Longitudinal assessment of trait emotional intelligence: Measurement invariance and construct continuity from late childhood to adolescence.

Abstract

Amid the growing efforts to promote positive youth development, trait emotional intelligence (TEI) has emerged as an important protective factor in the processes of resilience and adaptation. The inclusion of a brief form of the Emotional Quotient Inventory-Youth Version (EQi:YV-Brief) in the Canadian National Longitudinal Survey of Children and Youth (NLSCY) presents a unique opportunity to study the developmental dynamics of TEI during the transition from childhood to adolescence. However, before drawing any inferences about construct continuity and change, researchers must establish that the EQi:YV-Brief functions equivalently over time. This study tested configural, metric, scalar, and residual measurement invariance of the EQi:YV-Brief over a 6-year period from late childhood (age 10-11) to adolescence (age 16-17). Longitudinal mean and covariance structures models were fitted to the data from 773 NLSCY participants (51% girls) who completed the EQi:YV-Brief at 4 biennial cycles. Three of the 4 EQi:YV-Brief subscales were found to be fully invariant at ages 12-13 through 17-18 and partially invariant at age 10-11. Controlling for partial noninvariance, we also investigated patterns of rank-order stability and mean-level change in TEI. These exploratory analyses showed that individual differences in TEI became increasingly more stable with age and that changes in mean TEI levels followed a complex nonlinear pattern over time. The results supported the longitudinal utility of 3 of the 4 EQi:YV-Brief subscales used in the NLSCY, supporting their further use in research on the developmental dynamics of TEI.