Objective: This multicenter retrospective study aimed at examining the clinical accuracy of four 15-item versions (Woodard's O and E; Quall's Q and S) of the original 30-item Benton judgment of line orientation test in a mixed clinical sample of 260 patients. It is a test frequently used as a measure of visuospatial processing. It has the advantage of requiring minimal motor skills, while a major weakness is the lengthy administration time.Method: An archival search was conducted within four in- and out-patient clinics. The frequency and magnitude of score differences were calculated to examine the equivalence of the short forms. We then checked the clinical accuracy of the short forms concerning classification of impaired, borderline, and non-impaired performance, according to NEURONORMA norms. After that, we calculated sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and kappa coefficients to assess the classification outcome of the short versions compared to the long version, when using a dichotomous classification (impaired versus intact performance).Results: When applying NEURONORMA norms, specificity (99.1%), PPV (93.1%), and kappa coefficient (0.87) were highest for version E. NPV (99.4%) and sensitivity (95.5%) were highest for version S, but the PPV of this version was relatively low (67.7%).Conclusions: We suggest use of version E when a short test is needed, as specificity, kappa coefficient, and PPV are highest for this version, while maintaining a high NPV (97.8%). However, future research should develop new normative data for these short 15-item versions.
Keywords: Benton judgment of line orientation test; assessment; short form.