Centre for Learning and Teaching Stockholm University

We conducted this study to examine measurement equivalence and mean differences in identity status across 3 ethnic/cultural contexts: White American, Hispanic American, and Swedish. We used the Extended Objective Measure of Ego Identity Status II (EOM–EIS–II; Bennion & Adams, 1986), a commonly used instrument in the identity status literature. We conducted analyses to ascertain the extent to which the EOM–EIS–II functioned equivalently in 3 ethni-cally/culturally different samples. The internal structure of the measure was consistent across contexts. When we statistically controlled effects of age and gender, mean differences tended to be largely cross-cultural at the observed level of analysis but to be both cross-ethnic and cross-cultural at the latent level of analysis. This divergence in findings was found despite the limited age range represented in each of the samples. We therefore concluded that measurement error may have played a role in these differences and that data gathered using the EOM–EIS–II should be analyzed using latent variable methods. We discuss results in terms of using the EOM–EIS–II with diverse populations. Identity has been the subject of widespread study for several decades. Erikson (1950) described identity as, among othe