Depressive Symptoms in Adolescence: a Cross-sectional and Longitudinal Investigation of the Contingency-competence-control Model

Abstract

Depressive symptoms in school-aged children destabilise friendships and create significant family disruption and anguish (Angold, Messer, Stangl, Farmer, Costello et al., 1998; Clarke, DeBar, & Lewinsohn, 2003). They are related to increased risk of drug use and suicide (Gould et al., 1998; Rohde, Lewinsohn, & Seeley, 1991) and other psychiatric disorders (Angold & Costello, 1993). Furthermore, depression is the third most common cause of death in the adolescent population (Arias, MacDorman, Strobino, & Guyer, 2003). There are many theories of depression. The Contingency-Competence-Control Model (CCC Model; Weisz & Stipek, 1982) incorporates many aspects of the most well known theories and has been supported as relevant in predicting depressive symptoms (Weisz, Southam-Gerow, & McCarty, 2001). However, there have been few investigations of the CCC Model to date. A primary aim of this thesis was to extend and test the CCC Model both cross-sectionally and longitudinally as it applies to adolescent depressive symptoms. Other constructs shown to predict depression and depressive symptoms were included to provide a more rigorous examination of the CCC Model. These variables included gender, negative life events, daily hassles, and coping style. First, all variables were tested as predictors of depression, and second, as predictors of anxiety and life satisfaction, to test whether it predicted other domains of well-being. These aims were met over two interrelated studies, with Study 2 having two parts (cross-sectional and longitudinal). Study 1 involved the development of the Brief Adolescent Daily Hassles Scale (Wright, Creed, & Zimmer-Gembeck, 2010). Three phases led to the development and initial validation of a brief daily hassles scale that could be used with adolescents. Phase 1 drew on hassles identified in existing scales, hassles from the literature, a focus group with adolescents, and expert feedback to generate 69 daily hassle items. In Phase 2, the items were administered to a sample of 212 adolescents. Item and exploratory factor analysis were used to reduce the number of items to 14, which represented two homogenous and internally reliable subscales of parent hassles and friend/other hassles. In Phase 3, the scale was administered to a second sample of 236 adolescents. The initial structure was supported by confirmatory factor analysis. Testing the correlation between factor scores and measures of depression, anxiety and life satisfaction supported the measure’s construct validity. The purpose of Study 2 was to examine the cross-sectional and longitudinal relationships among the variables of interest. Adolescent students were surveyed at Time 1 for Study 2A. This sample consisted of 437 students: 240 girls (54.9%) and 197 boys aged 14 to 18 years. Participants completed measures of perceived competence, contingency, control and unknown control, number of negative life events experienced, frequency of daily parent and friend/other hassles and coping style, as well as measures of depression, anxiety and life satisfaction. Hierarchical multiple regression analyses showed that the CCC Model accounted for 56% of the variance in depressive symptoms. The only CCC Model variables to significantly contribute to depression were perceived competence and control. Unknown control also predicted depressive symptoms. Daily hassles were more important in predicting depressive symptoms than negative life events, which did not significantly predict depressive symptoms. Significant gender differences and mediational pathways were uncovered and are discussed. Overall, the model accounted for more variance in depressive symptoms than anxiety and overall life satisfaction. In Study 2B, 166 students, 100 girls (60.2%) and 66 boys, were re-administered the same survey at Time 2, three months after the Time 1 administration. First, the cross-sectional analyses were replicated using the Time 2 data. Results showed that several variables were reliably associated with depressive symptoms, anxiety and life satisfaction at Time 1 and Time 2. The data were then analysed using three longitudinal models (basic, change-in-outcome and change-change), providing a comprehensive examination of the model over time. Time 1 psychological well-being was the strongest predictor of Time 2 well-being across all three outcome variables. Gender, daily hassles with friends and others, and non-productive coping were the most robust predictors of depressive symptoms. Hassles with parents and perceived competence predicted depressive symptoms three months later, and perceived contingency predicted changes in depressive symptoms from Time 1 to Time 2. The model was not well supported in predicting anxiety and life satisfaction longitudinally. The CCC Model was only partially supported in predicting depressive symptoms in the context of other important variables. Coping style, daily hassles, and gender were the most important predictors of depressive symptoms. Perceptions of unknown control added to our understanding of adolescent depressive symptoms and may provide a useful addition to the CCC Model. Control-related perceptions partially mediated the relationship between daily hassles and depressive symptoms. The variables under investigation were stronger predictors of depressive symptoms than anxiety and life satisfaction. The theoretical and practical implications of the research are discussed.