Negative Attachment Cognitions and Emotional Distress in Mainland Chinese Adolescents: A Prospective Multiwave Test of Vulnerability-Stress and Stress Generation Models

Abstract

The present study examined the relation between attachment cognitions, stressors, and emotional distress in a sample of Chinese adolescents. Specifically, it was examined whether negative attachment cognitions predicted depression and anxiety symptoms, and if a vulnerability-stress or stress generation model best explained the relation between negative attachment cognitions and internalizing symptoms. Participants included 558 adolescents (310 females and 248 males) from an urban school in Changsha and 592 adolescents (287 female, 305 male) from a rural school in Liuyang, both in Hunan province located in mainland China. Participants completed self-report measures of negative attachment cognitions at baseline, and self-report measures of negative events, depression symptoms, and anxiety symptoms at baseline and at regular 1-month intervals for an overall 6-month follow-up (i.e., six follow-up assessments). Higher levels of negative attachment cognitions predicted prospective depression and anxiety symptoms. Furthermore, support was found for a stress generation model that partially mediated this longitudinal association. No support was found for a vulnerability-stress model. Overall, these findings highlight new developmental pathways for development of depression and anxiety symptoms in mainland Chinese adolescents.

Acknowledgments

The research reported in this article was supported, in part, by a research grant from the Canadian Psychiatric Research Foundation awarded to John R. Z. Abela.

Notes

¹Although negative attachment cognitions have been called insecure attachment cognitions in prior research, we refer to them as negative because of the methods used to test the present study's hypotheses that were self-report. Although self-report measures have been shown to be a reliable and valid assessment tool for cognitions and perceptions concerning attachment, they may not be able to capture other facets of attachment (see Brumariu & Kerns, 2010, for a review). Therefore, others have suggested to not refer to cognitions/perceptions assessed through self-report measures using the word insecure, as this describes an overall style that, although correlated to cognitions, is more encompassing in nature (West, Rose, Spreng, Sheldon-Keller, & Adam, 1998).

Note: CES-D = Time 1 scores on the Center for Epidemiological Studies Depression Scale; MASC = Time 1 scores on the Multidimensional Anxiety Scale for Children; ALEQ-DEP = Time 1 scores on the Adolescent Life Events Questionnaire, Dependent Stressors subscale; AAQ = Time 1 scores on the Adolescent Attachment Questionnaire; Gender = participant's sex (girl = 0, boy = 1); Site = participant's location (0 = urban, 1 = rural).

*p < .05. **p < .01.

Note: CES-D = Time 1 scores on the Center for Epidemiological Studies Depression Scale; MASC = Time 1 scores on the Multidimensional Anxiety Scale for Children; ALEQ-DEP = Time 1 scores on the Adolescent Life Events Questionnaire, Dependent Stressors subscale; FU = follow-up assessment.

Note: N = 4,577. DV = dependent variable; IV = independent variable; Anxiety = Level 1 scores on the Multidimensional Anxiety Scale for Children; Depression = Level 1 scores on the Center for Epidemiological Studies Depression Scale; Dependent Stressors = Level 1 scores on the Adolescent Life Events Questionnaire, Dependent Stressors subscale (ALEQ-DEP).

**p < .01.

²For anxiety symptoms, separate models were initially tested for physical symptoms, social anxiety symptoms, and separation anxiety symptoms. However, findings were consistent across symptom types. Therefore, results for the total anxiety score are reported in the manuscript. Findings for the specific anxiety subscales can be obtained by contacting the first author.

³These findings were consistent across models that controlled for comorbid symptoms and for those that did not.

⁴BCa stands for the Biased Corrected and Accelerated Confidence Interval; see Efron and Tibshirani (1993) for an explanation.