Decades of research have shown that depressive symptoms in children are associated with social skills deficits and problems with peers [1–4]. Nevertheless, not much is known about the mechanisms underlying these associations. In the current paper a conceptual model is tested which tries to explain the interdependent effects of social skills and peer relations on the development of depressive symptoms.
Generally speaking social skills have been defined as behaviors that affect interpersonal relations . In our work, we conceive of social competence as the ability to use social interactions to satisfy one's own goals and needs while at the same time considering the needs and goals of others. We differentiate between two dimensions: (a) self-oriented social skills which are aimed at satisfying one's own needs (e.g. assertiveness and social participation) and other-oriented social skills which are aimed at satisfying another's goals and needs (e.g. pro-social, cooperative and non-aggressive behavior) . We assume that these dimensions operate through different mechanisms on depressive symptoms. First, we suggest that deficits in self-oriented social skills are directly associated with children's depressive symptoms because children's social needs and goals remain unsatisfied as a consequence of their inability to initiate social contacts, to express their needs, to assert themselves or to set limits to others' demands. These children therefore experience non-rewarding social interactions. Second, we propose that the impact of deficits in other-oriented social skills on depressive symptoms is mediated through negative peer relations, such as peer victimization.
Social skills and depressive symptoms
Interpersonal theories of depression have tried to explain the observed associations between depression and relationship problems and deficits in social skills in children, adolescents and adults . For example Lewinsohn  has suggested that a lack of certain social skills in depressed persons reduces their experience of positive reinforcement by others because they do not engage in behavior that leads to rewarding consequences. Coyne's theory  proposes that over the long term, a vicious circle develops: Depressive symptoms trigger negative reactions from others (e.g. aggressive behavior), which hinder recovery from depression. Moreover, intervention studies have shown that social skills training is an effective means to reduce depressive symptoms . These latter results may even suggest that social skills play a causal role in the development of depression.
There is broad agreement that deficits in self-oriented social skills are associated with depressive symptoms. Withdrawal is considered as one of the main behavioral precursors of depressive symptoms in children  and it may exacerbate internalizing problems . The empirical data on the association between deficits in other-oriented social skills and depressive symptoms are somewhat contradictory. Interestingly, aggressiveness - considered a deficit in other-oriented social skills - is positively associated with self-oriented social skills, i.e. aggressive children are more assertive and more prone to engage in social interactions [12–14]. Therefore, aggressiveness might even protect children from depressive symptoms. However, the specific role of pro-social behavior in the development of psychopathology has received only limited attention. It has been suggested that low levels of pro-social behavior may place children at risk for externalizing problems, whereas high levels of pro-social behavior might be a risk factor for internalizing problems [15–17]. It might be that a lack of pro-social behavior could be considered an indicator of a deficit in other-oriented skills, whereas high levels of pro-social behavior could also reflect a lack of self-oriented skills because these children might be too considerate of the needs of others and neglect their own feelings and needs . In fact, cross-sectional and longitudinal studies have found that children who are overly concerned for the welfare of others, are highly cooperative or over-friendly, have elevated levels of emotional symptoms [19–21]. Perren and collaborators showed that at kindergarten age pro-social behavior predicted increases in emotional symptoms, but only in children who already had emotional problems at the first assessment point . On the other hand, some studies found pro-social behavior to be negatively associated with emotional symptoms [22, 23]. Pro-social behavior was also shown to be a protective factor in terms of peer acceptance in children with emotional symptoms . Therefore, considering simultaneously the impact of self- and other-oriented social skills on depressive symptoms might give further insights into these apparently controversial results.
Social skills and peer victimization
Deficits in certain social skills may lead to negative reactions by peers, such as rejection or victimization. Peer rejection may be a precursor of peer victimization  and may play a crucial role in stabilizing a child's victim role . In fact Ladd and Troop-Gordon  reported peer rejection to be predictive of later victimization and victimization of later rejection.
Peer rejection has consistently been associated with aggressive and withdrawn behavior , i.e. with deficits in other- and self-oriented social skills. Likewise, in bully/victim research two different pathways to victimization are suggested and the need to differentiate between two types of victims has been emphasized: (1) children who are aggressive and victimized (aggressive victims) and (2) children who are victimized without being aggressive (passive victims) [12, 14, 29, 30]. Accordingly, aggressive as well as withdrawn-submissive behavior patterns are related to peer victimization.
Submissiveness has been discussed as a hallmark of victimization. One explanation for this association is that bullies are looking for easy targets for their assaults . Several studies revealed that passive victims have problems defending themselves [14, 32] and that they are less assertive, for example using fewer persuasion attempts . However, Perren et al.  could not confirm submissiveness as being an overall predictor of peer victimization in kindergarten children and the study by Fox and Boulton  reported submissive behavior in school children to be longitudinally predictive of social exclusion only. Withdrawal behavior (also reflecting a deficit in self-oriented skills) may be associated with victimization by (1) suggesting vulnerability, (2) suggesting low risk of retaliation, and (3) hindering children to find supporting and protecting friends in the class . Also, withdrawn children are not salient and socially less rewarding for their peers. This, in itself, might lessen the chance that peers would help them when they become victimized.
However, in younger children, aggressive behavior seems to be a stronger predictor of victimization and rejection than withdrawn-submissive behavior [36–38]. Nevertheless, not all aggressive children are at risk for becoming victimized and different findings suggest that the most important difference between non-victimized aggressive children (bullies) and aggressive victims consists in their respective ability or inability to control their physical aggression. These uncontrolled aggressive children are fairly disturbing in the class and it seems rather obvious that peers could easily be influenced to assist the bullies .
In sum, deficits in self- and other-oriented social skills are associated with rejection and victimization and deficits in other-oriented social skills have been found to be stronger predictors than deficits in self-oriented social skills.
Peer victimization and depressive symptoms
Children with depressive symptoms have generally been reported to have poorer peer relations in terms of popularity, rejection or victimization [22, 39, 40]. Hawker and Boulton's  meta-analysis of cross-sectional associations between peer victimization and psychosocial maladjustment showed that victimization is most strongly related to depression and least strongly to anxiety. Peer victimization and exclusion may also increase children's depressive symptoms [11, 42–44] or even be causally related to the development of self-derogation and depressive problems [45, 46]. Peer victimization is also associated with health problems, suicidality, and poor school adjustment [47–50].
In sum, empirical findings consistently show that depressive symptoms are associated with negative peer relations and that peer rejection and victimization may play a causal role in the development of depressive symptoms.
The interplay of social skills, peer victimization, and depressive symptoms
As shown above, social skills deficits are not only associated with depressive symptoms but are a strong predictor of peer victimization. Therefore, we assume that these variables interact in systematic ways, especially when we differentiate between self-oriented and other-oriented skills.
As outlined by Bukowski and Adams , peer relations have been discussed as markers, mediators, or moderators for maladjustment in children and adolescents. Similarly, Ladd  suggested different "Child by Environment Models" which take into account the interplay between child behavior, peer relations, and the development of internalizing (and externalizing) disorders. Empirical results mainly support additive or mediation models. For example a four-year longitudinal study by Ladd  has provided support for the additive model. Withdrawn behavior and peer rejection were shown to be overlapping risk factors for the development of children's internalizing problems. A study by Dill and collaborators demonstrated that peer rejection and victimization mediate between children's withdrawn/shy behavior and negative affect . Similarly, a study among young adults showed interpersonal relationships to mediate the impact of social skills on well-being . The conceptual model to be tested in the current paper was partly tested in a cross-sectional study with 198 kindergarten children . The study confirmed the distinct contribution of self- and other-oriented social skills on children's peer victimization and emotional well-being. Deficits in self-oriented social skills predicted higher levels of emotional symptoms, whereas deficits in other-oriented social skills predicted higher levels of peer victimization. The suggested mediating role of peer victimization was partly confirmed.
In the present paper we aim to replicate the findings in a larger sample and most importantly, to include longitudinal data. We hypothesize that deficits in self- and other-oriented social skills are associated with both peer victimization and depressive symptoms. Furthermore, we hypothesize that the impact of deficits in other-oriented social skills on depressive symptoms is mediated through peer victimization whereas deficits in self-oriented social skills are directly associated with depressive symptoms. We also hypothesize that peer victimization is associated with an increase in depressive symptoms over time. We will examine whether the associations are moderated by the child's gender.
In our study we are adopting a dimensional approach of assessing depressive symptoms , i.e. we do not use clinical diagnoses of depression.