Abstract
The present research sought to investigate the negative affect structure among Chinese in Hong Kong, with particular reference to the emotional states of anxiety and depression. Correlational analyses and separate factor analyses on four self-rating scales including the BDI, BAI, DASS and PANAS were carried out on a community sample of normal Chinese adults in Hong Kong (N = 649). Good convergent and divergent validity were established for all the scales used. A three-factor model of depression (Cognitive, Functional, and Somatic) as measured by the BDI was found to be a more parsimonious structure to our sample. For the construct of anxiety as measured by the BAI, two relatively distinct components of “Autonomic Arousal” and “Subjective Fear” were delineated. The relationship between depression and anxiety was discussed in the light of the tripartite model proposed by L.A. Clark and Watson (1991b) which divides symptoms into three groups: general distress (negative affect) that are largely nonspecific, anhedonia and low positive affect that are specific to depression, and somatic arousal that are relatively unique to anxiety. The present results were in support of the various predictions of the tripartite model. Factor analysis of the DASS, however, was found to be less consistent with the original structure of depression, anxiety and stress as proposed by S.H. Lovibond and P.F. Lovibond (1995), but was found to bear some consistency with the three-factor model of depression, anxiety/apprehension, and fear/panic proposed by Barlow and his colleagues (Barlow, Chorpita, & Turovsky, 1996). As a step further, the hierarchical model of negative affect was also investigated by subjecting the pooled items from BDI and BAI to a higher order factor analysis. The results were in support of the hierarchical model in which the extracted second-order factor could be interpreted as a NA dimension, while both BDI and BAI retained adequate saliency to suggest first-order specificity dimensions that were unique to depression and anxiety. Similar higher-order factor analysis of the DASS produced a different picture in which the extracted second-order factor was found to account for most of the variance from the autonomic and hyperarousal dimensions. Depression remained a relatively more stable dimension which still retained some degree of specificity saliency after the higher-order factor analysis. Analysis of the BDI and BAI produced a structural pattern which was more in line with the hierarchical, tripartite model; while analysis performed for the DASS provided factor patterns which were more consistent to the non-hierarchical, three-factor model. Moreover, cognitive elements consistent to the themes proposed in Beck’s cognitive content-specificity hypothesis were found to emerge as the prominent components for the anxiety and depression measures. In sum, the overall results seem to be readily accommodated by both the cognitive approach and the phenotypic approach. Finally, the results tended to suggest that our Chinese participants did not endorse more somatic items as compared with psychological items. The question of whether the Chinese tend more to somatise as compared to Western participants could not be adequately addressed in the present study, but the overall results seem to suggest that there is a larger degree of universality or commonality of emotional constructs between the East and the West than what is generally believed. The results were further discussed in the light of methodological limitations in the present study.