رفتار فرزندپروری و خطر تبدیل شدن به یک قربانی و یک زورگو/قربانی: یک مطالعه متاآنالیز

Abstract Objective Being bullied has adverse effects on children's health. Children's family experiences and parenting behavior before entering school help shape their capacity to adapt and cope at school and have an impact on children's peer relationship, hence it is important to identify how parenting styles and parent–child relationship are related to victimization in order to develop intervention programs to prevent or mitigate victimization in childhood and adolescence.

The Hedge's g for each parenting behavior category is shown in Fig. 2, Fig. 3, Fig. 4 and Fig. 5. The studies included in the analysis with their descriptions are shown in supplementary Tables 3 and 4. Positive parenting behavior ( Fig. 2 and Fig. 4): The combined effect size showed that victims and bully/victims were significantly less likely to have authoritative parents (victims: Hedge's g = −0.19, 95% CI: −0.28, −0.11; z = −4.42; p < 0.001; bully/victims: Hedge's g = −0.39, 95% CI: −0.61, −0.18; z = −3.55; p < 0.001), good parent–child communication (victims: Hedge's g = −0.12; 95% CI, −0.20, −0.05; z = −3.13; p < 0.01; bully/victims: Hedge's g = −0.17, 95% CI: −0.30, −0.04; z = −2.62; p < 0.01), parents that were involved and supportive (victims: Hedge's g = −0.22; 95% CI, −0.29, −0.15; z = −5.97; p < 0.001; bully/victims: Hedge's g = − 0.30, 95% CI: −0.40, −0.20; z = −5.82; p < 0.001), receive supervision (victims: Hedge's g = −0.16, 95% CI: −0.21, −0.12; z = −6.81; p < 0.001; bully/victims: Hedge's g = −0.34, 95% CI: −0.54, −0.14; z = −3.31; p < 0.01) and warm and affective parents (victims: Hedge's g = −0.22; 95% CI, −0.30, −0.14; z = −5.17; p < 0.001; bully/victims: Hedge's g = −0.42, 95% CI: −0.54, −0.31; z = −7.21; p < 0.001). Overall, both victims and bully/victims were less likely to live in a family with positive parenting (victims: Hedge's g = −0.19; 95% CI, −0.23, −0.15; z = −9.65 p < 0.001; bully/victims: Hedge's g = −0.33; 95% CI: −0.41, −0.26; z = −9.07; p < 0.001). Negative parenting behavior ( Fig. 3 and Fig. 5): The combined effect size showed that victims and bully/victims were significantly more likely to have been abused or neglected (victims: Hedge's g = 0.31; 95% CI, 0.18–0.44; z = 4.53; p < 0.001; bully/victims: Hedge's g = 0.68, 95% CI: 0.44–0.92; z = 5.57; p < 0.001), or to have experienced maladaptive parenting (victims: Hedge's g = 0.27; 95% CI, 0.15–0.40; z = 4.31; p < 0.001; bully/victims: Hedge's g = 0.49, 95% CI: 0.23–0.75; z = 3.74; p < 0.001). In addition, victims were more likely to have overprotective parents (Hedge's g = 0.10; 95% CI, 0.03–0.17; z = 2.63; p < 0.01). Overall, both victims and bully/victims were found to experience negative parenting more often (victims: Hedge's g = 0.26; 95% CI, 0.16–0.36; z = 4.90; p < 0.001; bully/victims: Hedge's g = 0.48 95% CI: 0.26–0.70; z = 4.23; p < 0.001). Potential moderator variables The heterogeneity analyses for some of the categories were significant (see supplementary Tables 3 and 4) indicating potential moderating. Meta-ANOVAs of continent (Europe, America or other), age (4–7, 7.5–12 or over 12 years), assessment method (self, peer, teacher or mixed) and design (cross-sectional versus longitudinal) were employed for each parenting behavior category where moderation effects were detected. Supplementary Tables 5 and 6 show all moderation effects. Victims: Communication showed a significant moderating effect according to the assessment method as indicated by the Qb heterogeneity coefficient (Qb = 6.741; p < 0.05) suggesting that studies using peer nomination showed lower levels of communication between the parent and the child (mean ES = −0.494; p < 0.01; N = 1). Warmth and affection category showed significant moderating effects according to the age group (Qb = 7.193; p < 0.05). Children aged 12 years or more were less likely to have warm and affectionate families (mean ES = −0.305; p < 0.001; N = 11) compared to the other age groups. Lastly, supervision category showed moderating effects according to continent (Qb = 16.862; p < 0.001) with European studies finding less supervision for victims (mean ES = −0.311; p < 0.001; N = 1). Bully-victims: Parental involvement and support showed significant moderator effects according to the assessment method (Qb = 7.03; p < 0.05) suggesting that children who self-reported victimization (mean ES = −0.35; p < 0.001; N = 8) were less likely to have parents who are involved and supportive. Warmth and affection showed significant moderator effects according to continent (Qb = 6.678; p < 0.05), assessment method (Qb = 13.651; p < 0.01) and age group (Qb = 10.704; p < 0.01). Children from other continents (mean ES = −0.59; p < 0.001; N = 1), who self-reported victimization (mean ES = −0.58; p < 0.001; N = 3) or were over 12 years old (mean ES = −0.52; p < 0.001; N = 4) had parents with less warmth and affection. Moreover, maladaptive parenting and overall negative parenting behavior showed significant moderating effects according to continent (maladaptive parenting: Qb = 32.326; p < 0.001; overall negative parenting: Qb = 20.124; p < 0.001), other continents showed strongest maladaptive parenting and overall negative parenting behavior (maladaptive parenting: mean ES = 0.94, p < 0.001: N = 2; overall negative parenting behavior: mean ES = 0.92, p < 0.001: N = 2). Publication bias A failsafe N and the “5k + 10” benchmark were calculated for all categories (see Table 2 and Table 3). For victims, the meta-analysis of authoritative parenting and overprotection did not exceed the benchmark suggesting effects are open for future disconfirmations. The Begg and Mazumdar rank correlation results for overall negative parenting behavior showed that controversial results from small studies were less likely to be published. Egger's test showed significant results for parental involvement and support and overall positive and negative parenting behavior suggesting that publication bias might have influenced the estimates. Duval & Tweedie's trim and fill procedure resulted in slightly different effect sizes for supervision, warmth and affection, overall positive parenting behavior, maladaptive parenting, overprotection and overall negative parenting behavior. For bully/victims, authoritative parenting, communication and supervision categories did not exceed the “5k + 10” benchmark suggesting that the effect may change with future studies. The Begg and Mazumdar rank correlation results for all categories were not significant. Egger's test showed significant results for communication, maladaptive parenting and overall negative parenting behavior suggesting that publication bias might have influenced the estimates. Duval & Tweedie's trim and fill procedure resulted in slightly different effect sizes for parental involvement and support, overall positive parenting behavior, and abuse and neglect. Table 2. Publication bias analyses for victims. Victims Fail safe N ar = 0.05 “5k + 10” benchmark b Kendall's tauc Egger's testd Trim-and-fill (95% CI)e Authoritative 24 35 0.00 p = 0.50 β = 0.22 (−5.94, 6.38) p = 0.46 −0.19 (−0.27, −0.11) Communication 57 50 −0.25 p = 0.19 β = −1.93 (−6.53, 2.67) p = 0.17 −0.12 (−0.20, −0.05) Parental involvement & support 1896 140 0.19 p = 0.09 β = −3.34 (−4.76, −1.91) p < 0.001 −0.22 (−0.29, −0.15) Supervision 354 70 −0.17 p = 0.23 β = −0.21 (−2.44,2.02) p = 0.42 −0.16 (−0.21, −0.12) Warmth & affection 821 105 −0.02 p = 0.14 β = 0.39 (−1.90, 2.68) p = 0.36 −0.22 (−0.30, −0.13) Overall positive parenting behavior 10,003 355 0.09 p = 0.13 β = −2.45 (−3.29, −1.61) p < 0.001 −0.17 (−0.21, −0.13) Abuse & neglect 42 40 0.00 p = 0.50 β = 0.09 (−3.69,3.87) p = 0.47 0.31 (0.17, 0.44) Maladaptive parenting 3622 140 0.20 p = 0.07 β = −2.48 (−5.50, 0.54) p = 0.05 0.31 (0.19, 0.43) Overprotection 6 30 0.17 p = 0.37 β = 0.76 (−5.93, 7.44) p = 0.34 0.09 (0.03, 0.16) Overall negative parenting behavior 4837 185 0.26 p = 0.01 β = −2.39 (−4.74, −0.04) p = 0.02 0.30 (0.20, 0.39) a Rosenthal's failsafe number: the number of the studies that would be required to nullify the observed effect. b Tolerance level around a failsafe N (5 times the number of effect sizes plus 10). c Begg and Mazumdar rank correlation test. d Egger's regression intercept. e Duval and Tweedie's trim and fill method (trims the studies from one side to identify the unbiased effect). Table options Table 3. Publication bias analyses for bully/victims. Bully/victims Fail safe N ar = 0.05 “5k + 10” benchmark b Kendall's tauc Egger's test d Trim-and-fill (95% CI)e Authoritative parenting 24 25 0.00 p = 0.50 β = −0.97 (−35.36, 33.41) p = 0.39 −0.39 (−0.61, −0.17) Communication 2 25 0.00 p = 0.50 β = −0.07 (−13.13, 12.99) p = 0.02 −0.17 (−0.30, −0.04) Parental involvement & support 347 65 −0.11 p = 0.32 β = −0.49 (−2.76, 1.79) p = 0.48 −0.26 (−0.37, −0.16) Supervision 20 25 0.00 p = 0.50 β = −2.17 (−59.82, −55.48) p = 0.36 −0.34 (−0.54, −0.14) Warmth & affection 354 45 0.00 p = 0.50 β = 1.27 (−2.66, 5.20) p = 0.22 −0.41 (−0.52, −0.30) Overall positive parenting behavior 2065 140 −0.21 p = 0.07 β = −0.15 (−1.64, 1.34) p = 0.42 −0.27 (−0.35, −0.19) Abuse & neglect 30 25 0.00 p = 0.50 β = 0.12 (−27.96, 28.19) p = 0.48 0.64 (0.41, 0.88) Maladaptive parenting 2568 75 0.11 p = 0.29 β = −4.29 (−8.07, −0.51) p = 0.02 0.49 (0.23, 0.75) Overall negative parenting behavior 3306 100 0.04 p = 0.41 β = −4.15 (−7.00, −1.31) p < 0.001 0.48 (0.26, 0.70) a Rosenthal's failsafe number: the number of the studies that would be required to nullify the observed effect. b Tolerance level around a failsafe N (5 times the number of effect sizes plus 10). c Begg and Mazumdar rank correlation test. d Egger's regression intercept. e Duval and Tweedie's trim and fill method (trims the studies from one side to identify the unbiased effect).