Differential item functioning of the Acculturation, Habits, and Interests Multicultural Scale for Adolescents...

Acculturation is a complex multidimensional construct that plays in important role in various outcomes across many domains (e.g., psychological, behavioral, social, and developmental). While there are many acculturation measures for specific racial/ethnic groups, few multigroup measures exists. Even fewer have been psychometrically evaluated for differential item functioning (DIF; e.g., measurement invariance). This is a critical step in determining whether item properties are equivalent across groups and thus, whether scores obtained from different groups are in fact comparable.

The Acculturation, Habits, and Interests, Multicultural Scale for Adolescents (AHIMSA) was completed by first and second generation participants identifying as White, Hispanic, or Asian (n = 1,399). The scale can be used to compute scores on four acculturation orientation subscales: Assimilated, Separated, Integrated, and Marginalized. Racial/ethnic DIF was evaluated for the Assimilation, Integration, and Separation subscales.

Analyses revealed racial/ethnic DIF was present for two items on the Separation scale based on McFadden's pseudo R². No DIF was identified for the Assimilation and Integration scales. Supplemental analyses revealed no evidence of longitudinal DIF or by generation status and some DIF by sex for Assimilation and Integration. Based on evaluation of test characteristic curves and DIF plots, it was determined that DIF was trivial and could likely be ignored.

Psychometric properties were generally equivalent across the three racial/ethnic groups for each of the subscales evaluated. This indicates that scores obtained from each subscale could be compared across racial/ethnic groups. By establishing cross-group usability of the scale, racial/ethnic differences in acculturation orientation can be examined, and, more importantly, used in the context of research on related mental, physical, and behavioral health outcomes.