جداسازی و استرس مورد زورگویی واقع شدن

Approximately one-third of children report being victims of bullying, and this victimization has been linked to a number of negative psychological outcomes. In the present study, we examined the effects of perceived isolation on the link between victimization before and during high school and stress symptoms during college. Consistent with our predictions, victimization appears to do the most damage to those who felt isolated during high school. These results suggest that schools should reframe their approach to the bullying problem, and devote more resources to helping students feel less isolated.

Introduction Being bullied by peers in childhood is a stressful experience. A growing body of research has identified several risk factors and consequences of being bullied, including psychological maladjustment (for a review, see Hawker & Boulton, 2000). However, research has yet to examine moderators of the consequences of being bullied. In the present article, we argue that bullying is best understood as a chronic stressor, and examine individual differences in social support as a moderator of reactions to bullying. According to most estimates, roughly 30% of children report being victims of bullying at some point, and between 5% and 10% are victims on a regular basis (e.g. Perry, Kusel, & Perry, 1988; Rigby & Slee, 1991; Olweus, 1993; Nansel et al., 2001). Boys are more likely than girls to bully and to be bullied (e.g. Arora & Thomson, 1987; Slee & Rigby, 1993; Siann, Callahghan, Glissov, Lockhart, & Rawson, 1994; Nansel et al., 2001), although changing definitions of “bullying” to include more indirect aggression (e.g. teasing, ostracism) can reduce the gender difference (e.g. Crick, 1995). In a study of middle schools in Italy, girls’ experiences with bullying included name-calling, rumours, rejection, and teasing, whereas boys’ experiences included more physical harm, threats, and rejection (Baldry, 1998). Being a victim of bullying has been associated with a number of negative psychological outcomes, including anxiety (e.g. Perry et al., 1988; Craig, 1998; Bond et al., 2001), depression (e.g. Neary & Joseph, 1994; Craig, 1998), and poorer perceptions of self-worth and competence (e.g. Roland, 1989; Slee & Rigby, 1993; Neary & Joseph, 1994). A recent meta-analysis found that victimization was related most strongly to depression and least strongly to anxiety (Hawker & Boulton, 2000). Although females tend to report more of these symptoms than males, and although males and females are subjected to different types of bullying, there is typically not a sex difference in the consequences of bullying (i.e. no interaction; Perry et al., 1988; Roland, 1989; Slee & Rigby, 1993; Craig, 1998; for an exception, see Rigby & Bagshaw, 2001).

The results are presented as follows. First, we examined the frequency and duration of victimization. Next, we tested the relationship between frequency of victimization and stress symptoms (Hypothesis 1), followed by the relationship between duration of victimization and stress symptoms (Hypothesis 2). We then explored the relationship between victimization and isolation. Finally, we tested our hypothesis regarding the moderating role of isolation (Hypothesis 3). Frequency and duration of bullying Before high school, 33% of the sample reported being bullied occasionally, and 26% reported being bullied frequently. Males were more likely to report occasional victimization (χ2(1)=8.80χ2(1)=8.80, p=.00p=.00), whereas females were more likely to report no experience with victimization (χ2(1)=49.74χ2(1)=49.74, p=.00p=.00). See the top half of Table 1 for these percentages. Table 1. Frequency of victimization, by gender Reported frequency of victimization Not at all Occasionally Frequently Total Before HS Males 30.8% (107) 33.7% (117) 35.4% (123) 100% n=347n=347 Females 47.0% (238) 33.0% (167) 20.0% (101) 100% n=506n=506 Total 40.4% (345) 33.3% (284) 26.3% (224) 100% N=853N=853 During HS Males 57.1% (198) 28.8% (100) 14.1% (49) 100% n=347n=347 Females 72.7% (368) 21.5% (109) 5.7% (29) 100% n=506n=506 Total 66.4% (566) 24.5% (209) 9.1% (78) 100% N=853N=853 Table options During high school, 25% of the sample reported being bullied occasionally, and 9% reported being bullied frequently. Similar to the gender difference seen before high school, males were more likely to report frequent victimization (χ2(1)=5.13χ2(1)=5.13, p=.02p=.02), and females were more likely to report no experience with victimization (χ2(1)=51.06χ2(1)=51.06, p=.00p=.00). See the bottom half of Table 1 for these percentages. There was a strong relationship between bullying prior to and during high school (χ2(4)=266.39χ2(4)=266.39, p=.00p=.00), which held for both males (χ2(4)=109.30χ2(4)=109.30, p=.00p=.00) and females (χ2(4)=137.87χ2(4)=137.87, p=.00p=.00). Among those who were bullied frequently prior to high school, 67% were also bullied during high school. Likewise, among those not bullied prior to high school, only 11% were bullied during high school. Interestingly, among the 224 people bullied frequently before high school, roughly one-third were bullied frequently, one-third were bullied occasionally, and one-third were not bullied at all during high school. We return to these three groups below, to examine the effects of duration of bullying on stress symptoms. See Table 2 for frequencies of chronic victimization. Table 2. Chronic victimization Before high school During high school Total Not at all Occasionally Frequently Females Not at all 90.8% (216) 7.1% (17) 2.1% (5) 100% n=238n=238 Occasionally 68.3% (114) 31.1% (52) .6% (1) 100% n=167n=167 Frequently 37.6% (38) 39.6% (40) 22.8% (23) 100% n=101n=101 Males Not at all 85.0% (91) 13.1% (14) 1.9% (2) 100% n=107n=107 Occasionally 61.5% (72) 35.9% (42) 2.6% (3) 100% n=117n=117 Frequently 28.5% (35) 35.8% (44) 35.8% (44) 100% n=123n=123 Total Not at all 89.0% (307) 9.0% (31) 2.0% (7) 100% n=345n=345 Occasionally 65.5% (186) 33.1% (94) 1.4% (4) 100% n=284n=284 Frequently 32.6% (73) 37.5% (84) 29.9% (67) 100% n=224n=224 Table options Frequency and stress symptoms We examined the relations between frequency of victimization and stress symptoms in two ways. First, we examined the relationship between victimization prior to high school and total stress symptoms1 (Fig. 1, top panel). Second, we examined the relationship between victimization during high school and total stress symptoms (Fig. 1, bottom panel). Both were examined using a 3 (frequency of bullying: never, occasionally, frequently)×2 (gender) ANOVA, with total symptoms reported as the dependent variable. Total stress symptoms, by gender and frequency of victimization. Note: (Top ... Fig. 1. Total stress symptoms, by gender and frequency of victimization. Note : (Top panel) victimization prior to high school; (bottom panel) victimization during high school. Total number of symptoms reported across all subscales. Error bars=1bars=1 standard error. Figure options For the analysis of victimization prior to high school, as expected, there was a main effect of gender (F(1,847)=24.40F(1,847)=24.40, p=.000p=.000), with females reporting more stress symptoms than males. There was also a linear main effect of bullying (F(2,847)=23.87F(2,847)=23.87, p=.000p=.000), such that those who were bullied before high school more reported more stress symptoms. Post hoc tests using a Bonferroni correction revealed that the “occasional” victimization group was no different than the not bullied group (p=.946p=.946), but the “frequent” victimization group was significantly different from both of the others (p 's<.01). The interaction with gender was not significant. However, post hoc tests revealed the difference between “occasional” and “frequent” victimization was significant for females (p=.001p=.001), but not for males (p=.158p=.158). For the analysis of victimization during high school, as expected, there was a main effect of gender (F(1,847)=21.24F(1,847)=21.24, p=.00p=.00), such that females reported more stress symptoms than males. Also as expected, there was a linear main effect of bullying (F(2,847)=23.87F(2,847)=23.87, p=.00p=.00), such that those who were bullied during high school more reported more stress symptoms. Post hoc tests revealed that, unlike bullying before high school, all three groups were significantly different from one another (p 's<.01). The interaction with gender was not significant. However, post hoc tests revealed a slightly different pattern for males and females. Females bullied “occasionally” had higher stress symptoms than those not bullied at all (p=.01p=.01), but this differences was not significant for males (p=.16p=.16). The difference between “occasional” and “frequent” victimization was significant for both males (p=.03p=.03) and females (p=.02p=.02). Duration and stress symptoms To examine the relationship between duration of victimization and stress symptoms, we limited our analysis to those who were bullied frequently before high school (n=224n=224), and compared those bullied frequently, occasionally, and not at all during high school. This comparison was examined using a 3 (frequency of bullying: never, occasionally, frequently)×2 (gender) ANOVA, with total stress symptoms as the dependent variable. As expected, there was a main effect of gender (F(1,218)=18.69F(1,218)=18.69, p=.000p=.000), such that females reported more stress symptoms than males. There was also a main effect of high school victimization (F(2,218)=7.24F(2,218)=7.24, p=.000p=.000), such that those bullied more frequently reported more stress symptoms. The interaction was not significant. As seen in Fig. 2, those who experienced frequent bullying prior to high school, but no bullying during high school, reported fewer stress symptoms. Duration of bullying and stress symptoms. Note: Limited to those bullied ... Fig. 2. Duration of bullying and stress symptoms. Note : Limited to those bullied “frequently” before high school (n=224n=224). Total number of symptoms reported across all subscales. Error bars=1bars=1 standard error. Figure options Victimization and isolation Before examining the moderating role of isolation, we compared the frequencies of victimization for individuals at the extremes of the high school isolation scale. “Non-isolated” individuals were defined as those reporting a 1 or a 2 on the isolation scale. “Isolated” individuals were defined as those reporting a 4 or a 5 on the isolation scale. (These extreme scores were used for this analysis only.) If isolation and victimization are independent, then the ratio of “non-isolated” to “isolated” individuals should be roughly equivalent across levels of victimization. Prior to high school, the ratios varied slightly between those never victimized (94% non-isolated and 6% isolated), those occasionally victimized (96% and 4%), and those frequently victimized (90% and 10%), χ2(2)=8.41χ2(2)=8.41, p=.015p=.015. During high school, the ratios also varied slightly between those never victimized (95% and 5%), those occasionally victimized (92% and 8%), and those frequently victimized (85% and 15%), χ2(2)=12.39χ2(2)=12.39, p=.002p=.002. At both time periods, those victimized frequently were more likely to be isolated, but a majority of this group still fell into the “non-isolated” category. In fact, the amount of variance shared between isolation and victimization at either time period was less than 4% (r 's (848)=.13 and .17, p’s=.000p’s=.000). Thus, it appears possible to be victimized without feeling isolated, and vice versa. Isolation as a moderator To test our primary hypothesis that the consequences of victimization depend on one's perceived isolation, we conducted two hierarchical linear regressions on total stress symptoms. The first examined moderation of victimization prior to high school, and the second examined moderation during high school. In both analyses, gender (females=0females=0; males=1males=1), perceived isolation (1–5), and frequency of victimization (1–3) were entered on the first step. The two-way interactions were entered on the second step, and the three-way interaction was entered on the third step. Coefficients for both analyses are presented in Table 3, separated by time period. Table 3. Regression coefficients: predicting total stress symptoms from isolation and victimization B (SEB) ββ Pre-high school High school Pre-high school High school Step 1 Constant 43.47 (1.12) 42.81 (1.07) Gender −3.84 (.76) −3.96 (.75) −.172*** −.177*** Victimization 2.19 (.47) 3.57 (.57) .160*** .213*** Perceived isolation 1.62 (.41) 1.44 (.41) .133*** .118*** Step 2 Constant 45.36 (1.3) 45.31 (2.13) Gender −.84 (2.15) −.29 (2.18) −.038 −.013 Victimization 1.11 (.51) 1.36 (1.40) .081* .081 Perceived isolation −.65 (.77) −.24 (.98) −.054 −.020 Victimization×isolation 1.65 (.34) 1.41 (.57) .326*** .278* Isolation×gender −1.03 (.81) −1.05 (.82) −.104 −.107 Victimization×isolation −.97 (1.04) −1.20 (1.14) −.079 −.098 Step 3 Constant 45.53 (1.35) 46.61 (2.64) Gender −3.17 (3.24) −3.12 (4.04) −.142 −.139 Victimization 1.03 (.52) .43 (1.79) .076* .025 Perceived isolation −.90 (.81) −.91 (1.27) −.074 −.075 Victimization×isolation 1.81 (.38) 1.88 (.80) .357*** .371* Isolation×gender .36 (1.66) .40 (1.93) .036 .041 Victimization×isolation .49 (1.84) .71 (2.56) .040 .058 Three-way interaction −.85 (.89) −.94 (1.13) −.169 −.187 Note: “B ” represents unstandardized coefficients; “ββ“represents standardized coefficients. For pre-high school analysis: Step 1, R2=.061R2=.061***; Step 2, ΔR2=.032R2=.032**; Step 3, ΔR2=.001R2=.001. For high school analysis: Step 1, R2=.082R2=.082***; Step 2, ΔR2=.009R2=.009*; Step 3, ΔR2=.001R2=.001. *** p<.001p<.001; ** p<.01p<.01; * p<.05p<.05. Table options Prior to high school, all three variables significantly affected stress symptoms in Step 1. When the interactions were entered in Steps 2 and 3, only victimization and the victimization×isolation interaction were significant predictors of stress symptoms. During high school, all three variables significantly affected stress symptoms in Step 1. However, these effects were qualified by the victimization×isolation interaction, which emerged as the only significant predictor in both Step 2 and Step 3. This interaction showed the same pattern before and during high school, so only the latter is depicted in Fig. 3. As this figure shows, the relation between frequent bullying and stress symptoms was most pronounced among those who reported a higher degree of isolation. This is consistent with the hypothesis that being bullied does the most damage to those who are isolated. For comparison with the other results, Fig. 4 presents total stress means by victimization and the isolation grouping described above. Victimization, isolation, and stress symptoms. Note: Y-axis represents ... Fig. 3. Victimization, isolation, and stress symptoms. Note: Y-axis represents standardized predicted values from the regression equation described in the text, and shown in Table 3. Error bars around each point represent deviation from the predicted values. Figure options Total stress symptoms, by isolation level and victimization. Note: Total number ... Fig. 4. Total stress symptoms, by isolation level and victimization. Note : Total number of symptoms reported across all subscales. Error bars=1bars=1 standard error. “Non-isolated”=1 or 2 on isolation scale (N=680N=680); “midpoint”=3 on isolation scale (N=125N=125); “isolated”=4 or 5 on isolation scale (N=45N=45).