School social cohesion,student-school connectedness,and bullying in Colombian adolescents

Abstract

Background:

Student-school connectedness is inversely associated with multiple health risk behaviors, yet research is limited on the relative contributions of a student’s connectedness with school and an overall context of school social cohesion to peer victimization/bullying.

Purpose:

We examined associations of perceived school cohesion and student-school connectedness with physical victimization, verbal victimization, and social exclusion in the past six months in adolescents in grades 6–11 (N = 774) attending 11 public and private urban schools in Colombia.

Methods:

Cross-sectional data were collected via a self-administered questionnaire and analyzed using mixed-effects linear regression models.

Results:

Higher perceived school cohesion was inversely related with exposure to three bullying types examined (p < 0.05); student-school connectedness was negatively related to verbal victimization among girls only (p < 0.01). In full models, school cohesion maintained inverse associations with three bullying types after controlling for student-school connectedness (p ≤ 0.05).

Conclusion:

Enhancing school cohesion may hold benefits for bullying prevention beyond a student’s individual school connectedness.

Keywords

adolescents and youth assets/protective factors Latin America resilience school setting social support social networks violence prevention

Introduction

Antonovsky (1) once wrote, in response to the popular aphorism of the river of disease and the role of public health in preventing people from falling in the river: ‘… [W]e are always in the dangerous river of life. The twin question is: How dangerous is our river? How well can we swim?’ Decades of research have found that peer victimization, also known as bullying, is one of the many dangerous currents in adolescents’ rivers of life that they must learn to navigate. Bullying takes place when children or adolescents exploit power imbalances in order to dominate and harm others physically, socially, or emotionally (2,3). Victims of bullying report higher levels of depression, social anxiety and loneliness, and other psychosocial adjustment difficulties (4,5) as well as increased substance abuse, school absenteeism, physical injuries (6,7) and suicidal ideation (8). In recent years, adverse health consequences of bullying have been documented in adolescents in low- and middle-income countries (9).

Findings from 66 countries and territories from five continents indicate that bullying is highly prevalent worldwide, with just under one-third (32%) of children aged 13–15 years reporting having been bullied at least once in the past two months (10). In countries experiencing or emerging from violent conflict such as Colombia, bullying may be of heightened concern as children’s exposure to a violent social context has been found to increase their risk for developing aggressive behavior (11,12). The high prevalence of adolescent bullying, along with its harmful consequences, has sparked calls for greater global public health action to prevent bullying (13).

While bullying represents one of the many dangerous currents in adolescents’ rivers of life, a growing body of research has found that positive social relationships at school may serve as a buoy, or protective effect, against adolescent problem behavior. Social relationships at school have been conceptualized at both the contextual and individual levels. At the contextual level, school social cohesion, defined as how well students get along with each other and their teachers; how well teachers support students; and a general climate of caring, respectful and supportive relationships, is a key dimension of school-level constructs that include school climate, school context, and school-as-community (14 –18). Theoretical foundations rooted in the schools-as-communities perspective (14) posit that students engage in less school misconduct and risk behavior when schools are organized as caring school communities that allow for students to feel emotionally connected and supported by their peers and teachers.

At the individual level, constructs that include school belonging (19), student-school connectedness (20), school attachment (21), and school bonding (22) generally refer to a student’s personal connection to school and the people within school, which may include the degree to which students feel close to peers and teachers at school, feel an emotional link with their school, and are committed to and involved with their school (23). Theoretical perspectives rooted in the identification-participation model (24), social bond/social control theory (25), and the social development model (22) emphasize the importance of an adolescent’s strong social bonds to and participation in conventional society, including schools, for the prevention of deviant and health risk behavior and for the promotion of academic achievement and well-being. Previous research indicates that students who report higher positive school social climate (18,26 –30) and greater student-school connectedness (20,31) are less likely to report bullying and other forms of aggression.

Despite the theoretical and empirical foundation for the positive effects of supportive and cohesive school relationships on adolescent aggressive behavior, research is limited on the relative contributions of a context of higher school social cohesion vs. a student’s individual connectedness with his or her school on reducing exposure to peer victimization, especially from low- and middle-income countries where emerging evidence documents the global nature of adolescent bullying (10). The current study examines the potential protective effect of supportive school social relationships on bullying in a sample of adolescents attending secondary and high schools in Valle del Cauca, Colombia, by assessing the associations of perceived school cohesion—a measure of the general school social climate as defined by caring, supportive and cohesive relationships among and between students and teachers at school, and student-school connectedness—a measure of a student’s perceived connectedness and belongingness with his or her school, with the frequency of student exposure to three types of bullying: physical victimization, verbal victimization, and social exclusion.

Methods

Study design, participants and setting

This study is based on data from the secondary and high school sample of the ‘Intimidación Escolar y Efectos Psicologicos en Victimas, Victimarios y Testigos’ study (School Bullying and Psychological Effects in Victims, Bullies, and Bystanders) (32), a cross-sectional study of bullying and victimization among public and private school students in the Department (province) of Valle del Cauca, Colombia, led by researchers at the Pontificia Universidad Javeriana-Cali. Secondary and high schools were purposively selected from five municipalities that represent the largest urban population concentrations from Valle del Cauca. A total of 11 secondary and high schools (n = 7 public and n = 4 private) were invited to participate and comprised the sample. The current study focuses on adolescents between the ages of 10 and 19 years, the age defined as adolescence by the World Health Organization (33). Of the 1980 students invited to participate in the survey, 840 students aged 10–19 years agreed to participate, representing a response rate of 42%. Data were not available on reasons students opted not to participate. Of these students, complete data on the primary study variables were available for 774 students, who comprised the final analytic sample.

The study was reviewed and approved by the ethics review committee of the Pontificia Universidad Javeriana-Cali, and the Committee for the Protection of Human Subjects at the University of Texas School of Public Health. Student participants were invited to participate in the study via a verbal and written letter of invitation, and student assent and active parental consent were obtained for all participating students.

Measures and procedures

Data were collected via the School Bullying Questionnaire (Cuestionario de Intimidación Escolar or CIE), a self-administered questionnaire composed of two parts: CIE Form A, which measures bullying, and CIE Form B, which assesses school social cohesion and connectedness. The CIE Form A bullying measures, developed by Cuevas et al. (32) and adapted from previous measures of bullying (2,34), have evidence of construct validity in Colombian adolescents (35). CIE Form A items ask students to report the frequency with which he or she has experienced specific types of bullying over that past six months using a three-point Likert type response scale (never, a few times, a lot of times). Measures assessed three dimensions of bullying: physical victimization (α = 0.86, 16 items, scale range: 0–32) (e.g. ‘They hit me’), verbal victimization (α = 0.89, 18 items, scale range: 0–36) (e.g. ‘They call me names’), and social exclusion (α = 0.89, 22 items, scale range: 0–44) (e.g. ‘They ignore me.’). These scales demonstrated good internal consistency based on a previously established criterion of α > 0.60 (36).

CIE Form B consists of the Student Perceptions of School Cohesion (SPSC) Scale, a measure that comprises 12 items on different facets of school social cohesion for which respondents rate their agreement using a five-point Likert response scale. Items were based on the ‘Students’ sense of the school as a community’ subscale of caring and supportive interpersonal relationships (14) and measures of student-school connectedness from the National Longitudinal Study on Adolescent Health (20). The SPSC scale has been found to have good internal consistency (α = 0.84) and construct validity in assessing school social cohesion and student-school connectedness in Salvadoran adolescents (37). In this study, the ‘Disagree’ and ‘Do not at all agree’ categories of the scale were combined to create one category due to small cell size. Principal component analysis using varimax (orthogonal) rotation was conducted to determine the factor structure of the scale based on our study sample. We retained factors with eigenvalues greater than 1 based on the Kaiser criterion (38,39), and items with a factor loading of 0.30 or higher were assigned to a given factor. Internal consistency for the total scale (global school cohesion scale) was high (α = 0.89, 12 items, scale range: 0–36). Principal component analysis resulted in two subscales: perceived school cohesion (α = 0.85, six items, scale range: 0–18), a general measure of caring and cohesive relationships at school, with items such as ‘students support each other at my school,’ and student-school connectedness (α = 0.81, six items, scale range: 0–18), a measure of a student respondent’s perceived connectedness with his or her school, with items such as ‘I feel close to people at school’ and ‘I feel a part of my school’.

Prior to administering the CIE, measures were reviewed for face-validity and appropriateness for use with Colombian adolescents by a panel of nine expert judges in child health from the Javeriana Universidad. Trained data collectors administered the survey during class time to students with parental consent who were present on the day of the survey. Data were collected in 2009–2011 in accordance with the school year calendar and school availability.

Analysis

Descriptive statistics are presented as percentages for categorical outcomes, and means and standard deviations (SDs) for continuous outcomes. The dependent variables, the three bullying outcomes, were treated as continuous; the independent variables (global school cohesion, perceived school cohesion, and student-school connectedness) were categorical variables that ranged from low school social cohesion (not in agreement) to high social cohesion (high agreement with statements on social cohesion). These composite variables were created by dividing the total score by four (the number of response options) to mirror the original Likert scale categories. Independent t-tests were conducted to assess mean differences in key study variables by demographic characteristics. Mixed-effects linear regression analyses using a random-effects generalized least squares (GLS) model accounting for school-level clustering (40) were conducted to examine associations between school cohesion and student-school connectedness and the three bullying/victimization outcomes for the total sample and by gender. All models were adjusted for grade level, gender (total sample only), and school type (private vs. public) and treated school as a random effect. Models were also run to assess the association between school cohesion and the bullying outcomes while controlling for student-school connectedness. Lastly, we tested for possible interaction effects by including an interaction term between school cohesion and student-school connectedness and running separate regression models for the three bullying outcomes. Statistical analyses were conducted using Stata (version 12), with significance set at p < 0.05.

Results

The secondary school sample comprised 6th to 8th grade students (n = 453) and 9th to 11th grade students (n = 321) (Table 1) attending 11 public and private secondary and high schools in five municipalities in Valle del Cauca, Colombia. The mean age of the students was 13.9 years (SD: 1.88), and the sample had a slightly higher composition of male students (53.3%). A relatively equal proportion of students was represented among the five municipalities (~20% per municipality) (data not shown).

Table 1.

Demographic characteristics of sample and mean scores of peer victmization and school cohesion by gender, grade and school type.

	Study sample	Exposure to peer victimization (bullying) (dependent variables)			School cohesion (independent variables)
	Study sample	Physical victimization	Verbal victimization	Social exclusion	Global school cohesion	Perceived school cohesion	Student-school connectedness
	n (%)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
Total	774 (100%)	20.23 (4.39)	25.80 (6.28)	30.71 (7.04)	21.98 (9.14)	13.88 (6.09)	8.10 (3.89)
Boys	413 (53.4)	20.89 (4.61)^c	26.31 (6.31)^a	30.11 (6.89)	22.20 (9.34)	13.88 (6.12)	8.32 (3.96)
Girls	361 (46.6)	19.49 (4.01)	25.23 (6.22)	31.40 (7.16)^a	21.74 (8.91)	13.90 (6.07)	7.85 (3.80)
Grade
6th–8th	453 (58.5)	21.00 (4.97)^c	26.46 (6.97)^b	31.80 (7.56)^c	21.16 (9.02)	13.34 (6.07)	7.82 (3.85)
9th–11th	321 (41.5)	19.21 (3.21)	24.91 (5.10)	29.22 (5.96)	23.14 (9.19)^b	14.65 (6.05)^b	8.50 (3.92)^a
School type
Public (n=7)	637 (82.3)	20.38 (4.50)^a	25.75 (6.28)	31.25 (7.03)^c	21.71 (9.16)	13.70 (6.17)	8.01 (3.91)
Private (n=4)	137 (17.7)	19.54 (3.83)	26.02 (6.34)	28.31 (6.64)	23.26 (8.96)	14.74 (5.69)	8.52 (3.82)

School Bullying Study, Valle del Cauca, Colombia, 2009–11.^ap < 0.05; ^bp < 0.001; ^cp < 0.001. Analyses based on independent t-test.Scale ranges: physical victimization : 0 (low) to 32 (high bullying); verbal victimization : 0 to 36; social exclusion : 0 to 44.Global school cohesion: 0 (low) to 36 (high cohesion); perceived school cohesion: 0 to 18; student-school connectedness: 0 to 18.

School cohesion, student-school connectedness, and bullying by demographics

Boys reported a higher frequency of physical (p < 0.001) and verbal victimization (p < 0.05); girls reported a higher frequency of social exclusion (p < 0.05) (Table 1). Across bullying measures, 6th to 8th graders experienced more bullying than 9th to 11th graders (p < 0.05). Public school students reported higher mean scores of physical victimization and social exclusion compared to private school students (p < 0.05); no differences by public/private school were found for verbal victimization. While 9th to 11th graders reported higher mean scores of the three school cohesion measures compared to 6th to 11th graders, no differences in school cohesion measures were found by gender or public/private school attendance.

School cohesion, student-school connectedness, and exposure to bullying

School cohesion vs. student-school connectedness: findings from separate models

Physical victimization

Higher levels of global school cohesion were significantly related to less physical victimization after adjusting for grade, gender, and school type (p < 0.0001) (Table 2). Stratifying by gender, high global school cohesion significantly predicted less physical victimization for boys (p < 0.01) but not for girls. With regard to the global school cohesion subscales, perceived school cohesion was negatively related to physical victimization for boys (p < 0.05) but not for girls. No significant association was found for student-school connectedness and physical victimization by gender (Table 2).

Table 2.

Associations of global school social cohesion, perceived school cohesion, and student-school connectedness with student-reported physical victimization at school for the total sample and stratified by gender. Findings based on nine independent regression models.

	Total sample			Boys			Girls
	Adj. mean	B	SE	Adj. mean	B	SE	Adj. mean	B	SE
Global school social cohesion
High (very much agree)	19.97	−1.72^e	0.73	20.43	−2.39^c	0.99	19.40	−0.83	1.12
Moderate high	20.41	−0.81	0.62	21.11	−0.75	0.83	19.65	−0.90	0.95
Neutral	20.60	−0.54	0.66	21.13	−0.52	0.89	19.99	−0.53	1.00
Low (disagree/do not at all agree)^b	21.14			21.70			20.33
		R²= 0.081; Wald chi²(6)=51.19			R²=0.088; Wald chi²(5)=29.04			R²=0.032; Wald chi²(5)=10.28
		p = 0.0000			p = 0.0000			p = 0.068
Subscales
Perceived school cohesion
High (very much agree)	19.86	−1.86^d	0.63	20.33	−2.43^d	0.87	19.23	−1.13	0.93
Moderate high	20.47	−0.73	0.53	21.20	−0.3	0.74	19.71	−1.41	0.77
Neutral	20.50	−0.67	0.56	21.06	−0.11	0.78	19.85	−1.40	0.82
Low (disagree/do not at all agree)^b	21.18			21.70			20.55
		R²=.08; Wald chi²(6)=53.70			R²= 0.10; Wald chi²(5)=34.62			R²=0.038; Wald chi²(5)=12.50
		p = 0.0000			p = 0.0000			p = 0.029
Student-school connectedness
High (very much agree)	19.90	−1.09	0.72	20.48	−0.89	0.98	19.33	−1.44	1.07
Moderate high	20.78	−0.10	0.7	21.39	0.23	0.94	20.05	−0.58	1.06
Neutral	20.74	0.13	0.78	21.33	1.39	1.06	19.99	−1.50	1.17
Low (disagree/do not at all agree)^b	20.93			21.41			20.20
		R²=0.08; Wald chi²(6)=53.22			R²= 0.09; Wald chi²(5)=31.73			R²=0.04; Wald chi² (5)=13.32
		p = 0.0000			p = 0.0000			p = 0.021

Secondary and high school students (n = 774), School Bullying Study, Valle del Cauca, Colombia, 2009–11.^bReferent group. ^c < 0.05; ^d < 0.01; ^e***<.001. Adj.:adjusted for grade, gender (total sample only) and school; B:beta; SE:standard error. ^aBased on random effects generalized least squares (GLS) model with school treated as a random effect, adjusting for grade (6 –11), gender (total sample only), and school type (private vs. public). Independent models run for total sample, by gender, and for each social cohesion variable (global school social cohesion, perceived school-level cohesion, and student-school connectedness).

Verbal victimization

Higher levels of global school cohesion and perceived school cohesion were significantly related to lower frequency of verbal victimization for the total sample and across gender groups (p < 0.05). While higher levels of student-school connectedness were significantly related with lower verbal victimization in girls (p < 0.01), no significant differences were found for boys (Table 3).

Table 3.

Associations of global school social cohesion, perceived school cohesion, and student-school connectedness with student-reported verbal victimization at school for the total sample and stratified by gender. Findings based on nine independent regression models.

	Total sample			Boys			Girls
	Adj. mean	B	SE	Adj. mean	B	SE	Adj. mean	B	SE
Global school cohesion
High (very much agree)	24.56	−4.19^e	1.04	24.8	−4.20^d	1.36	24.27	−4.40^d	1.64
Moderate high	25.58	−2.83^e	0.88	26.04	−2.25	1.14	25.08	−3.70^d	1.42
Neutral	26.89	−1.47	0.94	27.27	−0.82	1.23	26.46	−2.53	1.50
Low (disagree/do not at all agree)^b	28.24			28.67			27.61
		R²=0.058; Wald chi²(5)=37.54			R²=0.06;Wald chi²(5)=18.12			R²=.048; Wald chi²(5)=16.01
		p = 0.0000			p = 0.003			p= .0068
Subscales
Perceived school cohesion
High (very much agree)	24.55	−4.01^e	0.90	24.80	−4.45^e	1.19	24.24	−3.41^c	1.40
Moderate high	25.48	−2.73^e	0.76	25.95	−2.50^c	1.01	24.98	−2.93^c	1.18
Neutral	26.81	−1.39	0.81	27.18	−0.66	1.07	26.39	−2.04	1.26
Low (disagree/do not at all agree)^b	28.09			28.45			27.65
		R²=0.06; Wald chi² (6)=40.68			R²=0.07; Wald chi²(5)=25.53			R²=0.042; Wald chi2(5)=14.55
		p = 0.0000			p = 0.0001			p = 0.013
Student-school connectedness
High (very much agree)	24.71	−3.22^d	1.00	25.04	−2.40	1.31	24.38	−4.58^d	1.40
Moderate high	26.48	−1.40	0.97	26.88	−0.69	1.27	26.00	−2.70	1.18
Neutral	26.95	−0.62	1.11	27.35	1.61	1.45	26.42	−4.10^c	1.26
Low (disagree/do not at all agree)^b	27.89			28.28			27.3
		R²=0.06; Wald chi²(6)=36.05			R²=0.07; Wald chi²(5)=21.04			R²=0.057; Wald chi²(5)=19.89
		p = 0.0000			p =0.0008			p = 0.0013

Secondary and high school students (n = 774), School Bullying Study, Valle del Cauca, Colombia, 2009–11.^bReferent group. ^c<.05; ^d<.01; ^e<.001. Adj.:adjusted for grade, gender (total sample only) and school; B:beta; SE:standard error. ^aBased on random effects generalized least squares (GLS) model with school treated as a random effect, adjusting for grade (6 –11), gender (total sample only), and school type (private vs. public). Independent models run for total sample, by gender, and for each social cohesion variable (global school social cohesion, perceived school-level cohesion, and student-school connectedness).

Social exclusion

Students who reported high levels of global school cohesion and perceived school cohesion reported significantly lower frequency of social exclusion for the total sample and across gender groups (p < 0.01). No significant associations were found for student-school connectedness and social exclusion (Table 4).

Table 4.

Associations of global school social cohesion, perceived school cohesion, and student-school connectedness with student-reported social exclusion at school for the total sample and stratified by gender. Findings based on nine independent regression models.

	Total Sample			Boys			Girls
	Adj. mean	B	SE	Adj. mean	B	SE	Adj. mean	B	SE
Global school cohesion
High (very much agree)	29.05	−5.28^e	1.16	28.26	−5.37^e	1.51	30.03	−5.06^d	1.84
Moderate high	30.85	−2.67^d	0.98	30.32	−2.77^d	1.26	31.43	−2.37	1.55
Neutral	31.30	−2.03	1.04	30.45	−2.36	1.35	32.26	−1.66	1.64
Low (disagree/do not at all agree)^b	33.24			32.61			34.14
		R²= 0.097; Wald chi²(6)=60.01			R²= 0.08; Wald chi²(5)=24.97			R²= 0.11; Wald chi²(5)=28.43
		p = 0.0000			p = 0.0001			p = 0.0000
Subscales
Perceived school cohesion
High (Very much agree)	28.92	−5.19^e	1.00	28.22	−4.53^d	1.33	29.84	−5.59^e	1.53
Moderate high	30.83	−2.49^d	0.84	30.30	−1.59	1.13	31.38	−3.23^c	1.26
Neutral	31.23	−1.88^c	0.89	30.47	−0.6	1.19	32.11	−2.87^c	1.35
Low (disagree/do not at all agree)^b	33.27			32.43			34.31
		R²= 0.10; Wald chi²(6)=64.50			R²= 0.08; Wald chi²(5)=29.85			R²= 0.12; Wald chi²(5)=31.33
		p = 0.0000			p = 0.0000			p = 0.0000
Student-school connectedness
High (very much agree)	29.86	−2.29	1.17	29.07	−2.023	1.52	30.64	−2.72	1.86
Moderate high	31.53	−0.18	1.14	30.93	0.20	1.47	32.25	−0.72	1.83
Neutral	31.42	0.13	1.28	30.8	1.50	1.63	32.22	−1.76	2.05
Low (disagree/do not at all agree)^b	31.60			30.85			32.73
		R²= 0.09; Wald chi²(6)=49.61			R²= 0.07; Wald chi² (5)=24.40			R²= 0.10; Wald chi² (5)=21.32
		p = 0.0000			p = 0.0002			p = 0.0007

Secondary and high school students (n = 774), School Bullying Study, Valle del Cauca, Colombia, 2009–11.^bReferent group. ^c< 0.05; ^d< 0.01; ^e< 0.001. Adj.: adjusted for grade, gender (total sample only) and school; B: beta; SE:standard error. ^aBased on random effects generalized least squares (GLS) model with school treated as a random effect, adjusting for grade (6 –11), gender (total sample only), and school type (private vs. public). Independent models run for total sample, by gender, and for each social cohesion variable (global school social cohesion, perceived school-level cohesion, and student-school connectedness).

School cohesion and bullying after controlling for student-school connectedness

In models with the full sample that controlled for student-school connectedness, perceived school cohesion maintained statistically significant inverse associations with verbal victimization (p = 0.002) and social exclusion (p < 0.001); the association between perceived school cohesion and physical victimization became borderline significant at p = 0.05 (data not shown). Lastly, no significant interaction effects were found when a student-school connectedness and school cohesion interaction term was included in separate models for the three bullying outcomes (p = 0.58 to p = 0.97) (data not shown).

Discussion

In responding to emerging evidence of the global nature of adolescent bullying (10), this study examined the associations of school social cohesion and student-school connectedness with bullying in a sample of adolescents from Colombia. While student-school connectedness was only significantly and negatively related to verbal victimization among girls, higher perceived school social cohesion was significantly related to lower exposure to the three types of bullying examined in the total sample, for physical victimization in boys, and for verbal victimization and social exclusion in boys and girls. Across models for the full sample, the mean difference in exposure to the three bullying outcomes among students reporting the lowest and highest levels of school social cohesion represents a 0.3 (physical victimization) to a 0.6 (verbal victimization and social exclusion) SD unit difference in bullying exposure based on Cohen’s d, indicating a small- (≥ 0.2) to medium- (≥ 0.5) effect size difference (41) in bullying exposure based on students’ perceived level of school social cohesion. These findings underscore an important role of school social cohesion in bullying exposure and contribute to an emerging evidence base from Latin America and the Caribbean on the importance of supportive school social relationships for reducing a range of adolescent health risk behaviors, including aggression, substance use, sexual risk behavior, and suicidal ideation (17,42 –44).

Recent evidence from high-income countries provides support for the importance of perceived school cohesion and school interpersonal climate in understanding exposure to bullying (18,27 –30). Our findings contribute to this evidence base by demonstrating that an individual’s school connectedness vs. social cohesion across students and teachers hold different relevance for understanding exposure to bullying. The general lack of association of student-school connectedness with bullying in this study may point to the complexity of school social relationships and their ‘double-edged’ nature. As Zaykowski and Gunter (18) point out, youth may get along with their peers and teachers, but still have conflicts with others in the school. In support of the double-edged nature of social relationships, recent research based on Flemish secondary school students indicated that higher peer attachment was related to increased school misconduct (i.e. vandalism, drug use, and breaking school rules), while interpersonal school climate was inversely associated with misconduct (45). Our findings of a generally consistent and inverse association of school cohesion across the three victimization types may point to the importance of school-level social forces for shaping exposure to victimization that work above and beyond a student’s direct connection to school.

Several recent United States (US)-based intervention studies provide evidence for the role that a socially cohesive school climate plays in reducing bullying. Findings from The Steps to Respect program in California indicate that positive classroom-level climate (e.g. how well students get along with each other) positively influenced student engagement in bullying prevention activities, which was subsequently associated with lower bullying-related outcomes (16). In Pennsylvania, the Olweus Bullying Prevention Program was recently found to enhance school climate and reduce bullying, and included school-wide approaches (staff training, creating a coordinating committee, establishing school rules), classroom approaches (discussion, class meetings, role playing), and a community coalition (46). Another model recently found to decrease bullying in Maryland is the School-wide Positive Behavioral Interventions and Support, which targets the school environment via creating school-wide behavioral expectations (e.g. respect others) promoted via direct instruction, positive reinforcement, and data systems to guide staff to areas of need (47). Our findings, taken together with recent experimental findings, provide a foundation for further research on the effects of creating a cohesive school context on reducing adolescent bullying in low- and middle-income countries such as Colombia.

Although perceived social cohesion was inversely related to exposure to all three types of bullying across the full sample, important gender differences were noted. While student-school connectedness was generally not related to bullying outcomes, girls who expressed higher student-school connectedness reported lower exposure to verbal victimization. In girls, individual connectedness with peers may provide a protective effect against more overt bullying such as verbal victimization, yet may have no effect against less overt bullying such as social exclusion. These findings support previous research on gender differences in bullying (5,26).

Limitations

Specific limitations of this study merit mention. First, given the cross-sectional nature of the research, we cannot infer causality of the association between school cohesion and peer victimization. This limitation is worth highlighting as we cannot rule out the potential for reverse causality, in which higher bullying results in lower levels of school social cohesion. Second, while our study sample was selected from five of the most populous municipalities representing Valle del Cauca, Colombia, and represents both private and public schools, external generalizability of our findings is limited given the purposive nature of our study sample; further research is needed with a population-based sample of schools. Third, our sample size of schools was insufficient for exploring a school-level measure of social cohesion, and thus we were limited to a measure of individual perceptions of school cohesion. Lastly, factors such as adolescent risk behavior and social influences outside of school were not assessed; these factors merit attention in future research to rule out potential confounding of the associations observed in this study.

Conclusion

This study extends the body of literature on adolescent bullying by documenting the inverse association between perceptions of school social cohesion and exposure to three forms of bullying in a sample of Colombian adolescents and by demonstrating that student-school connectedness and school social cohesion represent distinct constructs that hold different relevance for understanding exposure to bullying. Further qualitative research on the cultural context of bullying and school cohesion in Colombia and other low- and middle-income countries, with attention to diverse socioeconomic and geographic contexts, is needed to further contextualize bullying in these contexts and identify mechanisms through which school social cohesion may prevent bullying. In addition to building social competency skills for adolescents to navigate their respective rivers of life, findings from this study add to a growing literature base on the importance of making adolescent rivers of life more navigable by fostering positive and cohesive social relationship buoys across the school setting.

Footnotes

Acknowledgements

We gratefully acknowledge Ms. Monica Gonzalez and Ms. Maria Alejandra Marmolejo from the Pontificia Universidad Javeriana-Cali for their important contributions with overall management of the study, including data collection and preparation. We also recognize the valuable contributions and participation of our partner schools, school leaders, and students from Valle del Cauca.

Conflict of interest

None declared.

Funding

This original study was made possible by financial support from the Pontificia Universidad Javeriana-Cali; participation of Dr. Springer, Dr. Wilkinson and Ms. Case was made possible in part by support from the Michael & Susan Dell Foundation.

References

Antonovsky

. The salutogenic model as a theory to guide health promotion. Health Promot Int.1996;11: 11–18.

Olweus

. Bullying at School: What We Know and What We Can Do (Understanding Children’s Worlds). Massachusetts: Blackwell; 1993.

Smith

Brain

. Bullying in schools: Lessons from two decades of research. Aggress Behav. 2000; 26: 1–9.

Hawker

Boulton

. Twenty years’ research on peer victimization and psychosocial maladjustment: A meta-analytic review of cross-sectional studies. J Child Psychol Psychiatry. 2000; 41: 441–455.

Card

Stucky

Sawalani

Little

. Direct and indirect aggression during childhood and adolescence: a meta-analytic review of gender differences, intercorrelations, and relations to maladjustment. Child Dev. 2008; 79: 1185–1229.

Gini

Pozzoli

. Association between bullying and psychosomatic problems: a meta-analysis. Pediatrics.2009;123: 1059–1065.

Srabstein

Piazza

. Public health, safety and educational risks associated with bullying behaviors in American adolescents. Int J Adolesc Med Health. 2008; 20: 223–233.

Espelage

Holt

. Suicidal ideation and school bullying experiences after controlling for depression and delinquency. J Adolesc Health.2013;53 (1 Suppl): S27–S31.

Fleming

Jacobsen

. Bullying among middle-school students in low and middle income countries. Health Promot Int.2009; 25: 73–84.

10.

Due

Holstein

. Bullying victimization among 13 to 15 year old school children: results from two comparative studies in 66 countries and regions. Int J Adolesc Med Health.2008; 20: 209–221.

11.

Chaux

Molano

Podlesk

. Socio-economic, socio-political and socio-emotional variables explaining school bullying: a country-wide multilevel analysis. Aggress Behav. 2009; 35: 520–529.

12.

Villar-Márquez

. School-based violence in Colombia: Links to state-level armed conflict, educational effects and challenges. United Nations Educational, Scientific and Cultural Organization/Education for All Global Monitoring Report; 2010. Available from:http://unesdoc.unesco.org/images/0019/001912/191228e.pdf (accessed June 16, 2014).

13.

Srabstein

Leventhal

. Prevention of bullying-related morbidity and mortality: a call for public health policies. Bull World Health Organ.2010; 88: 403.

14.

Battistich

Hom

. The relationship between students’ sense of their school as a community and their involvement in problem behaviors. Am J Public Health. 1997; 87: 1997–2001.

15.

Simons-Morton

Crump

. The association of parental involvement and social competence with school adjustment and engagement among sixth graders. J Sch Health.2002; 73: 121–126.

16.

Low

Van Ryzin

Brown

Smith

Haggerty

. Engagement matters: lessons from assessing classroom implementation of steps to respect: a bullying prevention program over a one-year period. Prev Sci.2014; 15: 165–176.

17.

Springer

Parcel

Baumler

Ross

.Supportive social relationships and adolescent health risk behavior among secondary school students in El Salvador. Soc Sci Med. 2006;62:1628–1640.

18.

Zaykowski

Gunter

. Youth victimization: school climate or deviant lifestyles? J Interpers Violence.2012; 27: 4431–4452.

19.

Goodenow

. The psychological sense of school membership among adolescents: scale development and educational correlates.Psychol Sch. 1993; 30: 70–90.

20.

Resnick

Bearman

Blum

. Protecting adolescents from harm: findings from the national longitudinal study on adolescent health. JAMA.1997; 278: 823–832.

21.

Mouton

Hawkins

McPherson

Copley

. School attachment: perspectives of low-attached high school students. Educational Psychology. 1996; 16: 297–304.

22.

Hawkins

Weis

. The social development model: an integrated approach to delinquency prevention. J Prim Prev. 1985; 6: 73–97.

23.

Libbey

. Measuring student relationships to school: attachment, bonding, connectedness, engagement. J Sch Health. 2004; 74: 274–283.

24.

Finn

. Withdrawing from school. Rev Educ Res. 1989; 59: 117–142.

25.

Hirschi

. Causes of Delinquency. Berkeley: University of California Press; 1969.

26.

Klein

Cornell

Konold

. Relationships between bullying, school climate, and student risk behaviors. Sch Psychol Q. 2012; 27: 154–169.

27.

Elsaesser

Gorman-Smith

Henry

. The role of the school environment in relational aggression and victimization. J Youth Adolesc.2013; 42: 235–249.

28.

Eliot

Cornell

Gregory

Fan

. Supportive school climate and student willingness to seek help for bullying and threats of violence. J Sch Psychol. 2010; 48: 533–553.

29.

Henry

Farrell

Schoeny

Tolan

Dymnicki

. Influence of school-level variables on aggression and associated attitudes of middle school students. J Sch Psychol.2011; 49: 481–503.

30.

Goldammer

Swahn

Strasser

Ashby

Meyers

. An examination of bullying in Georgia schools: demographic and school climate factors associated with willingness to intervene in bullying situations. West J Emerg Med. 2013; 14: 324–328.

31.

Simons-Morton

Crump

Haynie

Saylor

. Student-school bonding and adolescent problem behavior. Health Educ Res.1999; 14: 99–107.

32.

Cuevas Jaramillo

Hoyos Hernández

Ortíz Gómez

. Prevalencia de intimidación en dos instituciones educativas del departamento del Valle del Cauca. Pensamiento Psicológico. 2009; 6: 153–172.

33.

World Health Organization. Adolescent health. Available from:http://www.who.int/topics/adolescent_health/en/ (accessed 21 August 2014).

34.

Piñuel

Oñate

. AVE: acoso y violencia escolar. Madrid: TEA Ediciones S.A.; 2006.

35.

Bejarano

Díaz

Valencia

Cuevas

. Prevalencia de violencia entre compañeros y efectos psicológicos para las víctimas entre segundo y once grado de un colegio privado de la ciudad de Cali. Universidad Javeriana-Cali; 2008.

36.

DeVellis

. Scale Development: Theory and Applications. 2nd ed. California: Sage; 2003.

37.

Springer

McQueen

Quintanilla

Arrivillaga

Ross

. Reliability and validity of the student perceptions of school cohesion scale in a sample of Salvadoran secondary school students.BMC Int Health Hum Rights. 2009; 9: 30.

38.

Catell

. The Scree test for the number of factors. Multivariate Behav Res.1966; 1: 245–276.

39.

Kaiser

. The application of electronic-computers to factor-analysis. Educ Psychol Meas.1960; 20: 141–151.

40.

Murray

. Design and Analysis of Group Randomized Trials.United Kingdom: Oxford University Press; 1998.

41.

Cohen

. Statistical power analysis for the behavioral sciences.2nd ed. New Jersey: Lawrence Erlbaum; 1988.

42.

Anteghini

Fonseca

Ireland

Blum

. Health risk behaviors and associated risk and protective factors among Brazilian adolescents in Santos, Brazil. J Adolesc Health.2001; 28: 295–302.

43.

Kliewer

Murrelle

. Risk and protective factors for adolescent substance use: findings from a study in selected Central American countries. J Adolesc Health.2007;40: 448–455.

44.

Blum

Ireland

. Reducing risk, increasing protective factors: findings from the Caribbean Youth Health Survey. J Adolesc Health.2004; 35: 493–500.

45.

Demanet

Van Houtte

. School belonging and school misconduct: the differing role of teacher and peer attachment. J Youth Adolesc.2012; 41: 499–514.

46.

Schroeder

Messina

Schroeder

. The implementation of a statewide bullying prevention program: preliminary findings from the field and the importance of coalitions. Health Promot Pract. 2012; 13: 489–495.

47.

Waasdorp

Bradshaw

Leaf

. The impact of school-wide positive behavioral interventions and supports on bullying and peer rejection: a randomized controlled effectiveness trial. Arch Pediatr Adolesc Med. 2012; 166: 149–156.