[System of evaluation, application manual and cut-off score for the cube copying test]

One of the difficulties that arises in copying figures in clinical practice and in research is that of how to evaluate performance quantitatively.

To implement a system for evaluating the copying of a solid cube in adults that is sensitive, simple to apply, easily reproducible and which establishes a cut-off score that distinguishes between a well-copied cube and others with alterations.

The system proposed here is descriptive and is based on the compulsory features a set of nine lines (three groups of three parallel horizontal, vertical and sloping lines) must have in order to form a cube that offers the visual effect of having three dimensions when put together. The elements that are taken into account are the number of faces, the relation between the sizes of the faces, the size of the copy and the characteristics of the groups of horizontal, vertical and sloping lines. The maximum possible score is 9 points (perfect). Four scorers performed a blind analysis of copies of cubes drawn by 215 subjects; copies ranged from normal to others with serious alterations.

In all a total of 860 evaluations were obtained (215 x 4). The kappa coefficient was 0.6013 and the Cronbach's alpha coefficient was 0.991. Cubes with a score equal to or above 6 were considered to have been copied correctly.

The system succeeded in evaluating all kinds of cubes and the data set was reliable and consistent. Taking into consideration the extent to which the copies can vary as well as the subjectivity inherent in any clinical method, we believe that the system displayed a high degree of inter-evaluator reliability. The cut-off score can help researchers and doctors with little experience in evaluating the results of the cube copying test.