SEARCH

SEARCH BY CITATION

Keywords:

  • epistemology;
  • methodology;
  • nursing;
  • qualitative research;
  • realism;
  • rigour;
  • trustworthiness;
  • validity

Abstract

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

Title. Validity, trustworthiness and rigour: reasserting realism in qualitative research

Aim.  This paper challenges recent sceptical approaches to the possibility of validating qualitative research and underlines the benefits of adopting a realist approach to validity.

Background.  In recent discussion about the methodological bases for qualitative research it has been argued that, because different methodologies take different approaches to validity, attempts to develop a common set of validation criteria are futile. On the basis of this sceptical view, a number of strategies for judging qualitative research have been proposed. These include suggestions that: it should be judged according to aesthetic or rhetorical criteria, rather than epistemological validity; responsibility for appraisal should move from researchers to readers; each methodology should be assessed individually according to its own merits.

Discussion.  None of these suggestions provide a viable alternative to validity, defined as the extent to which research reflects accurately that to which it refers. Because the form of research does not determine its content, replacement of epistemology by aesthetics is unsustainable. Because research reports mediate between writer and reader, a one-sided approach to this relationship constitutes a false dichotomy. If we accept the criterion of practitioner confidence as a means of judging methodological approaches, this involves rejection of judgement according to a methodology’s own merits.

Conclusion.  If qualitative research is actually about something, and if it is required to provide beneficial information, then a realist approach to validity holds out greatest promise.


What is already known about this topic

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References
  • Different methodological approaches to qualitative research contain different approaches to validity, making it impossible to develop a universally accepted approach to the validation of qualitative research.
  • It has been suggested that, as a consequence, individual studies should be assessed according to the individual judgements of readers.
  • It has also been suggested that we abandon traditional concepts of validity in favour of aesthetic judgement.

What this paper adds

  • Rather than attempting to build a set of criteria common to all methodological approaches, it is possible to judge the methodological approaches themselves.
  • The ‘confidence criterion’ can be used by nurses to judge which approaches are most likely to provide knowledge that will beneficially inform practice.
  • On this criterion, realist approaches to validity provide a more promising approach than individual assessment or aesthetic judgement.

Introduction

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

In this paper I examine the issue of validity in qualitative research. It is written in response to a paper by Gary Rolfe (2006) entitled ‘Validity, trustworthiness and rigour: quality and the idea of qualitative research’. Rolfe argues that, because of the absence of a unified qualitative paradigm, attempts to construct predetermined frameworks to judge the validity of qualitative research are futile. Following Sandelowski and Barroso (2002), he suggests the replacement of epistemic criteria for the judgement of qualitative research with aesthetic criteria. Moreover, he suggests that quantitative research should also be subjected to aesthetic, rather than epistemic judgement.

Rolfe’s paper merits careful consideration because it constitutes an eloquent reiteration by a European author of a position that has largely (although not exclusively) been the domain of North American methodologists (Morse et al. 2002). As such, it is emblematic of the increasing international influence of scepticism in nursing approaches to qualitative research. I wish to argue that there are good reasons for nurses not to accept sceptical arguments, on the grounds that their adoption would be highly detrimental to nursing research and its capacity to inform nursing practice.

What is qualitative research?

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

One of Rolfe’s most significant contributions is his uncovering of definitional confusion concerning qualitative research. He notes that there are two main ways of defining ‘qualitative’. One relates simply to method – qualitative research uses verbal and textual data, while quantitative research relates to numerical data. The other appeals to qualitative research’s grounding in particular ontological, epistemological and methodological assumptions, which he describes interchangeably as interpretivism and constructivism. These are contrasted with the realist or positivist assumptions associated with quantitative research. The problem with this second definition is that the relationship between method and methodology is tenuous, either because those using textually oriented methods are in fact ‘unreconstructed (post) positivists’ (p. 309), to use Rolfe’s somewhat Stalinesque dismissal, or because they adhere to alternative theoretical bases such as critical theory or feminism.

The problem in relation to validity is that each of these methodological positions entails the adoption of a different, and sometimes incompatible, approach to it. Thus, Rolfe argues that, in relation ‘to the question of why issues of validity are so contested in qualitative/interpretivist/naturalistic research, it would appear that whichever terminology and criteria we use to describe this paradigm, one or more of the so-called “qualitative” methodologies will always fall partially outside of it’ (pp. 307–308).

In response to the impossibility of judging knowledge claims according to commonly accepted criteria, Rolfe makes a number of not entirely compatible suggestions. On the one hand, he argues for the abandonment of ‘epistemic criteria’ in favour of judgement ‘according to aesthetic and rhetorical considerations’ (p. 308). On the other hand, he argues that each methodology, and indeed each study, must be appraised on its own merits according to unique criteria.

Devaluation of epistemology

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

Grounding itself in Sandelowski and Barroso’s (2002) work on the issue, Rolfe’s argument for the replacement of epistemic by aesthetic criteria is based on three interlinking premises: first, that normally the only site for the reader to evaluate research is the research report; second that the report is a ‘dynamic vehicle’ (p. 308) that mediates between writer and reader, rather than a factual account and third that the form and content of reports are inseparable, with form shaping content. On the basis of these premises, Rolfe concludes that we should appraise reports aesthetically, according to the way they are written, rather than epistemologically, according to their content. Moreover, he suggests that the move from epistemology to aesthetics should also encompass quantitative research.

Rolfe’s suggestion that judgement of quantitative research should be based on aesthetic considerations is probably the most novel and contentious part of his argument. As such, it will undoubtedly be viewed askance by those who have invested considerable effort in developing robust statistical methods to ensure validity [see, for example, recent discussions on recruitment bias in healthcare research in this journal and elsewhere: Badger and Werrett (2005), Junghans et al. (2005), Miller et al. (2003) and Parkes et al. (2006)]. However, the relationship between quantitative research and aesthetics will not be explored here beyond the observation that, if it is established that aesthetic considerations cannot be accepted as the predominant criteria for judging qualitative research, then the same arguments will hold, a fortiori, for quantitative research.

In order to establish whether or not aesthetics should be used to judge qualitative research, it is necessary to consider Rolfe’s argument step-by-step. The first premise, that the reader engages with research at the point where it is reported, can be readily accepted and indeed may even be described as a truism. The second premise is more problematic in that it contains a dubious claim in the form of ‘X therefore not Y’, X being the assertion that reports are dynamic vehicles of mediation, and Y being the assertion that they are factual accounts of events. The observation that research reports involve dynamic mediation between writer and reader can be readily accepted, in that they are written on the basis of the active interpretation of the researcher and are read on the basis of the active interpretation of the reader. However, this does not obviate the possibility that they contain, to a greater or lesser degree, factual accounts of events or attitudes. While accounts of events depend on the active interpretation of both the person giving and the person receiving the account, this does not mean that the account must therefore be nothing more than a fabrication of their joint interpretations, bearing no relationship to the events being accounted. Indeed, a core part of the interpretative work of research report writing involves writers presenting reports in such a way that they hope will persuade readers of their veracity, while a large part of readers’ interpretations of the reports will involve judging the degree to which they accept their veracity. The whole point of attempting to standardize criteria for judging validity is that doing so provides a point of mediation between readers and writers, which enables them to agree on the best interpretative tools to use in making and judging claims. Mutually accepted criteria provide the common language by which writers and readers can talk to each other in a way they both understand. In short, rather than undermining the possibility of research reports making factual claims, the fact that those reports are dynamic vehicles of mediation provides the very channels through which those claims can be communicated and judged.

The third premise concerns the inseparability of the form and content of reports, whereby ‘content is shaped, pruned and reworked to fit the form’ (p. 308). Such a premise, that there is a necessary connection between form and content, would not appear greatly contentious. Thus, for example, a formal limit on an article length of 5000 words can also be seen as a limit on the amount of content permitted. However, this does not mean that the relationship between form and content is deterministic. No matter how stringent formal requirements, there is always room for the expression of unique content.

Given the importance of aesthetics to Rolfe’s argument, perhaps an appropriate medium to examine the relationship between form and content is that most carefully shaped mode of communication – poetry. Probably the most formally restrictive variant of poetry is the traditional Japanese haiku, in which each poem consists of just 17 phonetic units divided precisely into a 5 – 7 – 5 pattern. Yet within these very stringent parameters lies a stupendous diversity of poetic imagination, with each haiku striving to evoke in the reader the experience of a unique and individual moment (Blyth 1963). If content cannot be reduced to form in the haiku, it certainly cannot be reduced so in the far more permissive forms that shape the writing of research reports and scholarly papers.

From these premises, Rolfe concludes that ‘judgements can only be made about the way the research is presented rather than directly about the research itself, and as we have already seen, such judgements are predominantly aesthetic rather than epistemological’ (p. 308). I wish to argue strongly that, because of the flaws in the premises leading to this conclusion, the conclusion itself is fatally flawed. If the content of research reports is not determined by their form, then it follows that epistemic criteria cannot be regarded as redundant. Moreover, if we accept that the dynamic nature of reports does not negate their capacity to convey factual information, and that imparting information is the primary purpose of reports, then it is not appropriate to allow aesthetic criteria to predominate over epistemic ones.

The aesthetic elite

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

The problems with Rolfe’s position are compounded by its elitist consequences. He argues that the move to aesthetics involves research appraisal being ‘subject to individual judgement based on insight and experience’ (p. 308), which means that research can only be judged adequately by those who have sufficient experience of performing research. Even novice researchers are ruled out of court because they have not had sufficient practice of research and cannot rely on the standardized guidelines of discernment that are part and parcel of the epistemic approach. Non-research-active clinicians are not even mentioned as possible recipients of research information. One wonders where this leaves evidence-based practice – it appears that it would be restricted to clinicians who had sufficient research experience to satisfy Rolfe’s criterion. Moreover, what practical use could these clinicians make of reports which were adjudged aesthetically pleasing, but whose validity was not taken into account?

While the vast majority of potential nursing research readers are deemed insufficiently prepared, the power and responsibility that Rolfe would bestow upon the cognoscenti is considerable. He argues that ‘responsibility for appraising research lies with the reader rather than with the writer of the report’ (p. 309), contrasting this with what he portrays as the traditional view of Morse et al. (2002), who he reports ‘emphasize that the responsibility for ensuring rigour lies solely with the researchers themselves rather than with the readers of the research report’ (p. 305). As we have already noted, this is a false dichotomy because, if research reports are dynamic vehicles of mediation, then the responsibility for rigour lies with both the writer and the reader. It is the writers’ responsibility to demonstrate that the research they are reporting has been conducted in a valid and rigorous manner, while the readers’ responsibility is to interpret the report to ascertain whether or not they are persuaded that the writer has indeed demonstrated rigour.

It is difficult to see how Rolfe’s one-sided privileging of the reader is sustainable, either in terms of logic or practicality. Logically, research is a form of communication, and communication, by definition, requires the active participation of at least two parties. Practically, a moment’s consideration of the likely consequences for the quality of research that would ensue from absolving researchers of their responsibility to establish rigour should give us cause to baulk at Rolfe’s suggestion. Indeed, it would appear that Rolfe himself has reservations because he goes on to commend to researchers what seem extremely onerous responsibilities:

It behoves researchers to leave a ‘super’ audit trail, recounting not only the rationale underpinning the research decisions taken en route, and the actual course of the research process rather than the idealized version that the reader is usually presented with, but also...ongoing self-critique and self-appraisal...including the moral, social and political stance of the researchers themselves (p. 309; emphasis in original).

It is difficult to read these injunctions without feeling that we have gone full circle, returning to a rather traditional concern with the importance of transparency to the good conduct of research.

Relative methodologies

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

Notwithstanding this recidivistic slide into standardized rigour, Rolfe concludes his paper with a call for relativism:

Rather than searching for an overarching set of criteria by which to judge the validity of qualitative research, we should perhaps acknowledge that there is a multiplicity of (so-called) qualitative paradigms, each requiring very different approaches to validity. Or, put another way, there is no qualitative paradigm at all, so that each research methodology (and perhaps each individual study) must be appraised on its own merits (p. 310).

As has already been intimated, the problem with this conclusion is that its laissez-faire approach to different methodologies (or indeed studies) sits ill with the prescriptive approaches concerning the replacement of the epistemic by the aesthetic and the writer by the reader. Put another way, if a particular qualitative methodology adopts the validity criteria of quantitative research, should its claims to validity be rejected on the grounds ‘that such rigour is illusory’ (p. 309), or should it be ‘appraised on its own merits’ (p. 310)?

Leaving aside the apparent contradictions within Rolfe’s argument, his contention that the multiplicity of methodological approaches makes it impossible to construct validation criteria that are applicable to all qualitative methodologies can be accepted. That there exists a number of contesting methodologies, each with their own distinctive approach to validity is beyond question. Indeed, it might be argued that Rolfe’s categorization of methodologies involves a conflation of different approaches which artificially reduces the number of voices in the Babel that is research methodology. For example, he identifies the paradigmatic grounding of quantitative research as realism yet realists have quite explicitly argued that their approach provides an equally appropriate grounding for qualitative research (see, for example, Miles & Huberman 1994, Porter 2002). Conversely, he conflates realism and positivism into a single quantitative approach. In fact, they are quite distinct (Porter 2001). Probably the most pertinent difference here is their approach to the ontology of cause. Positivism asserts the existence of determinate causal laws, whereas realism conceives of causality as consisting of generative mechanisms whose effects will differ in different contexts (Bhaskar 1989). So, while positivism is vulnerable to the relativist accusation that its quest to describe validly universal relations of cause and effect means that it cannot embrace alternative representations of the same phenomenon, realism is able to accept the existence of alternative but equally valid representations. Thus, for example, in realistic evaluation (Pawson & Tilley 1997) it is asserted that the question should not be simply ‘What works?’ (which is based on the positivist assumption of unilinear causality). Rather, it should be ‘What works for whom in what circumstances?’. This moves the focus of validity from the judgement of a single interpretation to a judgement of the degree to which the researcher has encapsulated the multiple perspectives pertaining in a given situation.

Similarly, his conflation of interpretivism and constructivism hides a host of different positions, ranging from Weberian social interpretivism, which simply observes that in order to understand social action we must understand the meanings actors hold that motivate them to act in the ways that they do (Weber 1968), through to the radical constructivism of poststructuralists such as Foucault (1977), who asserts that ‘power produces reality; it produces domains of objects and rituals of truth’ (p. 194).

Assessing methodologies

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

It might seem that this qualification simply reinforces Rolfe’s central point-because different methodological foundations of qualitative research espouse different interpretations of validity, attempts to develop agreed criteria for validity across methodologies will inevitably prove futile. However, rather than accepting that the multiplicity of approaches to validity negates the possibility of using criteria to test the rigour of research, the problem can be addressed through the application of judgemental criteria to the methodological foundations themselves. In other words, while accepting Rolfe’s demonstration of the impossibility of constructing a universal list of criteria incorporating all methodological approaches, I wish to argue that some approaches to validity, trustworthiness and rigour are better suited than others to the needs of knowledge-led practitioners such as nurses.

It is the relationship between knowledge and practice that provides the key to judging research. If the point of nursing research is to inform practice, it is of paramount importance that those acting on the basis of the knowledge it provides are confident that it accurately describes and explains the issues being addressed. Without such confidence, they can have no way of knowing whether actions predicated upon research results will enhance or undermine the health and well-being of their clients. This might be termed the confidence criterion, in that it refers to the degree to which practitioners can be confident that the knowledge claims with which they are presented will beneficially inform their practice.

The demonstration of validity is a core component of generating practitioner confidence. As Sandelowski (1997) herself notes in a paper aptly entitled ‘To Be of Use’, ‘Practitioners may move directly from the…products of qualitative research to using, applying, or otherwise informing their practice with this knowledge. Ensuring…validity is an integral component of well-crafted and credible qualitative research’ (p. 130).

Approaches to validity

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

If it is accepted that nurses need to be confident that the information they use to inform their practice is sufficiently accurate to ensure that their practice is appropriate and effective, then the confidence criterion can be used to interrogate the various approaches to validity. While Rolfe uses Hope and Waterman’s (2003) three-part model to categorize these approaches, Sparkes (2001) similar four-part framework provides a more comprehensive model.

At one end of the spectrum outlined by Sparkes is the ‘replication perspective’ which sees the concept of validity as equally applicable to qualitative as quantitative research, even though the procedures used to demonstrate it may differ. This approach roughly equates to the first position identified by Rolfe but does not assume a positivist foundation, merely an orientation to an external reality that is amenable to representation. Thus, the replication approach is exemplified by the ‘subtle realist’ position of Hammersley (1990), who defines validity as ‘the extent to which an account accurately represents the social phenomena to which it refers’ (p. 57).

Sparkes’ second position, ‘parallel perspective’, sees qualitative and quantitative research as paradigmatically different and therefore requiring different criteria. This equates exactly with Rolfe’s second position, and, in common with his characterization of it, identifies the work of Lincoln and Guba (1985) as central, in that they propose the replacement of the notion of validity with that of trustworthiness, and suggest techniques whereby this might be established in the qualitative world of multiple realities and ways of knowing.

Sparkes’ third position, ‘diversification of meanings perspective’, sees validity as socially constructed and therefore differing in meaning according to the discourse in which it is set. While this perspective is missing from Rolfe’s model, it would nevertheless appear to best encompass his conclusion that each methodology should be appraised on its own merits.

Finally, Sparkes’‘letting go perspective’ rejects the concept of validity altogether, seeking radically different, often aesthetic criteria (Sandelowski & Barroso 2002), such as evocation (Ellis 1995). This equates to Rolfe’s third position and encompasses his endorsement of Sandelowski and Barroso’s assertion that, because quality is immanent within the research report rather than a reflection of that which is revealed by the research, it is amenable to ‘the wise judgement and keen insight of the reader’ (p. 309), rather than to validation criteria.

Let us consider each of these perspectives in reverse order. First, the letting go perspective – on the criterion of being able to provide guidance for action about which nurses could be confident, this perspective’s rejection of epistemological validation would appear to rule it out as a viable foundation for the judgement of qualitative research. However, it would be unwise to reject out of hand the usefulness of aesthetic criteria such as evocation. If one of the functions of qualitative research is to provide healthcare professionals with insight into people’s perspectives on health, illness and care, then the capacity to evoke more vividly those perspectives would entail an improvement of the quality and usefulness of this type of research. However, while evocation may provide one criterion, this does not mean it should predominate over epistemic criteria. Vividly evoked perspectives are of no use to nurses if they do not accurately reflect the perspectives of research participants. In other words, even when we accept the usefulness of evocation, epistemic criteria to establish the validity of the research remain prior to aesthetic criteria.

Assessing the diversification of meanings perspective is more clear-cut – acceptance of the confidence criterion automatically entails rejection of the diversification perspective. If the confidence criterion is applicable in all cases irrespective of the discourse of a particular methodology, such universalism is incompatible with the relativism of diversification. Put another way, if a particular approach to validity is being judged on its capacity to inform practice beneficially, then it is being judged externally rather than according to its own internal criteria. Moreover, the profusion of exotic validities which diversification allows, such as ironic, paralogical, rhizomatic and voluptuous (!) validity, to cite but one author (Lather 1993), reinforces the argument for its rejection on the grounds of impracticality.

In their seminal exposition of the parallel perspective, Lincoln and Guba (1985) argue that conventional trustworthiness criteria, such as reliability and validity, are incompatible with the axioms and procedures of qualitative research. They argue that these need to be replaced by new axioms and procedures to establish trustworthiness for what they term ‘naturalistic inquiry’. The most important of these new criteria is that of credibility, which can be established using the procedure of member checking, whereby tentative results are shown by researchers to their research participants to assess the degree of correspondence and to incorporate members’ perspectives into the study’s findings.

While appearing to involve a commendable democratization of the research process, basing judgement of research on members’ perspectives rather than those imposed by the researcher, consideration of this approach exposes considerable problems. Thus, Bloor (1978) has noted that we cannot presume that lay members will have the ability or interest to comment productively on scientific discourse. More acerbically, Fielding and Fielding (1986) observe that ‘there are many reasons and interests that can lead members to misreport to the researcher, and it must be borne in mind at all times that they have different purposes from the researcher’s’ (p. 43). In other words, there is no reason to assume that members enjoy the kind of epistemologically privileged status that would warrant their role as research validators. This is not to say that member checking is without merit. While it may not provide the basis for validation, it can generate additional data and suggest interesting paths for further analysis (Bloor 1983).

It might be noted that Guba and Lincoln’s (1989) alternative criterion of authenticity, although not explicitly mentioned by Rolfe, is even more susceptible to these sorts of criticisms, in that it involves the assertion that part of the process of establishing research authenticity entails the involvement of members not just in the interpretation, but also the design of research.

It is not simply a matter of the inappropriateness of the putative validators; there is also a logical contradiction inherent in the procedure. On the one hand, Lincoln and Guba espouse a relativist approach which accepts that there are different and equally valid perspectives on reality; on the other hand, they posit an epistemological foundationalism that aspires to sort out more trustworthy interpretations of reality from those which are less trustworthy (Smith 1993). They cannot have it both ways; it is only to the extent that they abandon radical perspectivism and take on board realist assumptions that the notion of judging research credibility makes sense.

It might be argued in Lincoln and Guba’s defence that, from a phenomenological perspective, because it is people’s perspectives that constitute reality rather than the things they perceive, it is perfectly appropriate to use those differing perspectives as evidence by which to judge research. Unfortunately, as Sandelowski (1993) notes, if we accept reality as multiple and constructed by members, then ‘repeatability is not an essential (or necessary or sufficient) property of the things themselves’ (p. 3). From a relativist position, attempts to force consensus between research members and between members and researchers does violence to the multiplicity of reality and is therefore an incoherent project. Once again, to maintain coherence, we are forced back to acceptance of a realist position that some perspectives capture actions and events better than others and that robust criteria are required to judge between them.

The realist approach

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

The argument presented so far in relation to the various perspectives on validity has involved a process of elimination that has led us to the conclusion that alternatives to the replication perspective, by dint of their rejection of realism, cannot provide a sufficiently rigorous basis to inform nursing action. As Seale (1999) points out, ‘The attempt to use language to refer to, describe or explain aspects of the social world (even if these aspects are the uses made of language in certain contexts) is a basic commitment for qualitative researchers and must ultimately depend upon some modified form of realism’ (p. 157).

Thus far, the confidence criterion has been used as an exclusionary tool. As such, it is necessary but not sufficient. While the remaining option of the replication perspective may claim validity, can it deliver it? The short answer is that yes it can, to a degree – only to a degree, because it is not possible for research to be absolutely rigorous and accurate, so as to give those using it total confidence that the actions it predicates will have entirely foreseen consequences. The world is far too complex a place to allow for such naivety. Realists accept that, rather than there being the possibility of a single and unproblematically valid representation of a phenomenon, ‘multiple valid descriptions and explanations of the same phenomenon are always available’ (Hammersley 2004, p. 243), and that all knowledge is socially produced and therefore influenced by the power relations obtaining in the social matrices in which it is produced (Bhaskar 1989). However, they reject the assumption that all beliefs are equally valid and that there are no rational grounds for preferring one to another.

While there is no ‘golden key’ to judging validity or rigour, robust procedures have been developed to help knowledge-based practitioners ascertain whether or not a knowledge claim can provide them with sufficient confidence to base their practice upon it. For example, Pawson et al. (2003) have developed a set of criteria under the acronym of TAPUPAS that has the merit of not restricting itself to validity, but including other pertinent issues relating to rigour such as ethics and accessibility:

  • • 
    Transparency: is the process of knowledge generation open to outside scrutiny?
  • • 
    Accuracy: are the claims made based on relevant and appropriate information?
  • • 
    Purposivity: are the methods used fit for purpose?
  • • 
    Utility: are the knowledge claims appropriate to the needs of the practitioner?
  • • 
    Propriety: has the research been conducted ethically and legally?
  • • 
    Accessibility: is the research presented in a style that is accessible to the practitioner?
  • • 
    Specificity: does the knowledge generated reach source-specific standards?

It should be noted that one of the benefits of TAPUPAS is that it is not just applicable to the judgement of individual research reports (which for pragmatic reasons are often not used by practitioners to inform their practice), but can be equally applied in the conduct of meta-syntheses or systematic reviews of evidence (Coren & Fisher 2006), which are increasingly becoming an important source for evidence-based practice.

Conversely, there are also procedures to help qualitative researchers ensure the credibility of their findings. Thus, for example, Silverman (2000) recommends five approaches to improve the validity of qualitative work: application of the ‘refutability principle’, use of the constant comparative method, comprehensive data treatment, deviant-case analysis, and use of tabulations. Such procedures do not hold out the promise of absolute validity, but do provide the vehicles for bolstering our confidence about the degree to which ‘an account accurately represents the social phenomena to which it refers’ (Hammersley 1990, p. 57).

What of the fact that different people will have different perspectives on the same phenomena? As has already been noted, rather than seeing the existence of multiple voices as a barrier to validity, realists such as Pawson and Tilley (1997) make a virtue out of perspectivism. In recommending realism as the basis for evaluation research, they argue that a major aspect of promoting the validity of evaluation studies is to recognize that programmes or interventions will be viewed differently from the different perspectives of the different stakeholders involved. Thus, if we consider a healthcare intervention, then intervention formulators, policy-makers, managers, clinicians and clients will all have their different ‘take’ on its effectiveness and the factors that promote or inhibit that effectiveness. Validity requires all of these perspectives to be taken into account, while accepting the limitations of any single perspective. Moreover, rather than seeing an intervention in isolation as working or not, their depth realist approach allows Pawson and Tilley to recognize that interventions are human activities and that understanding them requires an understanding of the social mechanisms at work in the contexts within which interventions are implemented.

Conclusion

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References

While I have included some examples of procedures that can be used to produce and ascertain rigour and validity within the rubric of a realist approach, my purpose has not been to assert that these procedures represent a comprehensive and permanent solution to these vexing problems. Rather, it has been to point to approaches which permit us to take qualitative research rigour seriously. These approaches need to take into account at least two factors. First, unless we accept that qualitative research is about something, then it makes no sense; and if it is about something, then researchers and readers have a responsibility to ensure that its accounts of that something are as accurate as possible. Second, nurses (and other professions who engage in research-based action) have their own criterion by which to judge competing approaches to qualitative research, namely the capacity to beneficially inform action. Thus, the abandonment of knowledge in favour of aesthetics or multiple opinions are not approaches which can provide useful guidance either to researchers or practitioners.

References

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. Introduction
  5. What is qualitative research?
  6. Devaluation of epistemology
  7. The aesthetic elite
  8. Relative methodologies
  9. Assessing methodologies
  10. Approaches to validity
  11. The realist approach
  12. Conclusion
  13. References
  • Badger F. & Werrett J. (2005) Room for improvement? Reporting response rates and recruitment in nursing research in the past decade Journal of Advanced Nursing 51(5), 502510.
  • Bhaskar R. (1989) Reclaiming Reality: a Critical Introduction to Contemporary Philosophy. Verso, London.
  • Bloor M. (1978) On the analysis of observational data: A discussion of the worth and uses of inductive techniques and respondent validation. Sociology 12(3), 545557.
  • Bloor M. (1983) Notes on member validation. In Contemporary Field Research: A Collection of Readings (EmersonR., ed.), Little Brown, Boston, pp. 3245.
  • Blyth R. (1963) A History of Haiku, Vol. 1. Hokuseido Press, Tokyo.
  • Coren E. & Fisher M. (2006) The Conduct of Systematic Research Reviews for SCIE Knowledge Reviews. Social Care Institute for Excellence, London.
  • Ellis C. (1995) Final Negotiations: A Story of Love, Loss and Chronic Illness. Temple University Press, Philadelphia.
  • Fielding N. & Fielding J. (1986) Linking Data. Sage Beverly Hills, CA, USA.
  • Foucault M. (1977) Discipline and Punish: The Birth of the Prison. Pantheon, New York.
  • Guba E. & Lincoln Y. (1989) Fourth Generation Evaluation. Sage, Newbury Park.
  • Hammersley M. (2004) Some reflections on ethnography and validity. In Social Research Methods: A Reader (SealeC., ed.), Routledge, London, pp. 241245.
  • Hammersley M. (1990) Reading Ethnographic Research: A Critical Guide. Longman, London.
  • Hope K. & Waterman H. (2003) Praiseworthy pragmatism? Validity and action research. Journal of Advanced Nursing 44(2), 120127.
  • Junghans C., Feder G., Hemingway H., Timmis A. & Jones M. (2005) Recruiting patients to medical research: double blind randomized trial of ‘opt-in’ versus ‘opt-out’ strategies. British Medical Journal 331, 940942.
  • Lather P. (1993) Fertile obsession: validity after poststructuralism. The Sociological Quarterly 34, 673693.
  • Lincoln Y. & Guba E. (1985) Naturalistic Inquiry. Sage Thousand Oaks, CA.
  • Miles M. & Huberman M. (1994) Qualitative Data Analysis: An Expanded Source Book. Sage, London.
  • Miller K. McKeever P. & Coyte P. (2003) Recruitment issues in healthcare research: the situation in home care. Health and Social Care in the Community 11, 111123.
  • Morse J., Barrett M., Mayan M., Olson K. & Spiers J. (2002) Verification strategies for establishing reliability and validity in qualitative research. International Journal of Qualitative Methods 1(2), 119.
  • Parkes J., Kerr C., McDowell B. & Cosgrove A. (2006) Recruitment bias in a population-based study of children with cerebral palsy. Pediatrics 118, 16161622.
  • Pawson R. & Tilley N. (1997) Realistic Evaluation. Sage, London.
  • Pawson R., Boaz A., Grayson L., Long A. & Barnes C. (2003) Types and Quality of Knowledge in Social Care. Social Care Institute for Excellence, London.
  • Porter S. (2001) Nightingale’s realist philosophy of science. Nursing Philosophy 2(1), 1425.
  • Porter S. (2002) Critical realist ethnography. In Qualitative Research in Action (MayT., ed.), Sage, London, pp. 5372.
  • Rolfe G. (2006) Validity, trustworthiness and rigour: quality and the idea of qualitative research. Journal of Advanced Nursing 53(3), 304310.
  • Sandelowski M. (1993) Rigor or rigor mortis: the problem of rigor in qualitative research revisited. Advances in Nursing Science 16(2), 18.
  • Sandelowski M. (1997) ‘To be of use’: enhancing the utility of qualitative research. Nursing Outlook 45(May/June), 125132.
  • Sandelowski M. & Barroso J. (2002) Reading qualitative studies International Journal of Qualitative Methods 1(1), Article 5, Retrieved from http://www.ualberta.ca/~ijqm/ on 15 January 2007.
  • Seale C. (1999) The Quality of Qualitative Research. Sage, London.
  • Silverman D. (2000) Doing Qualitative Research. Sage, London.
  • Smith J. (1993) After the Demise of Empiricism: The Problem of Judging Social and Educational Inquiry. Ablex, Norwood, NJ, USA.
  • Sparkes A. (2001) Myth 94: qualitative researchers will agree about validity. Qualitative Health Research 11(4), 538552.
  • Weber M. (1968) Economy and Society, Vol. 1. University of California Press, Berkeley.