Psychometric Assessment of the 10-item Experience of Close Relationship- Revised for Adolescents (ECR-R-10-AD)

Abstract Background : The Experiences of Close Relationships-revised (ECR-R) is a widely used 36-item self-report measurement for measuring adult as well as child and adolescent attachment. However, some short versions of the ECR-R have been developed and tested psychometrically. Due to cultural differences, respondents from different cultural backgrounds may respond differently to specific items of the ECR-R. This study aimed to develop a new, shorter version of the Thai version of 18-item ECR-R used in Thai adolescents to make it feasible to use and to improve psychometric properties. Methods : This study included 433 adolescents, aged between 13-17 years old. Participants completed the 18-item ECR-R, Rosenberg Self-Esteem Scale, Perceived Stress Scale, and Relationship Questionnaire. Exploratory factor analysis was performed to determine the suitable number of items for a new scale. A correlation between the ECR-R-18 and the new scale was carried out. Additionally, correlations between the ECR-R-18, the new scale, and other measurements were also analyzed. The Cronbach alphas and corrected correlations between the anxiety and avoidance subscales of the old and new ECR-R were calculated. Confirmatory factor analysis of the first-order two-factor solution model with fit statistics was examined. Multigroup CFA was performed to investigate measurement invariance. Results : he new scale with 10 items, called “ECR-R-10-AD” was developed with good Cronbach’s alpha values of 0.859 and 0.824 for avoidance and anxiety subscale, respectively. The corrected correlation between the ECR-R-18 and ECR-R-10-AD was between .76 (p < .001) and .78 (p < .001). The data fitted the first order two factor solution model well indicating by the values of the comparative fit index (0.968) and Tucker-Lewis index 0.957) for the model, whereas the root-mean-square error of approximation was 0.063 (90%CI, 0.048-0.079), and standardized root mean square residual was 0.046, indicating that the model fits were excellent. The ECR-R-10-AD was shown to have convergent validity with other measurements. The ECR-R-AD seemed to be superior to the ECR-R-18 regarding criterion validity. Measurement invariance was successfully established across different sexes. Conclusion : The ECR-R-10-AD provided sufficient psychometric properties among Thai adolescents. Factorial validity, convergent validity and measurement invariance were established. As the ECR-R-10-AD is brief, it can be administered with less burden. Limitation and future research were discussed.