Understanding reliability and validity in qualitative research

The article suggests that reliability and validity are common in quantitative research and that they must be redefined so that they can be used in qualitative research.

The article starts off by discussing “What is Quantitative research” and suggests this style of research will use experimental methods and quantitative measures to test hypothetical generalizations. The emphasis is on measurement and analysis of causal relationships between variables.

Focus is on concepts and the generation of hypotheses which will be tested. The emphasis is on facts, causes of behavior, measuring and analysis using numbers and the use of statistical terminologies, normally through the use of some standardized instrument.(AL: I would agree that this overview is correct)


A discussion of reliability and validity in quantitative research follows. Reliability is concerned with reproducibility and stability of the measuring instrument. Validity is concerned with how close to the actual ‘truth’ the results are. The validity of an experiment may not be answered in the study, but rather in the follow up studies based on that experiment.

“Qualitative research” is then discussed as an approach that tries to understand phenomena in a ‘natural’ or ‘real-world’ setting. (Al: This reminds me of the observational studies such as case-control and cohort).

The author suggests that quantitative research is focused on determination, prediction and generalization of findings which qualitative research is focused on illumination, understanding and situation extrapolation. Nahid clearly see two different directional results coming from this two methods.( Al: Well, I see the authors point, but would not think of the distinction as that clear cut).


Reliability:
Reliability in qualitative research, as suggested by some authors, is focused on “generating understanding” rather than “purpose of explaining”. Some authors even suggest that reliability is irrelevant in qualitative research. ( Al: This is a bit difficult to accept)

Patton does suggest that reliability and validity are worthy of concern (Al: Makes much more sense).Lincoln and Guba equate the term reliability with “dependability”. Clont and Seale use the term “consistency” with verification through examination of (1) raw data, (2) data reduction products ( 3) process notes.

References from this section further suggests that reliability and validity are bond together ( AL: I agree).

Validity:
Validity is concerned to have a very broad and wide ranging description. Again some qualitative researches have argued that validity has not meaning in this type of research but do see the need for some qualifying check on their research. (AL: This of course would be necessary, especially if a third party was funding the project). Validity is seen as providing “confidence in the finding”

Both reliability and validity come together through the use of the term “Trustworthiness”.

Testing Validity and Reliability:
Triangulation (Typically used to obtain substantiation of findings through convergence of different perspectives) (AL: Interesting, in my own area of biostatistics this is something that I have not used).

Triangulation is also used in qualitative research methodologies and is suggested as a method of testing validity and reliability. Barbour does argue that triangulation be used with the understanding that any exception should lead to a disconfirmation of the hypothesis (as opposed to quantitative research were exceptions are useful and dealt with).

Triangulation (with the appropriate method selected on a per study basis) is then purposed as the method of choice to test reliability and validity. (Al: this has also the useful feature of using a technology that bridges qualitative and quantitative research).