Request Consultation
Brian Hazeltine, Ed.D.
Abstract
Bullying has been studied world-wide for many years, but there has been virtually no research on bullying in Christian schools. Yet understanding the extent and nature of bullying in Christian schools may lead to a better understanding of ways to address the problem in all schools. This study is an expanded follow-up to the original research done by Hazeltine (2012) and reported by Hazeltine and Hernandez (2015). In addition to addresssing differences in bullying behaviors related to gender, grade, and the number of years a student attended a Christian school, this study includes school size as an independent variable, and includes data related to bystanders and perpetrators as well as victims. Additional variables included the location of bullying, reasons for bullying, whether it was reported, and success of efforts to intervene. Data were analyzed from responses to the Hazeltine Bullying Survey for Christian Schools (2016). Pearson Chi-square tests, independent samples t-tests, one-way ANOVA, and ANCOVA indicated statistically significant differences in bullying behaviors based on gender, grade, and size of the school, but not years attending the school. The results indicate that female students were significantly more likely to witness verbal and social bullying and significantly more likely to be the victim of social bullying than males. Both genders were equally likely to bully others. Bullying was significantly linked to grade with victims, bystanders, and perpetrators all reporting less in the higher grades. Students in lower grades were more likely to report instances of bullying and intervene to stop it. Females were more likely to intervene to stop bullying, and males were more likely to be successful in those efforts. The primary cause of bullying was a response to what others had said or done.
Keywords: Christian school, bullying, social identity theory, gender
BULLYING IN CHRISTIAN SCHOOLS 3
Bullying in Christian Schools
My interest in bullying came about from a parent conversation. The parent was concerned about her child being bullied and urged me to start an “anti-bullying” program for all students in the school. I knew our students and our school, and I didn’t believe then, nor do I believe now that an anti-bullying program was the answer. However, to defend why I felt a school-wide program wasn’t justified I said, “Look, our school does not have a significant problem with bullying.” I learned a valuable lesson that day. While bullying may not be a widespread problem in a school, when it is your child, it is most definitely significant!
Bullying is a common problem in schools worldwide and researchers have examined the phenomenon in many public schools, but very few in private schools, and fewer yet in Christian schools. Yet, every parent, teacher, student, and administrator will acknowledge that bullying does happen in Christian schools, even the best ones. Unfortunately, the person who is best situated to address the problem is usually the least informed: the administrator. On the other hand, the ones with the best knowledge of what is happening are the students. Addressing the problem must begin by finding out what students know. As it turned out, no one had actually investigated the problem of bullying in a Christian school, so that became the topic of my dissertation.
The research by Hazeltine (2012) and reported by Hazeltine and Hernandez (2015) was based on data gathered in 2012 from 346 students in one Christian school. The number of student participants enabled us to begin the work of determing several statistically significant relationships; however, with its focus on one school, we recognized this as preliminary work that would require followup. There are important differences in the design of this study and the
BULLYING IN CHRISTIAN SCHOOLS 4
analysis of the data. Those who are interested in a deeper understanding of the topic generally are advised to review the earlier work.
This current study uses data gathered from over 2000 students in nine Christian schools from across the country. The larger number of schools and students should provide a greater level of confidence in the findings. In particular, the initial study suggested that years attending a Christian school did have some impact, but the smaller numbers limited the analysis. Since the original study in 2012 there have been over 10,000 peer-reviewed journal articles on bullying; however, only three include the words Christian school anywhere in the article. And given the fact that some of the results of the initial study departed from what was found in public schools, I felt that this remained a group worthy of further research.
The data for this research was collected at the end of 2016, but the information will still be highly relevant to those who work in Christian schools. The purpose of this follow-up study was to understand bullying within the Christian school context and to determine whether there are significant differences in bullying behaviors based on gender, grade, school size, and the number of years attending a Christian school.
Introduction
As a followup to the original research, it is worth noting the findings from that preliminary study. Those results indicated generally that
• the patterns of bullying in the Christian school are similar to those found in schools around the world.
• boys more frequently experienced the more obvious physical forms of bullying, • girls are more frequently victimized with a variety of forms including exclusion, rumors, sexual comments, and cyberbullying.
BULLYING IN CHRISTIAN SCHOOLS 5
However, there are also some striking differences that became apparent through that study.
• Bullying in a Christian school takes notably less physically aggressive forms. • Girls were much less likely to be bullied by boys than is typical in other schools. • There was a statistically significant relationships between bullying and years
attending a Christian school, but we were unable to describe the nature of the relationship, likely due to the small size of our study (p. 84)
By expanding the number of schools from one to nine and the number of participants from 346 students to 2018, I hoped to confirm or refute some of the findings of the initial study. Furthermore, with more schools in the study, we were also able to consider the influence of school size as a possible influencing factor on bullying.
Fletcher and Tan (2021) in a summary of the research on school size, noted that there have been some conflicting results. While some researchers argue that more students in the larger schools simply increased the chances of bullying being observed, most researchers found that bullying did increase in larger schools. In a study of charter schools, Hamlin and Li (2021) found that smaller schools were significantly related to increased student safety. They noted that smaller schools typically benefit from closer adult supervision and stronger student-teacher bonds. Goess and Smith (2018) theorized that schools with positive Organizational Citizenship Behavior (OCB) would display less bullying. That was indeed the case, and they found a statistically significant relationship between school size and OCB.
Research Qustions and Hypotheses
The research questions to be answered are as follows:
RQ1: What is the relationship between the independent variables of gender, grade, years
BULLYING IN CHRISTIAN SCHOOLS 6
attending a Christian school, and school size, and the dependent variables of the kind of bullying, its location, and whether it was reported or not.
RQ2: How are the experiences of bullying reported by bystanders, victims, and perpetrators?
Twelve null hypotheses were proposed to examine the individual relationships of each independent variable (gender, grade, years attending, school size) with each dependent variable (kind, location, reporting). Thus, as an example, the first hypothesis was as follows:
H01: There is no statistically significant relationship between gender and the type of bullying reported.
Assumptions, Limitations, and Delimitations
As with all surveys, the assumption is that participants will answer honestly and accurately. To encourage honesty, no personally identifying information was gathered. Furthermore, having the classroom teacher administer the survey provided a familiar environment where students would likely feel free to be honest. The primary limitation is that memories fade with time. However, people tend to remember traumatic experiences more vividly than mundane ones, so this will likely be a minor factor. It is also true that occasionally participants will deliberately sabotage a study, but again this is likely a small minority of participants and not likely to skew results significantly (Ripper et al., 2017). The study was restricted to students in grades three through twelve. Students in lower grades do experience bullying, but due to their limited reading skills, it is better to use observations or interviews when researching students in kindergarten through second grade.
BULLYING IN CHRISTIAN SCHOOLS 7
Method
This setting for this study was an online survey of students from nine Christian schools ranging in size from 55 to 434. Schools joined the survey by responding to an online inquiry in the Administrators discussion forum of the Association of Christian Schools International (ACSI). Each school was a member of ACSI and as such prescribed to its statement of faith. Schools administered the Hazeltine Bullying Survey for Christian Schools (2016) in December of 2016. Teachers administered the survey under standard testing conditions in their classrooms, and data was immediately available to the researcher to analyze. The independent variables were grade, gender, years attending a Christian school, and school size. The dependent variables were (a) the kinds of bullying behaviors (physical, verbal, and social including online), (b) the location where bullying occurred, and (c) how students reported these events to adults. Dependent variables were further categorized in terms of responses from bystanders, victims, and perpetrators.
Survey
While numerous bullying surveys exist, none were considered adequate to address all of the variables of interest to us, and some were quite cumbersome to administer. The Hazeltine Bullying Survey for Christian Schools (2016) was designed to specifically correct these problems. For example, locations where bullying might occur were expanded to include the bus, walking to and from school, after school events, online, at home, and at church. Students were asked about their personal sense of safety and asked to compare their experience in the Christian school with their experience in a public school. A number of questions were designed in a matrix format to reduce the amount of reading and make responding as simple and intuitive as possible. While most surveys inquire about the perspective of victims and bystanders, we included
BULLYING IN CHRISTIAN SCHOOLS 8
perpetrators as well. Finally, the survey was designed with skip logic that enabled students to skip questions that were personally irrelevant. For example, if a student indicated that they had not been bullied, then they were automatically skipped over questions related to location and reporting of bullying. This feature considerably reduces the time for students to complete the survey. Questions were further refined after input from students, teachers, and administrators. Participants
Table 1 displays a breakdown of the participants by gender and includes an additional summary by grade, ethnicity, and race. A total of 2,018 students in Grades 3 through 12 completed the survey, roughly evenly split by gender and across divisions and grades though with somewhat lower numbers in high school. A large majority of the participants (77.3%) were White, and the second largest group was Black at 11.8%. The nine participant schools were categorized for size from small (1-125, n = 3), medium (126-400, n = 4), and large (>400, n = 2). Table 1
Results of the Frequency Analyses for Demographic Characteristics of the Sample
Demographic Characteristics Frequency Percent Gender Male 984 48.8
Female 1032 51.2
Grade 3 205 10.2
4 225 11.2
5 228 11.3
6 198 9.8
7 240 11.9
8 287 14.2
9 167 8.3
10 154 7.6
11 150 7.4
12 162 8.0
Race White 1558 77.3
Asian 84 4.2
Black 237 11.8
BULLYING IN CHRISTIAN SCHOOLS 9
Multiple 76 3.8
Amer Indian 20 1.0
Native Hawaiian 12 0.6
Hispanic 27 1.3
Hispanic or Latin American 2 0.1
Grade Division Elementary School 671 33.3
Middle or Junior High School 721 35.8
High School 624 31.0
Response choices to the survey on bullying were measured on a 6-point Likert-style frequency scale, with 0 = It has not happened to me and 6 = It happens daily. Students were asked to constrain their responses to the frequency with which the incidents of bullying occurred over the previous 2-month period. There is no agreement as to the time frame used by various surveys; two months was considered long enough to ensure that a variety of behaviors would be observed but short enough that student memories would be reasonably accurate. Students were asked to comment on what they had seen as bystanders, what had happened to them personally as victims, and what they did to others as perpetrators. Questions asked students to indicate the location of bullying and whether or not they had reported it and whether or not they had intervened to stop it. Students were also asked about their perceptions of safety and their preception of teacher responses to bullying.
Results
Descriptive Statistics
The following data describe the similarities and differences in responses of the bystanders, victims, and perpetrators. In the previous two months, 43% of participants indicated that they had seen incidences of bullying, 29% indicated that they had been victims, and 11% stated that they were perpetrators. This difference is unsurprising for it is often the case that one incidence of bullying may be witnessed by many bystanders, and conversely much bullying
BULLYING IN CHRISTIAN SCHOOLS 10
appears to be done by a much smaller group. While the absolute numbers of responses to the survey questions varied depending on the group, the relative responses were remarkably consistent. Students, whether bystanders, victims, or perpetrators provided a fairly consistent picture of what was happening in their schools with some notable exceptions. Since victims are our primary concern and most likely to remember events accurately, the percents below are based on victims’ reponses unless otherwise noted.
Fully 71% of students reported that they had not been bullied in the previous two months. In terms of the kinds of bullying experienced by the other 29%, physical bullying was experienced the least with 38% indicating that it had happened more than once in the previous two months. Verbal bullying was somewhat more common at 43%, and social bullying was the most common form (52%).
The most common form of physical bullying was taking or hiding things (51%) and pushing or pulling (49%), and the least common forms were using a weapon (3%) or biting/scratching (10%). The most common forms of verbal bullying were teasing (63%), name calling (60%), and laughing at someone (58%). The least common forms were using social media (11%) and sexual comments (12%). The most common forms of social bullying were spreading rumors (55%) and refusal to include (53%), and the least common form was refusal to eat with someone (18%)
By far the most common reason someone was bullied was in response to something that was said or done (63%). This was nearly double compared to the next two which were athletic ability (32%) and personality/mannerisms (31%). Furthermore, bystanders and perpetrators all pointed to what was said or did as the chief cause for bullying.
The most common locations for bullying were in the classroom without the teacher
BULLYING IN CHRISTIAN SCHOOLS 11
present (41%), in the hallways (41%), in the classroom with the teacher present (40%), on the playground (38%), or in the cafeteria (37%); however, there were large differences due to grade, which is discussed in the hypotheses testing below. The least common locations included at church (4%), walking to or from school (5%), and on the bus (6%).
Most bullies (46%) were identified as males, 34% as female, and 20% of victims indicated that they were bullied equally by males and females. However, perpetrators identified their most frequent victims as male (50%), 22% as female, and 28% as both male and female.
The following numbers reflect responses from all students. When asked about efforts to stop bullying, 43% of students indicated that they had intervened. Of those students, 20% indicated that they had been successful and 65% stated that they had been partly successful. Students were also asked if they had reported bullying to someone else, and 80% indicated that they had not. Of the students who did report, they spoke most frequently with either a teacher (62%), a parent (54%), or friend (32%). When asked about the results of this reporting, 18% indicated that this had solved the problem. Another 44% stated that it had gotten better, 38% indicated no change, and 8% stated that the situation had gotten worse.
There were several questions that addressed student feelings about their school. In terms of safety, 37% said that they felt extremely safe, and 43% said they felt quite safe. Only 4% said that they felt slightly safe, and 2% said that they did not feel safe at all. When asked how much they liked their school, the responses were spread across the spectrum with 43% stating that they liked it a great deal or a lot and 22% saying that they liked it a little or not at all. The remaining 35% were in the middle, liking school a moderate amount. However, students reported that they did well in school with the most commonly reported grades being either an A (65%) or a B (29%).
BULLYING IN CHRISTIAN SCHOOLS 12
Nearly half (46%) of the students surveyed had attended their current school for four years or less. Thirty-four percent had attended between five and eight years, and just 20% had attended for nine to 13 years. Thirty-six percent of students had attended a public school at some point in their educational career. Of these students about half had attended a public school for 3.3 years or less, and half had attended for more than that. When asked to compare bullying in their current school with their previous public school, 75% of students said that it was either much less frequent (53%) or somewhat less than frequent (22%) than in the public school. On the other hand, 14% said that it was somewhat more frequent (8%) or much more frequent (5%).
In terms of addressing the problem, students indicated that 62% their Christian school teachers tried to stop bullying either somewhat more frequently (22%) or much more frequently (40%) than their public school teachers. Sixteen percent said that their Christian school teachers intervened either somewhat less frequently (8%) or much less frequently (8%). Finally, students were asked to compare the success of their teachers in addressing the problem. Compared to their public-school experience, students said that 59% of their teachers in the Christian school were either much more successful (34%) or somewhat more successful (25%). Evaluating Hypotheses
A total of twelve research hypotheses were addressed in this study. Pearson Chi-square test was used to evaluate Hypotheses 1-3 and 10-12, independent samples t-test was employed to evaluate Hypotheses 4-6, and ANCOVA was utilized to examine Hypotheses 7-9. Due to the multiple tests being conducted to address most of these hypotheses, there was the risk of increasing Type I error probability when interpreting the results of the statistical tests. The Bonferroni correction method was utilized to avoid this issue. Following this approach, the significance level for each hypothesis is determined by dividing the number of tests being
BULLYING IN CHRISTIAN SCHOOLS 13
conducted by the original significance level of .05 (Sedgwick, 2012). Furthermore, for each Chi square test, Cramer’s V value was reported to assess the magnitude of the association between the categorical variables under study when a significant association is identified. Values less than .20 were considered indicative of weak associations, .20-.40 as moderate, .40-.60 as relatively strong, and above .60 as strong associations. In addition, Cohen’s d values were calculated for each of the independent samples t-tests as a measure of the magnitude of the mean differences between the two groups under comparison. This measure was calculated by dividing the mean of the differences by the pool standard deviation of the differences between the two groups. Values around .20 were regarded as indicative of small effects, around .50 as moderate effects, and around .80 as large effects (Cohen, 2013).
Gender Correlations
Hypothesis 1
Null Hypothesis 1 is that there is no statistically significant relationship between gender and the type of bullying reported. To address this hypothesis, three Pearson Chi-squared tests were conducted to determine whether there were significant associations between gender and each of the dummy variables representing seeing bullying, perpetrator of bullying, and victim of bullying. By applying a Bonferroni correction, the significance level for these tests was determined at α = .017. The results of these analyses are summarized in Table 2.
It was found that only the dummy variable representing seeing bullying was significantly associated with gender (p < .017). These results suggested that females were more likely to state that they had witnessed an act of bullying than males. The Cramer’s V value indicated that the magnitude of this association was weak. On the other hand, there were no significant associations between gender and other types of bullying (p > .017).
BULLYING IN CHRISTIAN SCHOOLS 14
Table 2
Results of Examining the Associations Between Gender and Different Types of Bullying
Bullying Type χ2df p-value Cramer’s V Seeing Bullying 14.154 1 <.001 .084
Perpetrator 0.451 1 .462 .016
Victim 3.266 1 .071 .040
In addition to the Chi-squared tests, a series of independent samples t-tests were performed to examine whether there were significant differences in physical, verbal, and social bullying by gender. Multiple one-way ANOVAs were also conducted to compare these dependent variables by school size. Furthermore, Pearson correlation analysis was utilized to evaluate the correlations between these dependent variables and grade. These analyses were conducted based on the number of times students had seen, been victimized by, and been perpetrators of physical, verbal, and social bullying. Hence, a total of nine tests were needed to be conducted for each type of analysis mentioned above. Using a Bonferroni correction method, the significance level for these tests was set at α = .050/9 = .006.
The results of the independent samples t-tests are summarized in Table 3. From these results, it can be seen that there were statistically significant differences in the number of times students had seen verbal and social bullying, and the number of times students had been victimized by social bullying between male and female students. The effect size values for these variables indicated that the magnitudes of all these mean differences were very large. It can be concluded from these results that, on average, female students witnessed significantly more instances of both verbal and social bullying. In addition, the average number of times being socially bullied by other students was significantly higher for females than males. In terms of these three types of bullying, no further significant differences were identified by gender.
BULLYING IN CHRISTIAN SCHOOLS 15
Table 3
Comparing Physical, Verbal, and Social Bullying by Gender
Variable Equal Variances T df p valueMean Difference
Std Error Difference
Effect Size
Seeing Physical Assumed -1.338 2014 .181 -0.090 0.067 -1.343 Seeing Verbal Not Assumed -3.821 2013.849 <.001 -0.304 0.080 -3.800 Seeing Social Not Assumed -4.428 2011.26 <.001 -0.360 0.081 -4.444 Victim of Physical Assumed -0.111 2014 .912 -0.007 0.059 -0.119 Victim of Verbal Assumed -1.094 2014 .274 -0.068 0.062 -1.097 Victim of Social Not Assumed -3.321 1995.405 .001 -0.219 0.066 -3.318 Perpetrator of Physical Not Assumed 1.148 1960.054 .251 0.039 0.034 1.147 Perpetrator of Verbal Not Assumed 1.595 1904.629 .111 0.057 0.036 1.583 Perpetrator of Social Assumed 0.125 2014 .900 0.004 0.036 0.111
Hypothesis 2
Null Hypothesis 2 is that there is no statistically significant relationship between gender and the location of bullying. To address this hypothesis, a series of Pearson Chi-squared tests were conducted to determine whether there were significant associations between gender and each of the 15 dummy variables representing location of bullying for those who had witnessed bullying, perpetrators of bullying, and victims of bullying. This means that a total of 45 Chi squared tests were needed to be conducted. By applying a Bonferroni correction, the significance level for these tests was determined at α = .050/45 = .001.
The first series of Chi-square tests were performed to examine the associations between gender and the dummy variables representing location of bullying for seeing bullying. These analyses were performed based on the data collected from the students who reported seeing bullying. It was found that there was a significant association between the dummy variable representing hallways and gender (p < .001). The Cramer’s V value suggested that the magnitude of this association was weak. These results indicated that females were more likely than males to
BULLYING IN CHRISTIAN SCHOOLS 16
see an act of bullying in the hallways than males. No further significant associations were identified between gender and the dummy variable for bullying location (p > .001). Hypothesis 3
Null Hypothesis 3 is that there is no statistically significant relationship between gender and whether bullying was reported. To address this hypothesis, a Pearson Chi-squared test was conducted to examine the association between gender and whether bullying was reported. Only the data for those who reported being victims of bullying were included in this analysis. The results of this analysis are reported in Table 4. It was found that a significant association existed between gender and whether bullying was reported (p < .05). The Cramer’s V value indicated that the strength of this association was weak. These results suggested that females were significantly more likely to report being victimized by bullies than males. Table 4
Results of Examining the Associations Between Gender and Whether Bulling Was Reported for Victims of Bullying
χ2 Df p-value Cramer’s V
Bullying Being Reported 6.121 1 .013 .103
Grade Correlations
Hypothesis 4
Null Hypothesis 4 is that there is no statistically significant relationship between grade and the type of bullying reported. To address this hypothesis, three independent samples t-tests were conducted to examine whether there were significant differences in the mean of grade by the dummy variables representing seeing bullying, being a perpetrator of bullying, and being a victim of bullying. By applying a Bonferroni correction, the significance level for these tests was determined at α = .050/3 = .017.
BULLYING IN CHRISTIAN SCHOOLS 17
As displayed in Table 5, the results of the independent samples t-tests revealed that the mean of grade level significantly varied by all three different types of bullying (p < .017). The effect size values for these grouping variables suggested that the magnitudes of all these differences were very large. These results indicated that the mean of grade was significantly lower for those who had seen bullying, perpetrators of bullying, and victims of bullying compared to the other groups. It can be concluded from these findings that students in lower grades were significantly more likely to have witnessed bullying, be perpetrators of bullying, and be victims of bullying compared to students in higher grades
Table 5
Results of Comparing Grade Level by Different Types of Bullying
Grouping Variable Equal
Variances t Df p-value Mean Difference
Std Error Difference
Effect Size
Seeing Bullying Not assumed -3.238 2012.910 .001 -0.393 0.121 -3.248 Perpetrator Not assumed -2.546 314.350 .011 -0.448 0.176 -2.545 Victim Assumed -9.327 2014 < .001 -1.225 0.131 -9.351
In addition to the independent samples t-tests, a series of Pearson correlation analyses were conducted to examine whether there were significant associations between physical, verbal, and social bullying and grade. A total of nine tests were needed to be conducted. Hence, using a Bonferroni correction method, the significance level for these tests was set at α = .050/9 = .006.
As reported in Table 6, the results of the Pearson correlation analyses indicated that significant negative associations existed between grade and the variables of the number of times students had seen physical bullying, had been physically, verbally, and socially bullied by other students, and had physically bullied other students. On the other hand, grade was significantly positively correlated with the number of times students had witnessed verbal bullying. No further significant associations were identified.
BULLYING IN CHRISTIAN SCHOOLS 18
Table 6
Examining the Correlations Between Grade and Physical, Verbal, and Social Bullying
Variable Df Pearson Correlation p-value Strength of the Association
Seeing Physical 2014 -.117 < .001 Small Seeing Verbal 2014 .069 .002 Very Small Seeing Social 2014 -.002 .935 Very Small Victim of Physical 2014 -.200 < .001 Medium Victim of Verbal 2014 -.072 .001 Very Small Victim of Social 2014 -.124 < .001 Small Perpetrator of Physical 2014 -.066 .003 Very Small Perpetrator of Verbal 2014 .015 .506 Very Small Perpetrator of Social 2014 -.034 .129 Very Small
Hypothesis 5
Null Hypothesis 5 is that there is no statistically significant relationship between grade and the location of bullying. To address this hypothesis, a series of independent samples t-tests were conducted to determine whether there were significant differences in the mean of grade by each of the 15 dummy variables representing location of bullying for those who had witnessed bullying, perpetrators of bullying, and victims of bullying. Thus, a total of 45 t-tests tests were needed to be conducted. By applying a Bonferroni correction, the significance level for these tests was determined at α = .050/45 = .001.
The first series of t-tests were performed to examine whether there were significant differences in the mean of grade by each of the dummy variables representing location of bullying for seeing bullying. These analyses were performed based on the data collected from the students who reported seeing bullying. As reported in Table 7 the results of these t-tests indicated that there were statistically significant differences in the mean of grade level by the dummy variables representing classroom with the teacher present, playground, gym, hallways, online,
BULLYING IN CHRISTIAN SCHOOLS 19
phone, and after school even. The effect size values for these grouping variables indicated that the magnitudes of all these associations were very large. Furthermore, these results suggested that students in higher grades were significantly more likely to see bullying in the classroom with the teacher present, in the hallways, online, by phone, and after school events compared to students in lower grades. On the other hand, students in lower grades were significantly more likely to witness bullying on the playground and in the gym. No other significant differences were observed.
Table 7
Results of Comparing Grade Level by Different Locations of Bullying for Those Who Had Witnessed an Act of Bullying
Grouping Variable Equal
Variances T Df p valueMean Difference
Std Error Difference
Effect Size
Class-Teacher Present Assumed -6.068 951 <.001 -1.028 0.169 -6.083
Class- Teacher not Present
Not assumed -3.024 943.727 .003 -0.507 0.168 -3.018
Playground Not assumed 15.561 751.458 <.001 2.383 0.153 15.575 Gym Not assumed 3.580 825.886 <.001 0.596 0.166 3.590 Hallways Assumed -9.976 951 <.001 -1.611 0.162 -9.944
Restroom Assumed 2.610 951 .009 0.565 0.217 2.604 Cafeteria Not assumed -2.317 950.351 .021 -0.390 0.168 -2.321
Walk To or From School
Assumed 0.150 951 .881 0.047 0.314 0.150
Bus Assumed 1.870 951 .062 0.554 0.296 1.872 Online Not assumed -5.607 231.765 <.001 -1.212 0.216 -5.611 Phone Not assumed -5.836 228.521 <.001 -1.224 0.210 -5.829 After School Events Assumed -3.503 951 <.001 -0.685 0.196 -3.495 Home Not assumed 1.173 110.765 .243 0.376 0.321 1.171 Church Not assumed -0.025 41.471 .980 -0.012 0.493 -0.024 Other Places Not assumed -0.189 39.600 .851 -0.067 0.355 -0.189
The second series of t-tests were performed to examine whether there were significant differences in the mean of grade by each of the dummy variables representing location of bullying for perpetrators. These analyses were conducted based on the data collected from those
BULLYING IN CHRISTIAN SCHOOLS 20
who reported being perpetrators. As displayed in Table 8, it was found that there was a significant difference in the mean of grade level by the dummy variable representing playground (p < .001). The value of the effect size suggested that the magnitude of this difference was very large. These results indicated that students in lower grades were significantly more likely to be perpetrators of bullying on the playground than students in higher grades. No further significant differences were observed (p > .001).
Table 8
Results of Comparing Grade Level by Different Locations of Bullying for Perpetrators of Bullying Grouping Variable Equal
Variances t df p-value Mean Difference
Std Error Difference
Effect Size
Class-Teacher Present Assumed -1.321 233 .188 -0.447 0.339 -1.319 Class- Teacher not
PresentAssumed -1.084 233 .280 -0.376 0.347 -1.084 Playground Assumed 5.075 233 <.001 1.829 0.361 5.066 Gym Assumed -0.663 233 .508 -0.263 0.397 -0.662 Hallways Not assumed -2.98 175.079 .003 -0.971 0.326 -2.979 Restroom Not assumed -0.383 48.757 .703 -0.153 0.398 -0.384 Cafeteria Not assumed -1.595 145.326 .113 -0.545 0.342 -1.594 Walk To or From School Assumed -0.22 233 .826 -0.157 0.716 -0.219 Bus Assumed 1.239 233 .217 0.827 0.668 1.238 Online Assumed -1.543 233 .124 -1.19 0.772 -1.541 Phone Assumed -1.624 233 .106 -0.927 0.571 -1.623 After School Events Not assumed -1.468 46.117 .149 -0.594 0.405 -1.467 Home Assumed 1.797 233 .074 0.917 0.51 1.798 Church Assumed 0.77 233 .442 1.122 1.457 0.770 Other Places Assumed 0.315 233 .753 0.284 0.903 0.315
The third series of t-tests were conducted to examine whether there were significant differences in the mean of grade by each of the dummy variables representing location of bullying for victims of bullying. These analyses were conducted based on the data obtained from those who reported being victimized by bullies. As shown in Table 9 the results of the mean comparisons using the t-tests indicated that there were significant differences in the mean of
BULLYING IN CHRISTIAN SCHOOLS 21
grade level by the dummy variables for classroom with the teacher present, playground, hallways, cafeteria, online, and phone. The effect size values indicated that the magnitudes of all these mean differences were very large. These results suggested that students in higher grades were significantly more likely to be bullied in the classroom with the teacher present, in the hallways, in the cafeteria, online, and by phone. On the contrary, students in lower grades were significantly more likely to be bullied on the playground. No further significant differences were observed.
Table 9
Results of Comparing Grade Level by Different Locations of Bullying for Victims of Bullying Grouping Variable Equal
Variances t df p-value Mean Difference
Std Error Difference
Effect Size
Class-Teacher Present Assumed -5.613 589 <.001 -1.203 0.214 -5.621 Class- Teacher not
PresentAssumed -2.19 589 .029 -0.478 0.218 -2.193 Playground Not assumed 13.364 537.325 <.001 2.449 0.183 13.383 Gym Not assumed 0.549 383.65 .584 0.120 0.219 0.548 Hallways Assumed -6.359 589 <.001 -1.344 0.211 -6.370 Restroom Assumed 0.956 589 .340 0.304 0.318 0.956 Cafeteria Assumed -3.547 589 <.001 -0.784 0.221 -3.548 Walk To or From
School Assumed -1.082 589 .280 -0.543 0.502 -1.082 Bus Assumed 0.595 589 .552 0.265 0.446 0.594 Online Assumed -3.924 589 <.001 -1.271 0.324 -3.923 Phone Assumed -5.284 589 <.001 -1.636 0.31 -5.277 After School Events Assumed -2.376 589 .018 -0.650 0.274 -2.372 Home Not assumed -0.105 85.501 .916 -0.038 0.361 -0.105 Church Assumed -0.696 589 .487 -0.392 0.563 -0.696 Other Places Assumed -0.786 589 .432 -0.381 0.486 -0.784
To sum it up, the results of the independent samples t-tests determined there were statistically significant differences in the mean of grade by several indicators of bullying location.
BULLYING IN CHRISTIAN SCHOOLS 22
Hypothesis 6
Null Hypothesis 6 is that there is no statistically significant relationship between grade and whether bullying was reported. To address this hypothesis, an independent samples t-test was performed to test whether there was a significant difference in the mean of grade by whether bullying was reported. Only the data for those who reported being victims of bullying were included in this analysis.
As reported in Table 10, the results showed that there was a statistically significant difference in the mean of grade level by whether bullying was reported. The effect size value determined that the magnitude of this difference was very large. These results indicated that students in lower grades were significantly more likely to report that they were bullied than students in higher grades.
Table 10
Results of Comparing Grade Level by Whether Bullying Was Reported for Victims of Bullying Grouping Variable Equal
Variances T Df p-value Mean Difference
Std Error Difference
Effect Size
Bullying Being Reported Not assumed -3.717 523.946 <.001 -0.792 0.213 -3.718
Years Attending a Christian School Correlations
Hypothesis 7
Null Hypothesis 7 is that there is no statistically significant relationship between years attending a Christian school and the type of bullying reported. To evaluate this hypothesis, three ANCOVAs were performed to determine whether there were significant differences in the mean of number of years attending a Christian school by the dummy variables representing seeing bullying, being a perpetrator of bullying, and being a victim of bullying after controlling for grade. By applying a Bonferroni correction, the significance level for these tests was determined
BULLYING IN CHRISTIAN SCHOOLS 23
at α = .050/3 = .017.
As shown in Table 11, the results of these analyses indicated that there was a statistically significant difference in the mean of number of years attending a Christian school by being a perpetrator of bullying (p < .017). The effect size for this grouping variable indicated that the magnitude of this difference was small. It can be concluded from these results that after accounting for grade, students attending a Christian school for more years were significantly more likely than those attending for fewer years to be perpetrators of bullying. No further significant differences were observed (p > .017).
Table 11
Results of Comparing Years at a Christian School by Different Types of Bullying After Controlling for Grade
Grouping Variable df1 df2 F p-value Effect Size Seeing Bullying 1 2013 0.026 .871 <.001 Perpetrator 1 2013 9.986 .002 .005 Victim 1 2013 2.080 .149 .001
Hypothesis 8
Null Hypothesis 8 is that there is no statistically significant relationship between years attending a Christian school and the location of bullying. To evaluate this hypothesis, a series of ANCOVAs were conducted to determine whether, after controlling for grade, there were significant differences in the mean of number of years attending a Christian school by each of the 15 dummy variables representing location of bullying for those who had witnessed bullying, perpetrators of bullying, and victims of bullying. Hence, a total of 45 ANCOVAs tests were needed to be conducted. Using the Bonferroni correction approach, the significance level for these tests was determined at α = .050/45 = .001. No significant differences were identified in in
BULLYING IN CHRISTIAN SCHOOLS 24
the mean of number of years attending a Christian school by any of the dummy variables representing location of bullying for bystanders, victims, or perpetrators (p > .001). Hypothesis 9
Null Hypothesis 9 is that there is no statistically significant relationship between years attending a Christian school and whether bullying was reported. To evaluate this hypothesis, an ANCOVA was performed to determine whether there was a significant difference in the mean of number of years attending a Christian school by whether bullying was reported after controlling for grade. Only the data for those who reported being victims of bullying were included in this analysis. After controlling for grade, there was no significant difference in the mean of number of years attending a Christian school by whether bullying was reported, F(1,577) = 0.635, p = .426.
School Size Correlations
Hypothesis 10
Null Hypothesis 10 is that there is no statistically significant relationship between school size and the type of bullying reported. To address this hypothesis, three Pearson Chi-squared tests were conducted to test whether there were significant associations between school size and each of the dummy variables representing seeing bullying, perpetrator of bullying, and victim of bullying. Three chi-square tests were conducted to evaluate this hypothesis.By performing a Bonferroni correction, the significance level for these tests was determined at α = .017..
As can be seen from Table 12, the results of the Chi-squared tests revealed that there were significant associations between school size and all dummy variables representing different types of bullying (p < .017). The Cramer’s V values indicated that the magnitudes of these associations were weak. These results indicated that the students in the schools that fell into the
BULLYING IN CHRISTIAN SCHOOLS 25
category of medium were more likely to see bullying, commit bullying, and be bullied than students in small and large schools. In addition, students in small schools were more likely to witness bullying, commit bullying, and be bullied than students in large schools. Table 12
Results of Examining the Associations Between School Size and Different Types of Bullying
Bullying Type χ2 Df p-value Cramer’s V Seeing Bullying 50.908 2 <.001 .159 Perpetrator 19.754 2 <.001 .099 Victim 27.548 2 <.001 .117
In addition to the Chi-squared tests, a series of one-way ANOVAs were performed to examine whether there were significant differences in physical, verbal, and social bullying by school size. These analyses were conducted based on the number of times students had seen, been victimized by, and been perpetrators of physical, verbal, and social bullying. Hence, a total of nine tests were needed to be conducted. Using a Bonferroni correction method, the significance level for these tests was set at α = .050/9 = .006.
The results of the one-way ANOVAs comparing physical, verbal, and social bullying by school size are reported in Table 20. There were statistically significant differences in the number of times students had seen physical, verbal, and physical bullying, the number of times students had been physically and socially bullied, and the number of times they had physically, verbally, and socially bullied other students by school size. The magnitudes of these differences can be seen in Table 13. Tukey’s procedure was utilized for post hoc analysis in order to determine where these differences had occurred. The results of the post hoc analysis showed that the average number of times students had seen physical bullying was significantly higher (p < .001) in medium schools (M = 1.44, SD = 1.73) than in large schools (M = 0.83, SD = 1.32). The
BULLYING IN CHRISTIAN SCHOOLS 26
average number of times students had witnessed verbal bullying was significantly higher (p = .003) in small schools (M = 1.55, SD = 1.99) than in large schools (M = 1.16, SD = 1.75). The average number of times students had seen social bullying was significantly higher (p < .001) in medium schools (M = 1.69, SD = 1.91) than in large schools (M = 1.19, SD = 1.75). Moreover, the average number of times students had been physically bullied was significantly higher (p < .001) in medium schools (M = 0.91, SD = 1.56) than in large schools (M = 0.51, SD = 1.14). The average number of times students had been socially bullied was significantly higher (p < .001) in medium schools (M = 0.94, SD = 1.60) than in large schools (M = 0.66, SD = 1.38). Furthermore, the average number of times students had physically bullied other students was significantly higher (p < .001) in medium schools (M = 0.34, SD = 0.96) than in large schools (M = 0.14, SD = 0.52). The average number of times students had socially bullied other students was significantly higher (p = .001) in medium schools (M = 0.33, SD = 0.93) than in large schools (M = 0.18, SD = 0.68). No further significant differences were identified.
Table 13
Comparing Physical, Verbal, and Social Bullying by School Size
Variable df1 df2 F p-value Effect Size Magnitude of the Difference
Seeing Physical 2 2013 33.937 < .001 .033 Medium Seeing Verbal 2 2013 7.939 < .001 .008 Small Seeing Social 2 2013 15.898 < .001 .016 Medium Victim of Physical 2 2013 19.260 < .001 .019 Medium Victim of Verbal 2 2013 3.331 .036 .003 Small Victim of Social 2 2013 7.733 < .001 .008 Small Perpetrator of Physical 2 2013 16.094 < .001 .016 Medium Perpetrator of Verbal 2 2013 5.242 .005 .005 Small Perpetrator of Social 2 2013 7.074 .001 .007 Small
Hypothesis 11
BULLYING IN CHRISTIAN SCHOOLS 27
Null Hypothesis 11 is that there is no statistically significant relationship between school size and the location of bullying. To evaluate this hypothesis, a series of Pearson Chi-squared tests were conducted to determine whether there were significant associations between school size and each of the 15 dummy variables representing location of bullying for those who had witnessed bullying, perpetrators of bullying, and victims of bullying. Thus, a total of 45 Chi squared tests were needed to be conducted. Using the Bonferroni correction method, the significance level for these tests was set at α = .050/45 = .001.
The first series of Chi-square tests were performed to examine the associations between school size and the dummy variables representing location of bullying for seeing bullying. These analyses were conducted based on the data obtained from the students who reported seeing bullying. As reported in Table 14, the results of these tests showed that there were significant associations between school size and the dummy variables representing classroom without the teacher present, playground, bus, and phone (p < .001). The Cramer’s V value indicated school size was weakly associated with classroom without the teacher present and moderately associated with playground. These results showed that regarding school size, students in small schools were most likely to see bullying in the classroom without the teacher present, while students in large schools were least likely. In addition, students in medium schools were most likely to see bullying on the playground while students in large schools were least likely. Moreover, students in small schools were most likely to witness bullying on the bus whereas students in large schools were least likely. Lastly, students in large schools were most likely to witness bullying by phone whereas students in small schools were least likely. No other significant associations were found.
BULLYING IN CHRISTIAN SCHOOLS 28
Table 14
Results of Examining the Associations Between School Size and Different Bullying Locations for Those Who Had Witnessed Bullying
Bullying Location χ2df p-value Cramer’s V Class-Teacher Present 1.074 2 .584 .034 Class- Teacher not Present 26.170 2 <.001 .166 Playground 53.596 2 <.001 .237 Gym 1.689 2 .430 .042 Hallways 8.246 2 .016 .093 Restroom 6.926 2 .031 .085 Cafeteria 6.071 2 .084 .080 Walk To or From School 8.689 2 .013 .095 Bus 27.876 2 <.001 .171 Online 8.119 2 .017 .092 Phone 16.635 2 <.001 .132 After School Events 3.374 2 .185 .060 Home 6.999 2 .030 .086 Church 4.257 2 .119 .067 Other Places 0.818 2 .664 .029
The second series of Chi-square tests were performed to evaluate the relationships between school size and the dummy variables representing location of bullying for perpetrators of bullying. These tests were conducted based on the data collected from those who reported being perpetrators. As shown in Table 15, these results revealed that there were significant associations between school size and the dummy variables representing classroom without the teacher present, and playground (p < .001). The Cramer’s V value indicated that the magnitude of both these associations was weak.
These results revealed students in small schools were most likely to be perpetrators of bullying in the classroom without the teacher present, while students in large schools were least likely. Furthermore, students in medium schools were most likely to be perpetrators of bullying
BULLYING IN CHRISTIAN SCHOOLS 29
on the playground while students in large schools were least likely. No further significant relationships were observed (p > .001).
Table 15
Results of Examining the Associations Between School Size and Different Bullying Locations for Perpetrators of Bullying
Bullying Location χ2df p-value Cramer’s V Class-Teacher Present 6.647 2 .036 .168 Class- Teacher not Present 18.628 2 <.001 .282 Playground 18.957 2 <.001 .284 Gym 2.160 2 .340 .096 Hallways .064 2 .968 .017 Restroom 6.322 2 .042 .164 Cafeteria 4.665 2 .097 .141 Walk To or From School 9.601 2 .008 .202 Bus 1.348 2 .510 .076 Online 0.885 2 .642 .061 Phone 0.134 2 .935 .024 After School Events 0.436 2 .804 .043 Home 2.824 2 .244 .110 Church 2.247 2 .325 .098 Other Places 1.150 2 .470 .080
The third series of Chi-square tests were performed to examine the associations between school size and the dummy variables representing location of bullying for victims of bullying. These tests were performed based on the data obtained from those who reported being victimized by bullies. The results of these tests are provided in Table 16. It was found that there were statistically significant associations between school size and the dummy variables indicating playground, and walk to or from school (p < .001). The Cramer’s V values indicated that school size was moderately associated with playground and weakly associated with walk to or from school. These results indicated that students in medium schools were most likely to be victimized by bullies on the playground whereas students in large schools were least likely. Furthermore,
BULLYING IN CHRISTIAN SCHOOLS 30
students in medium schools were most likely to be bullied when walking to or from school while students in small schools were least likely. No other significant associations were identified (p > .001).
Table 16
Results of Examining the Associations Between School Size and Different Bullying Locations for Victims of Bullying
Bullying Location χ2 Df p-value Cramer’s V Class-Teacher Present 1.209 2 .546 .045 Class- Teacher not Present 9.265 2 .010 .125 Playground 24.978 2 <.001 .206 Gym 1.313 2 .519 .047 Hallways 12.334 2 .002 .144 Restroom 1.584 2 .453 .052 Cafeteria 2.755 2 .252 .068 Walk To or From School 15.615 2 <.001 .163 Bus 10.004 2 .007 .130 Online 2.045 2 .360 .059 Phone 3.864 2 .145 .081 After School Events 7.225 2 .027 .111 Home 13.233 2 .001 .150 Church 2.272 2 .321 .062 Other Places 3.541 2 .170 .077
To sum it up, the results of the Pearson Chi-squared tests determined that some indicators of bullying location were significantly associated with school size.
Hypothesis 12
Null Hypothesis 12 is that there is no statistically significant relationship between school size and whether bullying was reported. To evaluate this hypothesis, a Pearson Chi-squared test was conducted to examine the association between school size and whether bullying was reported. Only the data for those who reported being victims of bullying were included in this
BULLYING IN CHRISTIAN SCHOOLS 31
analysis. The results of this test indicated that there was no significant association between bullying being reported and school size (p > .05).
Examining the Relationship Between Bullying Intervention and Gender Two chi-square tests were conducted to determine whether significant associations existed between gender and whether students intervened as well as whether the intervention was successful. The results from these tests are summarized in Table 17. It can be seen that both associations were statistically significant. The Cramer’s V values indicated the strength of the associations evaluated by the first and second chi-square values were negligible and weak, respectively.
These results suggested that females were more likely than males to intervene. Of the 984 male students, 39.0% (n = 384) reported that they had tried to stop someone from bullying someone else. On the other hand, of the 1031 female students, 46.3% (n = 477) stated they had intervened. Moreover, it was concluded from the results of the chi-square tests that males were more likely to be successful than females. Of the 385 male students who responded to the question about whether their intervention was successful, 24.9% (n = 96) reported being successful, 15.8% (n = 61) stated being not successful, and 59.2% (n = 228) stated being partly successful. On the other hand, of the 479 female students, 15.7% (n = 75) were successful, 14.6% (n = 70) were not successful, and 69.7% (n = 334) were partly successful. Table 17
Results of Examining the Associations Between Gender and Whether Students Intervened as Well as Whether the Intervention was Successful
Bullying Type χ2 Df p-value Cramer’s V Intervene 10.789 1 .001 .073
Intervene Success 13.119 2 .001 .123
BULLYING IN CHRISTIAN SCHOOLS 32
Examining the Relationship Between Bullying Intervention and Grade Two one-way ANOVA’s were conducted to determine whether grade had significant effects on whether students intervened and whether the intervention was successful. The results from these analyses are summarized in Table 18. It can be seen that the effect of grade on both these variables was statistically significant. The magnitude of both these effects was found to be moderate. It can be concluded from these results that students in lower grades were statistically significantly more likely to try to stop someone from bullying someone else than students in higher grades. Furthermore, the results of the Tukey post hoc test for the second ANOVA revealed that the mean grade of students who were partly successful was significantly higher than students who were unsuccessful. These results suggested that students in higher grades were more likely to be partly successful and students in lower grades were more likely to be unsuccessful.
Table 18
Results of Examining the Effect of Grade on Whether Students Intervened as Well as Whether the Intervention was Successful
Grouping Variable df1 df2 F p-value Effect Size
Intervene 1 2013 72.759 < .001 .035
Intervene Success 2 861 6.728 .001 .015
Discussion
The purpose of this study was to expand on the original preliminary research by Hazeltine (2012) to explore bullying in Christian schools using a larger sample of students and more participant schools. Despite numerous studies on bullying, the Christian school context remains largely unexamined. Social identity theory argues that people adjust their behaviors to
BULLYING IN CHRISTIAN SCHOOLS 33
accommodate to the norms of the prevailing culture, and the original study did indeed suggest that there was a statistically significant correlation between bullying and years attending the Christian school. This study was designed to either confirm or refute the findings of the original study; to explore whether school size was an additional mediating factor; and to determine whether student responses differed markedly depending on their perspective as a perpetrator, victim, or bystander.
Gender
Girls see more bullying of any kind than boys. The difference is statistically significant, but the effect size is rather small. However, in terms of specific kinds of bullying, girls see significantly more verbal and social bullying, and they indicate that they are are more likely to be victims of social bullying. The effect size was very large. This is consistent with our previous study and other research. However, in contrast to other studies and our previous work, boys did not indicate that they experienced physical forms of bullying more than girls. Indeed, physical bullying was among the least prevalent forms of bullying reported with 63% of students saying that they had never witnessed any (41%) or had seen it only once (22%) or two to three times (21%) in the last two months. However, this is actually more than what was reported in the original study where 90% of the students reporting no incidents in the past “couple of months.” This raises the question as to whether this is a genuine increase in physical bullying over the four-year period between studies or whether it is the result of a different sample.
There was only one difference between the genders about where bullying happens, and that was that girl bystanders were somewhat more likely to notice it in hallways. The other difference between the genders is that girls were also a bit more likely to report bullying than
BULLYING IN CHRISTIAN SCHOOLS 34
boys. These findings were supported by all students, whether perpetrators, victims, or bystanders.
Grade, Division, and Years Attending
There was a statistically significant and very large correlation between grade and the occurrence of bullying, and this was noted by perpetrators, victims, and bystanders alike. As student progress through school, there is a clear decline in bullying behaviors by grade but not specifically by division. This contrasts with other studies that indicate that bullying increases through the grades. This is encouraging because it suggests that parents, teachers, and fellow students are doing an effective job of socializing children. The question we hoped to answer was whether the Christian school specifically could be identified as a factor.
Our previous study and the initial analysis in this study suggested that there was a correlation between years attending and the decline in bullying behavior. However, as the details were examined it became clear that the pattern in bullying behavior reported by those who had attended the longest was very similar to that reported by students in the upper grades. Once we completed a second analysis in which we controlled for grade, we could find no statistically significant relationship between years attending the Christian school and any bullying behaviors. The decline that had appeared to be due to years attending was simply due to the fact that only the older students could possibly have attended for, as an example, 10 to 12 years.
The fact that we could not find a direct correlation between years attending the Christian school and the decline in bullying does not mean that the school has no influence, but it does mean that we cannot identify that influence apart from the influence of supportive parents and a positive peer culture. These two factors are likely more significant in socializing young people
BULLYING IN CHRISTIAN SCHOOLS 35
than the school. At the same time, the Christian school works as a means of bringing together like-minded people to create a culture that appears to do a good job of socializing students. There was one additional finding about years attending that is quite striking. At the same time that students in upper grades indicate that they have seen less bullying and been victimized less, students in upper grades were more likely to report themselves as being perpetrators. However, beyond this one data point, nothing in the location of bullying or reporting of bullying experienced by victims and bystanders, or even by perpetrators provides any additional support to this. It is difficult to explain this particular finding, but one possibility is that while there are more perpetrators in higher grades, they victimize fewer students and do so more covertly. The second possibility is that as students mature, perhaps they are more honest and willing to admit to this kind of negative behavior.
School Size
As noted in Table 13, school size was significantly associated with verbal, social, and physical bullying for perpetrators, victims, and bystanders. However, the effect sizes were very small. Anecdotally, small schools are viewed as being like a family. Everyone knows everyone, and that is typically helpful for raising well-adjusted children. Small schools generally reported less bullying than midsize schools, but larger schools reported even less.
It is not clear why bullying tends to be somewhat more prevalent in a medium size school, but it may be due to the fact that these schools are losing some of the closeness that they once had. Not every teacher may know every student, for example. At the same time, a medium size school may not be large enough to hire a full-time counselor or provide other support systems for students. For example, two of the reasons given for why students bully are their grades and their athletic ability. Medium size schools have less flexibility about course offerings,
BULLYING IN CHRISTIAN SCHOOLS 36
and it may be that students who are forced into courses where they do not do well become targets.
There were a few exceptions to the above pattern when it came to the location of bullying. For example, students in small schools were more likely to bullying in the classroom when the teacher was not present, and those in larger schools reported more bullying by phone. While these were statistically significant differences, the effect size was weak, and there is no clear explanation for these differences. Finally, school size had no observable impact on whether students reported bullying.
Intervention
We did not generate a hypothesis regarding intervention, but the data revealed that females were more likely than males to intervene to stop bullying behavior. On the other hand, males were more successful in their efforts. The reason for these differences is not clear, but it may be that there is acceptance in the Christian school of traditional gender stereotypes in which the female is viewed as nurturing and the male is viewed as powerful. Girls in this kind of setting may see themselves as obligated to reach out to care for others while boys may prefer to use their size and strength to ensure that they get the desired results.
Students in lower grades were somewhat more likely to intervene to stop bullying than those in the uppoer grades. The reasons for this is unclear, but it might be due to elementary students having a stronger sense of community. They typically spend most of their day in one classroom with the same group of students. In contrast, high school students typically move from class to class, and depending on the size of the school, the number of sections per grade, and the electives chosen, they may encounter far more students in any given day. Although upper grades students were less likely to intervene, they were somewhat more likely to achieve positive results
BULLYING IN CHRISTIAN SCHOOLS 37
when they did. This too may be due to increased size and strength, but also to increased status in the school as upper grades students.
Location and Reporting of Bullying
This study confirmed the findings of the original research that bullying patterns are directly connected to school life. For example, elementary students typically spend more time on the playground, and high school students spend more time in the hallways, and high school students are more likely to have cell phones and be connected on social media. While victims and bystanders gave similar reports about where they had experienced bullying based on their grade, perpetrators only indicated playground as significantly more common for elementary students. Teachers and administrators should take note of the locations where students in various grades are most likely to experience bullying. See Tables 7-9.
Students in elementary grades are much more likely to report incidences to a teacher than those in the upper grades. Since much bullying therefore goes unreported, it is critical that teachers exercise careful supervision of the most common locations. A challenge for teachers is that perpetrators in upper grades seem to have developed skill at bullying in ways that go undetected by teachers. Notably, students in higher grades report being significantly more likely to be bullied in the classroom with the teacher present! However, since most students who reported bullying indicated that doing so either solved the problem or made it better, teachers should encourage students to speak up.
Attitude Towards School
It should come as no surprise that being bullied has an impact on a student’s attitude toward school. We found a clear relationship between a higher incidence of being bullied and a greater dislike of school. Students who felt threatened and unsafe at school due to the presence of
BULLYING IN CHRISTIAN SCHOOLS 38
bullies began to develop negative attitudes toward school. These negative attitudes may eventually impact student behavior or academic achievement.
Conclusion
The results of this followup study confirm much of the findings of the original research (Hazeltine, 2012). Gender and grade are both significantly associated with the three major kinds of bullying, verbal, social, and physical. Given the very large effect sizes for grade and gender, these are the factors that teachers and administrators should keep foremost as they attempt to address problems in their schools. The number of years attending a Christian school was not found to be significant, and the size of the school had only minimal impact. Bystanders, victims, and perpetrators largely agreed about the kinds of bullying taking place, where it happens, and the reasons for it. Armed with this information, administrators can focus their efforts on addressing verbal and social bullying in less supervised areas such as playgrounds, and teaching students about the impact of their words and actions and the need to respond with self-control.
Two very positive trends in Christian schools are that bullying tends to decline as students get older, and physical bullying is the least common experience. Students report generally positive attitudes toward their schools, and they report that bullying is less frequent and teachers more committed to stopping bullying it than in their previous public school.
The general agreement on many items among bystanders, victims, and perpetrators indicates that the survey instrument can be revised to focus mostly on victims. This will simplify the survey and reduce the time to administer it. Additionally, differences among individual individual grades are relatively small, so supervision may be safely designed by division.
Further research is needed to understand the reasons for some of the differences we noted. For example, it may be worth doing a qualitative study of schools to determine why
BULLYING IN CHRISTIAN SCHOOLS 39
medium size schools are somewhat more likely to experience bullying. The differences uncovered between the genders and among the grades regarding interventions also needs further study.
Finally, the comparisons to other schools regarding the amount of bullying and teacher efforts to quell it are based on student reports of what they remembered from their previous schools. It would be helpful to do a cross-sectional study of bullying in Christian schools, charter schools, and public schools for more sophisticated statistical analysis of the comparisons. The lack of statistically significant relationship between years attending a Christian school and bullying behaviors suggests that the impact of Christian school may be found more in forming student values and worldviews than in changing their behavior.
BULLYING IN CHRISTIAN SCHOOLS 40 References
Cohen, J. (2013). Statistical power analysis for the behavioral sciences. Routledge. https://ebookcentral.proquest.com/lib/cornerstone-ebooks/detail.action?docID=1882849. Fletcher, E. C., & Tan, T. X. (2021). A Comparative Analysis of School and Student Characteristics on Bullying in Urban Career Academies and a Large Comprehensive High School. Journal of School Violence, 20(2), 114–126. https://doi.org/10.1080/15388220.2020.1850463 Hamlin, D., & Li, A. (2021). Factors Mediating School Safety in Charter Schools: An Analysis of Five Waves of Nationally Representative Data. Journal of School Choice, 15(3), 371–392. https://doi.org/10.1080/15582159.2020.1783475
Hazeltine, B. (2012). The Extent and Nature of Bullying in a Christian School [ProQuest LLC]. In ProQuest LLC (Vol. 74, Issue 5). https://go.exlibris.link/bzpCK9J4
Hazeltine, B. C., & Hernandez, D. A. (2015). The Extent and Nature of Bullying in a Christian School. Journal of Educational Research and Practice, 5(1), 74–87.
Ripper, L., Ciaravino, S., Jones, K., Jaime, M. C. D., & Miller, E. (2017). Use of a Respondent Generated Personal Code for Matching Anonymous Adolescent Surveys in Longitudinal Studies. Journal of Adolescent Health, 60(6), 751.
Sedgwick, P. (2012). Multiple significance tests: The Bonferroni correction. BMJ, 344. https://doi.org/10.1136/bmj.e509