Psychological control and lack of warmth are widely assumed to cause internalizing symptoms in adolescents (Hunter et al., 2015; Steinberg, 2001). However, most research has been cross-sectional, and longitudinal findings have been mixed (e.g., White et al., 2015) or the used statistical methods were not optimal to support causal conclusions (Hunter et al., 2015). Only few studies have inspected child effects on parenting (Brenning et al., 2015). Thus, evidence is lacking on whether parenting style affects adolescent internalizing symptoms such as depression, loneliness, and poor self-esteem. Moreover, from a systems perspective, further factors should be explored such as adolescents’ and parents’ perceptions of each other, goals, and strategies to change their mutual relationship. This study examines bidirectional effects of all these facets of parent-adolescent relationships and parenting behaviors and adolescent internalizing symptoms.

Using two annual data collections (N = 1,281/1,274/824 at T1/T2/overlap, resp.) in a representative Swedish community sample of adolescents originally in grades 7-10 (Mage = 15.2, SD = 1.2), effects of perceived parenting (warmth, psychological control, behavior control, overcontrol), adolescent relationship satisfaction, goals (establishing autonomy, submission under parental authority), and strategies (disclosure, secrecy) on internalizing problems (depression symptoms, loneliness, low self-esteem) and vice versa were examined, controlling for the respective dependent variable at T1, gender, and school grade. Parental attitudes (e.g., perceived child depression, satisfaction, and feelings of giving up) were assessed at T2 in a sub-sample (N = 290), allowing for the prediction of these attitudes by T1 internalizing. In order to preserve as much information as possible, missing data were multiply imputed (20 datasets), reaching over 95% efficiency of analyses. Still, those analyses involving parent attitudes are tentative due to the lack of T1 measures and the large number of missing data, reducing power and introducing bias if data were not missing at random (e.g., non-response of dissatisfied parents being not entirely predicted by adolescent data).

Consistent with and expanding previous research, most parenting and parent-adolescent relationship variables were cross-sectionally correlated with adolescent internalizing symptoms (see Table 1). In most instances of significant within-time associations, also the predictions over time of the respective parenting and parent-adolescent relationship variables by teen internalizing symptoms were significant (Table 1). In contrast, only three effects in the opposite direction reached or approached significance: feelings of being overly controlled by parents increased depression and tentatively reduced self-esteem, and low child disclosure increased loneliness. Supporting a systems perspective, parent-reported feelings of giving up and of low relationship satisfaction mediated effects of adolescent depression on e.g. reduced warmth and increased psychological control over time.

Thus, the study has shown broad deteriorating effects of teen internalizing on parenting and parent-adolescent relationship quality and has provided first evidence of mediation by parent cognitions and feelings. However, parenting effects on internalizing were sparse and involved other than the expected variables. If adolescents felt overly controlled by their parents, they became depressed and their self-esteem was tentatively reduced. And if they did not disclose much information to their parents, they became lonelier over time.