Personal Constructs

Repertory Grid Technique

Repertory grid technique is by far the most widely used method of personal construct assessment (See entry on ‘Subjective Methods’ in this volume). A set of ‘elements’, or aspects of the person's experience, is first elicited from the subject. These are usually significant people and aspects of the self, but this is not necessarily so, and numerous other types of element have been used: for example, relationships, life events, holiday resorts, and therapy sessions. A sample of the subject's constructs is next elicited, usually by presenting a series of triads of the elements and asking, for each triad, how two of the elements are alike and thereby different from the third. Since the technique is extremely flexible, there are numerous alternative ways of obtaining constructs, including free descriptions of the elements, interviews, structured methods designed for children, and supplying of the constructs by the investigator. Although the investigator who is only interested in obtaining a sample of the subject's constructs may not proceed to this stage, the final step in grid procedure is for the subject to sort all of the elements, usually by rating or ranking, in terms of all the constructs.

While some understanding of the subject's construct system may be obtained from visual inspection of the grid matrix, a much more detailed assessment is possible by the use of various computer packages, most of which include cluster and/or factor analyses, and some of which are interactive, eliciting a grid from, and providing feedback on its results to, a subject. Information may be derived from the grid on such areas as:

• (i)
  the content of the subject's constructs, for which several coding systems are available;
• (ii)
  the structure of the construct system, for example how ‘tightly’ or ‘loosely’ constructs are interrelated;
• (iii)
  the particular relationships between constructs, which are assumed to indicate the personal meaning of these constructs;
• (iv)
  distances (or degree of perceived dissimilarity) between particular pairs of elements.

In addition, most computer packages provide a spatial representation of the subject's construct system; and some allow the comparison of pairs of grids or the derivation of ‘consensus’, or modal, grids of a group of subjects.

Since there is no standard form of the grid, general statements about its psychometric properties are fairly meaningless, and in any case it has been questioned whether models derived from mental test theory are appropriate for the evaluation of grids. Nevertheless, various studies attest to the reliability and validity of certain grid measures in particular domains.

...