Are reliability and validity important in all cases?

Home Forums Methodspace discussion Are reliability and validity important in all cases?

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
  • #917
    Saleh Ali

    Are reliability and validity important if:

    1- I use one of the quantitative methods and run sort of statistical analysis just to explain some of the findings that I have already come up with using thematic analysis.

    2- The epistemological stance I am adopting is constructivism.

    Your help is highly appreciated. 


    i guess it depends on which paradigm of research you’re more accustomed to. for a quantitative approach then yes, they’re both important. for a qualitative approach… well. validity is definitely an issue but reliability is a statistically-defined quantity i’m not entirely sure whether it makes sense or not outside of the paradigm of quantitative analysis. 

    Stephen Gorard

    Sorry, but this is a common confusion. Research is research (being curious and finding stuff out). We do what we need to get the job done (i.e. to answer pre-specified RQs). The pointlessly divisive Q- word terms such as ‘quantitative’ do not denote paradigms. I explain what a paradigm is and isn’t for students in the following (attached):

    Gorard, S. and Makopolou, K. (2012) Is mixed methods the natural approach to research?, pp.106-119 in Armour, K. and Macdonald, D. (Eds) Research Methods in Physical Education and Youth Sport, London: Routledge

    Every researcher will use numbers, text and probably a host of other forms of data. Their analyses will all require judgement. 

    You cannot just ‘adopt’ a stance. This is not a fashion pick! It is right to use numbers when doing thematic analysis. Just don’t make the mistake of using daft stuff like significance tests. But frequency counts, averages, correlation, logistic regression and many other techniques are at your disposal.

    Some commentators get very specific about validity and reliability. Some offer different terms for the same thing when using different approaches (needlessly confusing). But if you wonder whether your data is picking up the concepts you wanted, whether the participants understood the same as you…. this is about’ ‘validity’ – whether you call it that, nothing or something else. If you wonder whether the same participants would respond similarly again, this concerns reliability (not possible in reality since asking the first time could influence them on the next round). If you wonder whether another set of participants would respond similarly this is more to do with generalisability. The key issue for a researcher is to wonder these things. What to call them is trivial. 

    Research is easy. Don’t make it complicated for no reason. 


    I think validity and reliability are matter, because the use of different measures can produce different results. The modern measurement theory argues that good indicators should be reliable and valid. Reliability refers to the consistency of the measure and the degree to which the measure produces consistent results. Validity refers to whether the indicator measures what it is supposed to measure (see Kline, Rex B, 1998, Principles and Practice of Structural Equation Modeling). In other words, reliability is the absence of random unsystematic error and validity is being free from any systematic or non-systematic errors (see Bartholomew, David J. 1996, The Statistical Approach to Social Measurement).



    If you are writing a dissertation the qualitative counterparts of validity and reliability will be important to your committee.


Viewing 5 posts - 1 through 5 (of 5 total)
  • You must be logged in to reply to this topic.