I am currently doing research on validity and reliability of qualitative research. Appreciate if you could share with me how you validate your qual data and how you support the reliability of your data?
Over the years, there has been a split between those who take traditional quantitative criteria, such as reliability and validity, and apply them to qualitative research, versus those who feel that this kind of reasoning does more harm than good. Like Dean, I think the general consensus is currently against using literal ideas of reliability and validity in qualitative research.
For me, the most relevant argument is that qualitative research is generally used for generating theory rather than testing it. In quantitative research, validity can be about having accurate measurements (e.g., in surveys) or ways to assess cause and effect (e.g., in experimental designs). But neither measurement nor hypothesis testing is part of qualitative research. In other words, the criteria that were developed to work for quantitative purposes are probably not a good fit to the very different purposes of qualitative research,
As a general reference, the ideas of trustworthiness and credibility were introduced by Lincoln & Guba in their 1985 book "Naturalistic Inquiry." They have several recommendations for enhancing these quality criteria, including triangulation.
We tend not to talk about validity and reliability within qualitative research. They are terms more associated with quantitative research. Better to use constructs such as 'trustworthiness' or 'credibility'. That said, 'face' validity is sometimes used with qualitative studies.
The opinions on this subject are quite divided within qualitative research. However, I believe that research quality criteria like validity and reliability are equally applicable to qualitative research, if you understand and use them appropriately. For example, I find that triangulation across both data sources and researchers (analytical triangulation) is very helpful in increasing the credibility of your results. The same could be said for respondent validation, which I often use during interviews to gain respondents' view on findings from a previous interview or from observations I have conducted. When it comes to reliability, this can be achieved through systematic operation at the level of the research design (methods and techniques, interview protocol, and so forth) (see de Ruyter & Scholl, 1998). For example, the systematic steps taken towards developing a concept or theory.
Over the years, there has been a split between those who take traditional quantitative criteria, such as reliability and validity, and apply them to qualitative research, versus those who feel that this kind of reasoning does more harm than good. Like Dean, I think the general consensus is currently against using literal ideas of reliability and validity in qualitative research.
For me, the most relevant argument is that qualitative research is generally used for generating theory rather than testing it. In quantitative research, validity can be about having accurate measurements (e.g., in surveys) or ways to assess cause and effect (e.g., in experimental designs). But neither measurement nor hypothesis testing is part of qualitative research. In other words, the criteria that were developed to work for quantitative purposes are probably not a good fit to the very different purposes of qualitative research,
As a general reference, the ideas of trustworthiness and credibility were introduced by Lincoln & Guba in their 1985 book "Naturalistic Inquiry." They have several recommendations for enhancing these quality criteria, including triangulation.
Very well put David - I should have tried harder with my earlier tentative and half-hearted response!! Personally, I like Schou et al.'s (2012) [excuse the fact that it is nursing] VAKS tool (Danish anacronym for Appraisal of Qualitative Studies) which is based on Guba's (1981) four seminal criteria (credibility, auditability, fittingness, and confirmability). Through this work and others, trustworthiness can be viewed from six broad positions; using the criteria of quantitative research (which does constitute reliability and validity - but is one of the more minor positions of the six), parallel methodological criteria, multiple criteria, fresh and universal criteria, each study develops suitable, justifiable criteria - through to the last position which is 'no criteria is necessary'.
I agree with Dean Whitehead that reliability and validity are generally not qualitative terms and can be replaced with credibility, transferability and confirm ability Creswell, J. W., (2009). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches. Sage Publications is a good easy to read text. I also agree with David Morgan that triangulation of both Qualitative and Quantitative allows you to utilize all criteria's to substantiate your findings. Although, I would not suggest undertaking Quantitative Research unless well versed. It has destroyed many a researcher in an unprepared attempt.
I similarly approach questions of research quality and notions of trustworthiness in qualitative research in the ways discussed above by David and Dean. My recommendation is not to fall into the trap of 'operationalising' trustworthiness but to plan how trustworthiness can be achieved in your research design. There is a very useful chapter in Schwarz-Shea and Yanow's book on Interpretive Research Design that explores trustworthiness. They examine it in the context of research design (contrasting interpretive qualitative approaches with positivist research approaches).
Schwarz-Shea, P and Yanow, D (2012) Chapter 6 Designing for Trustworthiness. Knowledge Claims and Evaluations of Interpretive Research. In "Interpretive Research Design. Concepts and Processes. Abingdon: Routledge.
Hi, the concepts reliability and validity have often been viewed as alien to qualitative research and the concepts of rigour, trustworthiness, dependability, credibility... so on being used. I have found Miles and Hubermans (1994) book on Qualitative Data Analysis from a critical realist perspective very helpful in bridging this gap between the varying concepts. See also Sam Porters (2007) in the Journal of Advanced Nursing titled 'validity, trustworthiness and rigour...
As has already been pointed out, reliability and validity are terms most associated with quantitative research. The qualitative counterparts are as follow:
Thanks Ruth - I agree that comparisons can be made but, depending on context and situation, sometimes the comparison can feel like square pegs and round holes.
One slight oversight I've just noted. Despite the fact that it can be viewed as a 'loose' and subjective criteria 'face validity' is often used to evaluate qualitative findings. I'm not a fan myself. It's akin to 'if it looks OK - then it is' - which is hardly robust.
When repeating the experiment, researcher should get same or similar results which depends on items you used in your study. You do research using a target sample. You collecting data from this selected sample. When you collect sample for the second time, you can collect data from some other sample. Reliability means results from these two tests should be same or similar. In SPSS, reliability is how items are relaiable. This is calculated by Cronbach Alpha value that is higher than 0.6.
Validity refers to researh should measure what it wants to measure. There are face validity (content validity), criterion validity and construct validity. Face validity refers to questions or statements should be suitable for the research question and objective. Criterion validty refers to appropriate criterion variable should be selected along with personal profile variables.
Construct validity refers to constructs selected for the study should be congruent with each other. There are convergent validity and discriminating validity. In quantitative research, correlation and other techniques are used to describe these validities.
Mohamed - I don't think that you read the original question. Zaleha is enquiring about qualitative research methodology - you are purely discussing quantitative issues.
first and foremost, thank you so much dean, david, sindre, sandra, susan, michael, ruth, mohamed. Highly appreciate your participation. As a qualitative research, I find the words 'validity and reliability' are not suitable to descrip our kind of work. Too positivist, i think. Can I know your view on this and if you have similar view as me, can you suggest the most appropriate word(s) to describe qualitative work that could support the 'validity and reliability'. I prefer to describe my work with the term trustworthiness - just to keep pace with the rhetorical assumption, that's all.
Zaleha, a realist approach to qualitative research supports the use of the concepts of validity and reliability. Please see Miles and Huherman's work and the application of critical realism to research - notably, the works of Derek Layder; Margaret Archer and Pawson and Tilly; Anrew Sayer. I have used Layder's work for my PhD
Thanks Michael. Will look at the works of Derek Layder, never come across his name but interested to find out. Thank you so much.
One of the finding emerged in my research on validity and reliability of qualitative study is that 'the need to have some kind of accuracy measure to valid the data'. As a qualitative researcher, do you agree with that view? appreciate if I could get some feed back on that.
As has already been pointed out, the terms “reliability” and “validity” are terms that are associated with quantitative research and, when attended to, they contribute the “robustness” of quantitative research. Validity is not something that qualitative researchers are concerned with because their goal is to understand phenomena from the perspectives of the people who experience them, and not so much to generalize their findings (Trochim, 2006). Robustness in qualitative research has to do with the quality of the research, and its trustworthiness. In qualitative research, therefore, counterparts to validity are (a) credibility, which is referred to as “internal validity” in quantitative research; (b) transferability (external validity in quantitative research); dependability (reliability in quant research), and confirmability (construct validity in quant research) (e.g. Miles & Huberman, 1994). These address the integrity of the research.
Qualitative research seeks to understand phenomena from the perspectives of those who are experiencing them; therefore, research participants are the ones who determine the credibility of the research. If the research findings can be transferred to other settings or contexts, it is said to be transferable. Dependability has to do with how the researcher accounts for, and deals with, unexpected occurrences or situations that change during the course of the research. Finally, the degree to which qualitative research findings can be confirmed by other researchers, it is said to have confirmability (Trochim, 2006). Credibility, transferability, dependability and confirmability seek to add robustness to qualitative research, something that quantitative researchers maintain it lacks.
Please check out the Trochim site – it was recommended in one of my research classes and it was invaluable in helping me to understand these concepts. Hope this helps!
Ref.
Miles, M.B. and A.M. Huberman 1994, Qualitative data analysis: An Expanded Sourcebook of New Methods, (2nd ed.), Sage, Thousand Oaks.
Trochim, W.M.K. (2006a). Research methods knowledge base: Qualitative validity. Retrieved from http://www.socialresearchmethods.net/kb/qualval.php
Apologies for late reply Zaleha but I do agree with view that you can use a measure to support your qualitative data. As you maybe aware there are so many approaches and philosophical underpinnings that support the many qualitative approaches that a researcher can use and depending on your position (ontologically and epistemologically) you can defend such an approach to measure your propositional statements that arise from your analysis. A realist perspective,for example the work of Layder facilitates this approach. Also, the use of NVivo allows you to validate your propositional statements in terms of the data.This position is in contrast to purist qualitative approaches i.e constructivist/relativism.Such is the challenge for the researcher in the diversity of Qualitative research designs and the philosophies that support them.
Great discussion! Along with trustworthiness, credibilty and robustness - another idea to consider is transparency. Can you show how you arrived at your findings? Did you keep an audit trail of your methods and decisions - that are closely tied to the data? This recent paper by Kristi Jackson is a useful starting point
This is a great discussion! I am an expert in the field of qualitative research.
Reliability and validity are so important to consider! Qualitative methodologies, like any field, are evolving and ever changing. What was available to us previously is much different that what is available to us, present day.
In the social sciences, qualitative methods are designed to help researchers understand the qualities of, or meaning in, a statement, event, or other phenomenon. Qualitative methods tend to involve sets of somewhat standardized procedures, in which researchers attempt to identify, clarify, and (to greater or lesser extent) classify the qualities revealed in their data.
By doing so, they intend to promote greater understanding of the meaning and relationships revealed through the words of their participants. Phenomenology is one such qualitative approach, as are for example Content Analysis, Discourse Analysis, Ethnography, Grounded Theory, Participatory Action Research, and Queer Theory. Raven’s Eye accommodates research from any of these and other approaches, but it is built on a quantitative approach to phenomenology.
Check out this link to the laboratories at Raven's Eye.
May some of you know an alternative definition of: 'content validity' and 'construct validity (divergent and discriminant patterns)' for qualitative research.
My main interest is examining the validity of a qualitative tool and I find all the alternative definitions of Lincoln and Guba focused more on the findings and not on the methods (tools).
Both content validity and construct validity are terms from quantitative research that refer to how effectively a variable measures an underlying theoretical construct. As much, they are not directly relevant for qualitative research.