Neuropsychological test norms based on multiple samples: Effects of demographic variables, sample characteristics, and substance use disorders

This study uses a metanalytic technique to combine diverse samples into large samples representing a large number of participants. Samples were created for five neuropsychological tests: the California Verbal Learning Test, Category Test, Trail Making Test, Stroop Color Word Test, and Verbal Fluency Tests. Data were obtained from normative studies and experimental studies including normal control groups. A second set of samples included data from studies of neuropsychological performance in individuals with substance use disorders. The influence of demographic variables was investigated using regression analysis. Variables studied included age, education and gender. Characteristics of the samples were also investigated to determine their influence on performance. The age of the study, strictness of exclusion criteria, and the inclusion of college students versus community participants were studied. In addition, the form of the test was included in the analysis for two tests, the Category Test and Letter Fluency Test, which offer two forms of administration. Finally, the influence of substance use disorders on neuropsychological performance was examined. Results indicate that age and education influence performance on all neuropsychological tests examined with older age and lower education predicting worse performance. Gender was associated with performance on the Category Test, the Category Fluency Test, and two subtests of the CVLT. Performance was better for women on the CVLT and better for men on the Category and Category Fluency Tests. The use of strict exclusion criteria did not influence performance. Use of samples composed entirely of college students was associated with better performance on the Stroop Test, but had no influence on other tests. Better performance was observed in more recent studies for Part A of the Trail Making Test. Substance use disorders were associated with poorer performance on all but the Letter Fluency Test. Demographic variables and study characteristics influenced variability of performance on some tests, supporting the role of individual differences in normal performance. Mean performance in the metanorms was not significantly different from that in the currently available published norms.