validity

The term validity in research methods bears a strong relation to its every day meaning as something which is reasonable or justifiable in the circumstances, something which brings about the desired conclusion.

One sense in which the word is used is construct validity: do your methods address the research questions or hypothesis you set out to address. Very often construct validity will lead you into an exploration of a construct and its use in past research. For example the construct of engagement has been painstakingly explored over the years and different tests of engagement developed. The researcher needs to ask, ‘which if any of these tests are valid and what have past researchers said about these claims to validity?’.

Construct validity can be contrasted to reliability. An inventory designed to ask about learning styles will not be valid if you are using it to measure, say, intelligence. Construct validity is bound up with the ‘fitness of purpose’ of research methods. For example if you wanted to investigate the behaviour of learners using a new item of software you might learn a lot about their attitudes through surveys and interviews but will this help very much?

The term validity reappears in relation to the conclusions drawn within a study and inferences drawn as to interplay of variables and cause and effect. This is internal validity and is very much related to the concept of theorizing. In experimental studies cause and effect is often argued in relation to statistical tests of different kinds, the picture is more complicated in interpretive research where relations between factors may not so much be ‘proven’ or ‘shown’ but suggested. Interpretative findings tend to be so context rich that there might not be fewer grounds to expect generalisability though of course findings can still be compared to those reported by others.

As with so many of the concepts described in this book, validity will mean different things within different research traditions and is strongly aligned to ontological stance: if there is an objective world then validity may be seen as a state, this conclusion is valid, this is not. If the world is there to be interpreted, validity is perhaps a scale, this inference is defensible or credible rather than proven. Construct validity may further be questioned by interpretivists: fitness for purpose is clear cut but the purpose of the research unfolds over time and the concepts are always disputed. Researchers within this tradition have tended to speak of the credibility of their research – even if this does tend to revisit reliability and validity in a different guise. An account might be expected to be more credible if there has been a process of triangulation and moderation and, at least in regards to studies which lay claim to expressing the views of participants, a process of participant validation.