Factor structure and factorial invariance of the StateTrait Anxiety Inventory for Chinese children and adolescents

As previous research utilizing Spielberger's State‐Trait Anxiety Inventory (Form Y; STAI‐Y) has mostly involved adults or clinical groups, there have been relatively few reports assessing adolescents. This study is the first using data for children and adolescents in mainland China, from a large‐scale cross‐sectional survey in Beijing (Sample 1) and a longitudinal survey from the Wenchuan 512 earthquake (Sample 2), to clarify the factor structure and factorial invariance of the STAI‐Y, Mandarin Chinese version. As a result, only in Sample 1 did a comparison of 11 confirmatory factor analysis models indicate the best goodness‐of‐fit indices shown by a two‐factor structure for both state and trait anxiety, with both models reaching the selected cutoff criteria. These two optimal models were used in a subsequent simultaneous confirmatory factor analysis to test four conditions of factorial invariance, using eight participant groups divided on the basis of sex and school grade (fourth, fifth, sixth, and seventh to ninth). The two‐factor structure state and trait anxiety models achieved the cutoff criteria for factorial invariance, with the exception of male fourth graders. Further, it was clearly shown that in comparison with the early stage of puberty, as puberty advanced the absence of state and trait anxiety gradually decreased, while scores for the presence of anxiety gradually increased. At the same time, in the case of Sample 2, which had experienced a traumatic event, as the goodness‐of‐fit indices for none of the 11 models reached the cutoff criteria, the factor scores showed arbitrariness and a lack of objectivity. The authors conclude that cognitive structure with regard to the STAI‐Y may change with traumatic experience or the development of secondary sex characteristics at the onset of or in the stage of puberty. Also, computing the scores according to the STAI‐Y manual is problematic.