The Reliability of the African Neuropsychology Battery in Persons of African Descent.

Abstract

The African Neuropsychology Battery (ANB) includes eight culturally appropriate cognitive tests developed for use in the Congo and other sub-Saharan African populations. The current study examines the reliability of the ANB in three samples of participants of African descent. Subjects were recruited in the United States and the Congo to participate in three studies of ANB internal consistency reliability (Study 1), test-retest reliability (Study 2), and interrater reliability for the two ANB measures (i.e., Visuospatial Memory and Proverb Tests) requiring examiner ratings of response adequacy (Study 3). Subjects were administered ANB tests of visuospatial perception, language, memory, abstract reasoning, and problem solving. We calculated Cronbach's alpha, corrected item-total correlations and mean inter-item correlations for internal consistency, Pearson product-moment correlations and intraclass correlation coefficients for test-retest reliability, and intraclass correlation coefficients for interrater reliability. The ANB tests had acceptable internal consistency (Cronbach's alphas ranging from .37 to .93). Across subtests, test-retest reliability coefficients ranged from .39 to .91, and intraclass correlation stability coefficients (ICCs) ranged from .39 to .82. Of the two ANB tests requiring interrater reliability, only the Proverb Test had a low ICC of .13, (confidence intervals: -.29 to .52). The present study demonstrated that most ANB tests show adequate reliability in participants of African descent. However, the scoring criteria of the African Proverb Test require revision in order to improve the interrater reliability of the measure.