LAUSR

OBJECTIVE Tests of cognitive abilities are particularly susceptible to culture-based bias because these abilities are culturally bound. The specific purpose of this study was to examine the Semantic Interference Test, a clinical neuropsychological test, for culture bias. METHOD The sample included 415 community-dwelling participants (mean age 74, SD = 8.32; 308 were females) living in South Florida (USA). The sample included 72 African Americans, 93 Afro-Caribbeans, 77 Hispanic Americans, and 173 European Americans. An Item Response Theory analysis of bias was employed using the Differential Item Functioning (DIF) procedure. RESULTS Overall, the items appear to be invariant across gender, ethnicity, and education levels. Although the DIF identified several items that appear to differ across the two latter groupings, the multiple group confirmatory factor analysis (MG-CFA) suggests that these items have low impact on the overall measure. There were however, meaningful differences across age groups in the MG-CFA, suggesting that an age adjustment might be required. CONCLUSION The SIT can be considered a cognitive test that is not significantly affected by the participants' cultural background, at least within the culture range included in this study.