Quality
All research is rightly open to scrutiny from its readers. Health researchers too must not only consider the ‘truth value’ of their studies but also demonstrate that it is credible and valid for professional practice, and that it has quality. Ways of reflecting on validity have reached a variety of conclusions. No single or unitary concept of validity exists in qualitative research which is comparable to its meaning in quantitative inquiry, and as Onwuegbuzie and Leech (2007: 233) state: ‘to date, no one definition of validity represents a hegemony in qualitative research’; no one idea of validity dominates.
There are several distinct perspectives on the quality of qualitative research (Murphy et al., 1998), some of which are listed below. Some believe that
- qualitative and quantitative research should be evaluated by the same criteria;
- qualitative research should be evaluated by criteria that have been specially developed for it;
- criteriology should be rejected.
One group, for instance Maxwell (2005) and Silverman (2006) among many others, argues for the retention of the criteria of reliability and validity while arguing, at the same time, that these criteria cannot be directly translated from quantitative to qualitative research because qualitative inquiry has its own criteria by which it can be evaluated. Indeed validity in qualitative research has different implications and applications. Others (see Seale, 1999), reject evaluative criteria as inappropriate for qualitative inquiry and stress that contextualisation is most important. Although qualitative researchers are flexible and open minded, they are also advised to be systematic and well organised and through this the research gains validity.
Proponents of another group follow the ideas. Lincoln and Guba (1985 initially) and Guba and Lincoln (1989) developed the concepts of trustworthiness and authenticity as parallel and alternative criteria. Researchers will come across both groups of terms during their reading and therefore will have to know about them regardless of the terms they themselves apply. However, a simplistic stance is sometimes taken, and concepts developed here are complex. Different qualitative approaches often take a variety of viewpoints on criteria of quality (Whittemore et al., 2001). Sparkes (2001), too claims that there is no shared understanding of ‘good’ qualitative research. Researchers find difficulty agreeing on how to judge the ‘validity’ of qualitative research or how to present convincing evidence of its trustworthiness.
Conventional criteria
We will discuss the traditional criteria generally used in quantitative research, their meaning in qualitative inquiry and their alternatives. Trustworthiness and authenticity are more often used than validity and reliability in qualitative healthcare research, and they are discussed in detail later in the chapter.
- Rigour – trustworthiness
- Reliability – dependability
- Validity – credibility
- Generalisability (external validity) – transferability
- Objectivity – confirmability
Porter (2007) claims that standardising and establishing acceptable criteria for evaluating qualitative research enables writers to create mediating tools by which this type of inquiry can be judged by all readers including those who carry out quantitative research. This means that some form of ‘criteriology’, might be acceptable although criticised by some such as Seale (1999) or Sparkes (2001).
Rigour
The concept of rigour has its origin in science, and quantitative researchers use it because of its particular connotations with measurement and objectivity; hence it has a more appropriate place in quantitative research. In qualitative research rigour indicates thoroughness and competence. Sandelowski (1986, 1993) wrote two classic articles on rigour in qualitative nursing research. Her latter article recognises that the term rigour could imply inflexibility and rigidity and that researchers should not be too preoccupied with it. Instead she advises they should create ‘evocative, true-to-life and meaningful portraits, stories and landscapes of human experience … ’ (p. 1), and she criticises ‘the reduction of validity to a set of procedures’ (p. 2). Indeed, excessive rigour may hinder creativity and artistry (Bradbury-Jones, 2007).
Reliability
Reliability in quantitative inquiry refers to the consistency of the research instrument. It is also linked to replicability, that is, the extent to which the study is repeatable and produces the same results when the methodology is replicated in similar circumstances and conditions. As the researcher is the main research instrument in qualitative inquiry, the research can never be wholly replicable. Other investigators have different emphases and foci, even when they adopt the same methods and select a similar sample and topic area.
The researcher’s characteristics and background will also influence the research.
Validity
Validity in quantitative research is seen as the extent to which an instrument measures what it is supposed to measure. In qualitative research the concept is more complex. Description and interpretation by researchers and truth telling by participants are all important.
One of the threats to validity is posed when collecting incorrect or incomplete data. The field diary must therefore be detailed and extensive. In interpretation, researchers are in danger of imposing their own ideas or distorting the meaning of the participants’ accounts. Therefore it is important for the researcher to listen to the participants’ voices and let them speak. Researchers hope that the stories of the participants are true; they do occasionally make mistakes or tell deliberately lies, though the latter seems to be rare. However, this does not mean that there is no truth as the participants describe their world as they see it from their own perspective in the context of their time and culture as well as their own biography. Researchers generally trust their participants even if they cannot prove the ‘truth’ of their tales. The description by the researcher should not only be plausible but also trustworthy. Researchers set aside their own thoughts and preconceptions about the phenomenon under study at some stage. Alternative and rival explanations to the researchers’ own initial interpretation should be taken into account. Although researchers can never be fully certain that all threats to validity have been eliminated, awareness of these threats helps produce a valid piece of research.
To the term validity Hammersley (1998) adds that of relevance as a criterion for evaluating qualitative research. Relevance means that explanatory factors should have significance related to the purpose of the research and in solving the problems of practitioners in the discipline. The research must not only be meaningful but also useful for those who undertake it.
Internal validity is the extent to which the findings of a study are true, and whether they accurately reflect the aim of the research and the social reality of those participating in it. This can be established to an extent by taking the findings back to the participants (see the section on member check later in this chapter). The researchers can compare their own findings with the perception of the people involved and explore whether they are compatible. Bryman (2008: 376) adds ‘the match between the researchers’ observations and the theoretical ideas which they develop.’ External validity, also called generalisability, is described in the next section.
Generalisability or external validity
This is the most contentious concept linked to validity. For some authors generalisability is not an issue to be discussed at length for they believe it is not relevant as they speak of specific situations and cases. For others, however, it is problematic. Most funding agencies and research committees in the UK National Health Service demand that the proposed research be generalisable, and this is understandable. If large amounts of money are given to researchers, funding bodies wish to know whether the outcomes are of general use in clinical practice and not just the results of ‘blue skies’ research undertaken for its own sake or only applicable to specific situations.
Generalisability exists when the findings and conclusions of a research study can be applied to other similar settings and populations, i.e. when they can be generalised across a variety of settings. The term has its origin in quantitative research with its random statistical sampling procedures. Random sampling ensures that the results of the research are representative of the group from which the sample was drawn. It is clear that this type of generalisability cannot be achieved in qualitative research in which sampling is purposeful or, in grounded theory, theoretical.
Generalisability is difficult to achieve in qualitative research. Positivist and interpretive research differ in the sense that positivists seek law-like generalities while interpretivists focus on unique cases even though they might want to establish patterns. As much quantitative research – though by no means all – is carried out in the positivist tradition and uses deductive methods, it can be more easily generalised. Many qualitative researchers, however, do not aim to achieve generalisability as they focus on specific instances or cases not necessarily representative of other cases or populations. The case(s) may even be atypical. Indeed the concept of generalisability is irrelevant if only a single case or a unique phenomenon is examined. For instance, a nurse or physiotherapist may want to examine a particular phenomenon important for local practice and patients in a particular area rather than of interest to the whole country. However, the study can still be successful, because it highlights specific non-typical features that can be related and compared to those of other, more typical cases.
Many qualitative researchers attempt to achieve some generalisability, however, because they feel that their research should be useful beyond their own studies. Strauss and Corbin (1998) speak of the representativeness of concepts and applicability of theory to other situations. This means that qualitative research can have external validity through ‘theory-based generalisation’. Morse (1994) claims that theory contributes to the ‘greater body of knowledge’ when it is re-contextualised into a variety of settings. It involves the application of theoretical concepts found in one situation to other settings and conditions. If the theory developed from the original data analysis can be verified in other sites and situations, the theoretical ideas are generalisable. The findings from multi-site studies are, of course, more easily generalisable than those from a few unique cases from one setting.
Objectivity
This is a term often used in quantitative research. This means that the research is free of biases and relatively value neutral. Qualitative researchers do not find this concept very useful. Objectivity and neutrality are difficult to achieve; in fact, the values of researchers and participants become an integral part of the research, and they must openly acknowledge their own subjectivity. They do not conceal it but examine and then set it aside. Critical subjectivity, a term originally coined by Carr and Kemmis (1986) and later developed by other writers such as Reason and Heron (1995), is useful here. Although much knowledge is based on subjective experience it should not be accepted in a simplistic way but rooted in critical consciousness. Researchers do not disregard their subjectivity, they are aware of it and attempt to have self-reflexivity, so no prior assumptions can introduce bias in the study.
Validity in various approaches to qualitative research
The concept of validity is used in phenomenological research, but its meaning and the way in which it is ensured is less precise and prescriptive than in other forms of qualitative research. For instance, Dahlberg et al. (2001) state that the research report should not contain any internal contradictions if the researcher wants it to be seen as valid.
Research can be valid through intersubjective knowledge. Moustakas (1994) speaks of ‘intersubjective truth’. He states (p. 57) that according to Husserl ‘each can experience and know the other, not exactly as one experiences and knows oneself but in the sense of empathy and copresence’. Initially truth is based in the unique perspective of unique individuals and their self-knowledge. As individuals inhabit the world of self and others, there is also communication with others. This enhances intersubjective understanding. If the research is to have validity, its readers will have learnt something of the human condition as well as recognise and grasp the essence of the phenomenon under study. This form of ‘validity’ is similar to, though not the same as the concept of ‘ontological authenticity’ described by Guba and Lincoln (1989) or that of ‘thick description’ by Geertz (1973).
In phenomenology and a number of other approaches such as grounded theory, internal validity (being faithful to the ideas of the participants) is a complex concept as the researchers always transform the data and take them to a different level from that of the participants when they describe the phenomenon or interpret the ideas of the participants. The researchers’ ideas are more abstract and theoretical than those of the participants, and ultimately the researchers’ description and interpretation is presented to the readers of the account, though they are grounded in the participants’ thoughts and feelings. Lomberg and Kirkevold develop the ideas about validity in grounded theory, namely those of fit, relevance, and modifiability (see details in Lomberg and Kirkevold, 2003). Hope and Waterman (2003) discuss the re-conceptualisation of validity in action research, stressing the importance of contextualisation and rigorous application of the chosen approach. (It is not possible to develop ideas on validity in each approach, but we hope that researchers might gain more details from the references).
An alternative perspective: trustworthiness
It can be seen that the conventional terms used in quantitative research have different meanings in qualitative inquiry. Guba and Lincoln (1989), as stated before, go further than this and develop alternative terms and criteria. We will show how health researchers can attempt to demonstrate trustworthiness in the last section of this chapter.
Trustworthiness
Trustworthiness in qualitative research means methodological soundness and adequacy. Researchers make judgements of trustworthiness possible through developing dependability, credibility, transferability and confirmability. The most important of these is credibility.
Dependability
Lincoln and Guba (1985; Guba and Lincoln 1989) use the term dependability instead of reliability. If the findings of a study are to be dependable, they should be consistent and accurate. This means that readers will be able to evaluate the adequacy of the analysis through following the decision-making processes of the researcher. The context of the research must also be described in detail. To achieve some measure of dependability an audit trail is necessary. This helps readers follow the path of the researcher and demonstrates how he or she achieved their conclusions. It also guides other researchers wishing to carry out similar research. Although the study cannot be replicated, in similar circumstances with similar participants, it might be repeated.
Credibility
Credibility corresponds to the notion of internal validity (see p. 252). This means that the participants recognise the meaning that they themselves give to a situation or condition and the ‘truth’ of the findings in their own social context. The researcher’s findings are, at least, compatible with the perceptions of the people under study.
Transferability
Lincoln and Guba use transferability instead of generalisability. This means that the findings in one context can be transferred to similar situations or participants. The knowledge acquired in one context will be relevant in another, and those who carry out the same research in another context will be able to apply certain concepts originally developed by other researchers. It seems to us that the concepts of transferability and generalisability are not too different.
Confirmability
Confirmability has taken the place of the term objectivity. As the research is judged by the way in which the findings and conclusions achieve their aim and are not the result of the researcher’s prior assumptions and preconceptions, Lincoln and Guba demand ‘confirmability’. This again needs an audit or decision trail where readers can trace the data to their sources. They follow the path of the researcher and the way he or she arrived at the constructs, themes and their interpretation. For this, details of the research and the background and feelings of the researcher should be open to public scrutiny. When confirmability exists, readers can trace data to their original sources. Dahlberg et al. (2001) also demand intellectual honesty and openness from the researcher, as well as sensitivity to the phenomenon under study thus incorporating the idea of the audit trail although they do not explicitly call it this.
Authenticity
Trustworthiness, which relies on the methodological adequacy of the research, does not suffice according to Guba and Lincoln (1989), and therefore they add the concept of authenticity. A study is authentic when the strategies used are appropriate for the true reporting of the participants’ ideas. Authenticity consists of the following.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
