A Factor Analytic Study of the Frostig Developmental Test of Visual Perception

While many attempts have been made to devise methods of assessing deficits in visual perceprion with adults, there is a relative paucity of work in this area with young children. One notable exception is rhat of Frostig (1961) who has developed an easily administered test of visual perception for children between the ages of 4 and 8 yr. This test provides five subtest scores in the areas of eye-motor coordination, figure-ground perception, form constancy, position in space, and spatial relationships as well as an over-all score defined as a Perceprual Quotient. Frostig, Lefever, and Whittlesey (1961) have stated rhat these five areas of functioning as represented in the Frostig test were derived from their clinical experience. These areas are presumed to be relatively distinct from one another. In addition, these authors present some test data and state that children with learning difficulties obtained lower total scores and greater subtest scatter (range) than did normal children. Unforrunately, no tests of significance are reported to substantiate these statements. While there is little question that such a test would be of benefit to clinicians who work with children, no evidence is presented by these authors that it does anything other than measure age differences (or developmental changes). Frostig, et al. concluded that the greater degree of subtest scatter of the children with learning handicaps "suggests that distinct functions of visual perception can be disturbed independently and in varying degrees" (p. 392). This was the basis for a recommendation that retraining in specific areas of deficit diagnosed by the test be initiated. It could be just as cogently argued that those Ss with lower total scores and greater subtest scatter have their scores based on fewer test items and, consequently, that the increased scatter is a function of decreased reliability. Such issues became more apparent as the present writers began to use the rest. I t was noted, for example, that where subtest discrepancies occurred, Figureground and Form Constancy usually varied together. The tasks on these subtests are essentially the same--form discrimination in an overlapping context. The present study was devised and carried our to determine what common factors did in fact exist in the test scores and to determine what proportion of -