Evaluation Report
The Safe School Ambassadors
®
Program:
A Student Led Approach to
Reducing Mistreatment and Bullying in Schools
Submitted by:
Alexander White
Texas State University, San Marcos
Katherine Raczynski
University of Georgia, Athens
Chris Pack
Community Matters
Aijun Wang
University of Georgia, Athens
January 2011
Support for this program evaluation was provided in part by a grant from
the Kaiser Permanente Foundation Community Benefit Program
Northern California Region.
Safe School Ambassadors Program Evaluation 2011
Table of Contents
1. Introduction ............................................................................................................. 1
1.1 Overview of the Field .......................................................................................... 1
1.2 Description of the Safe School Ambassadors Program ...................................... 2
1.3 Design of the Evaluation ..................................................................................... 3
2. Part I ......................................................................................................................... 3
2.1 Methods .............................................................................................................. 3
2.2 Data Sources ...................................................................................................... 4
2.3 Results ................................................................................................................ 5
2.3.1. Ambassadors and Key Students ................................................................. 5
2.3.2. Ambassador/Key Student Survey ................................................................ 6
2.3.3. Climate Survey ............................................................................................ 6
2.3.4. Friends of Ambassadors and Key Students................................................. 6
2.3.5. School population ........................................................................................ 7
2.4 Discussion ........................................................................................................... 8
2.4.1. Learnings about Ambassadors .................................................................... 8
2.4.2. Learnings about Friends .............................................................................. 8
2.4.3. Learnings about Overall Climate ................................................................. 9
2.5 Challenges and Limitations ............................................................................... 10
2.6 Implications for Future Research ...................................................................... 11
3. Part II ...................................................................................................................... 11
3.1 Methods ............................................................................................................ 11
3.2 Data Sources .................................................................................................... 11
3.3 Results .............................................................................................................. 12
3.4 Discussion ......................................................................................................... 13
3.5 Implications for Future Research ...................................................................... 13
4. Conclusion ............................................................................................................ 13
5. References ............................................................................................................ 15
6. Tables and Figures ............................................................................................... 16
Table 1: Summary of participants and roles in the SSA program and evaluation ...... 16
Table 2: Summary of scales/items included in student self-report surveys ............... 17
Table 3: Summary of Friends Index ........................................................................... 18
Table 4: Slope Coefficients and p-values from the Friends analysis ......................... 18
Table 5: SSA Program Implementation Index ............................................................ 19
Table 6: Descriptive Statistics on Post-SSA Suspension Rate .................................. 19
Table 7: Descriptive Statistics on the Difference in Suspension Rate ....................... 19
Table 8: Administrators’ Reports of SSA Program Impact ......................................... 19
Figure 1: Estimated Marginal Means of Active Intervention with Non-Friend............. 20
Figure 2: Key Adults’ Observations of Active Interventions to Stop Mistreatment. .... 21
Figure 3: Change in Suspension Rate after SSA Program Implementation............... 21
7. Appendix I: The Logic Model of the Safe School Ambassadors program. ...... 22
Safe School Ambassadors Program Evaluation 2011
Executive Summary
Safe School Ambassadors (SSA) is a student-centered bystander education program
developed by Community Matters to reduce bullying/mistreatment and enhance school
climate. A two-part program evaluation was designed to assess the efficacy of the
program’s logic model, and its impact on school-level discipline indicators; several
statistically significant outcomes (p < 0.05) were obtained.
For the first part, a two-year evaluation was conducted in five middle schools in Texas
using a quasi-experimental pre-post-post design. After two years, rates of helpful
intervention were higher for male Ambassadors than for their controls. Ambassadors’
Friends noticed more helpful interventions and reported observing less mistreatment
than Friends of Key Students at the control schools. Process data indicated positive
effects on discipline and overall climate.
For the second part, suspension and other discipline data was gathered from nineteen
schools that implemented the SSA program with a high degree of fidelity to internal
benchmarks, and from demographically matched non-SSA schools that served as
controls. Analysis of suspensions and other discipline indicators at SSA schools showed
reductions averaging 33%, while indicators at matched control schools rose 10% during
the same years.
Safe School Ambassadors Program Evaluation 2011
Page 1
1. Introduction
1.1 Overview of the Field
An extensive body of literature has established that the rate of mistreatment – including
verbal and physical bullying, relational and electronic aggression (cyber-bullying) – that
students experience in connection with school is unacceptably high. Approximately 30%
of U.S. students in grades 6-10 report being involved in moderate or frequent bullying
(Nansel, Overpeck, Pilla, et al., 2001). A survey commissioned by the National Crime
Prevention Council (2007) found that 43% of youth ages 13-17 report experiencing
some form of cyberbullying in the prior year. These behaviors are harmful – to the
aggressor, the target, and others in the school environment – and they are also costly,
in terms of resources spent on discipline and physical security, as well as hours of lost
learning time due to distraction, absenteeism, and dropout rates.
Over the past decade, more than $10 billion has been invested to reduce violence and
improve safety in schools (U.S. Department of Justice, 2003. Packaged Facts, 2000).
The vast majority of those dollars have underwritten an “outside-in” approach, i.e., one
that has focused on physical safety and security, including fences, locks, cameras,
tough rules and zero-tolerance policies, and additional staff to monitor student behavior
(Knickerbocker, 2004). That approach has not yet produced the desired results, which
suggests the need to explore the reasons for this lack of success, as well as alternative
and/or complementary approaches.
A social-ecological theory has been used to describe the development of aggressive
behavior in school children. The theory posits that the interaction of individual, family,
peer, and community level factors influence an individual’s propensity to engage in or
refrain from aggressive behaviors (Espelage, 2004). Experts have identified
characteristics at the school level that protect against (protective factors) or contribute to
(risk factors) youth violence in schools. These characteristics relate to relationships
within a school (e.g., student-teacher, student-student), school policies and regulations
(e.g., rules against bullying), and the physical school building and grounds (e.g.,
supervised areas). Many prevention programs aimed at addressing school violence
have focused on improving these aspects of the school environment, although many
utilize an adult-centered approach.
Several studies have described the ways that social influence impacts bullying
behavior. A study by Cairns, Leung, and Cairns (1995) indicated that students tend to
bully about as much as their friends and that students whose friends bully tend to
increase their levels of bullying throughout the school year. This result indicates that
children form social groups with others who possess similar norms and behaviors, and
that these behaviors can spread throughout the group.
Researchers have further underscored that bullying is a context-specific occurrence,
i.e., the occurrence, type, and severity of bullying behavior may depend on
characteristics of the environment, including the absence or presence of a crowd of
witnesses. Researchers have found that at any given time, 70-85% of students at a
Safe School Ambassadors Program Evaluation 2011
Page 2
school are neither aggressors nor targets of mistreatment; instead, they are bystanders
who witness the aggression (Hazler, 1996. Craig & Pepler, 1995). Researchers have
articulated several bystander roles that students may play when bullying is occurring.
Salmivalli, et al. defined the roles of assistant of the bully, reinforcer of the bully,
defender of the victim, and outsider (1996). Bystanders can either encourage the
bullying to continue by supporting the bully or ignoring the bullying, or they can
contribute to reducing bullying by supporting the victim or expressing disapproval of the
bully’s actions. Slaby (2005) corroborates the key role of student bystanders in
determining whether bullying will escalate, be stopped, or be prevented. This and other
research, together with field experience (Vossekuil, et al, 2004), indicate the value of
incorporating “inside-out,” student-centered approaches to violence prevention efforts
as well.
1.2 Description of the Safe School Ambassadors Program
Beginning in 2000, the Safe School Ambassadors (SSA) program has been developed
by Community Matters, a 501(c)3 education and youth development non-profit based in
northern California. It has been implemented in approximately 900 elementary, middle
and high schools in 28 US states and 2 Canadian provinces. The program incorporates
established principles of youth development and prevention, as well as emerging
research and field experience. In 2005, the organization commissioned Denver-based
Omni Research and Training to conduct a review of the literature, which provided
evidence for the soundness of the underlying logic model (see Appendix I).
In the Safe School Ambassadors program, the socially influential opinion leaders of the
diverse cliques and interest groups on a school campus are identified, recruited, and
selected via a precise process involving staff and student input. It is expected that
through their collective social networks, they will have a large sphere of influence on the
student population’s behaviors, and thus on the climate of the school.
Through a 2-day, highly experiential training process, these student Ambassadors are
equipped to identify, prevent, and respond to student aggression and mistreatment, i.e.,
to act as proactive and helpful bystanders by:
• interrupting mistreatment as it occurs
• preventing mistreatment from happening by discouraging peers from committing
hurtful or violent acts
• supporting students who have been mistreated, and
• obtaining adult help when situations are too complex or dangerous for them to
handle by themselves.
Competence in these skills allows Ambassadors to wield their social influence to reduce
bullying and promote a healthy school environment.
After the training, ongoing support, supervision, and skill development are provided
throughout the year via regularly scheduled “Family Group” meetings of 7-10
Ambassadors facilitated by 1-2 specially trained adults who use a curriculum provided
by the program. Ambassadors periodically document the nature and frequency of their
interventions on Action Logs, which are analyzed by program adults.
Safe School Ambassadors Program Evaluation 2011
Page 3
1.3 Design of the Evaluation
To assess the quality of the SSA program’s logic model in practice, and to determine
the impact of the program on school discipline indicators (e.g. suspension rates), a two-
part evaluation was designed.
• Part I: in-place SSA programs were studied using a quasi-experimental pre-post-
post design that included surveys of students and adults, as well as an
examination of process data.
• Part II: schools with high-fidelity implementation of the SSA program were identified
and matched to demographically similar controls, and school discipline data and
surveys were used to assess program impact.
An Institutional Review Board (IRB) Exemption was granted by Texas State University,
San Marcos.
2. Part I
2.1 Methods
The primary research questions of Part I centered around potential changes in the
attitudes, behaviors, and school connectedness of Ambassadors and their closest
friends. We also sought to determine whether there were perceptible changes in school
climate and school discipline indicators.
Five middle schools in single school district in an urban/suburban area of central Texas
participated in this two-year evaluation. Three schools elected to receive the SSA
program and two control schools were selected to match these intervention schools with
regard to school size, student demographics (% receiving free/reduced lunch, ethnicity),
and prevalence of discipline incidents. Interventions took place over two years.
Participants were identified as described below and in Table 1.
In year 1, in each of the three treatment schools, staff and students identified 60-80
students in grades 6-8 as being the socially influential leaders of their school's diverse
cliques, i.e. the affinity groups that naturally form around common interests and beliefs,
such as football players, cheerleaders, band members, skaters, etc. Selection criteria
included high verbal skills, loyalty to peer group, and ability to discern right from wrong
even if they sometimes got into trouble. These prospective Ambassadors attended an
orientation to the SSA program and were invited to participate. Of those that chose to
participate, the adult Program Advisor (coordinator of the SSA program, typically a
school counselor, dean, assistant principal, or influential teacher) selected 40 to
represent the diversity of the school. The two day Ambassador training session focused
on the following objectives: (a) discovering commonalities and building bridges of
understanding across diverse groups; (b) becoming more aware of the problem and the
costs of mistreatment and violence on campus; (c) acquiring and sharpening skills for
preventing and responding to mistreatment and violence; (d) and developing motivation
to actually use those skills.
Safe School Ambassadors Program Evaluation 2011
Page 4
Treatment schools also chose six to eight adults to receive training, administer the
program, and conduct follow-up meetings with Ambassadors at two-week intervals
throughout the year. Small "Family Groups" of seven to ten Ambassadors met with one
to two of these Program Adults to discuss their successes and challenges, receive
suggestions and support, and to develop additional prevention/intervention skills.
Several meetings of all the Ambassadors at a school were also held each year.
In control schools, an administrator selected Key Students to match Ambassadors from
the treatment schools with regard to gender, grade, ethnicity, academic performance,
and socio-economic status. Key Students received no intervention (no training,
meetings, etc.) but participated in surveys and focus groups.
In year 2, returning Ambassadors, Program Advisors and Family Group Facilitators
were invited to continue; Key Students retained that designation. Another 30-40
influential students were selected to replace non-returning Ambassadors and increase
the "reach" of the program within the school; control schools also chose a matching
cohort of new Key Students. In treatment schools, the initial training and Ambassadors
meetings were replicated.
2.2 Data Sources
The following data sources were used in Part I of this program evaluation. Table 2
provides a description and α values for the student survey scales.
School Climate Survey: This confidential survey includes 109 items about perceived
level of bullying/mistreatment both as a target and as an aggressor (observed and
personally experienced) as well as school norms regarding mistreatment, school
climate, and school connectedness. It also included internal reliability measures (e.g.,
pairs of opposite questions). The format of each item was Likert-type or ordered
category; most items had five response categories. Demographic information (sex,
grade, race/ethnicity) and identifying information for linking surveys across time were
collected. In year 1, the School Climate Survey was completed by all students in
grades 6-7, plus any eighth graders who were named by Ambassadors/Key students as
"Friends" (see below). In year 2, all students in grades 7-8 completed it.
Ambassador/Key Student Survey: This survey is composed of 94 items designed to
measure Ambassador/Key Student perceptions of bullying and mistreatment, school
connectedness, climate, and norms regarding mistreatment. It too included internal
reliability measures. Ambassadors also responded to a mix of ordered category and
open-ended questions measuring aspects of the Ambassador program, such as
perceived helpfulness of the trainings/meetings, and perceived changes in school
climate/student norms as a result of the Ambassador program.
Friends: Each Ambassador and Key Student also identified as many as 10 students
who are their close friends. The School Climate Surveys of these “friends” were flagged
for further analysis, because it was hypothesized that these students would show the
earliest changes as the Ambassadors began to use their skills.
Safe School Ambassadors Program Evaluation 2011
Page 5
Key Adult Survey: This survey is composed of 42 items (ordered-response and open-
ended) designed to measure adults' perceptions of the frequency of student
mistreatment, school safety, connectedness and climate, and the effectiveness of other
bullying prevention programs implemented in the school.
School records data: discipline records were collected and analyzed for trends over
time.
Process data sources: To augment the statistical data collected, the following
information was also obtained and reviewed: a) questionnaires and interviews with
school principals and/or disciplinarians, b) Action Log Snapshots (i.e., records of
Ambassador interventions), c) Ambassador attendance at Family Group meetings, d)
reports from Program Advisors and Family Group Facilitators, and e) Year-End Surveys
of Program Advisors and Family Group Facilitators.
Focus Groups Sessions with small groups of students in control and treatment schools
were recorded. Students were asked about the surveys, the overall school climate and
where appropriate about the perceived success of the program.
Four waves of data collection occurred: a pre-test in October of Year 1, a post-test in
May of Year 1, a second pre-test of newly identified Ambassadors and Key Students in
October of Year 2, and a second post-test in May of Year 2. At each wave except Pre
2, three types of surveys were collected: (a) School Climate Survey, (b)
Ambassador/Key Student Survey, and (c) Key Adult Survey. At Pre 2, only the
Ambassador/Key Student surveys were administered.
2.3 Results
The research design includes three nested groups: 1) Ambassadors and Key Students,
2) friends of Ambassadors and Key Students, and 3) the school population at large.
Differences were evaluated in each of these groups across intervention condition.
2.3.1. Ambassadors and Key Students
The Ambassadors from the treatment schools and the Key Students from the control
schools form the smallest group of interest, approximately 80 at each school or 400
students in total. Because the program deals directly with Ambassadors, we expected
to see the largest impact in this group as measured by their survey responses.
Ambassadors and Key Students took two surveys in each data collection wave: the
Ambassador/Key Student survey and the School Climate survey. For each scale
described in Table 2, the difference between Ambassadors and Key Students were
examined using Pre 1, Post 1, Pre 2, and Post 2 data from these surveys. The data
from the Climate survey were analyzed using separate ANCOVA models with an
additional factor Program Participant (0 = not an Ambassador nor Key Student, 1 = first
year Ambassador or Key Student, 2 = second year Ambassador or Key Student) and
one covariate Friends Index (a weighted count of the number of times a student was
mentioned as a friend of an Ambassador/Key Student).
Safe School Ambassadors Program Evaluation 2011
Page 6
2.3.2. Ambassador/Key Student Survey
The data from the Ambassador/Key Student surveys were analyzed using separate
ANOVA models with three fixed factors: SSA (1 = treatment school, 0 = control school),
Gender (1 = male, 2 = female), and Grade (sixth, seventh and eighth). For Pre 1, Post
1, and Pre 2, there were no significant differences between Ambassadors and Key
students with regard to any of the scales listed in Table 2. However, at Post 2, male
Ambassadors reported significantly higher frequency of Active Intervention with Others
(students who were not amongst their closest friends) than male Key Students. On a
scale ranging from 1 (never) to 5 (always; one or more times a day) the average
frequency reported for male Ambassador was 3.18 vs. 2.45 for male Key students (p
=.007). Figure 1 shows the estimated marginal means by gender and intervention
condition. Similarly, male Ambassadors reported higher frequency of Active Intervention
with Friends than male Key Students, but the difference was not statistically significant.
2.3.3. Climate Survey
The data from the Climate survey were analyzed using separate ANCOVA models
adding an additional factor Year of Participation (0 = not an Ambassador nor Key
Student, 1 = first year Ambassador or Key Student, 2 = second year) to the ones used
above, and one covariate Friends Index that represented the student’s social proximity
to Ambassadors/Key Students. The factor Year of Participation in conjunction with SSA
was analyzed to determine what, if any, effect of the program on the Ambassadors
themselves was evident in the climate survey. It was expected to see the difference
between the responses of Ambassadors and Key students change over time. However,
there were no significant differences between Ambassadors and Key students with
regard to the scales listed in Table 2 for the Pre 1, Post 1 or Post 2
2.3.4. Friends of Ambassadors and Key Students
The second group of interest is comprised of approximately 900 Friends identified by
Ambassadors and Key Students. Based on the theoretical design of the Safe School
Ambassadors program, we anticipated that the Friends would be the first students from
the general school population impacted by the program.
The slope of the covariate Friends Index and its interaction with SSA in the ANCOVA
models described above was used to assess this impact. The index gave higher weight
to students who were named as Friends on the post-intervention surveys and to
students who were named as a friend by the same Ambassador/Key Student more than
once. Table 3 describes the Friends Index data. The estimated values for the slopes
across the intervention condition for each of the subscales from the climate survey is
listed in the first two columns of Table 4. The final two columns represent the p-values
for testing the significance of the overall slope and the interaction of the slope with the
intervention indicator SSA.
Examining Table 4, we see that for the School Climate scale, on the Pre 1 assessment,
the slope was negative in the treatment schools and positive in the control schools. This
Safe School Ambassadors Program Evaluation 2011
Page 7
indicates that, at the beginning of the study, the Friends of the Ambassadors reported a
significantly worse climate than their fellow students, while in the control schools the
Friends of the Key Students reported a better climate than their peers. At Post 1 and
Post 2, the slope for both groups was positive, indicating that the climate around the
Ambassadors had improved significantly from its pre-SSA-program condition.
Similarly, on the scale for Observed Active Intervention by peers to prevent/stop
mistreatment, the slope in the Ambassador schools is negative on Pre 1, positive but
not significantly different from 0 on Post 1, and small but significantly greater than 0 on
Post 2. This indicates that, over the two years of the study, as the program was “rolled
out” in treatment schools, the students closer to the Ambassadors reported seeing more
active interventions by peers than did the students who were not close to the
Ambassadors.
At Pre 1, for the scales for Mid-level Mistreatment, Passive Bystander Behavior, and
Active Bystander Behavior, the slope in the treatment schools was positive, indicating
the Friends of the Ambassadors reported higher levels of these negative behaviors than
their peers did. At Post 1 and Post 2, the slopes were not significantly different from 0,
which indicates that situation improved and the Friends of Ambassadors no longer
reported higher levels of these negative behaviors than their peers did.
2.3.5. School population
The largest group is the school population at large (6th and 7th grade students in year
one and 7th and 8th grade students in year two). For each observation on the climate
survey (Pre 1, Post 1 and Post 2), the data were analyzed using the ANCOVA models
described in the previous section. Overall, reported levels of mistreatment were higher
on Post 1 than at Pre 1 in both types of schools (e.g,. for mid-level mistreatment, the
overall mean increased from 2.05 to 2.36).
Students at SSA schools reported significantly higher rates of mid-level mistreatment at
Post 1 (2.52 vs. 2.24, p = .086) and at Post 2 (2.68 vs. 2.07, p =.000), but not at Pre 1
(1.99 vs. 2.12, p = .40). Also, students at SSA schools reported significantly higher
rates of low-level mistreatment at Post 2 than students at control schools reported (2.70
vs. 2.22 on a five point scale, p =.003), but not at Pre 1 (2.15 vs. 2.14, p = .94) nor at
Post 1 (2.47 vs. 2.40, p = .29). No other significant differences between treatment and
control schools were found with regard to the other subscales listed in Table 2.
In our preliminary analysis of the school discipline data, the annual rate of referral to an
Alternative Education setting (i.e. the number of students with at least one referral
divided by the total student population) was examined. Comparing the baseline of the
year prior to intervention (2006 - 2007) to the final year of the program (2008-2009), the
referral rate increased from 7.3% to 8.0% in the treatment schools and from 3.3% to
5.0% in the control schools. This represents a 9.6% increase in the treatment schools
as compared to a 51.5% increase in the control schools. In particular, the referral rate at
one of the three treatment schools decreased from 10.6% to 9.2%.
Safe School Ambassadors Program Evaluation 2011
Page 8
2.4 Discussion
2.4.1. Learnings about Ambassadors
We anticipated that Ambassadors would show the greatest amount of change, but the
survey data did not reveal significant differences between Ambassadors and Key
Students in their attitudes about human relations (e.g., diversity) as measured by the
Attitude scale, or in their school connectedness. However, at Post 2, male
Ambassadors did report significantly higher rates of active intervention than male Key
Students did. Since there was no difference between Ambassadors and Key Students
on this variable at Pre 1, this result indicates that these Ambassadors were
implementing the positive intervening behaviors that were taught in the SSA program.
Observations of the trainers and comments made during the focus groups suggest that
the intervention behaviors taught in the SSA program are more familiar and culturally
acceptable to the females than the males, so the change in the males was more
noticeable and is a result that warrants further investigation.
This finding is reinforced by a similar pattern noted in the Friends data: while Friends of
Ambassadors initially reported witnessing significantly less active intervention than
Friends of Key Students, after the program was implemented, Friends of Ambassadors
reported witnessing significantly more active intervening than their counterparts at
control schools.
This result was further replicated in Key Adult surveys, which underwent analyses
similar to the Ambassador and Climate surveys. The most interesting results are
depicted in Figure 2, which shows that on average, Key Adults in SSA schools reported
an increase in witnessing helpful student behaviors (e.g., speaking up for students who
were being put down) from Pre 1 to Post 1 and Post 2. In control schools, the rates of
these behaviors as observed by key adults decreased slightly over time.
The process data suggests that the surveys may not have detected all the changes that
happened with Ambassadors. One principal relayed an Ambassador’s comment: "I used
to be the one picking on other kids and starting fights, but now I am the one out there
protecting the kids from guys that are like I used to be." Additionally, the Program Adults
reported via Year-End Surveys that Ambassadors had shown improvements in self-
confidence, leadership, communication, empathy, tolerance and willingness to intervene
when they noticed mistreatment (mean = 2.17 on a 3-point scale ranging from 1 = “have
not changed" to 3 = "improved a lot"). These same adults also reported that
Ambassadors had become more willing to report to adults information about potential
problems or dangerous situations (mean = 2.30), an observation corroborated by
principals and other administrators at the treatment schools, one of whom noted, "My
administrative staff and I can confirm the flow of information from Ambassadors to our
offices was ongoing throughout the year."
2.4.2. Learnings about Friends
Additional positive results were also found in the Friends data. Although Friends of
Ambassadors initially reported significantly worse school climate than Friends of the
Safe School Ambassadors Program Evaluation 2011
Page 9
Key Students in control schools, after the implementation of SSA this pattern had
reversed; Friends in intervention schools rated school climate significantly better than
Friends in control schools. Furthermore, at Pre 1, Friends of Ambassadors reported
observing some types of mistreatment (e.g., mid-level mistreatment, passive bystander
behavior, active bystander behavior) more frequently than their counterparts in control
schools, but after the implementation of the SSA program, these differences were no
longer seen (i.e. the environment around the Friends of Ambassadors appears to have
improved). These changes suggest that the Ambassadors are actually using the
intervention skills they have learned through the program, which suggests that the logic
model at the core of the SSA program (socially influential students can use their
influence in positive ways to improve the behavior of peers) is working and deserves
further study.
2.4.3. Learnings about Overall Climate
Although the changes seen amongst the Ambassadors' Friends were promising, similar
results were not found in the school population at large. Overall climate at all schools
appears to have deteriorated over time, and this may be due in part to a feature of the
study design. The Pre-Post 1 group included 6th graders but no 8th graders. The 6th
graders reported less mistreatment than older students: a lot less on the Pre and not as
much less on the Post 1. This change matches the perception by adults that over the
course of their first year in middle school, 6th graders lose their "innocence" and come
to act out more like stereotypical middle school students. When the 6th graders became
7th graders and were measured at Post 2, the difference between grade levels had
disappeared.
There were also no clear trends that distinguished the SSA from the control
schools. Notably, the results of the entire school assessment did not replicate the
positive results from Ambassador, Friends, and Key Adult surveys regarding witnessing
increased rates of intervention at SSA schools. Though the differences reported in the
climate survey at Pre 1 were not significant, in each scale the difference favored the
control school (lower mistreatment, better climate). In some cases these differences
were magnified at Post 1 and Post 2. In particular, students at SSA schools reported
significantly higher rates of mid-level mistreatment at Post 1, and significantly higher
rates of low-level mistreatment at Post 2, than students at control schools reported. The
higher rates of lower-level mistreatment may reflect an increased sophistication in
students' mistreatment as they age: subtly snubbing rather than plainly pushing.
Regardless, this overall deterioration indicates that while teachers and friends of
Ambassadors noticed differences, the changes were not so great as to produce
measurable differences at the school-wide level. We anticipate that school-wide effects
for the SSA program would take longer to take hold, given the “ripple effect” logic model
of this program.
Safe School Ambassadors Program Evaluation 2011
Page 10
2.5 Challenges and Limitations
The nature of the preceding results appears to be due, at least in part, to some of the
complexities and challenges of conducting this type of research in schools. The
research design included many practices designed to increase the rigor of this
evaluation. However, some of these safeguards were implemented in a way that
reduced their intended benefits. For example, the original design of the study called for
one-to-one matching of Ambassadors and Key Students. In order to preserve
confidentiality, students were asked to write only their date of birth and the first three
letters of their last name on the survey booklet; however, a sizable minority of entries
were illegible or did not correspond to earlier data.
Treatment and control schools also appeared to differ initially on potentially important
characteristics. Although schools within the same school district were matched on
commonly used factors, this method of matching did not completely account for school
climate. In particular, discipline records, climate surveys, key adult surveys and school
walk-throughs by project staff indicated that control schools started off with more a
positive school climate than their matched treatment schools. Although we are able to
statistically adjust for initial differences on measured variables, these schools may be
different in other unmeasured ways, which may lead to skewed overall comparisons.
The withdrawal of a third control school additionally complicated comparisons.
This study also highlights the difficulty in quantitatively measuring mistreatment in
middle school students. First, when asking students to remember what happened to
them within a particular time frame (e.g., past two weeks), their recollection may be
inaccurate. Additionally, students may not know what to count as mistreatment, or may
need guidance regarding examples of mistreatment to assist in their recollection. We
also struggled with interpreting the result that some Ambassadors reported witnessing
more mistreatment after the intervention took place. It seems impossible to untangle
whether incidents of mistreatment actually increased, or whether Ambassadors were
simply noticing more mistreatment than before given that the intervention is designed to
sensitize Ambassadors to this behavior.
This project also underscored the complexity of collecting records from schools and the
importance of collecting information from multiple perspectives within the school. For
example, during the course of the project, district policies and school practices
regarding discipline changed, which influenced discipline statistics. Our conversations
with school administrators helped us understand how this change was implemented
within the school so discipline data collected for the evaluation could be interpreted
appropriately.
Furthermore, at some schools, staff were assigned responsibility for SSA activities
without additional time or resources to carry out these tasks adequately. Therefore the
quality of program implementation, even within a school, may have varied, and
impacted Ambassadors.
Safe School Ambassadors Program Evaluation 2011
Page 11
2.6 Implications for Future Research
One exciting line of inquiry for future research involves the linking of Ambassadors and
their friends. Using social networking software, it may be possible and informative to
examine the differential impact of individual Ambassadors (or other peer leaders) on
their social networks across time, and correlate that impact to predictor variables.
3. Part II
3.1 Methods
Since it was impossible to predict in advance the fidelity of program implementation at
the three intervention schools discussed above, Part II of this evaluation was designed
to assess program impact under conditions of high fidelity to internal program
benchmarks. The primary research questions focused on the presence of changes in
discipline indicators over time, and whether those were correlated to fidelity of program
implementation.
To identify high-fidelity implementations, the pool of more than 600 program launch
sites was scanned for schools that were in at least their third year of implementation
(suggesting sustained commitment and successful integration into school policy and
practice), had at least 2 documented trainings conducted by certified trainers
(suggesting sufficient numbers of students and adults with adequate knowledge and
skills), and had completed Year-End Surveys at least once (preliminary indicator of
fidelity). The 59 schools identified were surveyed to gather information about their
implementation; 31 schools responded and were scored according to the Program
Implementation Index developed from program benchmarks and described in Table 5.
Of these schools, 19 scored above the median (50) and indicated ability to provide data.
This group included elementary, middle and high schools in urban, suburban, and rural
areas of New York, Texas, Colorado and California. For each of these schools, potential
control (non-SSA) schools were identified from the applicable state Education Agency’s
list of similar schools based on state testing. A matching control school for each
treatment school was then chosen based on the following criteria: same grade levels,
overall school size (student population), racial/ethnic composition, percent of students
eligible for free/reduced price meals (a commonly used measure of socioeconomic
factors), and geographic proximity (with preference for schools in the same school
district, county, or region). Discipline and other data was then collected from the 19
treatment schools and 18 control schools, for a total N = 37 schools. For analyses done
by school year, the treatment schools represented N = 84 years of implementation of
the SSA program (average of 4.4 years at each school).
3.2 Data Sources
Implementation Data was collected using a 10-item survey developed from program
benchmarks. Items included total number of active Ambassadors, frequency of Family
Group Meetings and attendance of Ambassadors (over two years), frequency of
planning meetings and attendance of adults (over two years), degree of principal’s
support for the program, nature and frequency of communication with staff about the
Safe School Ambassadors Program Evaluation 2011
Page 12
program (as an indicator of staff support), and the methods used by Ambassadors to
document their interventions (as an indicator of Ambassador accountability and follow-
through).
The Impact Survey included sections on demographic, enrollment, attendance and
discipline data (e.g. office referrals, detentions, and suspensions), along with other
information about the school’s bullying prevention and climate improvement efforts.
Knowing that many factors influence discipline data, respondents at the treatment
schools were also asked about their perceived impact of the SSA program itself. When
schools were non-responsive, impact data about each matched pair was obtained via
state databases.
3.3 Results
To determine whether there are significant differences in school discipline before and
after implementation of the SSA program, discipline indicators from 84 different years of
SSA program implementation at the 19 treatment schools were analyzed. A paired t-test
was performed on a variable representing the average of pre-post change in the
school’s rate of suspension or other discipline indicator(s) during all years of SSA
program implementation, normalized by school population. Pre-SSA rates ranged from
0.076 to 3.3 incidents per student per year, and analysis showed an average change of
-0.203 after SSA program implementation, a statistically significant result (p = 0.045).
Office referral and detention data were then omitted and suspension rates were
analyzed at both treatment and control schools independently, using a paired t-test. The
data is described in Table 6 and Table 7. Across all years of SSA program
implementation, the mean change in suspension rate in treatment schools is -0.135 (an
average decrease of 33.1%), which is statistically significant (p < 0.0001). The same
analysis was performed for control schools, and the mean change in suspension rate for
control schools is 0.004 (an average increase of 10.4%), which is not statistically
significant. As shown in Figure 3, the treatment schools had higher baseline suspension
rates than the controls, and those rates decreased significantly after high-quality
implementation of the SSA program, while suspension rates at the controls remained
relatively unchanged during the same years.
The change in suspension rate from the baseline was calculated for treatment and
control schools for each year of implementation. An independent t-test showed that the
mean difference between treatment and control schools is -0.140, indicating that the
treatment schools have lower suspension rates than the matched set of control schools
after implementation of the SSA program. The result is statistically significant (p
<0.0001).
The relationship between fidelity of program implementation and change in discipline
incidents was explored using correlational analysis of a school’s Implementation Index
score (higher score = higher fidelity implementation) and the average change in
discipline indicator (the larger the number, the greater the increase in discipline
incidents). A one-directional correlation test for these 19 schools found a Pearson
Safe School Ambassadors Program Evaluation 2011
Page 13
Correlation of -0.318 (p = 0.092), suggesting a medium correlation between fidelity of
implementation of the SSA program and a decrease in discipline incidents.
3.4 Discussion
We hypothesized that there would be decreases in the rates of suspensions and other
types of discipline at the treatment schools after implementation of the SSA program.
We also anticipated that the SSA schools would over time have lower suspension rates
than the control group. The analysis confirmed both of these hypotheses.
The challenge in any program evaluation is attributing effect to the range of potential
contributing factors, including the program being studied. The Impact Survey used in
this evaluation included questions about the impact of the Safe School Ambassadors
program on school discipline, school climate, staff morale, finances and
learning/achievement; for example: What if any impact has the Safe School
Ambassadors program had on school discipline data (e.g. office referrals, detentions,
suspensions)? The 5-item response scale ranged from 1 = Very Negative to 5 = Very
Positive. As shown in Table 8, the 14 respondents (principals, deans, counselors,
teachers) from treatment schools reported an average impact of 4.26, indicating a
relatively strong perception that the SSA program has had a positive impact on these
aspects of their schools.
The impact survey also asked respondents to identify and rate the effectiveness of any
(non-SSA) curricula, extra-curricular programs or clubs, one-time events or activities,
rules and policies, and any other initiatives designed to reduce bullying and help
students get along. Respondents listed a range of 0-7 such items, with the average
number being 3.14. The average reported effectiveness of these other programs and
activities was 3.94, slightly lower than that reported for the SSA program. Further
analysis of these potentially confounding variables is beyond the scope of this
evaluation.
3.5 Implications for Future Research
An intriguing avenue to explore in future research would be an analysis of discipline
data that is more sensitive to Ambassadors’ influence (e.g. office referrals rather then
overall suspensions). It would also be helpful to study school climate at matched pairs
of treatment and control schools over time, using regular & standardized instruments
like the California Healthy Kids Survey, and to correlate that data to annual SSA
Implementation Index numbers as well as annual information about other (non-SSA)
programs and activities.
4. Conclusion
The underlying model of the Safe School Ambassadors program is consistent with the
findings of DeFur and Korinek (2010) and others before them – that students are a
valuable resource for resolving seemingly intractable school problems. The SSA
program is a departure from the more traditional outside-in, adult-driven approaches
that are based on rules and consequences, and from approaches that focus on the
students directly involved in the bully-victim dyad. It represents an innovative approach
Safe School Ambassadors Program Evaluation 2011
Page 14
more consistent with the ecological perspective on bullying discussed by Espelage
(2004) who noted that effective bullying interventions need to change the conditions in
the social environment that permit bullying to occur. Specifically, the Safe School
Ambassadors program seeks to improve that environment by harnessing the social
power of certain students to influence their peers to stop or refrain from engaging in
hurtful or violent behaviors, thereby over time shifting the underlying social norms that
govern those behaviors.
As noted by Smith et al. (2004), few studies evaluating whole school anti-bullying
programs have yielded unequivocal and sustained positive outcomes. This result is
unsurprising given the complexity of implementing, monitoring and measuring such
programs. However, several promising results from this evaluation of the Safe School
Ambassadors program indicate that it increases the frequency of student intervention to
stop mistreatment, which is corroborated by Ambassadors' reports of interventions. The
statistically significant reductions in discipline indicators seen at schools where the
program was implemented with high fidelity, augmented by school administrators’
reports, indicate that the program can have a significant positive impact on discipline
incidents and overall school climate. Taken together, these findings suggest that an
inside-out, student-centered approach may be an effective and more inclusive method
for reducing bullying/mistreatment and enhancing school climate.
Safe School Ambassadors Program Evaluation 2011
Page 15
5. References
Cairns, R. B., Leung, M. C., & Cairns, B. D. (1995). Social networks over time and
space in adolescence. In L. J. Crocket & A. C. Crouter (Eds.), Pathways through
adolescence: Individual development in relation to social contexts. The Penn
State
series on child and adolescent development (pp. 35-36). Hillsdale, NJ: Lawrence
Erlbaum.
DeFur, S. H. and Korinek, L. (2010) “Listening to Student Voices”, The Clearing House:
A Journal of Educational Strategies, Issues and Ideas, 83: 1, 15-19.
Espelage, D. L. (2004). An Ecological Perspective to School-Based Bullying Prevention.
The Prevention Researcher, 11(3), 3-6.
Hazler, R. J. (1996). “Bystanders: An Overlooked Factor in Peer on Peer Abuse,”
Journal for the Professional Counselor, 11(2), 11-22. Craig, W. M., and Pepler, D. J.,
(1995). Peer Processes used in Bullying and Victimization: an Observational Study,”
Exceptionality Education Canada, 5, 81-95.
Knickerbocker, B., “Five Years After Columbine, the Insecurity Lingers,” Christian
Science Monitor, April 20, 2004, sec 4.
Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W. J., Simons-Morton, B., & Scheidt, P.
(2001). Bullying behaviors among U.S. youth—Prevalence and association with
psychosocial adjustment. Journal of the American Medical Association, 285, 2094-2100.
National Crime Prevention Council, “Teens and Cyberbullying: Executive Summary of a
Report on Research,” February 28, 2007. P.1.
Packaged Facts, The U.S. School Security Market (Rockville, MD: 2000), 2.
Salmivalli, C., Lagerspetz, C., Bjorkqvist, K., Osterman, K. & Kaukianinen, A. (1996).
Bullying as a group process: participant roles and their relations to social status within
the group. Aggressive Behavior, 22, 1-15.
Slaby, R. G. (2004). Knowing the Facts: The Role of the Bystander in Preventing
Bullying. Health in Action, 3: 4.
Smith, J. D., Schneider, B. H., Smith, P. K., & Anaiadaou, K. (2004). The effectiveness
of whole-school antibullying programs: A synthesis of evaluation research. School
Psychology Review, 4, 547-560.
U.S. Department of Justice, Office of Community Oriented Policing Services. (2003).
COPS Office Announces $4.9 Million to Increase School Safety in 27 States [press
release]. Retrieved from http://www.cops.usdoj.gov/default.asp?Item=958
Vossekuil, B. and others, (2004). Final Report and Findings of the Safe School Initiative:
Implications for the Prevention of School Attacks in the United States. (Washington,
D.C.: U.S. Department of Education, Office of Elementary and Secondary Education,
Safe and Drug-Free Schools Program and U.S. Secret Service, National Threat
Assessment Center), 25-26.
Safe School Ambassadors Program Evaluation 2011
Page 16
6. Tables and Figures
Table 1: Summary of participants and roles in the SSA program and evaluation
Program Evaluation
SSA Role Definition
Treatment
(SSA school)
Control
(Non-SSA
School)
Instrumentation and
Data Collection
Students
Ambassador
Students trained to stop
mistreatment
Ambassador Key Student
Ambassador or Key
Student Survey.
Climate Surveys
flagged for analysis
as a group
N/A
Students named by
Ambassadors or Key
Students (via their respective
surveys) as being one of the
10 people they “hang out”
with the most.
Friends Friends
Climate Surveys
flagged for analysis
as a group
N/A All Students* All Students* Climate Surveys
Adults
Family Group
Facilitator
Adults trained to provide
support, supervision, and skill
development to
Ambassadors via small-group
meetings throughout the
year.
Family Group
Facilitator
N/A
Year-End Survey for
FGFs
Program
Advisor
Adult who provides overall
leadership, supervision, and
planning.
Program
Advisor
N/A
Year-End Survey for
PAs
N/A
Cross-section of adults who
have insights or knowledge of
student climate b/c of their
position (e.g. Assistant
Principal in charge of
Discipline, Campus
Supervisor or Yard Duty
personnel) OR because of
their relationship / personality
(i.e. students are naturally
drawn to this individual).
Key Adults Key Adults Key Adult Survey
* In year 1, the Climate Survey was completed by all students in grades 6-7, plus any 8th graders who
were named by Ambassadors or Key Students as "Friends". In year 2, all students in grades 7-8
completed it.
Safe School Ambassadors Program Evaluation 2011
Page 17
Table 2: Summary of scales/items included in student self-report surveys
Name of scale Description of scale Included in surveys
School Climate
11-item scale measuring caring social atmosphere at
school. Items are Likert-type with 4 response categories.
Higher values indicate better school climate. α = .776.
School Climate
Connectedness
Multi-item measure of Blum's concept of connectedness.
Items are ordered response type with 5 response
categories. Higher values indicate more connection to
school. Climate: 5 item scale with α=.834, Amb/Key: 4-
item scale with α = .646.
Ambassador/Key Student
School Climate
Low Level Mistreatment
Observed
Multi-item scale measuring frequency of observing low-
level forms of mistreatment such as leaving someone out
of a group. Items are ordered response type with 5
response categories. Higher values indicate more
mistreatment observed. Climate: 5 item scale with α = .774,
Amb/Key: 2-item scale with α = .734.
Ambassador/Key Student
School Climate
Mid Level Mistreatment
Observed
Multi-item scale measuring frequency of observing mid-
level forms of mistreatment such as hitting or pushing.
Items are ordered response type with 5 response
categories. Higher values indicate more mistreatment
observed. Climate: 3 item scale with α = .805, Amb/Key: 2-
item scale with α = .849 .
Ambassador/Key Student
School Climate
High Level Mistreatment
Observed
1 item measuring frequency of observing weapon-carrying
at school. Item is ordered response type with 5 response
categories. Higher values indicate more observed
weapon-carrying.
School Climate
Passive Bystander
Behavior Observed
3-item scale measuring frequency of observing peers being
passive bystanders and not intervening to stop
mistreatment, such as by watching a fight. Items are
ordered response type with 5 response categories. Higher
values indicate more passive bystander behavior observed.
α = .785.
School Climate
Active Bystander
Behavior Observed
2-item scale measuring frequency of observing peers being
active bystanders and encouraging mistreatment, such as
by encouraging a fight. Items are ordered response type
with 5 response categories. Higher values indicate more
active bystander behavior. α = .654.
School Climate
Active Intervention
Observed
6-item scale measuring frequency of observing another
student actively intervening to stop mistreatment. Items
are ordered response type with 5 response categories.
Higher values indicate more active intervention observed.
α = .899.
School Climate
Attitude
7-item scale measuring the self reported attitude toward
diversity and mistreatment. Items are ordered response
type with 5 response categories. Higher values indicate a
more positive and tolerant attitude. α = .611.
Ambassador/Key Student
Active Intervention with
Friends
7-item scale measuring the self reported frequency of
actively intervening to stop mistreatment among friends.
Items are ordered response type with 5 response
categories. Higher values indicate more active intervention
reported. α = .793.
Ambassador/Key Student
Active Intervention with
Others
7-item scale measuring the self reported frequency of
actively intervening to stop mistreatment among students
who are not friends. Items are ordered response type with
5 response categories. Higher values indicate more active
intervention reported. α = .871.
Ambassador/Key Student
Safe School Ambassadors Program Evaluation 2011
Page 18
Table 3: Summary of Friends Index
School Type
Number
Mean
St. Dev.
Maximum
SSA
204
1.61
1.24
7
Control
206
1.92
1.56
11.5
A student’s Friends Index = .5*Number of times s/he was named by an Ambassador or Key Student as
one of her/his friends on Pre 1 + Number of mentions on Post 1 and Post 2 + number of times mentioned
by same Ambassador on two surveys + number of times mentioned by same Ambassador on three
surveys
Table 4: Slope Coefficients and p-values from the Friends analysis
Scale
Time of
Measurement
Slope
SSA
Slope
Control
p-values
Friends Index
p-values
Friends Index*SSA
School Climate Pre1 -0.02 0.02 0.96
0.05
Post1 0.02 0.01 0.29 0.75
Post2 0.05 0.04
0.00
0.60
Connectedness Pre1 0.04 0.07
0.01
0.55
Post1 0.11 0.04
0.00
0.27
Post2 0.09 0.09
0.00
0.95
Low Level Mistreatment Pre1 0.03 0.03 0.13 1.00
(e.g. exclusion, insult) Post1 0.06 0.02 0.09 0.36
Post2 0.02 -0.04 0.74 0.12
Mid Level Mistreatment Pre1 0.07 0.01
0.04
0.09
(e.g. threaten, push, fight) Post1 0.03 -0.01 0.62 0.35
Post2 0.00 -0.03 0.42 0.55
High Level Mistreatment Pre1 -0.01 0.01 0.44 0.89
(e.g. weapon) Post1 -0.07 -0.02
0.02
0.27
Post2 0.01 0.00 0.76 0.80
Passive Bystander Behavior Pre1 0.09 0.01
0.02
0.03
(e.g. watch argument / fight) Post1 0.00 -0.01 0.77 0.88
Post2 0.04 -0.05 0.94
0.04
Active Bystander Behavior Pre1 0.07 0.02
0.02
0.14
(e.g. encourage argument / Post1 -0.02 -0.01 0.64 0.82
fight) Post2 0.03 -0.02 0.77 0.33
Observing Active Intervention Pre1 -0.04 0.06 0.63
0.03
(e.g. saw students speak up Post1 0.02 0.06 0.11 0.50
to prevent/stop mistreatment) Post2 0.06 0.04
0.03
0.57
Safe School Ambassadors Program Evaluation 2011
Page 19
Table 5: SSA Program Implementation Index
Component and Definition
Maximum Points
Ambassador Saturation: Ratio of Ambassadors to overall school population 25
Family Group Meetings: Regular meetings of 7-10 Ambassadors with 1-2
Program Adults
25
Program Management Team (adults) Meetings: Regular meetings or
communication to assess and plan
15
Principal Supports Program 10
Staff Communication: written and face-to-face communication about SSA program
activities, concerns, results, etc.
10
Ambassador Intervention: which methods are used by Ambassadors document
their interventions
15
Total possible points
100
Table 6: Descriptive Statistics on Post-SSA Suspension Rate
Group N Mean SD Min Max
Control 56
0.210
0.176
0.000
0.743
Treatment 66
0.175
0.187
0.019
0.816
Table 7: Descriptive Statistics on the Difference in Suspension Rate
Group N Mean SD Min Max
Control 56
0.004 0.078
-0.182 0.240
Treatment 66
-0.135 0.209
-0.895 0.068
This variable was calculated for each year of implementation, by subtracting the baseline (pre-SSA)
suspension rate from the suspension rate for that year.
Table 8: Administrators’ Reports of SSA Program Impact
Indicator
Average Score
a. School discipline data (Ex: office referrals, detentions, suspensions) 4.31
b. Overall social-emotional climate (Ex: feeling in halls, lunch, other common
areas; tension between cliques)
4.54
c. Staff Morale (Ex: fewer classroom discipline incidents allows teachers to focus
on teaching, teacher retention)
4.08
d. School Budget / Finances (Ex: costs for vandalism, suspension processing) 4.23
e. Learning and achievement (Ex: grades, test scores, student interest in learning) 4.15
Average Impact
4.26
Scale: 1 = Very Negative Impact, 3 = No Impact, 5 = Very Positive Impact
Safe School Ambassadors Program Evaluation 2011
Page 20
Figure 1: Estimated Marginal Means of Active Intervention with Non-Friend
Safe School Ambassadors Program Evaluation 2011
Page 21
Figure 2: Key Adults’ Observations of Active Interventions to Stop Mistreatment.
Key Adults were asked about how often in the previous month they
noticed students intervening to prevent or stop mistreatment. Y-axis
is frequency of response; scale ranges from 1 = never to 5 =
always.
Figure 3: Change in Suspension Rate after SSA Program Implementation.
Comparison of Change in Mean Suspension Rate
0
0.05
0.1
0.15
0.2
0.25
0.3
Pre SSA Post SSA
Mean Suspension Rate
control
treatmen
t
The decrease of 0.1174 at treatment schools was significant
(p<0.0001).
1.8
1.9
2
2.1
2.2
2.3
2.4
Pre Post Post2
Control
Treatment
Safe School Ambassadors Program Evaluation 2011
Page 22
7. Appendix I: The Logic Model of the Safe School Ambassadors
program.
Literature Review on Bullying and Its Prevention:
Implications for the Safe Schools Ambassadors® Program
Prepared by OMNI Research and Training, Inc for Community-Matters
November 10, 2004
This document serves to assess the literature in relation to the core components of the Safe
Schools Ambassador® (SSA) program. The SSA program is a variant of a student watch
program (Ross, 1996) that recruits and trains socially influential youth from the different social
groups that exist in schools to note and report instances of bullying and other antisocial behavior
(Community Matters, 2003). However, SSA is considered a second generation student watch
program because a much greater emphasis is placed on training youth bystanders on methods to
intervene as warranted by the situation rather than simply noticing and reporting bullying to
school staff. The overall goal of the SSA program is to improve the school climate by
empowering students who are neither bullies nor victims of bullies, but have a good probability
of being a bystander, to play a clear role in preventing episodes of bullying and related activities.
The aims of this preliminary review of the literature are to:
• Provide a brief overview of the bullying problem,
• Establish the importance of utilizing bystanders in the prevention of bullying,
• Discuss core factors in effective bystander interventions, and
• Assess the literature for outcomes of bullying prevention programs and bystander
interventions.
General Overview of the Bullying Problem
Violence as well as the fear of violence is prevalent in U.S. society. Although overall rates of
violent crime have declined from their peak in the 1980s, intentional violence still accounts for
one-third of all injury deaths (Hamburg, 1998), and violent injury and death continues to
disproportionately affect children, adolescents and young adults (US Department of Health and
Human Services, 2001). Over the last several decades, violence has come to be defined as a
public health issue and greater public attention has been paid to its prevention, not simply its
deterrence and control (Elliot, 1998).
During the 1980s and increasingly throughout the 1990s, major national initiatives to address
youth violence were mounted, including the passage of the Safe and Drug-Free Schools and
Communities Act of 1994. Violence also became an established field of research study. Over
Safe School Ambassadors Program Evaluation 2011
Page 23
time, these prevention and research efforts helped define bullying prevention as an important
area of violence prevention by drawing attention to the:
• Cycle of violence in which victimization begets future perpetration
• Violence prevention needs in the school environment
• Establishment of a bully prevention program as one of the ten Blueprints for Violence
Prevention (Elliot, 2000).
Today, bullying is considered to be one of the most common and pervasive forms of school
violence (Swearer & Doll, 2001). Much of what has been learned to date comes from studies
conducted in Scandinavian countries, Great Britain, Ireland, Spain, Australia, the Netherlands,
and Japan (Banks, 1997; Olweus, 1993). An important conclusion drawn from this research is
the universality of bullying behaviors (Macklem, 2003). Bullying is a problem in all schools and
countries around the world.
Throughout the 1990s, a number of studies were conducted to understand the prevalence and
incidence of bullying. The American Academy of Child and Adolescent Psychiatry (2004)
reported that approximately 50% of all children had been bullied while in school and that 10% of
those that had been bullied were victimized on a regular basis. A subsequent study, representing
the first nationwide research on bullying in the U.S., was conducted by the National Institute of
Child Health and Development. This study found that in grades sixth through tenth, 16% of
students indicated that others had bullied them during the current term. Almost 30% reported
that they had been involved as a bully, a victim, or both (Nansel, Overpeck, Pilla, Ruan, Simons-
Morton, and Scheidt, 2001). Ericson (2001) used these data to estimate that 1.6 million students
in these grades were bullied at least once a week. Moreover, based on the original study, the
National Institute of Child Health and Development found that bullying behaviors were a marker
for involvement in future violence-related behaviors. In fact, it reported that children and youth
that engaged in bullying behaviors were the group most at risk for engaging in violent behaviors
over time (Nansel, T.R., Overpeck, M.,D., Haynie, D.L., Ruan, W.J., & Scheidt , 2003).
Importance of Utilizing Bystanders for Bullying Prevention
Peers are present in 85% of bullying episodes in school settings (Bonds & Stoker, 2003, Craig &
Pepler, 1995) and up to 92% of elementary school students have observed instances of bullying
in their schools (Henderson & Hymel, 2002, as cited in Macklem, 2003). Because bystanders
are almost always present whereas adults rarely witness bullying, their participation in school-
based bullying programs is considered instrumental. Researchers using a variety of data
collection methods, including unobtrusive observation, describe various roles that bystanders
may take in a bullying episode (Craig & Pepler, 2000; Cowie & Wallace, 2000; O’Connell,
Pepler, & Craig 1999; Salmivalli, 1999). These include:
• Becoming an assistant of the bully,
• Defending the victim, or
• Remaining an outsider.
Safe School Ambassadors Program Evaluation 2011
Page 24
Bystanders can reinforce the bully by laughing or cheering on behavior. They can also pretend
not to notice what is going on but often are described by the victim as cooperating with the bully
because of their nonverbal behavior (Macklem, 2003).
While over half of surveyed children reported that they would intervene in a hypothetical
bullying situation, their reports do not match with observed playground behavior (O’Connell,
Pepler, & Craig, 1999). Actual intervention occurred by bystanders on average in about 11% of
the bullying episodes (Craig & Pepler, 1997). In a study conducted by O’Connell, Pepler, and
Craig (1999) 54% of the time children (ages 5-12) passively watched a bullying dynamic, 21%
of the time joined in on bullying, and 25% of the time intervened on behalf of the victim. Older
children are less likely to report intervening and more likely to report encouraging a bully than
younger children (Jeffrey, Miller, & Linn, 2001; Rigby & Slee, 1992).
According to Lazarus (2001), bystanders may not report abuse because they:
• Do not recognize behavior as bullying
• Fear getting a friend in trouble
• Fear alienation
• Fear retaliation
• Believe that adults will not help
Moreover, Khosropour and Walsh (2000) found that bystanders prioritized the intentions of the
bully over the victim’s feelings even after a bully prevention program that taught the opposing
lesson. Over 50% of sixth and seventh graders felt that the victim could control the reason for
being mistreated and many bystanders believed that the victim would learn something from the
encounter. In addition to blaming the victim, Sutton and Keogh (2000) showed that students
who seek to be accepted by a social group that supports bullying are less willing to become
involved or take action.
Prevention of bullying is further undermined by the fact that there is a strong taboo among
students against informing school staff when episodes of bullying occur (Crary, 2001 as cited in
Macklem 2003). Many students report concerns that they would not be protected in
unsupervised areas of the school (Suderman, Jaffe, & Schiek, 1996 as cited in Macklem 2003).
Some bystanders also feel relieved when they are not a target and often distance themselves from
the victim, reducing empathy and making it easier to walk away from the situation.
As bullying-prevention in schools begin to place greater emphasis on encouraging bystanders
during episodes of bullying to intervene rather than simply report, empowering students to act
becomes an important focal point. In addition to addressing the barriers described above, giving
youth the knowledge and skills for effective behaviors to engage in as well as providing a
network of support from peers and school staff is critical (Garrity, Jens, Porter, Sager, Short-
Camilli, 1994).
Core Factors in Effective Bystander Interventions
Group Processes that Facilitate Bullying
Safe School Ambassadors Program Evaluation 2011
Page 25
Children are highly motivated by the need to belong. Social groups start forming as early as the
preschool years. During this period, children may be excluded from groups because often the
inclusion of a new member causes deterioration in play. For example, two children who are
playing a make-believe setting and are interrupted by a third ‘outsider’ may lose focus of what
they were doing and that particular play activity may come to an end. According to Macklem
(2003), children learn early on that they can exclude others and use the concept of ‘friend’ to
control play activity.
By elementary school children are still largely motivated by wanting to belong, and many peer
groups become established by third grade. Children are now forced to find their place using
social power. Social groups have established rules of how children are supposed to act to be
accepted by a particular group. This includes how aggression is expressed, which accounts for
the finding that aggressive children hang out and play with other aggressive children (Cairns &
Cairns, 1991). These norms are often established by the most popular child in the group; the
child with the most social power. However, being popular is not the same as having many
friends. While having the ability to establish and maintain friendships is associated with social
competence and prosocial behaviors, popularity tends to be associated with dominance,
aggression, and power (LaFontana & Cillessen, 2002). By fifth grade, both boys and girls will
choose popularity over friendship and will reject a ‘friend’ to be accepted by a group (Macklem,
2003).
Because social groups have a hierarchy, the child with the most social power has tremendous
influence on the behavior of the other members and determines the status of others in the group.
Macklem (2003) describes these group dynamics as follows:
“The leader or leaders of cliques use their popularity to control the group. Clique leaders
may manipulate and rearrange the hierarchy by extending favor on a particular child who
is lower in the hierarchy. Leaders hang on to their power by manipulating the feelings
and status of other children. Leaders also manipulate clique members into bullying
others, and then withdraw themselves so that others take the blame” (pp. 110-111).
Macklem (2003) further describes that by fourth or fifth grade, students in schools with greater
than 80 individuals per grade are often split into four groups. The popular group makes up
approximately one third of all students in any given grade level. This group drives the school
climate, and is motivated to protect their power. There may be a select group of leaders within
the popular circle that determine the ever-changing status of the followers. The next group, the
‘wannabes’, makes up about 10% of students in a grade level. These are students who have their
own small group of friends, but would prefer to be in the popular group. The third group
includes about 50% of the students of a class that is made up of a middle group of students who
mostly operate on their own. These students are generally accepting of others, but do not wish to
be a part of the popular clique, and are also critical of the wannabes. The remaining 10% of the
class is made up of isolated children. These children lack support, and even tend to reject one
another.
Based on research conducted by Adler and Adler (1998), by late elementary school, students
have a strong sense of:
• Social status based on power and popularity,
Safe School Ambassadors Program Evaluation 2011
Page 26
• Pressures to conform to group norms, and
• Consequences that result from going against group norms.
During middle school, the social hierarchy is still rooted in popularity. Although popular peers
may not necessarily be well-liked, their influence on members within the group remains. They
effectively use both positive and negative behaviors to achieve their desired goals and often are
controlling and manipulate others. According to Macklem (2003):
“Social groups among adolescents have gate-keeping regulations which assure that new
members are selected because they are similar to others who already belong to the group.
The group helps the individual student establish identity and helps meet the individual’s
need to belong, but at the same time the group controls the student’s behavior,
particularly the expression of aggression… Students move up and down in status within
the larger group, but do not change from one large group to another. Once a new
member of the group is allowed to participate, further socialization takes place. Students
learn the values of the group and take on the social problems of the group.” (p. 113).
Since peers reinforce the harmful group dynamic of antisocial perpetration, intervention must
incorporate the youth network (Macklem, 2003). For this reason, effective bystander
intervention programs must focus recruitment strategies to select influential members that cut
across the social groups and cliques that exist is school settings. However, the difference
between ‘tattle telling’ and prosocial behavior must be taught and reinforced by peer networks.
Since social pressures dictate a tacit response, student participants need regular and ongoing
support.
Involvement of Influential Youth
Social groups in middle and high schools commonly referred to as cliques, have sophisticated
hierarchies. According to Greener (2000), popular children are more prosocial than other
children. Leader(s) of cliques use their popularity, often maintained by prosocial behavior, to
control the social group. Popular students may also manipulate others to bully particular
students. The dynamics and hierarchies of these cliques are fluid; social status can be
manipulated by leaders who work hard to maintain their heightened social status. Students who
are eager for acceptance do not stand up to bullies. If a group does not support intervening in a
bullying situation, these students hesitate to aid bully victims (Sutton & Keogh, 2000).
According to a study by Ginsburg and Miller (1981), the small number of boys who intervened
in playground fights were children who held high social status among their peers. Similarly,
studies have found that group leaders or children well-liked by their classmates were more likely
to self-report intervening in a potential bullying episode (Salmivalli, Lagerspetz, Bjorkqvist,
Osterman, & Kaukiannien, 1996). Research on modeling and social influence shows that greater
impact is achieved by individuals that are well-liked and who have high social status (Bandura,
1977; Petty & Cacioppo, 1981). These studies suggest that children with social influence are
more likely to intervene on behalf of a victim. These influential children, already more prosocial
than other children, are an untapped resource for modeling appropriate bystander intervention.
These leaders may influence their peers who are seeking acceptance to stand up bullies or
include marginalized youth. Therefore training students who have been identified as leaders to
intervene may be the most efficient and successful means to increase bystander intervention.
Safe School Ambassadors Program Evaluation 2011
Page 27
Because modeling prosocial behavior is a major function of bystander intervention programs, the
characteristics of individual members and the groups that they represent are quite important.
The relationship between prosocial youth and popularity may also encourage students to become
involved in programs such as the SSA programs. In other words, youth may be motivated to
enhance their prosocial skills and help bullying victims if they felt these skills might maintain or
enhance their popularity.
Empathy-building
To mobilize bystanders to intervene on behalf of a victim, youth must feel empathy toward
victims and guilt for not intervening (Eisenberg & Miller, 1987). Can prosocial behavior be
trained? Hoffman (1984) reports that there is modest evidence that practice in role taking in a
noncompetitive context can teach prosocial behavior. Similar studies found that empathy trained
children displayed more prosocial behaviors than that of the control group (Feshbach, 1982).
However, the literature concludes that children respond more empathetically to children of the
same race (Hoffman 1982, 1984). Moreover, empathy is mediated by gender. Boys are more
likely to feel sympathy for females than males (Pellegrini & Long, 2002).
Empathy is undermined by strong social and cultural norms concerning the attribution of blame
to victims in social situations (Brigham, 1991). If a bystander believes the victim caused the
perpetration, then empathetic discomfort is reduced and often replaced by anger. Victims with
bad reputations (immoral, deceitful, etc.) are often seen as deserving maltreatment. The most
compassionate student may not feel empathy, and therefore intervene, if the victim can be
blamed.
While students may benefit from empathy training, they also need to be taught to recognize
bullying, feel concerned about this behavior, and avoid blaming the victim. According to a U.K.
study, 50% of students reported sympathy for victims, 25% were neutral, and 25% reported no
sympathy (Macklem, 2003; Smith & Sharp, 1994). Jeffrey, Miller, and Linn (2001) reported that
9% of students did not care about the victim and 36% did not care about bullying episodes.
Apathy appears to increase with age, especially among boys, as myths about victims are
perpetuated and maintained. By grade eight, two times as many boys as girls reported victims
deserved being bullied. Over 50% of sixth and seventh graders felt that the victim could control
the reason for being mistreated. Bystanders who blame the victim can also feel anger toward the
victim.
School-wide Intervention
In addition to educating students, teachers need to be aware of their own misconceptions and
reinforcing behaviors. One study of elementary school students found that adults ignored 71% of
bullying episodes (Froschl & Gropper 1999). Ignoring adults believed that children should learn
to handle their problems when in reality bullying involves an imbalance and abuse of power that
cannot be solved by the victim alone (Garrity et al., 1994). When asked why victims and
bystanders do not report the maltreatment, victims reported fearing that school personnel will not
handle the event appropriately (Macklem, 2003) while bystanders believed that the adults would
not help (Lazarus, 2001).
Safe School Ambassadors Program Evaluation 2011
Page 28
In addition to peer support, strong relationships with trusted adults are important for empowering
bystanders to report bullying. School personnel must be trained and made available for support.
It is essential that students know and feel that this support is available. If students feel that
teachers don’t care about maltreatment (even if teachers have been trained to avoid reinforcing
bullying behavior), then even prosocial students may feel uncomfortable enlisting help from
adults in more serious bullying episodes. Students cannot and should not handle all bullying
episodes alone. Positive rapport and relationships between teachers and students must be
established for bystander intervention participants to successfully intervene on behalf of victims.
Concern about student’s welfare and the danger of antisocial behaviors must be palpable in the
school climate. Dedicated and trained peers will fail without adequate support from school
personnel, ongoing support from peers, and an overall change of the school climate.
School-wide interventions that are designed to change school climate are needed to change
normative responses to bullying behaviors. According to Macklem (2003), prevention efforts
should incorporate the following elements:
• Provide accurate information regarding bullying and the attitudes of peers about bullying to
dispel myths and minimizes misconceptions about bullying behavior, victims, and tacit
compliance
• Re-sensitize peers to bullying behavior. This includes learning to recognize bullying and
differentiating bullying behavior from innocuous banter.
• Promote empathy and healthy guilt.
• Reinforce helping behavior and distinguish between reporting behavior as opposed “tattle
telling”.
Outcomes of Bullying Prevention Programs and Bystander Interventions
Prompted by three suicides attributed to bullying, the National Campaign Against Bullying
headed by Dan Olweus and Erling Roland was implemented in Norway in 1983 and resulted in a
50% reduction in bullying episodes and fewer new victims in Bergen schools at one year follow
up (Ross, 1996). In addition to a reduction in episodes of bullying, there was a change in the
school climate and a reduction of bullying in the greater community. Subsequent evaluations of
the Olweus program in other countries have also found positive results, but to a much lesser
degree than the original 50% reduction (Elliot, 2000).
Two other major bully prevention programs that used core components of the Olweus Bully
Prevention Program, known as the Sheffield and Safe Cities projects, found positive results with
an average reduction of about 15% (Macklem, 2003). Although Olweus original evaluation
results pointed to a dramatic reduction in bullying, others have had much less success, with some
programs reporting negative results (Macklem, 2003; Ross, 1996). However, Olweus and other
researchers have concluded that these wide differences in outcomes are largely due to poor
program planning and implementation of individual programs (Macklem, 2003). For example,
Olweus provided a high level of support to the Bergen schools, which increased program
implementation fidelity that resulted in greater reduction in bullying (Smith et al., 1999).
Moreover, the evaluation results as a whole emphasize the importance of utilizing a multi-
pronged in bullying prevention (Macklem, 2003; Olweus, 1991).
Safe School Ambassadors Program Evaluation 2011
Page 29
While traditional programs that target instances of bullying tended to only focus on the bully-
victim dyad and not the social process of bullying, most researchers conclude that the
effectiveness of programs can be greatly enhanced using a whole school approach (Macklem,
2003; Olweus, 1994). These researchers argue that approaches that focus only on identified
victims and bullies underestimate the incidence of bullying, place undue pressure on the victim
to solve the problem, and perpetuate myths about bullying and negative attitudes concerning
victims. An approach that includes all students and incorporates adults increases the likelihood
that an enduring change in overall school climate will occur over time. To improve bully
reduction rates from the average 15% reported in work conducted since Olweus original
evaluation work, researchers need more innovative strategies. One such innovation is
incorporating an understanding of group dynamics into program development. Currently,
researchers are encouraging the mobilization of bystanders to reduce school bullying (Macklem,
2003; Smith, Twemlow & Hoover 1999; Salmivalli, 1999).
Macklem (2003) concludes that most researchers feel that soliciting help from bystanders is the
key variable in decreasing bullying. For example, Dr. Ronald Slaby has demonstrated that
bystanders can be effective agents for resolving conflicts and preventing future cruelty and
violence in school settings (Slaby & Roedell, 1982; Slaby & Guerra, 1988). However, much less
is known about the individual- and school-level outcomes that are realized from bystander
intervention programs such as SSA. Although peer interventions in general have been
established as effective in many domains (Deck & Einspruch, 1999), little empirical evidence
has linked bystander intervention programs to specific bullying-related outcomes. At the same
time, a meta-analysis of 143 drug prevention programs conducted by Tobler (1986) showed that
school-based bystander interventions can be quite effective in reducing substance use and abuse
among youth. Because there are a number of similarities between many of these drug prevention
programs and bystander intervention programs, these results are encouraging.
The SafePlace: Domestic Violence and Sexual Assault Survival Center that developed the Expect
Respect school-based program to address sexual and domestic violence state in their handbook
that “of greatest significance was the impact of the project on increasing students’ willingness to
intervene to help a target of bullying and to seek help from an adult on campus”. According to
Macklem (2003), evaluation of the Expect Respect Program found an increase of awareness
about bullying, knowledge of sexual harassment, and positive attitudes toward helping peers.
The change of bystander behavior appears vital to program developers from diverse disciplines
in changing school climate and reducing a variety of antisocial behaviors.
Within the field of school-based bully prevention, Maher (1987) provides a good example of the
effective use of a student watch program in a New York high school and reported that antisocial
behavior had decreased and overall school climate increased as a result of the program.
Specifically, Maher (1987) and others argue that these programs help to break down the code of
silence that protects the antisocial, by presenting “telling on” as positive manner. Equally
important, though one might feel that the potential for negative stigmatization for student
members is high, Maher (1987) showed that because membership was presented as a high status
activity, this did not occur. Additionally, to the extent that these programs effectively recruit
socially influential youth from a broad section of the student body, the possibility of retributions
or reprisals are greatly diminished.
Safe School Ambassadors Program Evaluation 2011
Page 30
Similar to Maher’s student watch program, peer support programs have recently been developed
to include students in monitoring peer behavior. In addition to monitoring, these programs also
encourage and teach students to play an active role in changing relationships and group
dynamics. Cowie (2000) and colleagues have found that programs that focus on students’ social
interactions through encouraging peers to befriend others, teaching conflict resolution skills, and
supporting informal counseling-based approaches, help protect victims and increase prosocial
attitudes and behaviors.
Menesini, Codecasa, Benelli and Cowie (2003) recently published a study that tested the peer
support model and reported promising results in two Italian middle schools. The curriculum for
this program included promoting awareness in bullies of their and others’ behavior(s), enhancing
the ability to support victims (befriending by peers), encouraging responsibility and involvement
of bystanders, and improving the quality of interpersonal relationships. This peer support model
resulted in negative attitudes and behaviors to remain stable in students who received the
intervention while these factors increased in a control group. Moreover, feelings of support for
the victim decreased in the control group and remained stable in the experimental group. The
preliminary findings of this research suggests that the implementation of a peer support and
intervention model in the middle schools years, when rates of bullying have been shown to
increase, is an effective bullying prevention strategy. Since a number of components of the
Italian intervention are similar to the SSA program, the results provide good support for the
underlying model upon which SSA is based. The authors in the Italian study believed that their
short-term intervention broke the code of silence, enhanced responsibility, and promoted
empathy among students at an age with a heightened risk for bullying behaviors.
Because the idea of empowering bystanders to intervene rather than monitor and report is
relatively new, the outcomes associated with this approach is quite limited. However, much
more is known about the deleterious effects of bullying on both victims as well as bystanders.
For victims of bullying, fear can become so ingrained that they adopt fugitive-like routines to
avoid places likely to be frequented by the bully (Macklem, 2003). This avoidance deprives the
target of essential formal and informal social experiences that are important for social
development (Ross, 1996). Victims also can experience active rejection (Perry, Kusel, & Perry,
1988) by peers who formerly were friends or at least friendly toward them. This climate of fear
from chronic bullying coupled with peer rejection leads to poor school performance among
victims (Turkel & Eth, 1997). For example, Hazler, Hoover, & Oliver (1993) reported a
significant drop in grades for 90% of the victims of bullying in their study. Overall, victims can
be caught in downward spiral leading to:
• Low morale and acute despair manifested in truancy (Reid, 1990).
• Chronic illness such as recurrent abdominal pain of unknown origin (Ross & Ross,
1988).
• Running away, and in extreme cases, suicide (Beck, 1986; Besag, 1989; Elliot, 1991).
Equally important is research that shows that bystanders, who are neither bullies, nor target of
bullies, are negatively impacted by witnessing episodes of bullying (Ross, 1996). Ross (1996)
reports that bystanders can be affected by bullying in the following ways:
• Anger and feeling helpless at not knowing how to help victim.
• Guilt for not intervening.
• Worry that they might be the next target.
Safe School Ambassadors Program Evaluation 2011
Page 31
Furthermore, Ross (1996) argues that when bullying is witnessed by others and goes unpunished,
it can create a climate of fear with widespread effects that:
• Inhibits learning and is a major distraction
• Undermines enjoyment of free-time periods.
• Creates fear of certain areas of school such as lunchroom.
• Elevates the sense of vulnerability on the way to and from school.
Because school environments that are characterized by a high incidence of episodes of bullying
clearly undermine the learning process for all students, improved academic achievement is an
important medium-term outcome to expect from implementation of an effective prevention
program.
Conclusion
The SSA bystander intervention program includes many core components that are thought to be
effective in promoting a positive school climate and reducing individual incidents of bullying.
Specifically, the use of bystanders and the guidance provided on recruitment strategies are
consistent with recent discussions on effective bullying interventions provided in the literature.
One challenge of implementing and evaluating bully prevention programs is the finding that
bullying is not evenly distributed over schools (Ross, 1996). Some schools have higher or lower
rates of bullying; therefore a program with one narrow focus will not be effective for every
school. The need for a multi-faceted approach is supported by the work of Olweus (Macklem,
2003). Comprehensive programs that intervene on different levels and can be adapted to the
needs of the individual school may be more successful in reducing antisocial behaviors and
victimization. This applies to the Safe School Ambassador program in that the program includes
several opportunities of prevention and intervention. During a bullying episode, the leader may
intervene as a bystander by defending the victim, distracting member(s) of dyads, or diffusing
the situation through other social techniques. This immediate response may reduce escalation of
a teasing or bullying episode to something more violent. The immediate response may also
model prosocial behavior for other bystanders, diffuse the power of the bully, and support the
victim. In this way, the effect of the leader can transcend beyond the immediate episode to
influencing other peers, changing the classroom, and hopefully the overall school climate.
Building trust with teachers and staff personnel allows students leaders to solicit support from
teachers when they do not feel safe handling situations alone and introduces another opportunity
to impact the school environment. The literature on successful outcomes of comprehensive
programs supports the multidimensional approach of the Safe School Ambassador program.
While the SSA program addresses a number of important antecedents and outcomes associated
with bullying, its overall effectiveness is somewhat dependent on the extent to which a
participating school is part of a more comprehensive and district-wide bullying prevention
strategy in place. The literature underscores the fact that the social dynamics that lead to
bullying need to be addressed very early on in the developmental process (Boxer & Dubow).
Because SSA is designed for middle and high school youth, the effectiveness of the SSA
program will be greatly improved by districts and communities that have early education and
Safe School Ambassadors Program Evaluation 2011
Page 32
elementary school bullying prevention programs in place. Additionally, complementing the SSA
program with programs and activities that impact other aspects of the school environment should
result in stronger outcomes. Examples include school staff awareness training and school-wide
student sensitization to the problem of bullying.
Because empowering students to intervene during episodes of bullying is a relatively new
approach, there is currently a lack of empirical evidence linking positive outcomes to these types
of interventions. However, the experts in the field appear to be encouraging the use of these
bystander interventions and preliminary evaluation results are encouraging. Ongoing
comprehensive evaluation is needed for all violence prevention programs to determine the best
practices for keeping children safe and facilitating the learning process.
References
American Academy of Child and Adolescent Psychiatry. (2004). Facts for family: Bullying (no.
80). [information sheet]. Lippincott Williams & Wilkins.
Adler, P.A., & Adler, P. (1998). Peer power: Preadolescent culture and identity. New
Brunswick, NJ: Rutgers University Press.
Bandura, A., (1977). Social Learning. New Jersey: Prentice-Hall.
Banks, R. (1997). Bullying in schools. Champaign, I: Clearinghouse on Elementary and Early
Childhood Education. (ERIC Digest, EDO-PS-97-17).
Beck, A.T., (1986). Hopelessness as a predicator of eventual suicide. Annals of the New York
Academy of Sciences, 487, 90-96.
Besag, V., (1989). Bullies and victims in schools. Milton Keynes: Open University Press.
Bonds, M. & Stoker, S., (2000). Bully Proofing your School: A Comprehensive Approach for
Middle Schools. Colorado: Sopris West.
Boxer and Dubow (2002) A social-cognitive information –processing model for school-based
aggression reduction and prevention programs: Issues for research and practice. Applied
& Preventive Psychology, 10, (3), 177-192.
Brigham, J.C. (1991). Social Psychology (2
nd
ed.). New York: HarperCollins Publishers, Inc.
Cairns, R., B., & Cairns, B.C., (1991). Social cognition and social networks: A developmental
perspective. In D.J. Pepler & K.H. Rubin (Eds), The development and treatment of
childhood aggression (pp. 249-278). Mahwah, NH: Lawrence Erlbaum Associates,
Publishers.
Cowie, H., & Wallace, P. (2000). What is peer support? IN H. Cowie & P. Wallace (Eds.), Peer
support in action: From bystanding to standing by (pp.5-22). London: Sage Publications.
Safe School Ambassadors Program Evaluation 2011
Page 33
Craig, W. M., & Pepler, D.J. (1995). Peer processes in bullying and victimization: An
observational study. Exceptionality Education Canada, 5 (3&4), 81-95.
Craig, W. M., & Pepler, D.J. (1997). Observations of bullying and victimization in the school
year. In W. Craig (Ed.), Childhood social development: The essential readings. (pp. 117-
136). Oxford, UK: Blackwell Publishers
Craig, W. M., & Pepler, D.J. (2000). Observations of bullying in the playground and in the
classroom. School Psychology International, 21(1), 22-36.
Community Matters (2003). Safe School Ambassadors: Program Advisors Handbook. Santa
Rosa, California.
Deck, D.D. & Einspruch, E.L. (1999). Outcomes of Peer Interventions. In Deck, D.D. (Ed.)
Intervening in Adolescent Substance Abuse: An evaluation of Washington’s Prevention
Intervention Program. Portland, OR: RMC Research Corporation
Eisenberg, N. and Miller, P. (1987). The relation of empathy to prosocial and related behaviors.
Psychological Bulletin, 101(1), 91-119.
Elliot, D. S. (1998). Prevention Programs that Work for Youth: Violence Prevention.
Charleston, SC: Aspen Institute’s Congressional Program.
Elliot, D.S. (Ed.) (2000). Blueprints for violence prevention: Bullying prevention program.
Boulder, Colorado: Institute of Behavioral Science, Regents of the University of
Colorado.
Elliot, M. (1991). Bullying: A practical guide to coping for schools. Harlow, Essex: Longman.
Ericson, N. (2001, June). Addressing the Problem of Juvenile Bullying. Fact Sheet. Washington,
DC: The Office of Juvenile Justice and Delinquency.
Feshback, N. (1982). Sex differences in empathy and social behavior in children. In N. Eisenberg
(Ed.) The Development of Prosocial Behavior. New York: Academic Press.
Froschl & Gropper (1999).Fostering friendships, curbing bullying. Educational Leadership,
56(8), 72-75.
Ginsburg and Miller (1981), Altruism in children: a naturalistic study of reciprocation and an
examination of the relationship between social dominance and aid-giving behavior.
Ethology and Sociobiology, 2, 75-83.
Garrity, Jens, Porter Sager, Short-Camilli (1994). Bullying proofing your school: a
comprehensive approach for elementary schools. Longmont, CO: Sopris West.
Greener, S.H. (2000). Peer assessment of children’s prosocial behavior. Journal of Moral
Education, 29 (1), 47-60.
Safe School Ambassadors Program Evaluation 2011
Page 34
Hamburg, M. A. (1998). Youth violence is a public health concern. In D. S. Elliott, M. A.
Hamburg, & K. R. Williams (Eds.), Violence in American schools: A new perspective
(pp. 31–54). New York: Cambridge University Press.
Hazler, R.J. Hoover, J.H., & Oliver, R. (1993). What do kids say about bullying? Education
Digest, 58 (7), 16-20.
Hoffman, M.. (1982). Development of prosocial motivation: empathy and guilt. In N. Eisenberg
(Ed.) The Development of Prosocial Behavior. (pp. 281-313). New York: Academic
Press.
Hoffman, M. (1984). Empathy, its limitations and its role in a comprehensive moral theory. In
Kurtines, W., Gewirtz, J. (Eds.) Morality, Moral Behavior, and Moral Development..
New York: John Wiley and Sons.
Jeffrey, Miller, Linn (2001). Middle school bullying as a context for the development of passive
observers to the victimization of others. In R.A. Gefner, M. Loring, & C. Young (Eds.)
Bullying behavior: Current issues, research, and interventions. New York: The Haworth
Press, Inc.
Khosropour & Walsh (2000). That’s not teasing-That’s bullying: A study of fifth grades
conceptualization of bullying and teasing. Paper presented at the annual conference of the
American Educational Research Association, Seattle, Washington.
LaFontana, K.M., & Cilessen, A. (2002). Children’s perceptions of popular and unpopular peers:
A multi-method assessment. Developmental Psychology, 38(5), 635-647.
Lazarus (2001). Breaking the code of silence; what schools can do about it. Communique,
29(7),28-29.
Macklem, G. (2003). Bullying and Teasing: Social Power in Children’s Groups. New York:
Kluwer Academic.
Maher, R. (1987, Fall). Students watch out for their own. School Safety, 26-27.
Menesini, E., Codecasa, E., Benelli, B. and Cowie, H. (2003). Enhancing Children‘s
Responsibility to Take Action Against Bullying: Evaluation of a Befriending Intervention
in Italian Middle Schools. Aggressive Behavior 29(1): 10–14.
Nansel, T.R., Overpeck, M.,D., Haynie, D.L., Ruan, W.J., & Scheidt, P.C. (2003). Relationships
between bullying and violence among US youth. Archives of Pediatrics & Adolescent
Medicine, 157, 4, 348-353.
Nansel, T.R., Overpeck, M.,D., Pilla, R.S., Ruan, W.J., Simons-Morton, B., & Scheidt, P.C.
(2001). Bullying behaviors among US youth: Prevalence and association with
Safe School Ambassadors Program Evaluation 2011
Page 35
psychosocial adjustment. Journal of the American Medical Association, 285 (16), 2094-
2100.
O’Connell, P., Pepler, D., and Craig, W., (1999). Peer involvement in bullying: insights and
challenges for intervention. Journal of Adolescence, 22, 437-452.
Olweus, D. (1993). Bullying at school. What we know and what we can do. Oxford, UK:
Blackwell Publishers.
Olweus, D. (1994). Bullying at school: Long-term outcomes for the victim and an effective
school based intervention program. In L.R. Huesman (Ed.). Aggressive behavior: Current
perspectives (pp. 97-130). New York: Wiley.
Pellegrini & Long. (2002). A longitudinal study of bullying, dominance, and victimization
during the transition from primary school through secondary school. British Journal of
Developmental Psychology, 20(2), 259-280.
Perry, D.G., Kusel, S. J., Perry, L.C., (1988). Victims of peer aggression. Developmental
Psychology, 24(6), 807-814.
Petty, R.E., Cacioppo, J.T. (1981). Attitudes and persuasion: Classic and contemporary
approaches. Dubuque, IA: Wm. C. Brown Publishers.
Reid, K. (1990). Bullying and persistent school absenteeism. In D.P.Tattum & D.A. Lane (Eds).
Bullying in schools (pp.89-94). Hanley, Stoke-on-Trent: Trentham Books.
Rigby and Slee (1992).Bullying among Australian school children: reported behavior and
attitudes toward victims. The Journal of School Psychology, 131, 615-627.
Ross, D.M. (1996). Childhood Bullying and Teasing: What School Personnel, Other
Professionals, and parents Can Do. Alexandria, VA: American Counseling Association.
Ross, D.M., & Ross, S.A. (1988). Childhood pain: Current issues, research, and management.
Baltimore: Urban & Schwarzenburg.
Salmivalli, C. (1999). Participant role approach to school bullying: Implications for intervention.
Journal of Adolescence, 22(4), 453-459.
Salmivalli, C. Lagerspetz, C. Bjorkqvist, K., Osterman, K & Kaukiannien, A. (1996). Bullying as
a group process: participant roles and their relations to social status with in the group.
Aggressive Behavior, 22, 1-15.
Slaby, R. & Guerra, N. (1988). Cognitive Mediators of Aggression in Adolescent Offenders.
Developmental Psychology, 24 (4), 580-588.
Safe School Ambassadors Program Evaluation 2011
Page 36
Slaby, R. & Roedell, W. (1982). The Development and Regulation of Aggression in Young
Children. In J. Worell (Ed.) Psychological Development in the Elementary Years. New
York: Academic Press.
Smith, P.K. Morita, Y., Junger-Tas, J., Olweus, D., Catalano, R., & Slee, P. (Eds) (1999). The
nature of school bullying. A cross national perspective. Routledge. London.
Smith & Sharp (Eds.) (1994). School bullying: Insights and perspectives. London: Routledge.
Smith J., Twemlow S.W., Hoover D.W. (1999). Bullies, victims and bystanders: a method of in
school intervention and possible parental contributions. Child Psychiatry and Human
Development, 30(1), 29-37.
Sutton, J., & Keogh, E. (2000). Social Competition in School: Relationships with Bullying,
Machiavellianism, and Personality, British Journal of Educational Psychology, 7, 443-
56.
Swearer, S.M., & Doll, B. (2001). Bullying in schools: An ecological framework. Journal of
Emotional Abuse, 2 (2/3), 7-23.
Tobler, N.S. (1986). Meta-Analysis of 143 Adolescent Drug Prevention Programs: Quantitative
Outcome Results of Program Participants compared to a control or comparison group,
Journal of Drug Issues, 16(4), 537-567
Turkel, S.B., Eth, S. (1997). Psychological responses to stress: Adjustment disorder and post
traumatic stress disorder in children and adolescents.; In: Eugene. A.L, ed. Childhood
stress. (pp.27-38). Oxford, England: John Wiley & Sons.
US Department of Health and Human Services (1999). Youth Violence: A Report of the Surgeon
General. Rockville, MD: Author.
