Journal of Youth and Adolescence, Vol. 31, No. 6, December 2002, pp. 405–417 ( C  2002) Factors Contributing to Gender Differences in Depressive Symptoms: A Test of Three Developmental Models Inge Seiffge-Krenke 1 and Mark Stemmler 2 Received February 17, 2001; revised October 3, 2001; accepted April 11, 2002 In a longitudinal study on 115 early adolescents, the impact of major events, relationship stressors, and coping style in interaction with biological changes on depressive outcome in late adolescence were ex- plored. Three developmental models proposed by Nolen-Hoeksema and Girgus (1994, Psychol. Bull. 115: 424–433), which attempt to explain the emergence of gender differences in adolescent depressive symptoms, were tested. Multiple regression analyses conducted to test Model 1 identified two etiolog- ical factors associated with gender differences of depression, namely, stress in the adolescent–mother relationship and a more negative body image. Tests of Model 2 revealed etiological factors which correlated with depression in adolescent girls but not adolescent boys. The correlational patterns of avoidant coping and body image emerged as relevant factors for Model 2. Finally, tests of Model 3 searched for factors which were more likely to be observed in adolescent girls than boys and which would put girls at risk should they interact with a third etiological factor. Early maturational timing emerged as a factor for Model 3. KEY WORDS: depressive symptoms; body image; physical maturity; major life events; relationship stressors; avoidant coping; gender differences. An increase in depressive symptoms in early adoles- cence has been well established (for a review, see Offord et al., 1987). In addition, more serious depressive symp- toms in adolescent girls compared to adolescent boys have been consistently documented in many countries (Ohannessian et al., 1999; Seiffge-Krenke, 1998). The gender difference in incidence is apparent around the world, with female–male ratios ranging from 1.6 in Beirut, Lebanon, and Taiwan to 3.5 in Munich, Germany (Wolk and Weisman, 1995). Higher rates in adolescent girls have been documented for clinical levels of depression, e.g., de- pressive disorders, as well as for subclinical levels such 1 Professor of Psychology at the Department of Psychology, University of Mainz. She received her Ph.D in 1981 from the University of Giessen. Her major research interests are in stress, coping, and close relation- ships in adolescents. To whom correspondence should be addressed at University of Mainz, Department of Psychology, Staudingerweg 9, 55091 Mainz, Germany; e-mail: seiffge@mail.uni-mainz.de. 2 Assistant Professor at the Department of Psychology, University of Erlangen-Nuremberg. He received his master’s degree from the Tech- nical University Berlin and his Ph.D from the Pennsylvania State Uni- versity. His research interests encompass developmental psychology and methodology. as depressive symptoms and depressive mood (Cicchetti and Toth, 1998). However, much controversy remains about the ex- act time frame for the emergence of gender differences (Hankin et al., 1998), the actual proportions of gender differences (Compas et al., 1997), and the impact of ethnic differences (Facia and Batistuta, 2001; Finkelstein et al., 2001). Several theories have been put forth to explain the emergence of the gender differences in adolescence, in- cluding those which take biological and psychosocial fac- tors into account. Nolen-Hoeksema and Girgus (1994) described three models to explain how gender differences in depression might emerge in early adolescence. Accord- ing to the first model, the same factors cause depression in both adolescent boys and girls, but these factors become more prevalent in girls than in boys during early adoles- cence. A second model posits that the factors leading to depression are different for girls and boys, and that the risk factors leading to depression in girls become more common in early adolescence, i.e., risk factors for adoles- cent girls increase much more than those for adolescent boys. Like the first model, the third model posits that the 405 0047-2891/02/1200-0405/0 C  2002 Plenum Publishing Corporation