FEATURE26 January 2023

Tips on triangulation: A strategy to generate insights

Features Insight Alchemy Trends

In our series on how insight alchemy can produce deeper insight, Dr Marie-Claude Gervais, Director at The Foundation and a member of the MRS Main Board, explores how to triangulate data, perspectives, and methodologies to access a golden seam of insight.

Marie-claude crop

Thousands of books and articles are written on data collection and analysis. Yet, the process of transcending ‘data’ to reach ‘insight’ – that is, ideas that have genuine explanatory power and can galvanise action – remains somewhat mysterious.

There is no simple strategy to guarantee insight. But, in my experience, triangulation – the combination of multiple perspectives, data sources and approaches – significantly increases the likelihood of revealing true insights.

Triangulation: from a positivist to a social constructionist point of view

The original use of triangulation was as an aid to navigation in the Middle Ages. Three different data points were combined to infer a ship’s position. By multiplying different sources of information, one could arrive at a reasonably accurate, single truth about one’s position in the world. If that position proved wrong, one might fall off the edge of the world.

When initially transposed in the context of social research, triangulation maintained its positivist focus on increasing the accuracy, or validity, of findings. By ‘playing off each research method against the other, [one could] maximise the validity of field efforts’ (Denzin, 1978; 304 ), mainly by minimising the reactivity inherent in each method.

However, it soon became apparent that putting the whole picture together was not so easy: ‘what goes on in one setting is not a simple corrective to what happens elsewhere – each must be understood in its own terms’ (Silverman, 1985; 21 ).

From this perspective, triangulation is less about achieving a single valid truth, and more about generating insight into how people construct different versions of reality, in different contexts, to achieve different aims. Delving deep into each version of reality – and paying attention to both what is salient and absent in each context – is hugely thought-provoking.

Some practical steps

Researchers can take practical steps to understand how different versions of reality get constructed. Ideally, triangulation takes place in four complementary ways, by purposefully combining:

  •        Researchers: Involving at least two researchers means that each person can challenge the other to think better about the same data. This is less about achieving ‘inter-rater’ reliability, and more about the helpful process of questioning one another’s assumptions, about articulating what the data means, and leading to a deeper understanding of the people we aim to fathom out.
  •        Audiences: To notice what is specific to any one group, it is helpful to be able to compare it with others. In qualitative research, purposive or theoretical sampling – that is, non-probability sampling that relies on expert knowledge of the population first to decide which characteristics are important to be represented in the sample – can be used to build comparative groups even in a small sample. For instance, when researching ethnic minority communities, you might wish to compare the experiences of people from different generations since migration, not just different age groups; or people who live in affluent suburban areas as well as in deprived inner-city areas.
  •        Methods of data collection: Having a clear awareness of the assumptions built into each method is key to combining approaches in ways that can produce insight. Exploring a person’s lived experiences, attitudes and opinions is best done through individual (such as semi-structured, episodic or narrative) interviews. Exploring how attitudes are formed and how resilient they are to group dynamics – as well as the arguments that individuals summon in defence of their views – is best done through focus groups. Teasing out non-conscious beliefs, exploring ideas that may not be socially desirable, or understanding nebulous concepts such as brand perceptions, for instance, is best done through projective techniques that short-circuit over-rationalised views, such as word associations, drawings, mood boards, personification, role play, etc. Generally, combining approaches that generate verbal, visual and behavioural data is also useful to trigger insight.    
  •         Approaches to data analysis: As a social psychologist, I am always interested in people’s motivations, needs and goals. Implicitly or explicitly, I tend to look at any data set through various analytical prisms: a functionalist approach, which focuses on practical needs and functional benefits; an identity approach, which looks at issues of personal and social identity, self-presentation and expression, ego defence, and so on; and a focus on ‘interests’, which asks what, and whose, interests are being advanced or defended by acting in a certain way.

Every practitioner will have honed their own approach. Ultimately, though, insight is about more than just method. It is a state of mind. It comes from being curious. It comes from tolerance of uncertainty: being prepared to live a little while with things that don’t seem to make sense. It’s about resisting the urge to close down meaning too quickly and remaining open to what is messy and challenging. And then, the alchemy happens.

See the programme and headline speakers for Insight Alchemy 2023 – MRS annual conference.