There are certainly quantitative indices that can be investigated (in addition to what you appear to be intending to perform in the way of content analyses of the construct changes). Following are some examples: (1) Interelement distance scores between the self and important others (or between other pairs of elements) can be compared from grid to grid even with different constructs (as long as the grids are the same size each time). This would provide information on how the developmental events (and perhaps changes in construct content) have impacted identification patterns. (2) The intensity of the grids (overall level of intercorrelatedness of the constructs) can be compared (again with identically-sized grids) to give some indication of how integrated versus complex the successive grids are. (3) As an alternative measure of complexity/simplicity, the amount of variance accounted for by the first component of a principle components analysis could be compared from administration
to administration. Adolescents are notorious for having huge first factors. One might expect this to lessen over time, particularly when psychological/psychosocial development is accellerated by the knowledge of foreshortened future. (4) Self-esteem indices (variously calculated) should still be deriveable and comparable from grid to grid. (5) Landfield's FIC/Ord indices could be calculated to assess changes in the integration and complexity dimensions as reflected by those measures.
These are just a few suggestions; I hope they are helpful.
Kenneth Sewell
***
Kenneth W. Sewell, Ph.D.
Associate Professor and Director of Clinical Training
University of North Texas
Psychology Dept., Box 311280
University of North Texas
Denton, Texas 76203-1280 USA
Phone: 940-565-2640
Fax: 940-565-4692
sewellk@unt.edu
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%