Prediction of Boston Naming Test performance from vocabulary scores: preliminary guidelines for interpretation

Patients with limited education or underdeveloped vocabulary skills may perform below the normal range on the Boston Naming Test when compared to the original published norms, even in the absence of brain damage. To reduce the frequency of false positive dysnomic classifications of patients with limited vocabulary skills, we developed a score adjustment to account for the significant shared variance between scores on this test and the WAIS-R Vocabulary subtest. Vocabulary significantly predicted performance on the Boston Naming Test (r = .65, p < .0001) in a sample of 62 outpatients who had no objective evidence of brain damage. Linear regression was used to derive expected performance on the Boston Naming Test from Vocabulary scaled scores. Relative to the original published norms, scores based on the Vocabulary subtest cut-offs produced fewer false positives and more accurately classified group membership for patients with and without objectively verified brain damage. These performance predictions are offered as tentative guidelines to assist clinicians in evaluating the presence of naming deficits by controlling for the variance associated with knowledge of vocabulary.