Product Information:-

  • Journals
  • Books
  • Case Studies
  • Regional information
Request a service from our experts.
Visit the JDAL journal page.

How to... use a repertory grid

Options:     Print Version - How to... use a repertory grid, part 3 Print view

Analysing a repertory grid

The power and attraction of the repertory grid technique is that it can be analysed in so many different ways. Skilled interviewing can yield a wealth of qualitative data which reveals the interviewee's thoughts and beliefs. For those who prefer the greater certainty of numbers, the ratings will give plenty of opportunity for statistical analysis.

Any analysis is a matter of going from first impressions down to ever deeper levels. The first stage should therefore be to gain an overall impression, through a descriptive analysis. After that, the fun with the numbers can start. However, for most research projects, a number of repertory grid interviews will be carried out, so it is important to have a strategy for analysing multiple grids. We shall consider these topics in turn.

For information on software packages, please see section 4 – Some examples, and further sources of information.

Basic descriptions

Having obtained a grid, make some quick notes on general observations you made on the interview – e.g. how did it go?/did the subject find it easy to talk? – as well as on the actual grid. Here are a few techniques for overall observation:

  • Process analysis: what was actually going on during the interview, as opposed to the constructs captured on the grid? How did the interviewee react to the topic? The elements? If the latter were elicited, what can you tell from the interviewee's choice? What were the constructs like? How easy was the rating process?
  • Eyeball analysis: this is where you turn from the process to looking at the grid itself, in order to form an overall impression. What strikes you about the elements and constructs? Is there anything obvious with the ratings?
  • Construct characterization: do a quick overall summary of the constructs. How many are what is termed core, i.e. of great significance to the interviewee, as opposed to being of peripheral or marginal significance? Is there any particular type of construct that dominates, any theme running through? How are the constructs related?

Remember: first impressions count!

Analysing relationships within a single grid

Some relationships between elements or constructs may be apparent at a glance, others you may need to dig a little deeper for.

Comparing relationships between constructs and elements

If you do not have access to a statistical software package, or if you want to reflect on these relationships with your interviewee, you can do a low level number crunching exercise looking at relationships between elements and constructs. This is just a matter of:

  • For elements, summing the difference between each element pair across all the constructs, and comparing the sums to see where there is the greatest similarity or difference.
  • For constructs, conducting a similar exercise for a similar effect across all construct pairs across all elements.

Cluster analysis

This technique is one which you need a software package for, and it helps you to highlight the relationships in a grid and see them in diagrammatic form. It works out the similarity scores as described above, and then reorders the columns and rows so that the ones that are most similar are closest together. The picture is then shown in the form of a dendrogram, which portrays the relationships in the form of branches on a tree. You will then be able to see close similarities between different constructs and elements from where "clusters" of branches occur.

Principle component analysis

This is another technique requiring a software package, which also shows in diagrammatic form the similarities and differences between elements and constructs. It works by isolating the components where there is the greatest variant, and plotting them on a graph with the first component as the horizontal line (x-axis) and the second component as the vertical line (y-axis). The constructs and elements are then plotted against the two component lines, or x-axis and y-axis, and their relative distance against the latter shows the extent to which the components represent them.

In "Improving team performance using repertory grids", Boyle provides a useful summary of the differences between these two techniques.

Analysing multiple grids

The nature of research being as it is, it is highly likely that you will be using not one, but a number of repertory grid interviews. In order to do this, it is important that you design the research so that the interviews have stable components: either the elements or the constructs should be the same. As the constructs are more commonly elicited, this will more often be the elements.

In "Articulating appraisal system effectiveness based on managerial cognitions", Wright and Cheung discuss the importance of keeping either the elements or the constructs constant in situations where you are trying to obtain comparisons, and refer to the literature.

If there is no constant component, it is still possible to obtain meaningful comparison providing the number of constructs is constant over the different interviews. The reason for this is to ensure that there is no variation in the sum of differences.

Content analysis

This is a common method for analysing qualitative data, using words rather than figures, and has been covered elsewhere on this site. (See "How to... analyse qualitative data".

One of the main decisions to make with content analysis is what is the unit of analysis; with repertory grid technique, it is always the construct. The procedure is then much the same as with any content analysis:

  • Determine the categories to be used. These may be drawn from the literature, or they may be compiled on the hoof, from the constructs themselves, by comparing these and putting each new theme into a new category.
  • Tabulate the categories and the constructs.
  • Summarize the table: reflect on the meaning of the categories, find examples of each, and reflect on the frequency with which they occur.

If part of the research is to consider the difference between particular sub-groups, for example according to gender, then it's important when collecting the data to have some sort of code to indicate these sub-groups, so that when you are carrying out the analysis you can create different columns to show them.

It is also important to establish reliability: one way of doing this is by getting someone else to replicate your analysis and see if you come up with the same categories. Another way is to have two researchers working independently.

The problem with the method outlined above is that it just uses verbal components, i.e. the constructs and the elements, whilst ignoring ratings. If you want also to be able to use the ratings, follow the technique suggested by Peter Honey in "The repertory grid in action", Industrial and Commercial Training, 1979, Vol. 11 Nos. 9-11, pp. 452-459. Part of his method is to ascribe an overall summary construct for each interviewee.

In the 2004 article, "Leading with integrity: a qualitative research study", Journal of Health Organization and Management, Vol. 18 No. 6, pp. 415-434, Storr describes how in a study of the relationship between leadership and integrity, she applied the repertory grid technique to managers in the UK's National Health Service. She interviewed managers from three levels – executive and non-executive, senior manager and supervisor. In analysing her findings she was able to draw conclusions which related to the management level.

In "Improving team performance using repertory grids", Boyle uses a range of techniques to analyse the constructs obtained from the interview.

In Hankinson (2005), "Destination brand images: a business tourism perspective", Journal of Services Marketing, Vol. 19 No. 1, pp. 24-32, the constructs are set out in clusters with frequencies.