Abstract
These studies examined the degree to which racial, religious, and national aspects of individuals' sense of ethnic identity stand as interrelated, yet distinct, constructs. Results of exploratory factor analyses in Study 1 (n = 272) revealed that a three-factor model specifying racial, religious, and national identities yielded optimal fit to correlational data from an expanded, 36-item version of the Multigroup Ethnic Identity Measure (MEIM; Roberts et al., Citation1999), although results left room for improvement in model fit. Subsequently, results of confirmatory factor analyses in Study 2 (n = 291) revealed that, after taking covariance among the items into account, a six-factor model specifying exploration and commitment dimensions within each of the racial, religious, and national identity constructs provided optimal fit. Implications for the utility of Goffman's (Citation1963b) interactionist role theory and Erikson's (Citation1968) ego psychology for understanding the full complexity of felt ethnic identity are discussed.
ACKNOWLEDGMENTS
The authors thank Tanya Batra, Rajdeep Bawa, Poonam Benawra, Jordelle Blackstock, Jaspreet Chana, Sarah Chouman Charife, Emma Eden, Aneeka Ghela, Sandeep Gill, Zara Lemuel, Fatema Lubna, Rajdeep Mahil, Gurshinder Mann, Caroline Meara, Lauren Meltzer, Amena Nadeem, Phuong Nguyen, Jodi Nicholls, Ayo Parchment, Laura Pearson, Samrah Shah, Farrah Sheikh, Navreen Soor, and Neriman Yetkil for their contributions to data collection and data entry for these studies.
Notes
Note. Items shown in bold are items that loaded together.
Note. Factor loadings shown in bold. Factor loadings in the confirmatory factor analyses were reported up to two decimal places in LISREL 8.72 (Jöreskog & Sörbom, Citation2005a). For comparison purposes with Study 1, we added a third decimal place of 0 to the factor loadings in Study 2.
Note. EXP = Exploration; COM = Commitment. Factor loadings shown in bold. Factor loadings in the confirmatory factor analyses were reported up to two decimal places in LISREL 8.72 (Jöreskog & Sörbom, Citation2005a). For comparison purposes with Study 1, we added a third decimal place of 0 to the factor loadings in Study 2.
*p < .05. **p < .01. ***p < .001.