In the last three decades, the research activity concerning school bullying has increased enormously. Smith (2014) describes this research expansion as consisting of four main phases: Origins (1970-1988); establishing research programs (1989-mid 1990s); establishing international research programs (mid 1990s-2004); and cyberbullying (2004-present). In accordance with those phases, intervention programs were developed to reduce school bullying, appearing first in the Origins phase, and now the interventional efforts are ongoing in cyberbullying phase.
School bullying is a world-wide problem (Smith et al., 1999). Terms similar to bullying are used to describe the problem with different nuances and approaches to tackle them in each country (Smith et al. 2002). However, the basic nature of the problem includes “systematic abuse of power” (Smith & Sharp, 1994: p. 2), “relationship problems” (Pepler, 2006) or “repetitive aggression in relationships” (Toda, Strohmeier, & Spiel, 2008; Toda, 2011).
In Japan, bullying has been one of the serious problems in Japanese schools since mid-1980, occasionally evoking a media focus on suicide cases (Toda, 2016). But it took rather long in Japan until bullying prevention law was enacted (Japanese Ministry of Education, 2013). However, following and learning from Western intervention programs, such as Social and Emotional Learning, social skills training and so on, some pioneer schools have introduced and practiced bullying prevention programs (Koshi, 2013). In addition, school counsellors have been struggling to support victims of bullying and supervising teachers to handle serious cases since 1990’s (Yagi, 2008). In order to prevent victimization from bullying effectively, it is important to intervene not only at the school level and classroom level but also at the individual level (Smith, Ananiaduou, & Cowie, 2003). That is why coping strategies for victims have been developed (e.g., Kochenderfer-Ladd & Skinner, 2002; Kristensen & Smith, 2003 ). These coping strategies were compared between Eastern and Western countries (e.g., Kanetsuna & Smith, 2002 ). However, the coping strategies pupils use against bullying have been relatively unexamined in considering bullying prevention.
There are many strategies that pupils use against bullying in Japan but one strategy would be telling others about acts of bullying and seeking help from an appropriate person (Smith & Shu, 2000). Tenenbaum, Varjas, Meyers, and Parris (2011) interviewed 102 pupils and found that seeking social support were the most commonly reported problem-focused coping strategies used by victims of bullying.
In school settings, friends and teachers are the most common helping resources (Mizuno & Ishikuma, 2004). Talking about bullying with teachers is an effective help-seeking strategy (Smith & Shu, 2000). Eliot, Cornell, Gregory, and Fan (2010) reported that pupils who perceived their teachers to be supportive were more likely to endorse positive attitudes toward seeking help. This tendency has been confirmed in Japanese junior high school pupils (Yamanaka & Hiraishi, 2015). But Boulton (2005) pointed out that almost one in 10 pupils said that they would “definitely not” seek help for bullying from their teachers. Dowling and Carey (2013) reported that teachers are considered to be one of the most difficult sources to talk to about being bullied, and were unlikely to make the victims feel better. Some pupils feel that telling a teacher will even result in worse situations of bullying (Fekkes, Pijpers, & Verloove-Vanhorick, 2005).
This finding is also applicable to Japanese pupils. Japanese children are more likely to seek help from their friends, rather than teachers or counsellors (Yamaguchi, Mizuno, & Ishikuma, 2004). Therefore, the establishment of a peer-support system appears to be one method for dealing with problems of bullying. Victims of bullying value genuineness and empathic understanding, which is typical of the type of support that close friends offer (Boulton, 2005).
Studies in Western societies show that it is not difficult to seek help from friends (Oliver & Candappa, 2007). In Japanese pupils, Nagai (2009) points out that children who report severe peer victimization are less likely to seek help from their friends. So in this Japanese sample, the victims of bullying are less likely to seek help even from their friends. As Nakai (1997) explains, the victims of bullying become isolated and are forced to learn helplessness through violence. Help-seeking behaviors show individual differences, due to gender, challenge appraisals, emotions and so on (Hunter, Boyle, & Warden, 2004). Especially, gender is one of the strongest predictors of help seeking (Glover, Gough, Johnson, & Cartwright, 2000; Hunter & Boyle, 2002).
In Japan, research concerning help seeking behaviors in counseling filed stared in 1999 (Mizuno & Ishikuma, 1999) and the scale of help seeking behaviors is developed by Tamura and Ishikuma (2001). This scale was assessing two dimensions of help seeking expectations. One is positive side of help seeking concerning how helper is responsible for seeking help. And the other side of help seeking is fear and concerns of accessibility of help seeking.
We focused on school bullying and help-seeking expectations, and hypothesized a linear/non-linear relationship between those constructs. Moreover, we hypothesize that this relationship differs between girls and boys, controlling for age. Because age and gender are the two main factors that affects help seeking behaviors (Mizuno & Ishikuma, 1999).
We recruited Japanese adolescents from 16 classes of one elementary and one junior high school situated in a Japanese city in 2008. In Japan, elementary school comprises grades 1 to 6 and junior high school comprises grades 7 to 9. Consistent with standard ethical procedures in Japan then, the school principals and teachers approved the study before data collection. The anonymous survey was conducted by homeroom teachers, after the approval of school principals in place of parental consent. Students were told they could leave the study at any time and received post-survey lecture on coping bullying.
A total of 532 Japanese pupils (53% boys), mean age 12.33 (SD = 1.78) years, participated in this study. Altogether, 29 records (5.5%) were incomplete. The percentage of missing values across the 19 variables used in the present study varied between 0.0 and 1.7%. Multiple imputation under the missing at random (MAR) assumption was used to deal with missing data (Rubin, 1987). Incomplete variables were imputed under fully conditional specification (van Buuren, Brand, Groothuis-Oudshoorn, & Rubin, 2006) including not only variables used in the present study but several potential auxiliary variables (Collins, Schafer, & Kam, 2001). A total of 20 imputed data sets were extracted during the imputation process. Calculations were done in R (R Core Team, 2014) using the mice package (Van Buuren & Groothuis-Oudshoorn, 2011).
1) Help-seeking expectations scale. Help-seeking expectations were measured with items adapted from a scale developed by Tamura and Ishikuma (2001). Four items measured responsibility expectations, e.g., “My friends would understand my emotional feelings when I tell my problems” and three items measured accessibility expectations, e.g., “I feel ashamed in consulting with my friends”. All items were presented using a five point Likert-Scale: 1 (strongly agree), 2 (somewhat agree), 3 (neutral), 4 (somewhat disagree), and 5 (strongly disagree). The Cronbach’s α coefficient for responsibility expectations was α = .78 (α = .76 for girls and α = .78 for boys) and for accessibility expectations α = .58 (α = .57 for girls and α = .60 for boys). Items for accessibility expectations were reversed to enhance interpretability. Thus, high values indicate high accessibility expectations.
2) Relational and overt victimization. Relational and overt victimization were measured with items adapted from the peer nomination measure originally developed by Crick and Grotpeter (1995). Seven items measured relational victimization, e.g., “How often during the last six months have you been excluded from play or another activity by one or more classmates?” and three items measured overt victimization, e.g., “How often have you been hit by one or more classmates during the last two months?” Answers to all questions were given on a five-point response scale: 0 (not at all), 1 (once or twice), 2 (two or three times a month), 3 (once a week), and 4 (nearly every day). The Cronbach’s α coefficient for relational victimization was α = .86 (α = .83 for girls and α = .88 for boys) and for overt victimization α = .82 (α = .74 for girls and α = .82 boys).
The items for all the scales used in the present study are shown in Appendix A.
2.4. Measurement Model
Confirmatory factor analysis (CFA) using Mplus 7.11 was conducted to investigate the measurement model for help-seeking expectations and relational and overt victimization in the whole sample and for girls and boys assuming strong measurement invariance. For help-seeking expectations, the two-factor model (responsibility expectations and accessibility expectations) with continuous indicators yielded an acceptable model fit for the whole sample (χ2(13) = 46.168, p < .001, CFI = .936, RMSEA = .069) and across girls and boys (χ2(36) = 84.734, p < .001, CFI = .905, RMSEA = .071). The path diagram of the measurement model is depicted in Figure 1.
For victimization, a two-factor model (overt and relational victimization) with ordered-categorical indicators yielded a good model fit for the whole sample (χ2(34) = 108.153, p < .001, CFI = .976, RMSEA = .064) and across girls and boys (χ2(99) = 169.081, p < .001, CFI = .979, RMSEA = .052).
Factor scores based on the measurement models assuming strong measurement invariance were extracted and subsequently used in the main analysis of the present study.
2.5. Analytic Strategy
In order to test the main hypotheses of the study, we conducted regression analyses separately for girls and boys to predict responsibility expectations and accessibility expectations controlling for age. A linear and a quadratic effect of overt and relational victimization was included in the model to account for non-linear relationships between help-seeking expectations and victimization.
Figure 1. Measurement model of help-seeking expectations: Unstandardized parameter estimates for girls and boys.
The predictors for overt and relational victimization were centered and subsequently squared to compute the quadratic terms.
3.1. Descriptive Statistics
Means, standard deviations and zero-order correlations of all study variables are presented separately for boys and girls in Table 1.
3.2. Relationship between Help-Seeking Expectations and School Bullying
The results of the regression analyses for girls and boys for the criterion variables responsibility expectations and accessibility expectations are displayed in Table 2.
For responsibility expectations, the regression model was statistically significant for both girls (F(5, 249) = 2.465, p < .05) and boys (F(5, 281) = 7.428, p < .001). Thus, the model is useful for predicting responsibility expectations explaining 5.1% of the variance for girls and 11.8% of the variance for boys. The linear term of relational victimization was statistically significant for predicting responsibility expectations (b = .297, p < .05) for girls, controlling for all other predictors in the model. Higher relational victimization corresponded to higher responsibility expectations for girls. However, the quadratic term of relational victimization was statistically significant for predicting responsibility expectations (b = .307, p < .05) for boys. This indicates a u-shaped relationship between relational victimization and responsibility expectations. More specifically, there is a negative relationship between relational victimization and responsibility expectations below the mean of relational victimization, but an increasing positive relationship above the mean of relational victimization. See Figure 2(Panel A)
Table 1. Means, standard deviations, and zero-order correlations among study variables for girls and boys.
Note. Correlations for girls are shown in the lower triangle, correlations for boys are shown in the upper triangle; statistically significant correlations at α = .05 are shown in bold.
Table 2. Multiple regression: Results for girls and boys for responsibility expectations and accessibility expectations.
Note. SE = standard error; 95% CI = 95% confidence interval; statistically significant results α = .05 are shown in bold; relational and overt victimization and age are centered.
Figure 2. Model implied relationships between relational victimization and criterion variables; responsibility expectations in Panel A and accessibility expectations in Panel B. Note: Overt victimization and age are statistically held constant.
for the relationship between relational victimization and responsibility expectations for girls and boys separately. All other predictors in the model were not statistically significant.
For accessibility expectations, the regression model was statistically significant for both girls (F(5, 249) = 4.132, p < .01) and boys (F(5, 281) = 4.936, p < .001). Thus, the model is useful for predicting responsibility expectations, explaining 8.0% of the variance for girls and 8.6% of the variance for boys. The linear term of relational victimization was statistically significant for predicting responsibility expectations for girls (b = .281, p < .05) and for boys (b = .421, p < .05), controlling for all other predictors in the model. In addition, the quadratic term of relational victimization was statistically significant for predicting responsibility expectations (b = −0.277, p < .05) for boys. Thus, there is a reversed u-shaped relationship between relational victimization and accessibility expectations. More specifically, there is a positive relationship between relational victimization and responsibility expectations below the mean for relational victimization, but an increasing negative relationship above the mean for relational victimization. See Figure 2(Panel B) for the relationship between relational victimization and accessibility expectations for girls and boys separately. All other predictors in the model were not statistically significant.
The main goal of the present study was to investigate the relationship between help-seeking expectations and relational and overt victimization. First, by conducting Structural Equation Modeling, we found that help-seeking expectations are composed of two kinds of expectations: responsibility expectations and accessibility expectations.
We hypothesized that experiencing school bullying would predict help-seeking expectations. The most important finding from the results of the present study is that the relationship between victimization and help-seeking expectations is complex, especially among boys. Regarding the relationship between relational victimization and responsibility expectations, the results show that relational victimization for girls is positively associated with help-seeking expectation. The linear term of relational victimization was statistically significant for predicting responsibility expectations for girls, controlling for all other predictors in the model. Higher relational victimization corresponds to higher responsibility expectation for girls.
However, the results suggest some complexity for boys. The quadratic term of relational victimization was significant for predicting responsibility expectations. More specifically, there was a negative relationship between relational victimization and responsibility expectations below the mean for relational victimization, but an increasingly positive relationship above the mean for relational victimization. Thus, boys who experience victimization have higher scores for responsibility expectations, which indicates a positive attitude toward help-seeking.
Regarding accessibility expectations, the linear term of relational victimization was statistically significant for predicting accessibility expectations for both girls and boys. In addition, the quadratic term of relational victimization was statistically significant for predicting responsibility expectations for boys. Thus, there is a reversed u-shaped relationship between relational victimization and accessibility expectations. More specifically, there is a positive relationship between relational victimization and accessibility expectations below the mean for relational victimization, but an increasingly negative relationship above the mean for relational victimization.
Regarding responsibility expectations, boys who experience severe relational victimization may seek help from friends. However, for accessibility expectations, boys who experience severe relational victimization may not seek help from friends. Some research studies are in conflict with the results of our present study (Mouttapa, Valente, Gallaher, Rohrback, & Unger, 2004; Wilson, 2010). Nevertheless, our results support Nagai (2009)’s study of Japanese elementary school students. Furthermore, the results of the present study are a little more complicated. Much more victimized boys may expect their friends’ help (responsibility expectations), but their accessibility expectations may be rather pessimistic.
One possible explanation for this finding is gender role conflict theory, which suggests that because boys have a masculine gender identity they may hesitate to seek help because seeking help is seen as feminine behavior (Good, Dell, & Mintz, 1989). Not surprisingly, Japanese male pupils score higher on a masculinity scale than female pupils (Sugihara & Katsurada, 2000). Therefore, in our study, boys might have had a stronger inclination not to seek help because of their masculine identity, which might have reduced accessibility expectations for those who experienced relational victimization. One key factor that interferes with help-seeking by those who have a strong masculine identity is emotional openness. Boys with a strong masculine identity tend not to be open about their emotions. This may explain the findings for the boys in the present study. Boys are reluctant to be open about their emotional experiences with friends. That is why boys who experience relational victimization need to be open about their emotions when they seek help from friends (Good, Dell, & Mintz, 1989).
Another possible explanation is that boys who experience severe relational victimization may have some kind of difficulty in seeking help. O’Neil (2013) reviewed research concerning gender role conflicts and asserted that men who had a high score on a gender role conflict scale reported greater difficulty coping with problems. Boys, who had been experiencing relational victimization, may experience gender role conflict, which might explain their hesitancy to ask friends for help. But further research is needed to explore why boys showed these complex results.
The implications for intervention are that it may be necessary to take two different approaches to deal with bullying for boys and girls. A suggested intervention for boys to increase help-seeking expectations about teachers and counselors should have more focus on bullying. One possible solution for victimized boys should take into account the boys’ help-seeking styles. When bullying is occurring, the intervention should consist of improving the classroom climate, such as the creation of a harmonious classroom climate. This harmonious classroom climate may influence boys to seek help. Promoting the attitude that bullying is unacceptable may lead to more boys seeking help (Williams & Cornell, 2006). It is generally essential for schools to create an environment that discourages bullying, and for teachers to provide support for children. However, support for boys in particular should include careful observation to identify bullying victimization since our findings suggest that boys are less likely to seek help from peers than girls. The approach to bullying/victimization should consider disclosing the presence of bullying cases at school, and thereby guiding boys to express their negative emotion about victimization. One of the potentially effective solutions would be to provide boys with opportunities to reflect on their expression of feelings in a safe environment. Incorporating Social-Emotional Learning (SEL) into the work of teachers and counselors might serve as an effective solution helping boys to reflect their emotional feelings and try to express it in safer circumstances. Though, further research is needed to identify effective ways to support boys who are experiencing victimization from bullying.
As for limitations, this study derived from a questionnaire research administered at only one time point. As a next step, a longitudinal study should be considered. Second, as the pupils are asked about how they seek help from their friends, the answers may reflect their social networks, which should be considered in the further research. Finally, this study analyze data from Japanese context. However, the questionnaire did not include scales which reflect Asian values or collectivism (cf. Triandis & Gelfland, 1998 ). The results, especially about boys, may have been influenced by Asian values that discourage disclosing emotional feelings (Lim, 2016). Thus, research into help seeking related to bullying of Japanese pupils should be conducted in the context of Asian value and culture in the future.
1) My friends would understand my feelings when I tell them my problems.
2) My friends will keep secret what I talked to them about, if we make a promise.
3) Talking with my friends triggers solutions of my problems.
4) My friends are kind enough to solve my problems with me, if I ask them.
1) I feel ashamed in consulting with my friends.
2) My friends may feel bothered if I consult them about my small concerns.
3) It would take too much of my friends’ time to listen to my concerns.
1) During the last 3 months, how often have you been left out on purpose by a young person or a group of young persons when it was time to play or do an activity?
2) During the last 3 months, how often has a young person or a group of young persons who were mad at you got back at you by not letting you be in their group anymore?
3) During the last 3 months, how often has a young person or a group of young persons told lies about you to make other kids not like you anymore?
4) During the last 3 months, how often has a young person or a group of young persons told you they won't like you unless you do what they say?
5) During the last 3 months, how often has a young person or a group of young persons tried to keep others from liking you by saying mean things about you?
6) During the last 3 months, how often has a young person or a group of young persons suddenly stopped talking and stayed silent when you approached them?
7) During the last 3 months, how often has a young person or a group of young persons treated you as if you were not there?
1) During the last 3 months, how often have you been hit by a young person or a group of young persons?
2) During the last 3 months, how often have you been pushed or shoved by a young person or a group of young persons?
3) During the last 3 months, how often have you been kicked or have been pulled by the hair by a young person or a group of young persons?