576305
research-article2015
PED0010.1177/1757975915576305Original ArticleA. E. Springer et al.
Original Article School social cohesion, student-school connectedness, and bullying in Colombian adolescents Andrew E. Springer1, Maria Clara Cuevas Jaramillo2, Yamileth Ortiz Gómez3, Katie Case1 and Anna Wilkinson1
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
1. Michael & Susan Dell Center for Healthy Living, University of Texas School of Public Health-Austin Regional Campus, Austin, TX United States. 2. Department of Social Sciences, Pontificia Universidad Javeriana-Cali, Cali, Colombia. 3. Grupo de Investigación Salud y Calidad de Vida, Pontificia Universidad Javeriana-Cali, Cali, Colombia. Correspondence to: Andrew E. Springer, Michael & Susan Dell Center for Healthy Living, 1616 Guadalupe St., Suite 6.300, Austin, TX 78701, USA. Email: [email protected] (This manuscript was submitted on 16 September 2014. Following blind peer review, it was accepted for publication on 11 January 2015) Global Health Promotion 1757-9759; Vol 0(0): 1 –12; 576305 Copyright © The Author(s) 2015, Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1757975915576305 http://ghp.sagepub.com Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
2
A. E. Springer et al.
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-ascommunities 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 studentschool 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 middleincome 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
IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
3
Original Article
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 IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
4
A. E. Springer et al.
(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).
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). 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). 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).
IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
637 (82.3) 137 (17.7)
20.38 (4.50)a 19.54 (3.83)
21.00 (4.97)c 19.21 (3.21) 25.75 (6.28) 26.02 (6.34)
26.46 (6.97)b 24.91 (5.10)
25.80 (6.28) 26.31 (6.31)a 25.23 (6.22)
31.25 (7.03)c 28.31 (6.64)
31.80 (7.56)c 29.22 (5.96)
30.71 (7.04) 30.11 (6.89) 31.40 (7.16)a
Mean (SD)
21.71 (9.16) 23.26 (8.96)
21.16 (9.02) 23.14 (9.19)b
21.98 (9.14) 22.20 (9.34) 21.74 (8.91)
13.70 (6.17) 14.74 (5.69)
13.34 (6.07) 14.65 (6.05)b
13.88 (6.09) 13.88 (6.12) 13.90 (6.07)
Mean (SD)
8.01 (3.91) 8.52 (3.82)
7.82 (3.85) 8.50 (3.92)a
8.10 (3.89) 8.32 (3.96) 7.85 (3.80)
Mean (SD)
Student-school connectedness
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 type Public (n=7) Private (n=4)
453 (58.5) 321 (41.5)
Grade 6th–8th 9th–11th
20.23 (4.39) 20.89 (4.61)c 19.49 (4.01)
774 (100%) 413 (53.4) 361 (46.6)
Total Boys Girls
Mean (SD)
Mean (SD)
Mean (SD)
n (%)
Perceived school cohesion
Global school cohesion
Social exclusion
Physical victimization
Verbal victimization
School cohesion (independent variables)
Exposure to peer victimization (bullying) (dependent variables)
Study sample
Table 1. Demographic characteristics of sample and mean scores of peer victmization and school cohesion by gender, grade and school type.
Original Article 5
IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
−0.81 −0.54
20.41 20.60 21.14
−0.73 −0.67
20.47 20.50 21.18
−0.10 0.13
20.78 20.74 20.93 R2=0.08; Wald chi2(6)=53.22 p = 0.0000
−1.09
19.90
R2=.08; Wald chi2(6)=53.70 p = 0.0000
−1.86d
19.86
R2= 0.081; Wald chi2(6)=51.19 p = 0.0000
−1.72e
B
19.97
Adj. mean
0.7 0.78
0.72
0.53 0.56
0.63
0.62 0.66
0.73
SE
21.39 21.33 21.41
20.48
21.20 21.06 21.70
20.33
21.11 21.13 21.70
20.43
Adj. mean
Boys
R2= 0.09; Wald chi2(5)=31.73 p = 0.0000
0.23 1.39
−0.89
R2= 0.10; Wald chi2(5)=34.62 p = 0.0000
−0.3 −0.11
−2.43d
R2=0.088; Wald chi2(5)=29.04 p = 0.0000
−0.75 −0.52
−2.39c
B
0.94 1.06
0.98
0.74 0.78
0.87
0.83 0.89
0.99
SE
20.05 19.99 20.20
19.33
19.71 19.85 20.55
19.23
19.65 19.99 20.33
19.40
Adj. mean
Girls
1.06 1.17
1.07
0.77 0.82
0.93
0.95 1.00
1.12
SE
R2=0.04; Wald chi2 (5)=13.32 p = 0.021
−0.58 −1.50
−1.44
R2=0.038; Wald chi2(5)=12.50 p = 0.029
−1.41 −1.40
−1.13
R2=0.032; Wald chi2(5)=10.28 p = 0.068
−0.90 −0.53
−0.83
B
and high school students (n = 774), School Bullying Study, Valle del Cauca, Colombia, 2009–11.bReferent group. c < 0.05; d < 0.01; e***
research-article2015
PED0010.1177/1757975915576305Original ArticleA. E. Springer et al.
Original Article School social cohesion, student-school connectedness, and bullying in Colombian adolescents Andrew E. Springer1, Maria Clara Cuevas Jaramillo2, Yamileth Ortiz Gómez3, Katie Case1 and Anna Wilkinson1
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
1. Michael & Susan Dell Center for Healthy Living, University of Texas School of Public Health-Austin Regional Campus, Austin, TX United States. 2. Department of Social Sciences, Pontificia Universidad Javeriana-Cali, Cali, Colombia. 3. Grupo de Investigación Salud y Calidad de Vida, Pontificia Universidad Javeriana-Cali, Cali, Colombia. Correspondence to: Andrew E. Springer, Michael & Susan Dell Center for Healthy Living, 1616 Guadalupe St., Suite 6.300, Austin, TX 78701, USA. Email: [email protected] (This manuscript was submitted on 16 September 2014. Following blind peer review, it was accepted for publication on 11 January 2015) Global Health Promotion 1757-9759; Vol 0(0): 1 –12; 576305 Copyright © The Author(s) 2015, Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1757975915576305 http://ghp.sagepub.com Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
2
A. E. Springer et al.
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-ascommunities 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 studentschool 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 middleincome 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
IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
3
Original Article
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 IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
4
A. E. Springer et al.
(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).
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). 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). 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).
IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
637 (82.3) 137 (17.7)
20.38 (4.50)a 19.54 (3.83)
21.00 (4.97)c 19.21 (3.21) 25.75 (6.28) 26.02 (6.34)
26.46 (6.97)b 24.91 (5.10)
25.80 (6.28) 26.31 (6.31)a 25.23 (6.22)
31.25 (7.03)c 28.31 (6.64)
31.80 (7.56)c 29.22 (5.96)
30.71 (7.04) 30.11 (6.89) 31.40 (7.16)a
Mean (SD)
21.71 (9.16) 23.26 (8.96)
21.16 (9.02) 23.14 (9.19)b
21.98 (9.14) 22.20 (9.34) 21.74 (8.91)
13.70 (6.17) 14.74 (5.69)
13.34 (6.07) 14.65 (6.05)b
13.88 (6.09) 13.88 (6.12) 13.90 (6.07)
Mean (SD)
8.01 (3.91) 8.52 (3.82)
7.82 (3.85) 8.50 (3.92)a
8.10 (3.89) 8.32 (3.96) 7.85 (3.80)
Mean (SD)
Student-school connectedness
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 type Public (n=7) Private (n=4)
453 (58.5) 321 (41.5)
Grade 6th–8th 9th–11th
20.23 (4.39) 20.89 (4.61)c 19.49 (4.01)
774 (100%) 413 (53.4) 361 (46.6)
Total Boys Girls
Mean (SD)
Mean (SD)
Mean (SD)
n (%)
Perceived school cohesion
Global school cohesion
Social exclusion
Physical victimization
Verbal victimization
School cohesion (independent variables)
Exposure to peer victimization (bullying) (dependent variables)
Study sample
Table 1. Demographic characteristics of sample and mean scores of peer victmization and school cohesion by gender, grade and school type.
Original Article 5
IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
IUHPE – Global Health Promotion Vol. 0, No. 0 201X
Downloaded from ped.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on February 17, 2016
−0.81 −0.54
20.41 20.60 21.14
−0.73 −0.67
20.47 20.50 21.18
−0.10 0.13
20.78 20.74 20.93 R2=0.08; Wald chi2(6)=53.22 p = 0.0000
−1.09
19.90
R2=.08; Wald chi2(6)=53.70 p = 0.0000
−1.86d
19.86
R2= 0.081; Wald chi2(6)=51.19 p = 0.0000
−1.72e
B
19.97
Adj. mean
0.7 0.78
0.72
0.53 0.56
0.63
0.62 0.66
0.73
SE
21.39 21.33 21.41
20.48
21.20 21.06 21.70
20.33
21.11 21.13 21.70
20.43
Adj. mean
Boys
R2= 0.09; Wald chi2(5)=31.73 p = 0.0000
0.23 1.39
−0.89
R2= 0.10; Wald chi2(5)=34.62 p = 0.0000
−0.3 −0.11
−2.43d
R2=0.088; Wald chi2(5)=29.04 p = 0.0000
−0.75 −0.52
−2.39c
B
0.94 1.06
0.98
0.74 0.78
0.87
0.83 0.89
0.99
SE
20.05 19.99 20.20
19.33
19.71 19.85 20.55
19.23
19.65 19.99 20.33
19.40
Adj. mean
Girls
1.06 1.17
1.07
0.77 0.82
0.93
0.95 1.00
1.12
SE
R2=0.04; Wald chi2 (5)=13.32 p = 0.021
−0.58 −1.50
−1.44
R2=0.038; Wald chi2(5)=12.50 p = 0.029
−1.41 −1.40
−1.13
R2=0.032; Wald chi2(5)=10.28 p = 0.068
−0.90 −0.53
−0.83
B
and high school students (n = 774), School Bullying Study, Valle del Cauca, Colombia, 2009–11.bReferent group. c < 0.05; d < 0.01; e***
