Testing the measurement invariance of the eating disorder inventory in nonclinical samples of Hispanic and Caucasian women

ABSTRACT Objective The factor structure of the Eating Disorder Inventory (EDI) has not been thoroughly tested in Hispanic populations, yet researchers commonly use this instrument in Hispanic samples. Thus, it is important to establish the validity of the EDI in Hispanic populations. This article investigated measurement invariance of the EDI's three eating‐ and weight‐related (eat/wt) scales because they are the most frequently used and are often used in isolation. These scales include Drive for Thinness, Bulimia, and Body Dissatisfaction. Method Female undergraduates were recruited for a study on body image. The final sample (N = 688) included participants categorized as Hispanic (N = 385) or Caucasian (N = 303). They completed the EDI‐3 and a measure of acculturation. Results Measurement invariance analyses of the EDI‐3 in Caucasian and Hispanic samples were conducted. The configural model provided an acceptable fit, providing support for the three‐factor structure of the eat/wt scales in both the Caucasian and the Hispanic sample. However, weak invariance of the three‐factor structure was not supported. When measurement invariance analyses were conducted on the three eat/wt scales separately, Drive for Thinness was the only scale to demonstrate measurement invariance. Discussion The theoretical three‐factor structure of the EDI eat/wt scales was supported in both ethnic groups. Furthermore, the Drive for Thinness scale can readily be used to make group comparisons across nonclinical samples of Caucasian and Hispanic women, but researchers should be cautious when using the other two eat/wt scales to make comparisons across these two groups. © 2014 Wiley Periodicals, Inc. (Int J Eat Disord 2015; 48:262–270)