Despite well documented associations of peer victimization with loneliness [1–4] health related effects of loneliness among victimized children have not been extensively studied [5]. Loneliness is a hurtful feeling [6, 7] that has been attributed to a discrepancy between desired and achieved levels of social contact [8]. Most children have an intuitive understanding of loneliness [9, 10], and both being alone and sadness are included in their understanding [9].

The results of many studies have suggested that loneliness is associated both with anxiety and depression among children [1, 11–14]. Also, lonely children appear to be less accepted [15] and more rejected by their peers [9, 16–18]. Compared to popular children who have many friends, the lonely children have fewer, and children with no friends appear as the most lonely [16, 19]. Intervention studies that aimed to increase the students’ attachment or belongingness to their school have shown reduced loneliness among the participants [20, 21].

Victimization (being bullied) is a harsh form of peer rejection, and includes being the target of aggressive behaviour, repetitive negative acts and imbalance of power [22–24]. There is consensus that children who are subject to bullying are at increased risk of mental health problems [25, 26], psychosomatic illness [27], and psychosocial maladjustments [28, 29]. Further, it has been suggested that rejected, anxious, or depressed children in the next turn more easily are bullied by peers than children without internalizing or adjustment problems [30, 31].

Previous studies have shown that friendship among peers may modify and protect against the adverse effects of victimization [32, 33]. Whether loneliness also has a modifying effect, has scarcely been studied. Therefore, we have used population data among school children to assess whether the health related effects of victimization could be modified by loneliness. We hypothesized that children who report both victimization and loneliness would have a higher prevalence of health symptoms than victimized children who do not report loneliness.

Separate analyses for boys and girls did not demonstrate any substantial differences between the genders in the multivariable analyses, except for headache. For boys, both victimization and loneliness showed statistically non-significant associations with headache, but for girls, loneliness was related to headache (odds ratio, 1.5, 95% CI 1.0 to 2.3) whereas victimization showed a weaker association.

In cross-sectional data on 419 school children, the impact of loneliness among victimized children was assessed in relation to self-reported health symptoms. The main finding was the adverse and modifying influence of loneliness among victimized children. As hypothesized, victimized children who also felt lonely had higher prevalence of health symptoms than victimized peers who did not report loneliness. The unexpected finding, however, was the large gap in prevalence of health symptoms between the two groups. Children who experienced both loneliness and victimization had three to seven times higher prevalence of internalizing or somatic symptoms compared to children who reported no victimization and no loneliness (the reference group). In contrast, victimized children who reported no loneliness were no more likely to have any health symptoms than the reference group. Moreover, lonely children who reported no victimization, showed approximately the same prevalence of internalizing symptoms and headache as lonely children who were victimized.

The study was conducted in rural communities, ranging from inland to coastal areas. All children attended schools in the Norwegian public school system. The population base and the very high participation are strengths of the study, but it is a weakness that the data do not include children from urban settings. The convenience sampling of schools may also be a limitation. The reported prevalence of victimization and the decline across school grades are, however, in line with results from other relevant studies [24, 36–39] and may support the external validity of the findings. By carefully following the questionnaire, school nurses interviewed the youngest children, whereas older children completed the questionnaire themselves. Although the nurses were trained for this task, we cannot exclude the possibility that different procedures could have influenced the collected information and introduced systematic differences in results between younger and older children. As described in the methods the variables applied in the study were dichotomized never/seldom versus sometimes or more often. The chosen cutpoint was preferable as regards the dispersion of data (Table 1) and power of the analyses, and also it was important to have a reference group of children who were neither victimized nor lonely. The cross-sectional design is a limitation of this study. Compared to longitudinal designs that allow investigating causal effects the cross-sectional design limit the researcher to report on associations. The findings must therefore be interpreted with caution.

It may be argued that the Victim-not-lonely group of children had lower proportions of weekly or daily victimization than the Victim-and-lonely group, and this may explain some of the differences between the two groups as higher frequencies of victimization are known to have stronger associations with health symptoms than lower frequencies [40–42]. On the other hand, victimization at lower frequency, i.e. sometimes, has also shown strong and consistent associations with health symptoms [41, 42]. One marked difference between those studies and the present study is, however, the exclusion of lonely children in the Victim-not-lonely group. Further, lonely children had a high prevalence of health symptoms regardless of reporting victimization or not. This finding is rather surprising since an additive effect of victimization could be expected for the lonely children who were also victimized. Possibly, the finding reflects loneliness as unique and painful experiences with strong individual relations to health symptoms. This is in line with research that presents strong associations of loneliness with depression and anxiety [11, 13, 43]. The relation between loneliness and somatic symptoms are, however, scarcely explored.

The two groups Victim-and-lonely and Victim-not-lonely showed a decrease across school grades, and these findings correspond to previous results on victimization by bullying that have reported far more children to be victimized during the first years in school than in later years [36, 37]. For the Lonely-not-victim group, another pattern was revealed with no significant downward or upward trends across the ten school years. To our knowledge, no other studies have reported measures on loneliness across ten school grades (from 7 to 16 years of age), but our findings are supported by publications that have reported approximately the same prevalence of loneliness in US children from preschool to sixth grade [9, 16]. On the other hand, Greek and Finnish studies have reported a remarkably higher prevalence of loneliness among primary school children [10, 44, 45].

The strong association of loneliness with internalizing as well as somatic symptoms calls for attention in schools and health care. It is important to search for strategies that prevent the development of loneliness and additionally, protect against painful feelings related to loneliness. Findings in previous studies may lead to some suggestions of strategies including friendship, participation in class, and belongingness to school. After training of the students’ social and emotional skills intended to increase the students’ belongingness to school, the prevalence of loneliness among students was reduced [20, 21] and also, belongingness to school may be a buffer against depression among lonely children [46]. Further, loneliness may be related to shyness and indirectly to passivity [47] and it has been suggested to work with participation in the classroom instead of working directly with loneliness [47]. This strategy may be supported by a study reporting negative associations of loneliness with competence and support from peers [48]. Thus, circles of passivity, sadness, rejection, and isolation can be turned to positive loops of participation, skilled interactions, increased popularity, and more friends [19–21, 47]. Furthermore, training of social and emotional skills and participation in the classroom can be included in the daily activities at school, but there is a need for studies designed to evaluate such pedagogical practice.

Our findings indicate that some children who report victimization may be little influenced by the experience as far as health symptoms are concerned. Moreover, it may be hypothesized that victimization is harmful to health only when the experience is linked to hurtful thoughts or feelings that may be present in loneliness. For peer victimized school children, loneliness may therefore be especially harmful. In addition, loneliness may be harmful among children without experiences of peer rejection or other peer harassment. This indicates that we need to be aware of loneliness at school – among all the children – and pedagogical practice that aims to promote inclusion and prevent loneliness should be highly acknowledged. The relation between loneliness, victimization, and children’s health needs to be further explored, also in longitudinal studies.