Early Predictors on Adolescents’ Depressive Symptoms: A 9-year Longitudinal Study

• In Event: Poster Session 13
  In Poster Session: PS 13 Topic - Parenting & Parent-Child Relationships

Abstract

Background. Identifying early predictors for adolescents’ depressive symptoms is crucial for preventing growing mental health issues among adolescents over time (Karevold et al., 2009; Stoolmiller et al., 2005). The current longitudinal study builds on past work examining the effects of family and school factors on the curvilinear growth trajectory of depressive symptoms in Taiwanese adolescents over 6 years (Yi et al., 2009). We examine the non-linear developmental trajectory of depressive symptoms over 9 years, and focus on the long-term impacts of early parental warmth, parental harsh discipline, and child self-esteem.

Method. The study sample comprised of 2,690 Taiwanese adolescents followed from aged 12 at baseline in the longitudinal study, “Taiwan Youth Project” (TYP). TYP followed these youths annually since the first wave in 2000 (age 12; 9th graders) until the ninth wave in 2009 (age 21; college seniors). At each wave, participants completed self-reports of depressive symptoms. At baseline, they reported on parental warmth, harsh discipline and self-esteem. Latent Growth Curve Modeling (LGCM) was used to examine the prospective change in adolescents’ depressive symptoms and identify early predictors. Because the data spans over multiple developmental transitions, we first fit an unspecified trajectory growth model to parsimoniously and flexibly represent the complex trajectory of change in depressive symptoms. Second, we tested three baseline predictors of change in depressive symptoms: parental warmth, parental harsh discipline, and child self-esteem.

Results. The unspecified trajectory model fitted the data well (χ2=379.76 [p<.001]; RMSEA=.063; CFI=.91; IFI=.91). Figure 1 shows the free estimated slope loadings (λ), representing the growth rates at each timepoint between the second to eighth waves. The youth depressive symptoms gradually increased from the first (λ= 0) to third waves (λ=1.67) during middle school, followed by decrease at the fourth (λ=-.83) and fifth waves (λ=-1.16) in high school. Depressive symptoms reached a peak at the sixth wave (λ=1.81), the final year in high school, decreased at the seventh wave (λ=-.34) and increased again at the eighth (λ=1.46) and ninth waves (λ=1). Overall, a significant linear slope mean indicated the average annual increase in depressive symptoms was .27 (p<.001). As shown in Figure 2, parental warmth (β=-.16, p<.001), parental harsh discipline (β=.16, p<.001), and self-esteem (β=-.23, p<.001) were significantly associated with adolescent’s initial level of depressive symptoms. Moreover, higher parental harsh discipline predicted a steeper increase in depressive symptoms (β=.17, p<.05) and higher self-esteem predicted slower growth (β=-.24, p<.01) over the 9 waves.

Conclusions. Our findings suggest that early parenting and self-esteem predicted the initial level and longitudinal change of depressive symptoms among adolescents. Over nine years, the growth rate of depressive symptoms were accelerated by early harsh discipline, but slowed down by early self-esteem. The findings indicated that early parental warmth wasn’t a protective factor in depressive symptoms. These results have potential implications for understanding the longitudinal trajectories and prevention of adolescent depression.

Authors

• Shou-Chun Chiang, Pennsylvania State University, University Park
  Presenting Author

• Sunhye Bai, Pennsylvania State University, University Park
  Non-Presenting Author