Analysis of repertory grids in clinical practice

This paper illustrates the use of several different forms of analysis of repertory grid data, using a case study of a client who completed repertory grids at various stages of therapy.

Participants in a survivors' group completed grids before and after therapy and at 3- and 6-month follow-up. Grids from one of the participants, who had been sexually abused as a child, are presented using different levels of analysis: visual inspection of the raw grid; analysis of dissimilarities between pairs of elements or constructs; hierarchical cluster analysis of elements and constructs both separately and combined; principal components analysis; biplot representation of elements and constructs combined; multidimensional scaling analysis and unfolding analysis. The latter two methods also provide ways of presenting a combined analysis of a number of grids. Computer packages are reviewed, and SYSTAT commands are presented to perform the analyses.

The review of methods illustrates how different levels of analysis can usefully back up clients' own accounts of progress through therapy, from detailed analysis of individual ratings in the raw grid, through analyses that pull out structural patterns at the expense of detail, to the broad sweep provided by methods that combine data from a number of grids.

Analysis of repertory grids completed at various stages of therapy can provide useful qualitative information on progress, but can also provide some simple quantitative measures (such as Self-Ideal Self discrepancy as a measure of self esteem) to track progress. Different forms of analysis can be informative, highlighting different aspects of progress, but also allowing checks on adequacy and goodness-of-fit of particular analyses.