ^chers we undertake to report the findings/outcomes of research. It is 165 d to imagine how this might be done, when using qualitative interviewing, ^what is said is not to be repeated.
Rather than assuring complete confidentiality, qualitative researchers offer nfidentiality within limits which are made explicit to participants in the rformed consent process. Key to this is normally the use of anonymisation -jncealing identities through the use of code numbers, pseudonyms or similar.
Confidentiality of the data
When participants agree to be interviewed they have the right to expect that the data as a whole will be handled with due respect and discretion. Participants do not expect their interview data to be available for general consumption, unless they have agreed for it to be archived for the use of other researchers, as part of the informed consent process. They should be able to rely on their data being kept securely, with identifying information removed and known only to the researcher, or research team if appropriate. A process for managing the separation of personal information from the data needs to be put in place before the collection of data. This may be achieved by numbering interviews (Participant 1, 2, 3 etc) or using pseudonyms, and keeping a record of actual names in a secure and separate location that can only be accessed by the researcher (lockable filing cabinet or password-protected computer). Indeed you should consider whether you have a good reason to maintain a record of participants' actual names. You may need to do so if you have agreed to enable participants to look through the transcript (a practice we will return to later), or you may need to contact them for follow-up interviews. Another reason for retaining personal details is to send participants a summary of findings, or some other type of dissemination; if this is your only reason for keeping contact details then you could do so without keeping a list matching them to pseudonyms, thereby enhancing anonymity. As well as anonymising participants, it is normal to similarly anonymise other individuals referred to in an interview, and sometimes organisations and geographical locations too.
Data protection regulations in most countries expect personally identifiable data not to be kept longer than necessary, except where archiving has been consented to, or in rare instances where there is an overriding public interest in so doing. As a result, ethics committees often ask how long the data will be stored, and the expectation is that unless agreed otherwise the data will
46 )   Interviews in Qualitative Research
Ethics   ( 47
be destroyed once the stated period has expired. We can hear qualitative researchers protesting at the loss of data that could be reworked and reana-lysed. Such measures do not preclude the archiving of qualitative data that can be made available for further or secondary analysis; for example, in the the Economic and Social Research Council (ESRC) provides access to data for secondary analysis from projects it has funded, via the ESRC Qualitative Data Archival Resource Centre. Actually, it could be argued that the principle relating to the length of time personal data is kept does not relate to data that have been anonymised because the data are no longer 'personal'. This is a view that many researchers take and indeed one that makes logical sense. However, serious consideration needs to be given to those data that are not so easily anonymised. Equally, we need to think about notions of good practice in scholarly research. If the data are destroyed too early, how might findings be validated if challenged? As you can see, it is hard to give categorical advice on this issue, particularly when thinking about work such as Catherine Riessman's 'A Thrice-Told Tale' (2004). In this work Riessman returns many years later to previously used data, analysing them for a third time and offering new insights into the experiences of someone living with multiple sclerosis. To suggest that such data should have been destroyed seems almost barbaric. What we are saying is that it is the researcher's duty to be aware of the responsibilities surrounding the storage of interview data and that these should be at the forefront of any decisions that are made with regard to confidentiality of the data. In addition, participants should be as fully informed as possible about how data will be anonymised and managed, including how long they will be stored.
Transcription
Transcription is always a time-consuming and demanding task and often it is contracted out to people with the essential skills. Of course there is the consequential impact that you do not develop the same level of familiarity with the data if someone else does the transcribing. Nonetheless, realistically time constraints may mean you need to employ others to do this task. This is not necessarily a problem, and can greatly reduce the demands placed on qualitative researchers. Often those who take on this work have experience working with confidential data. Even so, the researcher must ensure that the transcriber is aware of confidentiaUty issues and agrees to respect the confidentiality of the data. In a similar vein, researchers are very enthusiastic about their work and novice researchers can become eager to share their fieldwork experiences. Discussing the data professionally (with co-researchers, colleagues) is fine, but relating this in a way that identifies individual participants is not and researchers need to quickly develop the ability to use, at all times, the assigned pseudonyms. We would recommend that wherever possible, pseudonyms should be assigned as part of the transcription process - that way, researchers are less likely to find themselves-talking about participants by their real names.
Anonymising the data
rally as already mentioned, the advice given in textbooks is to use pseu-Ci°e replacing the participant's name with an alternative one. This is of don) m^oun^ advice and does go part way to making sure that participants are ^"ulentifiable. There are, however, many instances in qualitative interviewing tj^s simply will not suffice. For example, in a narrative interview a par-' • ant will describe in great detail their personal experiences, relationships ° d life events. The story that is told, its structure and fabric, remains trans-nt - knowable to others. The name of the participant is therefore not the ojjy way in which the participant can be identified. In a semi-structured inter-with a 'key informant' (someone who has specific information relating to c research, e.g. a head-teacher or public health manager) the position that che person occupies and/or other attributes and characteristics will be both relevant to the interview and identifiable to others. Take the example of interviewing a group of young mothers where one of the participants has three children, while all the other participants have one or two children. The mother with three children is distinct in relation to the other young mothers participating in the research. If this specific information is included the young women participating in the research, and others, reading the subsequent research report may be able to directly link quotes to the mother with three children. Evidently the use of a pseudonym would not suffice to protect anonymity in this instance. A decision needs to be made about the use of such identifying information. Its removal may in many instances have profound relevance in terms of research aims. The young mother referred to has her experiences and understandings rooted in the fact that she has three children, and therefore omitting this information from your analysis may have an impact in terms of contextually locating the data.
There are no simple solutions to such dilemmas, and it is the responsibility of the researcher to comprehensively think through the impact that participation might have for people taking part. It might be argued that the nature of qualitative interviewing makes anonymity a highly challenging concept. Participants are invited to share their personal thoughts and opinions, and these are always set within their own lives, which have distinctive features. Therefore the researcher needs to have considered in more detail how to anonymise, and perhaps even explore with participants if anonymisation is possible and/or desirable. You could also enter into a process of negotiation with participants with regard to the use of specific quotes. Participants will have insight into what might be fine to use and which quotes may hold potential dangers with regard to the disclosure of identity.
Regrettably, all too often issues around anonymity are left until the later stages of the research when the costs to both the participants and the research have escalated. Removal or omission of data can profoundly affect the overall outcome of research, yet inclusion without anonymity may have reverberating consequences. Our advice is always to be upfront and clear when discussing
Interviews in Qualitative Research
the potential benefits and effects of participation when seeking 'informed con-sent'; this should most certainly relate to anonymity if this is a concern. Als0 it is important to consider that there exists a growing awareness of how some research participants may want to be identified and not anonymised in research outputs, often because they wish to claim ownership of their stories (e.g. Grinyer, 2002), or because they feel their stories are already (at least in part) in the public domain. An interesting example of the latter is Linda Asquith's research with genocide survivors (Asquith, 2015). She makes the point that participants may have felt that being anonymised is another form of erasure and dehumanisation, and therefore gave participants the right not to be anonymised - an option some took. Even if this is the case, it is still the researcher's responsibility to decide if such a strategy is ethically sound. In these circumstances, researchers are advised to obtain written consent that an individual wishes to waive their right to anonymity.
Ownership of the data
Participants, commissioners and other interested parties all want to be represented in a positive light and at times this may bring about serious clashes and disagreements around who actually owns the data. Even when consent has been undertaken as a process of negotiation, there may still be points of conflict around ownership of the data. Based on the 'democratic principle' (Simons, 1984) data is the property of the interviewee and they have the right to negotiate what information is made public. In some cases, researchers take the view that it is therefore appropriate to give interviewees copies of their transcripts to check before analysis starts, though this is far from a universal position. If you take this route, you need to be clear what you are (and are not) asking of participants. An ethical case can be made for allowing them to ask for comments to be deleted with which, on reflection, they feel uncomfortable. However, sometimes participants want to correct their own grammatical mistakes or rephrase colloquial or dialect terms; this may be methodologically undesirable as research is often concerned to capture participants' 'natural' talk.
Ownership of the data may also be claimed by the hinders of research, and there are certainly instances when findings have not been disseminated because the commissioners have withheld permission. It is therefore important to consider the contractual claims that might be placed upon actual raw data and research insights when embarking on funded research projects. Leaving aside the contractual aspects, there may be other challenging moral issues. What if handing over data has implications for interviewees? While for some researchers there may be codes of conduct that bind people not to release information without the permission of interviewees, this may be neither robust nor generally applicable. The advice is to be circumspect when first embarking on research, try to think through and negotiate how data will not
Ethics
•ored and anonymised but also the potential for coming into conflict
ooly be s .l:0 This way you can take action to secure the appropriate oVer ownersiuf t _
ose
r dally dangerous error that should always be addressed. Thankfully, with ^--traduction of more rigorous ethical processes such concerns have
the m'
of interview data
physical safety and welfare of the researcher
the importance and complexity of issues relating to the welfare and tonomy of participants, it is possible for the physical safety and welfare of researcher to be overlooked when planning a research study. This is a
1
C_-
come more formalised. Let us first critically consider the physical safety of the researcher. In our personal lives we are cautious about entrusting our safety to others. For example, we check out the background of those who care for our children; if someone uses a dating app they are advised to ensure that any first meeting is in a public place. Similar levels of care need to be taken when undertaking research. The researcher should first and foremost be encouraged to consider the potential dangers when meeting participants who are often also 'strangers'. Of course we would not want to overstate the point; qualitative interviewing is primarily a fulfilling and enjoyable method of collecting data. Even so, being safety conscious from the start can ensure that problems do not arise.
Qualitative interviews can be undertaken in many different settings, with the researcher often able to negotiate the location. If this is possible then the interviewer can ensure that they control the environment, being able to access a telephone, notify another person that they are interviewing, and making that person aware of where they are and when they will be finished. On the other hand, sometimes the interviewee may be unable to travel, or it may just be convenient that the interview will take place in their home or another place that provides easy access for the interviewee. When making these arrangements the researcher needs to consider if there are safety issues. To merely arrange an interview without any strategies in place to ensure researcher safety is irresponsible in the extreme. It is essential to have a safety protocol (a set of practices) as an effective way of dealing with such issues. Here we offer what we believe to be a useful but far from exhaustive safety protocol:
• Always carry a mobile phone. Make sure it is charged and has credit. Do not be careless and think it will be OK this time if the battery is running low.
• Always inform someone of the time and place of the interview and let them know that you will call once the interview is complete. The person you inform and call can be your PhD supervisor, the research principal inves-tigator/co-investigator or a friend. The most important point is that they know when to expect your call and are alerted to the need to act should you not call.
Interviews in Qualitative Research
• If you inform your contact by e-mail check that your message has been received and can be acted upon if necessary. Often the person you designate as your contact will be a busy person; if they do not read your message your safety may be compromised.
• Throughout the interview be aware of- safety issues and reflect on whether you feel safe enough to continue. If you feel unsafe politely suggest re-arranging. Once you have terminated the interview you can consider the next course of action; possibly a different interviewer, or the decision may be that this interview needs to be abandoned.
• Always call your contact once the interview has been completed. Ingham et al. (2000) suggest making a phone call to your contact in the presence of the participant so that they are aware of the precautions being taken. This may indeed be good advice if you have specific reasons to be concerned, for example if interviewing particular target groups with a history of challenging or confrontational behaviour. Whatever the timing of the call, forgetting to ring will, and should, have consequences. You would expect your contact to take action and to be very angry to find that you merely forgot to get in touch after the interview. A further consequence may be that if you again fail to ring they will be unsure whether this is just another instance of forgetfulness and may delay taking action, thus further compromising safety.
Finally, issues of researcher safety also relate to the potential personal impact of qualitative interviewing on the interviewer. Earlier we referred to debriefing in relation to participants; it can often be the case that researchers themselves need to debrief. By its very nature qualitative interviewing can place researchers in situations where they hear and learn about experiences and events that are not easily put aside when the day's work has ended. It is important that researchers themselves are able to discuss and talk through what may have happened in the course of an interview. This should be done in a way that takes on board all we have been discussing in relation to confidentiality and anonymity. Therefore the researcher/interviewer debrief should most probably be an arrangement with co-investigators or research supervisors. While it would be naive to suggest that the safety of all those involved in qualitative interviewing can be completely and easily assured, the strategies we have suggested can uphold your confidence in having undertaken qualitative interview research that is ethically considered and responsible.
Ethics
©
de meeting the more functional demands when undertaking qualitative al°ngS1 our account, situating ethics in qualitative interviewing within a interv^^ epjsremological frame of reference, does place a great deal of
'tal <___C*-:11      yo imT^ri-oT"!*- ti~, rwv-iCTTiic*» l-haf nftpT, cm
inte mo
sibility on the researcher. Still, it is important to recognise that often an '^'"'compassing solution to some of the ethical dilemmas you will come up a1' enCr°will not be feasible. Ethical codes of practice can be of help, as can the 3Sm" ce and support of other experienced researchers and wider research ^vernance systems. Be that as it may, it is essential to constantly have a criti-S°|Vstance towards the whole research process, knowing that ultimately it is h researcher who is accountable. When we say this it is not to instil a level o/trepidation but rather to encourage the thoughtful ethical practice of qualitative interviewing where care and respect are intertwined.
Recommended reading
Hollway, W. and Jefferson, T. (2012) Doing Qualitative Research Differently: A Psychosocial
Approach, 2nd edn. London-. Sage. In addition to a very useful chapter specifically on ethics (Chapter 5), throughout the book the authors provide numerous detailed examples that help bring to life the kind of challenges qualitative researchers can face.
Miller, I, Birch, M., Mauthner, M. and Jessop, J. (eds) (2012) Ethics in Qualitative Research,
2nd edn. London: Sage. A wide-ranging collection on key topics relating to ethics and qualitative research.
Christians, C.G. (2017) Ethics and politics in qualitative research. In N.K. Denzin and Y.S. Lincoln (eds), The Sage Handbook of Qualitative Research, 5th edn. Thousand Oaks, CA: Sage.
Particularly useful for considering the principles behind different approaches to research ethics and their implications.
Conclusion
The ethical practice of qualitative interviewing is a wide-ranging and often demanding enterprise. Throughout this chapter it has been our intention to be inclusive, aiming to offer theoretical and conceptual ethical understandings,
Designing an Interview Study   ( 53
Framing your research question
Designing an Interview Study
In this chapter we outline the typical stages you are likely to pass through in designing a research study based on qualitative interviewing. In light of the discussion in the previous chapter, concerning a variety of positions within qualitative research, it should come as no surprise that there cannot be a single universal protocol to follow for developing a qualitative interview study. In the latter part of this book some of the distinctive features of particular methodological approaches are examined: phenomenology (Chapter 11), discourse analysis (Chapter 12) and narrative (Chapter 13). Nevertheless, we would argue that there is sufficient commonality among many traditions of qualitative research to make a generic account of the project development process at least a useful starting point.-
The chapter is organised around the following main tasks in the development of a qualitative research study:
• framing your research question « choosing the type of interview
• defining your sample and recruiting participants ° developing an interview guide
framing a research question that is appropriate to a qualitative study, there ^ several issues that you need to take into account. The first is the type of ^estion you should use. By this we mean the kind of knowledge that the ^U archer seeks to produce from analysis of interview data. The second is the ^ope of the question: how broad or narrow a range of experience is the study seeking to examine? A third issue is the need to avoid presuppositions in the
estion that might distort the research process. Finally, you need to consider jj,e extent to which the research question itself might change in the process of carrying out a qualitative study.
Type of research question
One of the most common and potentially damaging mistakes made by novice qualitative researchers is to frame their research question in a manner requiring a type of 'answer' that qualitative research cannot provide. This includes questions that ask about simple causal relationships. To give a real example, a student approached one of us for supervision, saying that they wanted to use qualitative interviews to find out 'What causes young women to develop eating disorders'. If you think this is a legitimate kind of question to ask (and see Chapter 2 for discussion of the problems many qualitative researchers have with conventional notions of causality), you need to address it using quantitative methods within the hypothetico-deductive tradition. You could, for instance, carry out an analysis of epidemiological data, or utilise a survey design, but qualitative interviews would never enable you to answer a question like this. There are qualitative approaches that are interested in explanatory questions, particularly within the limited realist tradition; for example, work drawing on the realist evaluation tradition (Pawson and Tilley, 1997) that seeks to identify how specific contexts and mechanisms produce particular outcomes. However, this is a much more nuanced and interpretive enterprise than the direct causal question we highlight above.
Another mistake that can be made in the type of the research question is to seek to establish general trends in the phenomenon under consideration. To extend the previous example, you might ask 'Are women more strongly influenced than men by media representations of body image?'. While this question is not seeking to uncover underlying causes of behaviour, it is trying to produce a highly generalised understanding of the differences between two very wide categories of person ('young women' and 'young men'). Qualitative researchers differ in the extent to which they permit any attempt to generalise, or 'transfer', understanding from a specific study to a wider context (Murphy et al., 1998; Williams, 2002), but even those who argue for some degree of transferability would not see this as an appropriate question for a qualitative study.
The research question for a qualitative interview study should not, therefore, fnmc r*n pctoKlioUinft -~1~*:---u:— — -------1: —1 ----' - r
Interviews in Qualitative Research
behaviour. What it should focus on is meaning and experience, with reference to a particular group of participants. So, for example, you might ask 'How do young women view the presentation of body image ideals in magazines and newspapers?'. And while qualitative research questions should not seek to establish causality, they may very well focus on perceptions of causality from the perspective of research participants: 'How do people diagnosed with anorexia make sense of why they have developed the condition?'
Scope of the research question
Even when a qualitative research question has the right kind of focus, it may still be inappropriate in terms of its scope. Questions that are very broad in scope are problematic because of the emphasis in qualitative research on understanding people's lives in context. If a study tries to encompass experiences from too wide a set of social contexts, the findings are likely to present a scattering of unrelated snapshots, from which it is impossible to draw any kind of conclusion. The revised research question on body image, stated in the previous paragraph, would be likely to suffer this weakness, as the category 'young women' is almost certainly too broad. The researcher here would be best advised to narrow the scope somewhat, perhaps in terms of characteristics such as specific age groups, class, occupation, and so on.
While research questions that are too broad may in effect prove 'unanswerable', those that are too narrow are likely to produce findings that are simply not very interesting or useful. Qualitative research is interested in how people differ in relation to a particular phenomenon, as much as in what they have in common. A very narrow research question can result in a highly homogeneous sample that does not enable diversity of meaning and experience to be revealed. Also, such a question may generate findings so localised in their relevance that they cannot contribute to the intellectual debate around the topic in question.
When deciding on the scope of your research question, a key factor to bear in mind is the level of resources available to you. On the whole, broader questions will require larger-scale studies to address them effectively. Researchers who are new to qualitative approaches may be prone to over-reaching themselves in terms of scope, feeling uncertain about the value of narrower questions. If this describes your situation, remind yourself that qualitative research is fundamentally concerned with the particular rather than the general; on that basis we would advise that if in doubt, err on the side of narrowing the scope of your research question.
Avoiding presuppositions
Texts on qualitative interviewing stress the importance of avoiding leading questions in interviews, and offer various tips as to how this may be done. The present volume is no exception, as you will see in Chapter 5. However, it is
Designing an Interview Study
important to note that it is possible for the research question in itself to be leading, such that it may blinker the way in which you go about exploring your topic with your participants. Take the question 'What are the perceived benefits to the victims of street crime of a self-help website?'. The question seems suitably focused on meaning and experience (here in the form of 'perceptions') and realistic in scope. But there is an in-built presupposition that the website does have benefits for victims, which may lead you to neglect probing properly for any negative experiences associated with its use. A better form of phrasing would be to refer to 'perceived benefits and costs', but even this may tend to encourage you to seek a clear dichotomy of good and bad that may not correspond to the way participants see things. Better still would be 'What experiences do users have of a self-help website for victims of street crime?'.
The shifting research question
In qualitative studies it is not uncommon for the researcher to feel that the research question is shifting as the study progresses. While this would be a cause for great concern in a positivistic quantitative study, it is not necessarily a problem in a qualitative one. Qualitative research always has (to some degree) an exploratory character, and as such it is inevitable that sometimes a project will move in directions that are of relevance to the research topic but outside of the scope óf the original research question(s). If we return to the study of street crime victims, the researcher might find that participants consistently want to comment on their experiences of a helpline accompanying the website rather than just on the website itself. In deciding whether to allow this kind of redefinition of the research question, any researcher would need to consider a number of conceptual and practical issues for their project. These include the following:
• 'Would the change to the research question undermine the coherence of the study as a whole? In our example, we might feel that to examine the helpline as well would not substantially alter the underlying concern of the study with experiences of using self-help resources for victims. In contrast, we would be reluctant to extend the study to look in detail at victims' experiences of individual psychotherapy, as this represents a very different kind of resource from the website and helpline.
Would the change stretch the resources of the project to an unmanageable degree? In our example, incorporating a detailed exploration of responses to helpline use might extend the interviews by 20 minutes or so. This will have a knock-on effect on the time taken to transcribe and analyse interviews that the researcher would need to take into account.
Are key stakeholders in the project happy with the change? Significant decisions about changes to a research project are rarely just the concern °f the individual researcher. In a master's or doctoral thesis, the student's
56 )   Interviews in Qualitative Research
Designing an Interview Study
©
supervisor will want to be sure that any change does not undermine th \ intellectual quality of the work, and that it will not result in unacceptabf delays to completion. In externally commissioned research, funders 3re ' also likely to be concerned with completion times, and there may be p0ljtj cal or ethical considerations as well.
Choosing the type of interview to use
In some cases it is apparent from the start that a particular form of qualitative interviewing is the most appropriate, because of the nature of the topic to be studied and/or the requirements of the methodological and theoretical stance to be taken. For instance, if you wanted to follow a life-story approach within the narrative tradition (see Chapter 13), you would of necessity use individual interviews. Very often, though, there are several types of interview that could be employed. You may, for instance, weigh the pros and cons of individual and group interviews. You may consider whether it is essential that you use face-to-face interviews, or whether telephone or internet interviews offer a viable alternative (see Chapter 7). You may want to incorporate visual methods into your interview process (see Chapter 8). Subsequent chapters of this book will, we hope, tell you enough about these different forms of interview to enable you to make an informed choice for your research. At this point we simply want to urge you to bear in mind that when designing your study, you think about the different ways that qualitative interviews can be conducted, rather than automatically taking the 'default' option of the individual face-to-face format.
Sampling and recruitment
Defining your sample
In quantitative studies, and especially surveys, recruiting a sample that is statistically representative of the population to be studied is of central importance, because of the need to establish the generalisability of the conclusions drawn from research. For example, if a researcher wanted to test the hypothesis that attitudes to risk-taking in men's driving behaviour were associated with attitudes towards their own masculinity, he would require a sample that was representative of the male driving population as a whole. Qualitative research, in contrast, does not seek to make this kind of generalisation and therefore does not normally use sampling strategies aimed at producing statistical representativeness. However, as we saw in Chapter 2, qualitative research very often is concerned to achieve different forms of generahsability or transferability. As Mason (1996) and May (2002) point out, this means that a purely! ad hoc, opportunistic sampling strategy is not appropriate; rather, the sample needs to relate in some systematic manner to the social world and phenomena;, that a study seeks to throw light upon.
riterion most commonly proposed for sampling in qualitative studies The c f^gsearchers seek to recruit participants who represent a variety of ' ^^'Tn relation to the research topic, of a kind that might be expected to ^f^ht on meaningful differences in experience. To continue the example a*the previous paragraph, if we wanted to carry out a qualitative study of
* en perceive risk-taking in driving, we might consider that age, years of
• ^experience, and family status may be important, and therefore seek to vin^ arQcipants who vary on these aspects. (This kind of targeted sampling
^fT  referred to as 'purposive'.) Of course, the effectiveness of such a sam-' ° strategy will depend on the choice of aspects (dimensions or categories) which to select participants. This choice will in most cases draw upon a or"    0f the researcher's knowledge of the academic literature, personal j^ovvled^e, and anecdotal information from those who have some involve-ment with the topic.
The philosophical position of a study is a crucial influence on sampling strategy> as>s tne methodological approach taken. Qualitative neo-positivist and limited realist positions (see Table 2.1 in Chapter 2) are concerned to ensure that interpretations correspond to a reality that exists outside of the research process and the researcher's position; as such, they tend to require larger samples than contextualist and especially radical constructionist studies, and often need to define a range.of relevant 'stakeholder' groups from whom participants will need to be drawn. In contrast, a phenomenological study will not only require a relatively small sample (often in single figures or the low teens) in order to enable sufficient depth of analysis, but also one more homogeneous in terms of participants' experiences of a common phenomenon. Even within this approach, though, there is variation, with interpretative phenomenological analysis favouring highly homogeneous samples (Smith et al., 2009) while descriptive phenomenological approaches argue for a higher level of diversity in the sample to enable 'essential' aspects of the phenomenon to be robusdy identified (Giorgi and Giorgi, 2008). Robinson (2014) provides a useful discussion of how both theoretical and practical issues should shape these kinds of sampling decisions.
It is important to bear in mind the real-world constraints on sampling arising from the scale of most qualitative interview studies. If your planned master's dissertation research is likely to involve around 20 interviews, it would clearly be unwise to try to select a sample on the basis of differences in ten different aspects. It is generally best to 'fix' one or two key aspects that define the group you are looking at, and then seek diversity in other aspects. Gerson and Horowitz (2002: 205) argue that:
By choosing a sample that controls for one consequential aspect of lived experience (e.g. age or generation), but varies on others deemed important in the theoretical literature (e.g. gender, race, class), the aim is to discover how similar social changes are experienced by different social groups.
Interviews in Qualitative Research
In qualitative research, sampling and recruiting participants may occur several stages in the course of a project. Thus an initial sample may ^ recruited and interviewed, and on the basis of preliminary analysis of thejt data, a further sample defined to address particular emerging issues. Thi kind of strategy is probably best known in grounded theory (e.g. Corbin and Strauss, 2015), in the form of 'theoretical sampling', although it may be used in other approaches too. In a study by the first author and Anne Little, w examined the users' experiences of a community gym in a highly deprived predominantly South Asian area of a large northern English town. For our initial interviews, we recruited a sample that varied in age and gender; our initial analysis suggested that a particular subgroup - women over 50 - were especially interesting in relation to the issues our funders were concerned with. We therefore recruited an additional sample just from this group (King and Little, 2017).
Recruiting participants
In this section we will look at some of the challenges that arise in the process of recruiting participants once you have defined your sample. We will consider how you may go about gaining access to potential participants, and the kind of information you need to provide for them, in order that they may make a decision about participation. Inevitably, this discussion will raise ethical issues, such as the need to avoid coercion and to ensure proper informed consent. Some of the practical consequences of these matters will be considered here, but a much fuller examination of the ethics of qualitative interviewing was presented in Chapter 3.
Gaining access
The precise nature of the tasks involved in gaining access to participants can vary enormously from study to study. In some, the main challenge may be that the kind of experience you are interested in is a very uncommon one - such as winning a major lottery prize. In others, you may face the difficulty that your topic is a painful or emotive one, which people may be reluctant to talk to a stranger about - such as the experience of sexual assault within marriage. Alternatively, access may be problematic because it requires the approval of several gatekeepers in a large and complex organisation, perhaps with political sensitivities to contend with too. The British National Health Service is a classic example of this kind of setting. These are only a few of the more common challenges regarding access that you may face. It is impossible to offer advice for every eventuality, so we will concentrate here on a set of issues that in our experience are quite commonly encountered in relation to gaining access for a piece of qualitative research: working with gatekeepers, using insiders to assist with recruitment, and advertising for participants.
Designing an Interview Study  ( 59
Working with gatekeepers
research studies, potential participants must be reached through one In man)'     peepers. We are defining a 'gatekeeper' here as someone who has 7 ^tv to grant or deny permission to access potential participants, and/ lC ulirv to facilitate such access. Examples could include health profes-
. rhe ability to
or Ui j-or access to patients, semor managers for access to their employees, Si0'!f 5 achers for access to schoolchildren, and so on. Sometimes different k| will play different gatekeeping roles for your project. You may need to ^C°^>- the overall permission for access from the managing director of a acquire but have to negotiate the details of recruitment with one or more <fT1 omental head(s). Sometimes, while it may not be obligatory to go through
department
both in terms of identify-
tekeeper, tnere may be advantages in so doing ■ f   and facilitating contact with participants, and reassuring them as to your ^edibility and trustworthiness.
The first step in working with a gatekeeper to recruit participants is to be sure you have actually identified the appropriate person. This may sound self-■ident, but it is possible to go awry by assuming, for instance, that because $omeone holds a particular position in an organisation,1 they are automatically the correct person to help you with access. You need to be sure you have a good working knowledge about how the organisation is structured and how it functions before you begin recruitment. In planning your research, you should budget for time to explore how the organisation works, especially where it is relatively urrramiliar to you.
You will normally need to provide gatekeepers with a range of information to enable them to grant you permission for access. They will require an overview of the project, summarising its aims, methods, anticipated outcomes, and clearly stating the time commitments required from participants. While you may provide this verbally, it is best to also have a written version available, using language that avoids jargon as far as possible. You should provide them with copies of any written information that is going to be given to potential participants, including the consent form (where one is to be used). Gatekeepers need to know the level of anonymity that is being promised - for instance, is the organisation as a whole to be anonymised, and will particular subsections of it be identifiable? It is important that they recognise your ethical obligations to individual participants, above all the need to maintain participants' confidentiality in relation to other members of the organisation (including the gatekeeper).
'Insider' assistance with recruitment
In some studies, researchers not only gain access through a gatekeeper within the organisation, but also use one or more insiders to actively assist in recruiting participants. Such a person might identify organisational members who meet the sampling criteria of the study, pass project information sheets and
Interviews in Qualitative Research
letters requesting participation to them, and forward queries to the research team. This kind of insider assistance can have real advantages. "Where atl organisation is located at a considerable distance from the researcher (0t research team) it may not be possible within time and budgetary constraints to visit in person to deal with the nitty-gritty of recruitment. In these circus, stances, you are much more likely to be successful with recruitment if atl insider is actively helping you, than if you merely send information and requests to individual members. Also, if the request is coming through a known and trusted colleague, people are more likely to give it proper consideration than if it had arrived from a stranger - where it might be seen as jas» another form of junk mail (this may be especially likely to happen with 'cold' requests received via e-mail).
Alongside these advantages there are some significant risks associated with using insiders to help with recruitment. It is possible they may be overtly biased, consciously choosing participants likely to hold (or not hold) certain views. Perhaps a greater threat is that there will be an unintentional distortion stemming from reliance on personal networks within the organisation. There is also an ethical danger that insiders may exert pressure on people to participate that would deny them genuine free informed consent. In balancing these risks and advantages the following guidelines may help you to determine whether and how to use insider assistance with recruitment:
• Select insider assistant(s) carefully - consider whether they may have axes to grind on the topic, and whether they are likely to be seen as trustworthy by those you seek to recruit. This generally means spending some time getting to know potential assistants before you ask them for their help.
• Make sure insider assistants are briefed thoroughly about the study.
e Keep in regular contact during the recruitment process. This will allow you to discuss how best to deal with any problems in recruitment as they occur, and to determine whether there is anything you can do to facilitate the process - for example, providing clarification or additional information to potential participants.
• When potential participants have received their invitation to take part in the study, where possible they should respond directly to the research team, not via the insider assistant. This will enable their participation to be kept confidential if they so wish.
Box 4.1 provides an example of the process of gaining access through organisational gatekeepers and insider assistants.
Designing an Interview Study
Box 4.1   Accessing participants through gatekeepers
. led by the first author, we examined experiences of allotment gardening among nt plot-holders (sometimes known as 'allotmenteers') on three sites in a northern 'fh town (King, 2012). The sites had received development support from a local charity, • h employe an a"otment development worker (Hassan). We sought to recruit a 1,0 sectjon of users in terms of age, gender and gardening experience. The process of
C""S frig access is sflown in ttie f'ow cfiart be'0W:
HEALTHY LIVING PARTNERSHIP MANAGER Overall permission for study
ALLOTMENT DEVELOPMENT WORKER
Access to allotment site secretaries (one plot-holder at each site responsible for administration and liaison with local authority that owns the site)
ALLOTMENT SITE SECRETARIES
Provided names and information about potential participants, in the light of our selection criteria, including in some cases personal introductions for the researchers
Note that in relation to the allotment site secretaries, the roles of 'gatekeeper' and 'insider assistant' are blurred. It would-have been possible to recruit participants without going through the secretaries, and it was not an ethical, legal or administrative necessity to do so. However, the secretaries had not only the most up-to-date knowledge about who held plots and was actively working them, but also personal acquaintance with the other allotmenteers on their site. This meant they could give considerable help in suggesting people to approach in order to collect the kind of diverse sample we sought. Furthermore, they were also participants themselves and were interviewed early in the project, and thus able to reassure others about the research process.
Two main issues arose from using the site secretaries in the recruitment process. Firstly, it was important that they did not exert any pressure on others to take part. We did not feel that this was a problem in reality, because their role did not place them in any substantial position of power over other allotmenteers. (It should be noted that some allotment sites do operate in quite a hierarchical manner, but these three were strongly egalitarian in ethos.) Secondly, there was a danger that site secretaries might select those to approach who they expected to take Particular positions. We can be less certain that this did not occur, although by emphasising the need for a diverse sample we hoped that such bias might be minimised. Certainly the interviews themselves did not suggest participants taking any kind of 'party line'.
62 )   Interviews in Qualitative Research
Designing an Interview Study   ( 63
Snowball sampling
In snowball sampling, the researcher uses the initial few interviewees (often recruited opportunistically) to recommend other potential participants who fit the inclusion criteria for the study. They in turn will be asked to suggest further contacts, and so the sample builds up. This strategy inevitably introduces a form of bias into the sample. Participants may, for example, tend to recommend people who share their view of the phenomenon under investigation. If one is concerned with some degree of generahsability or transferability, then snowballing would not be the preferred option (Gerson and Horowitz, 2002) Certainly, where snowballing is used simply as a convenient way of recruiting participants, it is really no more than a form of convenience sampling.
There are circumstances, however, where snowballing may be an appropriate strategy - namely where the population to be sampled from is especially hard to access and quite tightly defined. Langdridge (2004) and Howitt and Cramer (2005) give examples such as drug addicts, street gangs, and even banjo players! In Turley's research investigating the lifeworlds of bondage/ sado-masochism practitioners she used snowball sampling as a way in to a relatively self-enclosed community (Turley et al., 2017). Where snowballing needs to be used with these kinds of groups, some of the biasing effects may be reduced by giving very clear instructions about the characteristics sought. For instance, a researcher may ask a participant to recommend others of a particular age, gender or with other specified characteristics in order to enhance the diversity of the sample.
Advertising for participants
An alternative to contacting potential participants through organisational gatekeepers or insider assistants is to use some form of public advertising. This could include notices in public places (universities, doctors' surgeries, community centres, etc.), adverts in newspapers or magazines, on social media, or messages to internet discussion groups and other online communities. There are a number of issues to bear in mind if you are considering using such strategies.
• Be sure to get proper permission before you put up notices in a public place. This includes permission to use an online discussion group for recruitment purposes.
• Effectively target the population from which you wish to recruit. If you were looking for sufferers from a rare chronic illness, it would not be very efficient just to plaster the corridors of your university with posters. In such a case, finding a relevant internet discussion group would be more sensible, or alternatively identifying a self-help group which might be willing to distribute flyers to its members.
• Think carefully about the design of your advert. You need to provide enough information to let people know what the research is ahnnf who
u are, and what kind of commitment is required of them, but not so y° cri ^at they might feel overwhelmed or intimidated. You need to' use lay language, without being condescending. For material such as posters or flyers, you must think about the visual impact - font size and style, the use of colour and graphics.
When recruiting through public advertisement, you will be accessing people who in most cases will be complete strangers to you, and will not even have come to you through someone else's recommendation (unlike all the other methods described above). You therefore need to be especially mindful of your own safety when meeting them. We discuss the issue of researcher safety in more detail in Chapter 3.
Methodologically, the major pitfall of recruiting participants through advertising is that the sample is highly self-selecting. This may mean that you do not get the balance of participants that you would ideally have liked. One way to mitigate this effect is to use a purposive sampling frame to select the sample you need from all those who respond positively. For example, you might have 20 people coming forward in response to the advert, of whom 15 are female and five male. If your sampling frame stipulated up to five participants of each gender you would pick all the men and five of the women. This of course relies on you getting substantially more positive responses than you need. If you do use this strategy, you should reply to those you are not going to interview, thanking them for making the effort to contact you and briefly explaining why you could not include all respondents. This process does reduce some of the biases that may be inherent in a volunteer sample; however, there may still be the difficulty that some subsets of the population are particularly unlikely to respond and therefore will not be represented.
Developing an interview guide
Flexibility is a key requirement of qualitative interviewing. The interviewer must be able to respond to issues that emerge in the course of the interview, in order to explore the perspective of the participant on the topics under investigation. This means that the traditional interview schedule used in quantitative survey research - with fixed questions in a predetermined order - is inappropriate. Instead quautative interviews use an 'interview guide' that outlines the main topics the researcher would like to cover, but is flexible regarding the phrasing of questions and the order in which they are asked, and allows the participant to lead the interaction in unanticipated directions. The precise format of interview guides varies enormously, reflecting the needs of different methodological traditions in qualitative research as well as the personal preferences of individual researchers. Below we address some of the key questions you will need to consider when deciding on the form and content of your own interview auide.
0
Interviews in Qualitative Research
Designing an Interview Study (^T)
Upon what do I base my guide?
We would suggest that there are three main sources you can draw on to identify topics to include in your guide. Firstly, you can think about your own personal experience of the research area - both first-hand experience and stories and anecdotes told by people you know. For example, if you are studying responses to the experience of being burgled and have yourself been a victim of this crime you could reflect upon your own responses. You could also think about accounts you have heard from family, friends and colleagues of being burgled Secondly, you may consult the. research literature on the subject, to see what previous research suggests. (Note, though, that there are some methodologies that discourage you from reviewing the literature before doing your interviews to minimise presuppositions that might distort your research - grounded theory (Corbin and Strauss, 2015; Charmaz, 2014) is the best-known example of this.) Thirdly, you may carry out some informal preliminary work to focus yout thinking about the area. Thus, in the research the first author carried out on the psychological and social impact of allotment gardening, he spoke to several colleagues and family members about their experiences as well as to some of those involved in initiating the allotment development scheme that was being evaluated. This last point highhghts the necessity to bear in mind the requirements of external funders when deciding upon the scope of your interview guide.
How comprehensive should I be in covering topics relevant to my research area?
When you have drawn on the sources noted above, you will probably have an extensive list of topics that you could cover in your guide. It is important to now think about whether you want to comprehensively cover most or all of the potentially relevant topics, or whether you are going to be more selective - perhaps choosing three or four broad areas that you are sure you will want to address in the interview. (Of course, you can choose to adopt a strategy anywhere between these two extremes.) This is really a question about the extent to which you as a researcher want (or need) to lead the direction of the interview. If you have a very comprehensive interview guide, there will be a danger that you do not allow sufficient opportunity for the participants to bring up perspectives that may be unanticipated but actually of real interest to your research. If you go for a minimalistic interview guide, you may fail to address important issues, should the participant lead you into lengthy digressions from your research focus. In deciding how comprehensive to make your interview guide, you must reflect on the aims of your study and your methodological position. A realist evaluation study, for example, is likely to use quite a comprehensive guide, in order to ensure that key aspects of the programme or intervention being evaluated are covered. A narrative study, seeking to elicit life stories from a particular group of participants, would use a much more minimalistic guide.
What types of question should I ask?
(2015) argues that there are six types of question that can be asked in P^0". Jtjve reSearch interview, each seeking to elicit a particular kind of a 1ua 1 -on from the participant. In reality, these categories may not always ^"distinct as he describes, and we would certainly not suggest that you take 35 han»stic approach to including a certain number of each type in every
• r"eC uide- Nevertheless, they can be a helpful way to think about the differ-^'kinds of response you need within the course of an interview and about JoW you can best facilitate them.
Background/demographic questions. These are straightforward descriptive questions about personal characteristics of the participant that you might need to be aware of in your analysis - for instance, about their age, gender and occupation. It sometimes makes sense to collect at least some of this information on a simple form at the same time as obtaining written consent.
' • Experience/behaviour questions. These questions focus on specific and overt actions that you could have observed were you yourself present at the time. Examples could be: 'What did you do when the doctor told you the diagnosis?'; 'What happened when you cautioned the suspect?'; 'What did you and the other candidates do while you were waiting for your interviews?'.
• Opinion/values questions. These are questions that ask what the participant thinks about the topic at hand, and/or how their thoughts relate to their values, goals and intentions. They might ask such things as: 'What did you hope to achieve by doing that?'; 'What do you think is the best way to deal with that kind of situation?'.
• Feeling questions. These questions focus on participants' emotional experiences. Patron warns that they often get confused with the previous category, because we habitually use the question 'how did you feel about ...?'inavery loose way that can mean 'what is your opinion about
as well as 'how did you respond emotionally to If you particularly want to explore emotional responses, you must phrase your question in a way that makes this clear to the interviewee. 'What feelings did this provoke in you?' would be a better formulation than 'How did you feel about it?'.
• Knowledge questions. This category relates to questions about factual information the participant holds. The distinction between knowledge and opinions/values is a difficult one; the important thing to remember is that you are concerned here with what the participant believes to be a 'fact' and not with whether it is actually true in any objective sense. Knowledge questions might be formulated in such terms as 'What do you know about the systems for referring students for pastoral support?' or 'How well did you know this patient's history?'.
66 )   Interviews in Qualitative Research
Sensory questions. These axe questions about sensory aspects of experience - what the participant saw, heard, touched, tasted or smelled in any given situation. While conceptually they might be seen as a subset of experience questions, they tend to be quite distinctive in form, asking the participant to recollect a very specific sensory impression in a specific setting. They are particularly important in studies where embodied aspects of experience are of central importance (see Chapter 11 for further discussion of embodiment in interview studies).
Can I change my interview guide in the course of my study?
Not only is it permissible to change your guide in the course of your study, it is generally advisable. Remember, the aim of a qualitative interview is to elicit participants' accounts of aspects of their experience, rather than to collate answers to specific questions as if they were variables in a survey. As such, any insights you gain in the process of carrying out your first few interviews should inform subsequent ones; for instance, you may note a probe question that worked particularly well, or recognise that an aspect of participants' lives you had overlooked initially may be important to the phenomenon you are studying. So long as you remain aware of the way your interviewing practice developed over the course of the project, you should be able to avoid such changes distorting the analysis of the data.
How should I format the questions or topic areas on the guide?
Interview guides vary in how they lay out the questions to be asked. One style is to formulate full questions, written in proper sentence form. The opposite approach is to just include short phrases or single words as reminders of the topics to try to cover - perhaps in bullet points or similar. Box 4.2 shows two different styles of interview guide that could be used for the same study.
There are pros and cons to either style of interview guide. The advantage of using full questions is that it forces the researcher to think carefully about question formulation - to avoid the kinds of leading question, endorsement of participant opinions and so on that can happen if the interview drifts into a style that is too conversational (Willig, 2008). The disadvantage is that with full questions.stated on the guide, the interviewer may tend not to use it as flexibly as they should. The pros and cons of the topic heading format are, of course, the mirror image of those just stated: it encourages flexibility but does not help guard against inappropriate phrasing of questions.
Designing an Interview Study (67^)
Box 4-2   Examples of different styles of interview guide
jS an extract from the community gym study referred to earlier in this chapter
Blowand Little-, 2017). It is presented in two different formats: 'full questions' and (King
•Key Points ■
[xtract from Interview Guide for Community Gym Users - Full-Question Format 1 starting at the Community Gym How did you first hear about the Community Gym? When did you start attending? Why did you start attending the Community Gym? Had you used a gym before? Do you go to any other gyms now? If 'yes': Which? How does it compare to the CG? If 'no': Why not? Do you take part in any other regular sport or exercise now? If 'yes': What do you do? When did you start this? If 'no': Why not? 2. Overview
What (if anything) has been the best thing about attending the Community Gym for you?
What (if any) changes or improvements would you like to see made to the Community Gym?
Would you recommend the Community Gym to a friend? If 'yes': Why? If'no': Why not?
If the Community Gym closed, what (if anything) else would you do instead?
(Continued)
68)   Interviews in Qualitative Research
Designing an Interview Study   ( 69
(Continued)
Extract from Interview Guide for Community Gym Users - Key-Point Format 1. Starting at the Community Gym
• How heard about it?
• When?
• Why?
• Used a gym before?
• Other gyms now?
If tyes'.-Which? Compare to CG? If W: Why not?
• Other regular sport/exercise? If 'yes': What? When started? If 'no': Why not?
2. Overview
Best things about attending CG? Desired changes or improvements? Recommend to a friend? If Ves': Why? If'no'.-Why not? Alternatives, if CG closed?
As can be seen, the 'key point' format is much more succinct; it would be very easy to glance at momentarily to remind yourself what you want to cover during an interview. However, there are several places where having the full question before you might help you to avoid inappropriate formulation of questions. For instance, in the 'Overview' section, the first question asks 'What (if anything) has been the best thing about attending the Community Gym for you?' The parentheses here seek to remind the interviewer not to take it for granted that the participant thought there was any 'best thing' about the gym. A similar formulation is used for other questions in this section, for the same reason-, to avoid implicit leading of the participant. Using the 'key points' format, an interviewer -especially an inexperienced one - might fail to recognise this danger.
Two factors may be useful to consider when choosing how to format the gmde Firstly, the experience of the interviewer is relevant. On the whole we would recommend that relatively inexperienced qualitative interviewers opt for the full-question format, as the skill of phrasing questions appropriately in
ch you are using may influence your choice. As we noted above in rela-JfPr ^ comprehensiveness of the guide, a key issue is the extent to which n°n(I)eChodology requires the researcher to take a directive role in the inter-^e Those approaches that seek to minimise interviewer directiveness - such "'most narrative approaches - may be better served by the topic heading style a$ -jg We would suggest that whatever the format of the questions, it can ° useful to take two versions of the guide to interviews. To minimise distrac-^ "from focus on the interaction, it is helpful if a very brief version (ideally single page if that is possible) is used - perhaps just noting topics with bullet nts. Alongside this, the researcher can have a fuller version of the guide with them, so that should they find themself seriously losing their way (it happens occasionally, even to the most experienced of us) they can reorient themself.
What are probes and prompts and how do I include them in my interview guide?
Sometimes you may see these two terms used interchangeably, but we feel it is useful to make a distinction here. We take probes to be follow-up questions that encourage a participant to expand on an initial answer, in order to obtain more depth in their response. In contrast, we take prompts to be interventions that seek to clarify for the interviewee the kind of information a question is seeking to gather, usually used where the interviewee has expressed uncertainty or incomprehension about the initial question.
You will often need to formulate probes and prompts in the course of an interview, as you seek to obtain the fullest account possible from an individual interviewee. However, it is often reasonable to anticipate that certain probes or prompts are likely to be needed at specific points in an interview, and in such cases it makes sense to include them in the interview guide. This may be done at the outset, when designing the guide, and probes and prompts may also be added in the light of experience as the study progresses. For example, returning to the topic guide in Box 4.2, we may anticipate the need for probes and prompts to follow up the question asking why the participant started attending the community gym. For instance, we may want to be sure that they reflect on a variety of possible reasons for joining, and find out whether there had been a particular 'trigger' to their decision. However, we may be concerned that the notion of a 'trigger' may need some clarification. This part of the topic guide might therefore be laid out as below:
Why did you start attending the Community Gym?
Probe: Health, fitness, social reasons?
70 )   Interviews in Qualitative Research
(Prompt: e.g. following advice from a health professional, influence of friends or family, an incident that made you concerned about your fitness or health?)
Care must be taken in the way probes and prompts are formulated, so as to avoid them 'leading' participants such that they feel a particular answer is required. We will discuss this issue in detail in Chapter 5.
Conclusion
The success of a qualitative interview study is not just based on how well you ask questions and how skilfully you analyse the data. Rather, the decisions that you make at the very start of the research process, when you are designing your study, can have a major impact on its outcomes. In this chapter we have provided advice on each of the main steps on the journey from framing your research question to deciding on the areas to include in your interview guide. We would stress again that doing all this effectively requires you both to be aware of your theoretical position and its underlying philosophical assumptions and to be pragmatic about what you can achieve within the resources available to you.
Recommended reading
Flick, U. (2014) Designing qualitative research. In An Introduction to Qualitative Research, 5th edn. London: Sage.
A good overview of issues to consider when choosing a qualitative research design.
Patton, M.Q. (2015) Qualitative Research and Evaluation Methods: Integrating Theory and Practice, 4th edn. Thousand Oaks, CA: Sage.
An accessible and highly practical overview of all the main stages of designing and carrying out qualitative projects. Strong applied focus.
Note
Note that we use the word 'organisation' in a very wide sense, to include small, informal groupings such as self-help groups, clubs and interest groups, as well as larger, more formal bodies.
Carrying Out Qualitative Interviews
Producing a good research project based on qualitative interviewing relies on much more than just 'interviewing skills', as we stressed in the previous chapter. Nevertheless, however thoroughly you understand your methodological position, and however well you design your study, what happens in the interviews themselves is crucial. The present chapter therefore focuses on the process of actually carrying out a qualitative interview providing guidance on the following areas:
• the interview setting
• building rapport
• how (not) to ask questions
• starting and finishing interviews
• managing 'difficult' interviews
Since we cover group interviews, remote interviews (telephone and via the internet) and interviews incorporating visual methods elsewhere (Chapters 6, 7 and 8, respectively), this chapter will focus on individual facerto-face interviews.
19
0
Interviews in Qualitative Research
Carrying Out Qualitative Interviews
The interview setting
The physical space in which an interview is located can have a strong inflUe on how it proceeds. Three aspects of the physical environment are especia(] ] important: comfort, privacy and quiet. We will look at how you might achie *! each of these in turn, and then in the light of this consider the pros and co^' of the types of location you are most likely to use for your interviews.
The physical environment
The first requirement of the interview environment is that it should be as com-fortable as possible for the interviewee and for you. This means not only physical comfort but also (perhaps more significantly) psychological comfort ' If participants feel tense and unsettled, it may be reflected in stilted and under-' developed answers to your questions. Avoid sitting so that you face the interviewee over a desk or table; this can feel too formal and may remind the ' participant of a job interview - not an association you want to raise for therft. Sitting around the corner of a desk is a better option, or you may be able to arrange two easy chairs with perhaps a small table between them - again at an angle that encourages informality. Ensure you are close enough to hear them clearly and read non-verbal communication, but not so close that you are intruding on their personal space.
Privacy for the interview setting is vital. You want to avoid as far as possible any danger of being overlooked or interrupted during the interview. If you are using a room in the participant's or your own organisation, you should try to have phones switched off or diverted, and a note on the door asking people not to disturb you. Also, if you are booking a room, make sure you book it for longer than you anticipate you will need for the interview, to allow for introductions and briefing, and breaks the participant might want to take, and so on. We would suggest that you allow at least an hour and a half for what you would expect to be a 45-60-minute interview.
In some circumstances, participants may not want others to know that they are taking part in your research. Perhaps they may fear censure from their organisation, their family or others, because of the nature of the interview topic. Where you anticipate that this could be the case, you should have an option available to locate the interview where there is minimal danger of their participation being discovered by those they wish to conceal it from. This may often mean holding it on 'your' territory, such as a room in your university.
Try to use a location that is relatively quiet. This will not only help make the environment relaxing, but will also reduce the likelihood of problems with the audibility of recordings. Human attentional capacity means that we can quite easily filter out background noise when Ustening to someone talking in close proximity to us. Recording equipment does not have the same ability -indeed, background noises often sound surprisingly loud on tape. We have all
where recordings have been rendered almost inaudible by such PDSta"rafnc noise (even through a closed window), office machinery arid
* nee birdsong! Even where you can avoid background noise, you nsider the acoustics of the room you plan to use. Large, high-
° rooms - especially if fairly empty - can produce an echo to the voice riously impedes the quality of recording.
Choosing the interview location
erally good practice to ask participants where they would like the jew to be held, and more often than not they will select somewhere on
• * territory - such as their workplace or home. This of course limits your *^wv to arrange the physical space in an optimal manner, but you can at least f   'anticipants know what is required of the space in terms of comfort, pri-
and quiet. If you are carrying out a series of interviews in an organisation ov^r one or more day(s) it may be possible to negotiate access to a room that u can have control over for the duration of the data collection, and thus have more opportunity to arrange it as you wish. Noble et al. (2015) used this Strategy very effectively in a study of a specialist community palliative care service that required more than 50 interviews with a wide range of patient, carer and professional stakeholders.
Where you are carrying out interviews in participants' homes, the presence of other family members or friends can create difficulties. You need to think in advance about whether you. are happy to allow another person to sit in on the interview, as it is not uncommon for a family member to request this (and occasionally to insist on it). If the presence of another person is completely unacceptable for your research, you need to make this very clear to participants in advance, and reiterate it at the start of the interview. Equally, if you would be willing to allow another person to be present, you need to explain to them the level of participation permissible. In some projects it may be a positive benefit to have a joint interview, in others it may be vital that the second person contributes minimally to the interaction. When interviewing in participants' homes, your own safety is something you must think seriously about and discuss with supervisors or colleagues (we have addressed this issue in some depth in Chapter 3).
Sometimes participants may prefer to be interviewed away from their home or workplace - perhaps because of concerns about privacy as noted above. In these cases, your best choice is likely to be to find a suitable room in your university or other workplace, enabling you to organise the space yourself to suit the requirements of the interview situation. Another option could be a public space (indoors or outdoors). Public spaces can have the advantage of being comfortable and relaxing, and their neutrality may be an advantage. With indoor spaces such as cafes the possibility of being overheard needs to be borne in mind, and you should let owners or managers know what you are to avoid suspicion. A park or similar outdoor space can also be a
©
Interviews in Qualitative Research
Carrying Out Qualitative Interviews
0
comfortable environment and one with less danger of being overheard n. I major drawback is likely to be a practical one - it can be hard to obtain a g I recording of a conversation in the open. At the very least, you need to try °0<i \ your equipment in situ before agreeing to use a particular outdoor location you live in a climate such as ours in the North of England, the weather is a| a factor you will have to consider.)
Recording
In most qualitative research traditions it is strongly preferable, if not essential to have a full record of each interview. Usually, this means utilising some form of audio-recording, although in a small proportion of studies video-recording is used. Keeping a detailed hand-written record can be necessary - for instance when a participant refuses consent to be recorded or where there is a technical failure with equipment. In this section we will examine how to use each of these recording media optimally.
Before we consider the practicalities of recording, however, it is worth considering how the presence of an audio- or video-recorder may impact on the interview process. As Warren (2002) states, recording equipment inevitably has meaning for the interviewee; furthermore, it is likely to have different meanings for different people. For example, to a young offender it may be a reminder of procedures in the criminal justice system, and evoke suspicion or hostility. For others it may be seen as a sign of the 'serious' nature of the project, and encourage them to make an effort to provide the interviewer with 'what she wants'. While you can never be certain in advance how any one person will react to being recorded, you can often anticipate likely responses and seek to take action to alleviate potential problems. Where you might reasonably expect suspicion you need to make a particular effort to emphasise the confidentiality of their participation and explain how that is going to be achieved - where digital recordings will be stored, who will have access to them, how transcripts will be anonymised and so on. In contrast, when participants seem overly concerned to provide you with the 'right' answers for your needs, you should explain clearly that they can best help you by giving as full an account as possible of their own views and experiences.
Sometimes the inhibiting effect of recording only becomes apparent when you switch the machine off and the interviewee immediately opens up with some highly relevant material. One option here is to ask the interviewee whether they would be willing to repeat the comment on tape. If they are not, or the circumstances are such that you do not feel comfortable asking them, you can still keep a written record of the key points they have made, but you should check first whether they are happy for you to use the material or whether it is to be entirely 'off the record'. We discussed how to deal with 'off the record' comments in more detail.in Chanter 3
Audio-recording
•at majority of qualitative researchers now use digital recorders to " nterview data - they produce high-quality recordings than can very be transferred to a desktop or laptop computer for transcription and Increasingly, interviewers use their mobile phones, via either built-in ^i,emo' or sin^lar functions, or specialist apps that may offer a range ■a features. Alternatively, there is a wide range of digital recording ces on the market, varying in price, quality and functionality. We would ays recommend that you use the best-quality recording device available. ■ worth noting that research ethics and governance bodies increasingly ect recordings to be at least password-protected, and often encrypted, so should be borne in mind when purchasing recording equipment or artphone apps.
Whatever the technology you use, you must make sure you are thoroughly iliar with how it operates before you begin interviewing. If at all possible pou should test out the equipment in the room you will be using, to see whether there is a problem with acoustics and work out exactly where you will dace the recorder and external microphone(s) if using them. Even though you hould have made it clear when recruiting participants that you wish to record be interview, and must obtain consent so to do, people commonly feel uncom-ortable about being recorded. We would suggest that one way to address this j to switch the recorder on as early as possible - for instance, when you start laining the interview procedure to them. Our experience suggests that once idcipants get talking about a topic that matters to them, most quite quickly 6vt less self-conscious about being recorded than they were at the start — so the longer they have to get used to the machine the better. During the interview avoid the temptation to constantly check that the machine is functioning correctly. Where possible, place it in such a way that you can clearly see the 'record' light illuminated.
Video-recording
Video is more often used to capture 'naturalistic' interaction in approaches such as conversation analysis and discourse analysis (see, for example, Heath, 1997; Rusk et al., 2015) than for recording formal research interviews. However, there are occasions where you may find it worthwhile. A video-recording can be very useful in focus groups (see Chapter 6), to help you identify speakers and to enable non-verbal interaction among participants to be drawn upon in the analysis. Equally, any study that has a strong emphasis on the bodily aspects of experience may benefit from the availabiHty of a visual record of the interview.
Inevitably, video-recording places more constraints on where you carry out your interviews than audio-recording. In order to obtain good data from
Interviews in Qualitative Research
well need two cameras and ideally expert technical help with setting them u in the most effective possible positions. This means that it is usually prefeta^L to carry out video-recorded interviews in a setting such as a university, Wh equipment and technical support will be available. Transcribing video so thv you can match non-verbal with verbal interaction is an extremely ardu0Uj process, even where you are not seeking the level of detail required j conversation analysis and similar traditions.
As with audio-recording, you must let participants know from the start that you wish to video the interview, and must obtain their explicit consent to do so. Again, we would advise that you set the video-recorder(s) going early ijj the proceedings, to enable participants to get used to the fact that they are being filmed. Note, though, that participants are not necessarily more self-conscious about video-recording than audio-recording. Murray (2008) argues that interviewees quickly become accustomed to filming, so long as they have an interesting and meaningful topic to talk about.
Note-taking
It is always preferable in qualitative research to obtain an audio- (or video-) recording of the interview, and for some methods it is absolutely essential. You could not, for example, carry out conversation analysis or discourse analysis without a full, accurate record of what the participant said, and it is hard to see how a phenomenological or narrative analysis could be completed without a verbatim transcript. Nevertheless, there are circumstances where you may be forced to rely on written notes, and where such notes can be of value in the absence of a recording. You may also find it useful to take some notes to accompany audio-recording. Whatever the circumstances, we would always suggest you explain to participants why you need to take notes.
One situation in which note-taking becomes essential is where a participant refuses to allow audio-recording at the start of the interview. This is often due to suspicion about what may be done with recordings and who may have access to them, so if a participant declines to be recorded it is worth checking that they have fully understood your arrangements to preserve confidentiality - though of course you should not exert any pressure on them to change their mind. If they persist with the refusal, but are still happy to continue with the interview, you will have to rely on the notes you take, and it is as well to be prepared for such an eventuahty. The challenge is to balance the need to attend to what is being said and to framing your questions in response, and the need to keep the interview flowing reasonable smoothly. Unless you are an expert in shorthand, you should accept that you can only keep notes indicating the main points covered. Do not constantly ask the participants to repeat themselves but rather at key junctures, you might ask them to sum up their preceding points. Only
Carrying Out Qualitative Interviews
©
rnake a comment that immediately strikes you as especially impor-I 11Z ht yoU as^ ^aem t0 rePeat   again _ and you should keep such requests ^F011. • num. We have found it a useful technique to leave plenty of space ^the notes you have made; then, as soon as possible after the interview, BT   as much further detail as you can recall. Incidentally, where you have HL^udio-recording, always check that the machine has recorded properly; if E,   0t y°u can use i™111^316 post-interview notes to replace a failed record-F*7   to assist in the transcription of one that proves to be of poor quality, vjote-taking during an audio-recorded interview may serve either of two 'n purposes. Firstly, you can provide yourself with brief written reminders to Mlow up issues raised by the participant at a later point, instead of interrupting JP    ■ mid-flow. Secondly, you can keep a written record of non-verbal behaviours where these may be essential to a full and accurate transcription. For nstance' you ma^ want to note 8estures or facial expressions that suggest 0ng emotions. Of course, you cannot possibly keep a running commentary on the participant's non-verbal communication throughout the interview. The aim, rather, is to pick out those incidents where it appears particularly powerful jn conveying meaning. When you take notes during an audio-recorded interview it is often useful to do so on a copy of the interview guide itself, as this can help you subsequently to locate the notes within the course of the interview.
Building rapport
Building rapport with your participant is widely seen as a key ingredient in successful qualitative interviewing. It is important, though, to be clear what we do and do not mean by 'rapport'. Starting with the negative, building rapport is not about ingratiating yourself with your participant. If you try too hard to be liked by them, you may find yourself reinforcing their opinions in a way that can become leading. Rapport is essentially about trust - enabling the participant to feel comfortable in opening up to you. There are no guaranteed recipes for rapport, but there are things you can do to encourage a positive relationship that enables trust to develop. Much of the advice in the rest of this chapter is relevant to this goal - for instance, the sections on how to formulate questions and how to manage difficulties in the interviewer-interviewee relationship - but here we want to concentrate on how you present the research project and yourself to your participants. Getting these wrong can undermine anything else you do to build rapport.
Introducing your project
In the process of recruiting participants (see Chapter 3), you will have provided them with information about your project, most commonly in the form
0
Interviews in Qualitative Research
of a written information sheet. You may well have had more personal contaJ with them, via e-mail or a telephone call, for instance. Such communicatj0 • not merely an administrative process to ensure you have someone to interv' "* who is able to give informed consent to their participation in your research H also serves to set up expectations in the participant-to-be that can impact the way your relationship with them develops over the interview.
Institutional ethics review bodies will generally insist that information prj vided prior to interview is clear and comprehensible to the intended participant group, and provides an honest account of why the study is being undertaken ari(j how data from it will be used. However, it is dangerous to assume that because people have been given an 'approved' information sheet and are willing to sign a consent form, they share your understanding of the purpose of the research and the nature of the interview process. As Warren (2002: 89) states, 'there are many indications in the literature on qualitative interviewing that the [participant's] understanding may not match the interviewer's from the start, may shift over time, or may be "confused"'. If this is not recognised and responded to, it can harm the rapport-building between interviewer and interviewee. It is therefore vital that you do not assume that because the participant has 'been told' what the research is for and has signed a consent form, no further explanation is needed on your part. Always take time before you start the interview proper to check that the participant has an adequate understanding of what is about to happen and why, and if necessary revisit this during the interview.
It can be useful to arrange to meet participants and introduce your project in detail before the actual appointment to carry out the interview. This ensures that you are not building the relationship with the participant from 'cold' at the start of the interview. While practical concerns of time and distance may not always make this possible, we would particularly recommend this as a strategy for research on highly sensitive topics. For instance, in her research with children with Hfe-limiting illnesses and their parents, Alison Rodriguez visited participating families in advance of interviews both to discuss practical and ethical aspects of the research but also to give a chance for participants (especially the children) to get to know her (Rodriguez, 2009).
Self-presentation
In addition to what you tell participants about your project (in writing and verbally), how you present yourself will impact on your relationship with a participant. Self-presentation includes what you wear, your use of non-verbal communication, and the kind of vocabulary you use. These aspects do not just convey something of your personal qualities, such as whether you are warm and friendly or cold and distant, but perhaps more importantly provide the participant with information about your identity. For example, a PhD student of the first author was conducting group interviews with community nurses about some important and contentious organisational changes that were happening. At her first eroun. ahf » -
Carrying Out Qualitative Interviews
If   d suspicious towards her, despite her reassurances about confiden-^B* d"her independent status. On reflection she wondered whether the aIJiad dressed rather formally for the interview might have aligned her K; • • ants' minds with management. For subsequent interviews she Hrf Qte casually and emphasised her student status more clearly: this ^^e!, "to contribute towards a more open and trusting atmosphere in the
How (not) to ask questions
his section we will focus on the process of actually asking questions in an >w Our main concern will be with how questions are formulated, but will also address the issue of managing your non-verbal communication in ,hc interview setting.
Formulating questions: treating.words carefully
We have seen in earlier chapters that interviewing can be seen as a special form | of conversation. It differs from ordinary spontaneous conversation not only in that one party does most of the questioning and the other most of the answering, but also in that the former (the interviewer) has to be very careful about the way in which her questions are formulated. Adding to the challenge for the interviewer, her need to take care in her choice of words is coupled with the need to appear relaxed and comfortable in order to put the interviewee at their ease -an issue we explored further in the section on 'building rapport'.
In order to clarify what constitutes good practice in questioning, it can be helpful to examine the opposite. Look at the interview extract in Box 5.1, in which a researcher is asking a teacher about his experience of dealing with disruptive behaviour in the classroom.
Box 5.1   Interview extract: how not to ask questions
The extract below illustrates some of the pitfalls that can occur when asking questions in an interview.
1. I: So can you give me an example of a recent incident where a pupil was disruptive in your lesson?
2. T: Well, there'was this lad - I'll call him 'Jack' - who's always... well, you know, at best he's just disinterested but very often he's messing about, disturbing the kids next to him and all that.
3. Anyways, yesterday he was up to his usual tricks - leaning back on his chair to talk to his mate behind him, throwing stuff at the girls in front, nothing REALLY awful, just constant low-level disruption.
fjäcT)   interviews in Qualitative Research
Carrying Out Qualitative Interviews fs^)
(Continued)
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
I told him to stop two or three times, got nowhere, so I made him move over to the isolation desk-out on the front beside me on his own. He's not been there more than five minutes when the whole class starts laughing and t spin round to catch him pulling a daft face at me! I: Didn't that make you furious?
T: Well, to be honest it was a mixture - a mixture really -1 WAS annoyed but also I kinda saw the funny side. I: Really? I would have been mad as hell.
T: Thing is, he's not at all a nasty lad, just not very academic and a bit of a joker. Hmmm (pause) perhaps I should have been angrier. But anyway, I sent him out into the corridor, told him to calm down there for ten minutes and if I heard so much as a peep out of him he'd be at the Head's office before his feet could touch the ground. I: What did the Head say to him? T: No, I didn't actually have to send him there in the end. He quietened down - came back for last quarter of an hour and actually seemed to get a bit of work done.
I: OK. Let's imagine though that he had carried on messing around, or if he - or another pupil - had behaved in a more seriously disruptive way-at least what YOU would see as more serious - in that case, what might be the sort of tipping point that would have you taking the next step, whether that was the Head's office or whatever? T: Sorry, you lost me a bit there.
Are we still talking about that specific incident or what I might do in general? I: Er, that incident I think.
What kind of thing might have led you to sending him to the Headmistress? And what, other than the Head, might have been the next escalation if you like in punishment?
T: Well, the choice could be between the Head and putting him on report. I guess I thought sending him to Mrs Whitlow might be a kind of short, sharp shock. And to be honest, it doesn't involve all the admin that going on report entails.
extract illustrates several of the kinds of mistake that can all too be made when formulating questions in a qualitative interview. **'Vr   i|v  it includes instances of leading questions, over-complex uh'P'e quest^ons' judgemental responses and failure to listen to the tviewee.
Leading questions
•on is leading when its wording suggests to the interviewee the kind of - nse that is anticipated. In Box 5.1, the interviewer's question 'Didn't that ""ake you funous?' (line 12) is leading. It suggests to the interviewee that anger •the appr0Pr'ate resPonse to tne situation he has described. The danger is that the interviewee may feel some degree of pressure to conform to what appears be expected by the interviewer. This does not mean that people are likely to •ve a knowingly false response, rather they may play up the extent of their conformity or play down aspects of their experience that go against the perceived expectations. Equally they may just have the direction of their account sidetracked from that which it might otherwise have taken.
. Over-complex and multiple questions
The wording of questions should be kept as simple, clear and direct as possible. The question that begins 'OK. Let's imagine...' (line 25) shows what can happen when an interviewer fails to do this. A rather long and convoluted hypothetical question results in the participant losing the thread and having to ask for clarification. If this kind of questioning style persists, the smooth flow of the interview can break down, with an impact on the quality of the data obtained.
In this instance, the interviewer follows up with another example of a poorly formed question - namely a multiple question. This is where the interviewer asks two or more questions in combination, which can be confusing for the participant. The result is often that only one part of the question is addressed, as happens in our example - the interviewee fails to answer the part of the question asking about the circumstances that might have led to him sending the pupil to the headmistress (final 6 lines).
Judgemental responses
The interviewer should try to avoid responding to what the interviewee says in a way that suggests she is making a judgement about their position. Judgemental comments are problematic for two reasons. Firstly, they may have the same effect as a leading question, in that they signify the kind of answer that will be deemed appropriate. Secondly, they may harm rapport, by putting the interviewee on the defensive. In our example, the interviewer's comment
82 )   Interviews in Qualitative Research
Carrying Out Qualitative Interviews  f 83
that the example of pupil misbehaviour would have made her 'mad as n. (line 15) is somewhat judgemental, suggesting that the teacher's own niotj ambiguous response might not be normal. The latter's subsequent renvjJ ('perhaps I should have been angrier') shows he is beginning to question rk appropriateness of the response he gave previously. While a single milni < judgemental comment such as this is unlikely to derail the whole interview \t\ the interviewer persisted in the same tone it could well begin to have »| deleterious effect.
Failure to listen
Failing to listen to the participant's response can lead to inappropriate ques.' tioning, potentially leaving him or her frustrated or irritated by the interviewer. Our interviewer demonstrates this on line 21 where she has missed the fact' that the participant has just told her that he only threatened to send the pnpi to the headmistress. In reality it is almost inevitable that in the course of a long interview your attention may lapse occasionally - especially as you are having to both listen to what is being said and think about the overall progress of the interview. If you realise that you have missed something you may ask the interviewee to repeat or clarify their last point, particularly if you seem to be at an important place in the interview. However, if you repeatedly have to make such requests the interviewee may conclude that you are not really very interested in what they have to say.
One way to ntinirnise the danger of lapses in attention is not to overburden yourself with your schedule of interviews. Carrying out three or four interviews back-to-back is tiring and demanding. If the practicalities of your project mean you do need to carry out multiple interviews on a single day (e.g. because of access issues), try to take substantial breaks between them.
Non-verbal communication
If you appear tense and nervous, the interviewee is unlikely to feel at ease. Most of us have our own personal forms of non-verbal 'leakage' that can reveal any tension or anxiety we may be feeling - persistent foot-tapping, fiddling with jewellery, biting nails and so on. Reflect on your own non-verbal 'habits' and try to be conscious of them in order to minimise them during interviews. It can be useful to carry out a mock interview with a friend or colleague and either video-tape it or ask a third party to observe and note any non-verbal behaviours that could be distracting for a participant.
Probing
We defined 'probes' as a specific type of question in Chapter 5, in relation to their inclusion in your interview guide. However, although you can sometimes
and prepare useful probes, in most cases the majority will need to 'P3 j -n the course of the interview. In general terms, probing seeks to add ^ interview data; in reaching this goal it is possible to identify a range ■■  .cc roles that probes can play. Drawing on Patton (2015) and Rubin
J*"ubin (2012), we would suggest three main types of probe. Elaboration ■ encourage the participant to keep talking in order to gather more detail 'topic at hand. Clarification probes seek explanation - either of specific and phrases, or of more substantial sections of the account that the • ,ewer has not fully understood. Completion probes ask the interviewee i finish 3 story or explanation that seems to the interviewer to have broken if before its 'natural' end. Box 5.2 provides examples of the three types of robe, and considers how they might be combined.
Box 5.2   Types of probe
■ Tom has been interviewing a youth sports coach (Sarah) about how she tries to motivate teenagers to maintain commitment to their sport activities. Tom has just asked her what she does if she feels a member of the club is losing motivation. She replies as follows:
Sarah: Well, let me think. If it's someone who's been here a while, normally a good performer- actually I can think of an example like that'. What I did with this girl, I thought she needed more of a challenge, so basically I got her aiming at Regionals.
Following this response, Tom could use each of the three types of probe. To encourage Sarah to say more on this topic he could use an elaboration probe - either a non-verbal/ paralinguistic cue or an actual question:
Tom:    Aha (nodding head)
Tom:    And what if the person losing motivation was a less competent performer?
To make sure he understands what Sarah has told him, he could use a clarification probe:
Tom:    Could you just briefly explain to me what 'Regionals' are?
•f Tom is interested in the specific story Sarah has brought in as an example, he might use a completion probe:
(Continued)
Interviews in Qualitative Research
Carrying Out Qualitative Interviews   ( 85
(Continued)
Tom:    And was that successful for this girl?
In this example, ideally Tom would use all three types of probe to maximise the depth of data obtained, but of course he could not use them all at once or we would have the problem of multiple questions. He therefore needs to think about the order in which to use the different probes. There is no absolute right and wrong in this, but the goal of facilitating the easy flow of the interview should guide his choice. In this case we would start with the clarification probe, as knowing what 'Regionals' are might guide his ' subsequent questioning. Also, it is likely to produce a fairly simple response in the form of a definition and not lead the interview off on a tangent. We would then tend to favour using the completion probe, in order to keep this particular story going. Finally, we could use the 1 elaboration (question) probe, to get Sarah to discuss other strategies for motivating the ' young people in her charge. This part of the interview would then look as follows:
Tom:
So what do you do if you think one of the kids is losing interest in what they do at the club, losing motivation?
Sarah: Well, let me think. If it's someone who's been here a while, normally a good performer - actually I can think of an example like that. What I did with this girl, I thought she needed more of a challenge, so basically I got her aiming at Regionals.
Tom:    Could you just explain to me what the 'Regionals' are? [clarification]
Sarah: Sorry-just took it for granted ... we have regular official competitions for all the sports we do, but Jenna's an athlete so I was thinking of athletics. In athletics ' its very structured - District, Regional, National. I thought she probably had the< ability to get through to the Regionals in sprinting or long jump.
Tom:    And what happened? Was she enthused by this? [completion]
Sarah: Yes! One of my success stories - probably why Jenna came to mind! She got through to the finals of the 200m at Region, came fifth which is a real achievement, and she's determined to improve on it next year. She just might. ■
Tom: That's great. What about kids who are less able than Jenna, who you don't feel have the talent to compete at the higher levels? What strategies might you use with one of them who was beginning to lose motivation? [elaboration]
Good use of probes requires good listening skills, in line with our comments in the previous section ('How (not) to ask questions'). We would reiterate the point that not overburdening yourself with numerous interviews
•   should help you to remain alert and sensitive to opportunities for ^'robing- Where a participant's response invites a number of different 'f/m y°u - as n*PPens        commonly - you may find it useful, once '   decided which to use first, to jot down reminders so you can return ave ■      This is particularly the case where the answer to your initial ^ ° «.« further probing, perhaps resulting in three or four 'layers' of IT and response on a specific topic. ^° Ytfhile probing is essential to obtaining real depth in interview data, you T it too much. One example would be where you probe in enormous HP.,   ' an area of the interview that is of hmited relevance to your research resulting in restricted time to spend on more crucial areas. Another h'ere'y°u spend a great deal of time seeking clarifications of unfamiliar * ^ el technical language, professional jargon, slang words or whatever. If ,CrmS pear too ignorant of the participant's world and the topic at hand, he y°U limply decide that it is not worth his while making an effort to describe detail to you. In such circumstances, try to identify the terms that you 'Vsolutely must clarify in order to make any progress with the interview and concentrate your clarification probes on these. You may get a chance to ask for further clarifications later, and if (as is likely) you are audio-recording the interview, you should be able to enquire about others after transcription.
Starting and finishing interviews
You need to give special forethought to how you are going to start and finish vour interviews. The way you start can have a significant impact on how your rapport (or lack of it) with the interviewee develops. It is normally seen as good practice to start with relatively unthreatening and simple questions, such as asking for descriptive information about the participant in relation to the topic under investigation. For instance, in a study of teacher experiences of classroom discipline of the kind used in Box 5.2, you might start by requesting an outline of their career history, and continue by asking them about school policies regarding discipline. Neither of these areas requires self-disclosure on sensitive topics, but the sequence does ease the participant into the main focus of the interview.
As we have noted, it is commonplace as an interview develops to move into more difficult - perhaps emotionally charged - areas of questioning. This can create a problem in terms of how you bring an interview to a conclusion, as you do not want to say your goodbyes and leave immediately after the participant has reached the most sensitive part of their account. Just as it is usually good practice to ease the participant into the interview, you should plan a strategy for easing them out. Try to plan closing questions that move away from self-disclosure, and hand as much control as possible to the participant. One form that is often helpful is a question focusing on desired future changes or developments. Thus we could ask our hypothetical teacher what their top priority would be if they could influence government policy on discipline in schools. To close, it is generally good practice to ask the participant
86 )   Interviews in Qualitative Research
Carrying Out Qualitative Interviews (87
if there is anything else they want to tell you, and invite them to ask any qu? tions they may have about the research project and/or their part in it. TJnJJ the participant asks you to turn it off, leave your recording equipment runni ! until the very end of the interview, as it is not uncommon for interviewees mention something of interest and significance at this stage. Should this ha*.' pen, do not miss the opportunity to reopen more detailed questioning, so l0n 1 as the participant is happy to continue.
Managing 'difficult' interviews
Any qualitative interview can lead you in unexpected directions to face unexpected challenges, so that no amount of experience or preparation can provide you with a 'stock' response to every circumstance. Nevertheless there are certain types of interview situation where you can reasonably anticipate particular difficulties and with some forethought at least be prepared to face them. We will consider four such situations here: interviews where there are significant status issues, interviewer role conflicts, interviews on emotionally sensitive topics, and dealing with under- and over-communicative interviewees.
Status issues
If an interviewee perceives an interviewer to be of markedly higher or lower status than herself, this can have an impact on the quality of the interview. In the qualitative research literature, there tends to be more consideration of situations where the interviewer is perceived to be of higher status than the interviewee than vice versa. This is often in the context of discussions about power in the interview, and the desire to equalise this between the parties (Briggs, 2002; McKie, 2002). However, perceived status differences are just as likely to be in the opposite direction, either because the researcher is relatively junior (as in the case of a postgraduate student) or because the participant belongs to a group who would generally be seen as 'elite' in society - such as senior professionals and managers (Harvey, 2011). Status differences can del-eteriously effect the course of an interview in two ways: directly, by inhibiting interviewees from discussing particular topics (e.g. for fear of appearing ignorant or losing face), and indirectly, by preventing the building of rapport between interviewer and interviewee.
Interviewees may see themselves as lower in status than the interviewer because of differences in social class, education level, occupation, age and so on. Sometimes such differences may be compounded by interviewer role conflicts; for instance, where a qualified nurse undertaking a master's degree programme is interviewing an unqualified health-care assistant. We will look more closely at the role conflict issue shordy. In general terms, where you anticipate perceived status differences in this direction, vou should seek to
nem through the way you present yourself and your research. How ""^rhis must always take into account the specific context, although 1 areon and overly academic language is always likely to be especially *t Inviting participants to attend the interview at a university or jjte should probably be avoided, and if your participants may be una-unwilling to be interviewed at home, you should try to identify (P .   settings that will be familiar and comfortable for them, r'^jgh-status participants may be difficult to interview because they are | to being in control in their interactions with others - to asking rather answering questions. This can result in the kind of situation we discuss low in relation to the over-communicative interviewee. Rarely, more serious ttblems can develop where the interviewee experiences the reversal of roles threatening, and seeks to assert his normal authority by undermining'the fidence of the interviewer. He may do this by trying to show up the tarcher's lack of knowledge of the topic under investigation ('Did you really , |(now that?'), by questioning the validity of her methods ('You can't learn thing from just chatting with people'), or the credibility of her discipline if course, psychology is not a proper science').
To reduce the likelihood of problems like these when interviewing high-arus participants, you should avoid challenging their authority in their own Jd, but remain sure of your own expertise in yours. After all, it is most ikely they know more about qualitative methods than you do. Make sure 5U have done some background research into your area of study, so you are at least as knowledgeable as a well-informed layperson might be expected to be. For instance, if you were interviewing school head-teachers about discipline, you should ensure you have a grounding in the key points of relevant recent legislation. Should you be unlucky enough to face an interviewee who is intent on beHttling your role, your best option is to calmly and politely acknowledge your difference of opinion with them, but not get pulled into an argument, and try to move the interview forward. If you are carrying out research with ehtes from different nationalities than your own, it is crucial you learn as much as you can about their cultural norms and expectations, to avoid causing unwitting offence or missing important nuances in their accounts (Mikecz, 2012).
Interviewer role conflicts
In some cases researchers can face a potential conflict between their role as an interviewer and other roles they may have (or be perceived as having) in relation to the interviewee. Probably the most common example of this is where the researcher is also a practising health or social care professional and is interviewing members of the service user group they work with in their professional life. Similar circumstances can occur in other areas of practice and research too, such as education and the criminal justice system, though they are less fremientlv covered in the literature than health and social care.
88 }   Interviews in Qualitative Research
Carrying Out Qualitative Interviews   ( 89
If you are in such a position, the key requirement is to be clear from the start where the boundary lies between your researcher and professional roles. In general, our advice would be to tell participants explicitiy that you are not in a position to deal personally with any health or social care problems they may talk about in their interviews as you are there only in the capacity of a researcher. You can, however, bring with you a range of contact information for services you may anticipate as being relevant to your participants, which you can offer if necessary at the end of the interview. However, some methodological approaches would draw the role boundary differently, encouraging more of a blurring of roles; this would be the case in participatory approaches where the researcher is actively seeking to encourage change and empower participants. Jack (2008) provides a useful discussion of role-conflict issues in nursing research. Even here, though, there are limits in terms of what health-related interventions a nurse-researcher could legitimately offer; it is essential such issues are carefully thought through prior to data collection, and we would normally expect them to be covered in ethical protocols. Some professionals may be bound by professional codes of conduct that legally override other ethical commitments -for instance, responsibility to report cases where there are strong suspicions of chdd abuse or where serious professional malpractice may have occurred. Again, it is important that you make any such responsibilities clear to participants before the interview commences.
The other circumstance in which you might face role conflicts is where you are carrying out interviews with close friends and/or family members. We would advise that if you are doing this you think very carefully beforehand about any issues that could emerge that might create problems in your relationship with the participant beyond the interview situation. After that, it is generally a good idea to have an informal discussion with the person so that they are sure about what the interview is going to cover, and can think in advance about what they are and are not happy to discuss. You should take special care in protecting confidentiality and anonymity in interviews like this, as whatever you publish may be read by other family or friendship-group members, to whom the interviewee could be easily identifiable (see Chapter 3 for more on this issue).
Dealing with sensitive topics
Any qualitative interview can raise issues that the interviewee finds upsetting, although of course some topics are more likely than others to evoke strong feelings from participants: serious illness, bereavement, conflict or harassment at work, criminal victimisation and so on. You need to think ahead about how you will deal with situations in which a participant becomes distressed. What you should certainly not do is decide to immediately terminate the interview without consulting the participant. This may simply give the message that you
not cope with their feelings. In any case, the fact that a participant becomes x ressed does not necessarily mean that they are finding the interview experi-
e a negative one. Especially where they have consented to take part in rch on a sensitive topic (of the kind listed above), participants often
ort that they appreciate the chance to discuss a difficult subject with a
pathetic listener. In most cases, your best response if a participant becomes distressed is to calmly and gently offer them a range of options about how to
ceed. por instance, you might say: 'I can see you're finding this difficult. If vou want to move on to a different topic or take a break, that's fine. Or if
yOU
want to end the interview now, that's entirely up to you.' The issue of your responsibilities regarding the psychological well-being of participants and research ethics is discussed in Chapter 3.
Under- and over-communicative interviewees
• Some interviews can be difficult because the interviewee says too little or too much. Given the aim in qualitative interviewing to explore experiences in depth, the former is generally more of a problem than the latter. If you have a participant whose responses rarely go much beyond the monosyllabic, you must consider what might be holding them back. One possibility i is that despite everything you have told them before the interview, they are | still concerned about confidentiality and/or anonymity. If you have the impression that this may be the case, it could be fruitful to reiterate what you will be doing with the data and how you will protect their identity. Sometimes people may feel they are helping you by getting the interview over with quickly. You can try to counter this by using frequent probes that encourage them to 'tell me more'. Of course, you should not do this so much that it feels like you are harassing the participant, so it can be particularly effective to use silence to coax a response. When the interviewee gives a very short, superficial response, just refrain from responding yourself for a few seconds. Often this will serve as a cue to the participant that it is still their 'turn' to talk, and they may then expand on their answer without the need for you to say anything.
At the other extreme, the fact that a participant has a great deal to say is not in itself a problem - quite the opposite! Similarly, if someone seems to go off on a tangent, it is usually best to let them run with it for a while; they may bring you to perspectives on your research topic that you had not considered before. But if they are spending a great deal of time on matters of minimal significance to your research, you will need to try to guide them back on track. ! One tactic is to recall the last relevant section of the interview, and once they pause for breath, ask them to return to that issue and elaborate further. Alternatively, you can thank them for what they have just told you, say you ^e mindful that you have much you would like to cover with them, and Politely ask them to move on to your next question.
Interviews in Qualitative Research
Carrying Out Qualitative Interviews
0
Using multiple interviews
Interview-based studies do not necessarily have to consist of just a si J interview with each participant. While the single-interview design is the njTJ common, many researchers have used two or more interviews with participant, as part of either a longitudinal design that tracks changes ov time, or a multiple-interview approach (usually carried out in short successinJ to each other) to allow more depth of discussion and deeper engager^ ' with participants. For example, Cresswell and Eklund (2007) carried out I series of interviews over a year with New Zealand rugby players to look a ' the dynamic experience of burnout over time. Schilder et al. (2005) carried I out extended life-story interviews over four occasions with gay men, t0 explore the relationship between drug use and unsafe sex; multiple inter. views here were used principally to 'develop rapport and elicit sensitive I information' (p. 341).
Some writers have criticised single-interview research designs for often producing shallow data and as potentially exploitative of participants. Chamberlain (2012) coined the term 'drive-by interviewing' to characterise research in which the researcher dips briefly into the participant's life, extracts what they need, and then (from the participant's point of view) disappears. We feel that this critique has usefully cautioned researchers about taking the single-interview design as the default position, and has helped to focus attention on participants' experiences of the research process. However, we would strongly refute any blanket dismissal of single-interview designs on either methodological or ethical grounds. For some studies, it is not practical to carry out more than one interview, either in terms of researcher resources (time, travel costs) or participant availability. Equally, it is often not necessary to use multiple interviews to obtain the depth of data required, especially where the focus of the interview is on a relatively narrow topic. On the ethical issue, there is no reason why single interviews need be exploitative, so long as researchers think carefully about ensuring consent is fully informed (see Chapter 3 above) and provide feedback on research outcomes that makes clear the value of participants' contributions.
Turning to the practicalities of using multiple interviews, for longitudinal designs the crucial issue is to choose a time interval between interviews and an overall study duration that is likely to capture the kinds of changes that are of interest. Where there are likely to be points of significant transition or change in the phenomenon under investigation, it may make more sense to organise interviews around these rather than hold them at fixed intervals - see, for example, Murray et al.'s (2009) advice regarding serial interviews with patients. It is often a good idea to spend some time at the start of second or subsequent interviews recapping the main points from the previous interview(s), unless of course your research topic is such that it would be disadvantageous to provide any kind of reminder. When multiple
are used as effectively a single data collection point, an important f*eV/S.    js novv the interviews in the sequence relate to each other. In s the separate encounters with each participant add up to a single as was the case in Schilder et al.'s (2005) study referred to earlier, 'vely, subsequent interviews may be designed to build on the previ-,ne(s) PernaPs Pr0Dm8 m more depth, or approaching issues from a ent perspective. This may well require some very preliminary analysis • after each encounter, in order to define the focus for the subsequent or at least, a careful listening through of the recording. In Wengraf's aphic-narrative interpretive method, outlined in Chapter 13, life-story ,jews are divided into three subsections with the (optional) third taking after the initial two, which are normally concurrent (Wengraf and nberlayne, 2006).
Conclusion
lerviewing is without doubt a skill that improves with experience, perhaps iove all in managing the potential tension between listening closely and aintaining a sense of where you are - and where you are going - in the inter-EW as a whole. Equally, the more you have carried out, the less likely you are be thrown by some surprise occurrence in the course of an interview, hatever your level of experience, though, good preparation can considerably ance the quality of your interviews - and lack of preparation undermine an. Thinking carefully in advance about the issues we have covered in this apter - how you frame your questions, use probes and prompts, start and ush interviews, handle sensitive topics and status issues, as well as technical .apects of recording - should put you in a good position to gain the most from ihe method. Finally, it is crucial to recognise that even for the most experienced, best-prepared interviewer things will sometimes not go as planned: you suddenly realise your participant has fundamentally misunderstood a whole line of questioning. You are constantly interrupted by the interviewee's colleagues, or child, or dog. You hear your own voice turning your lovingly crafted question into incomprehensible gobbledegook. The trick is not to panic - take a breath, collect your thoughts and you will usually be able to find a way to deal with the situation. The key feature of the interview is that it is an extended encounter with another person (or persons); as such, it gives you the time to recover from any difficulties that arise and get back on track.
Recommended reading
Brinkmann, S. and Kvale, S. (2014) Interviews: Learning the Craft of Qualitative Research
Interviewing. Thousand Oaks, CA: Sage. Latest edition of a classic text on interviewing methodology.