Occupational therapists are often involved in the assessment and treatment of children, adolescents and adults with visual perceptual dysfunction. The Test of Visual-Perceptual Skills (non-motor) Upper Level (TVPS-UL) (Gardner, 1992b) is a standardized, norm-referenced test commonly used by occupational therapists working with older children. However, to date it has only been standardized on an American population. Furthermore, even though Gardner (1992b) claimed that the TVPS-UL was not biased according to the gender of the participants, the impact of gender on visual perception is controversial. The purpose of this study was to determine if TVPS-UL could be confidently used by occupational therapists with children outside its standardized population and to identify the factors that may impact on performance on the TVPS-UL. This study sample consisted of 105 Hong Kong and 69 Brisbane children aged 12 to 14 years. The TVPS-UL was administered to these 174 participants individually in their own native language. The aims of the study were to determine: (1) if the scores for 12 to 14 year old Hong Kong students (N = 105) and Brisbane students (N = 69) differed from the 12 to 14 year old American population recorded in the Test of Visual-Perceptual Skills (non-motor) Upper Level (TVPS-UL) manual; (2) the difficulty level of test items by removing the ceiling from the scoring procedures for the participants in this study (N = 174); and (3) if performance on the TVPS-UL for the participants (N = 174) in this study was affected by age, gender or culture.
Each participant completed all the items in each subtest of the TVPS-UL regardless of the ceiling. Two sets of raw scores were obtained for statistical analyses: one set was the total number of correct items (with ceiling removed in scoring); the other set was scored with ceiling level observed. The means of the subtest scores with ceiling level observed in scoring were compared with the smoothed medians of the American standardization sample. Separate t-tests indicated significant differences in performance on all subtests, except the Visual Memory subtest, between participants in this study and the American norms, though not for all age groups. Highly significant differences were found consistently in the performance on the Visual- Spatial Relationships and the Visual Figure-Ground subtests in all three age groups. Matched t-tests indicated highly significant differences between scores with ceiling observed and with ceiling removed for all the seven subtests. Visual Discrimination (t (173) = 5.116, p < .001), Visual Memory ( t (173) = 4.643, p < .001 ), Visual-Spatial Relationships ( t (173) = 2.833, p = .005 ), Visual Form Constancy ( t (173) = 6.396, p < .001 ), Visual Sequential Memory ( t (173) = 10.153, p < .001 ), Visual Figure-Ground ( t (173) = 3.929, p < .001 ), and Visual Closure ( t (173) = 7.948, p < .001 ) subtests. This implied that items might not be arranged progressively according to their levels of difficulty for the participants in this study. Standard multiple regression analysis failed to indicate significant variations in scores (with ceiling removed) on the TVPSUL on the basis of age, gender or culture. However, there were differences between the two sample groups in the strategies they reportedly used to solve some of the items in the Visual Sequential Memory and Visual Figure-Ground subtests. The results of this study are discussed in relation to current clinical practice and also with respect to the future refinement of TVPS-UL.