Reducing Stigma in High School Students: A Cluster Randomized Controlled Trial of the National Alliance on Mental Illness’ Ending the Silence Intervention

Abstract

Beyond education and contact program components, existing research on how to design a successful adolescent stigma reduction intervention has been inconclusive. This study evaluated the effectiveness of a school-based mental health (MH) stigma reduction and health promotion program, Ending the Silence (ETS), developed by the National Alliance on Mental Illness (NAMI). A diverse sample of 206 high school students in New York City participated in the current study. Using a cluster randomized controlled trial design, 14 ninth–12th grade classes (Grade 9–12) were randomly assigned to the ETS program or an active control presentation on careers in psychology. Students completed 4 surveys throughout the study (pre, immediate post-presentation, 4 weeks post, 8 weeks post). Prospective results (over 2 months) and qualitative feedback were analyzed. Prospectively, mixed effects modeling indicated significant effects in favor of the ETS group for reduced negative stereotypes, improved mental health knowledge, and less anticipated risk for disclosing to a counselor. There were also trends in favor of the ETS group for reductions in intended social distancing and negative affect, and improvements in help-seeking intentions. Other predictors of stigma included mental health knowledge, gender, race/ethnicity, prior contact with mental illness, and grade level. Qualitative feedback indicated positive impressions of ETS overall, but suggestions for more interactive activities and discussion. Relatively brief programs such as ETS appear to be a practical vehicle for stigma reduction. Future research is warranted on longer-term programs and adolescent development variables.

Full text

Running head: HIGH SCHOOL STIGMA INTERVENTION 1
©American Psychological Association, 2020. This paper is not the
copy of record and may not exactly replicate the authoritative
document published in the APA journal. Please do not copy or cite
without author’s permission. The final article will be available, upon
publication, via its DOI: https://doi.org/10.1037/sah0000235
Reducing stigma in high school students: A cluster randomized controlled trial of the National Alliance
on Mental Illness’ Ending the Silence intervention
Author Note
Joseph S. DeLuca, Department of Psychology, John Jay College of Criminal Justice, City
University of New York (CUNY) and CUNY Graduate Center.
Janet Tang, Department of Psychology, John Jay College of Criminal Justice, CUNY.
Sarah Zoubaa, Department of Psychology, John Jay College of Criminal Justice, CUNY.
Brandon Dial, Department of Psychology, John Jay College of Criminal Justice, CUNY.
Philip T. Yanos, Department of Psychology, John Jay College of Criminal Justice, CUNY and
CUNY Graduate Center.
Correspondence concerning this article should be addressed to Joseph DeLuca, Department of
Psychology, John Jay College of Criminal Justice, 524 W 59th St, New York, NY, 10019. Email:
jsaldeluca@gmail.com; fax: 212-237-8930

HIGH SCHOOL STIGMA INTERVENTION 2
Abstract
Beyond education and contact program components, existing research on how to design a successful
adolescent stigma reduction intervention has been inconclusive. This study evaluated the effectiveness of
a school-based mental health (MH) stigma reduction and health promotion program, “Ending the
Silence” (ETS), developed by the National Alliance on Mental Illness (NAMI). A diverse sample of 206
high school students in New York City participated in the current study. Using a cluster randomized
controlled trial design, fourteen 9th-12th grade classes (Grade 9-12) were randomly assigned to the ETS
program or an active control presentation on careers in psychology. Students completed four surveys
throughout the study (pre, immediate post-presentation, 4-weeks post, 8-weeks post). Prospective results
(over two-months) and qualitative feedback were analyzed. Prospectively, mixed effects modelling
indicated significant effects in favor of the ETS group for reduced negative stereotypes, improved mental
health knowledge, and less anticipated risk for disclosing to a counselor. There were also trends in favor
of the ETS group for reductions in intended social distancing and negative affect, and improvements in
help-seeking intentions. Other predictors of stigma included mental health knowledge, gender,
race/ethnicity, prior contact with mental illness, and grade level. Qualitative feedback indicated positive
impressions of ETS overall, but suggestions for more interactive activities and discussion. Relatively brief
programs such as ETS appear to be a practical vehicle for stigma reduction. Future research is warranted
on longer-term programs and adolescent development variables.
Keywords: stigma, adolescence, mental health, Ending the Silence, National Alliance on Mental
Illness

HIGH SCHOOL STIGMA INTERVENTION 3
Reducing High School Stigma: A Cluster Randomized Controlled Trial of the National Alliance on
Mental Illness’ Ending the Silence Intervention
Introduction
Adolescence is a key period for personal and social development, and mental health conditions
can significantly alter the trajectory of an adolescent’s life. The median age of onset for a mental health
condition is fourteen (Auerbach et al., 2018) and approximately 75% of all lifetime mental health
conditions worldwide begin by the mid-20s (Kessler et al., 2007). Formal service utilization and help-
seeking tend to be low (e.g., Merikangas et al., 2011) and are impacted by a host of factors, including
stigma (Cauce et al., 2002; Corrigan, Druss, & Perlick, 2014; Spencer, Chen, Gee, Fabian, & Takeuchi,
2010; World Health Organization, 2005). Stigma refers to a process of labeling that can lead to
stereotyping and discrimination within a context of power (Goffman, 1963; Link & Phelan, 2001). Stigma
can lead to a lack of engagement in mental health treatment and inhibit full inclusion in society (Corrigan,
Watson, Byrne, & Davis, 2005). Stigma remains understudied among youth, but recent reviews and meta-
analyses have documented that stigma is prevalent among youth (Silke, Swords, & Heary, 2016), impacts
help-seeking (Nam et al., 2013), and overlaps significantly with developmental processes such as peer
group formation and identity development (DeLuca, 2019). Overall, adolescents in the stage of middle
adolescence appear to be particularly ideal targets for stigma reduction, as this is the period in which
cognitive differences with adults relevant to stigma begin to diminish (Corrigan et al., 2007). Further,
adolescents typically have less information and more tentatively formed attitudes about people with
mental illness than adults (Corrigan et al., 2005), which also makes this population particularly conducive
to stigma change. Thus, the stigma process can be disrupted in this period before it forms more fully.
Youth Mental Health Stigma Reduction
Coordinated efforts to reduce youth mental health stigma in the US are in the early stages. A
recent systematic review (Salerno, 2016) of fifteen school-based (grades 5-12) programs in the US found
that most improved mental health knowledge, reduced negative stereotypes, and improved help-seeking
outcomes in the short-term. Salerno (2016) concluded that more studies on program implementation and

HIGH SCHOOL STIGMA INTERVENTION 4
long-term effects are needed, specifically by collecting more socio-demographic information and using
randomized designs and long-term follow-ups. The few researchers who have used long-term follow-ups
have found mixed findings related to maintained stigma reductions (Corrigan, Michaels, & Morris, 2015;
Perry et al., 2014; Pinto-Foltz et al., 2011; Thornicroft et al., 2016; Yamaguchi, Mino, & Uddin, 2011).
Overall, several research teams have concluded that they cannot make any firm recommendations for
school-based stigma reduction programs due to inconsistent or null results that may stem from poor
reporting quality, a dearth of randomized trials/lack of control groups, sample heterogeneity, program
structure heterogeneity and lack of fidelity measurement, different outcome measurements, and
inadequacy of stigma measures for youth (Austin & Schwartz, 2018; Koller & Stuart, 2016; Mellor, 2014;
Schachter et al., 2008; Wei, Hayden, Kutcher, Zygmunt, & McGrath, 2013).
Program structure and stigma measurement. In regard to structure, most stigma reduction
programs include education (e.g., dispelling myths about mental illness; providing mental health
education) or contact components (e.g., presentations by persons living with a mental illness who share
their stories of recovery), or a combination of both (Corrigan et al., 2015). The nature of the education
(e.g., diagnosis-specific) and contact (e.g., age of person, via video or in-person), however, differs from
program to program (Schachter et al., 2008). Similarly, outcome measures vary from study to study and
include a range of stigma dimensions, though help-seeking and disclosure carryover-related outcomes
tend to be the most understudied (Clement et al., 2015; Salerno, 2016). For example, only five out of the
forty studies (13%) in Yamaguchi and colleagues’ (2011) review of interventions evaluated personal
mental health and help-seeking outcomes. Hartman and colleagues (2013) conducted the first known
study to evaluate the impact of a short stigma reduction program (75 minutes) on non-help seeking
adolescents’ self-stigma, using a no-control group, pre/post-test design. These authors found reductions in
self-stigma of seeking help among Canadian high school youth after the program, but more research is
needed to generalize these findings.
Corrigan, Morris, Michaels, Rafacz, and Rüsch (2012) conducted the largest meta-analysis to date
of stigma reduction studies (N = 72) for adults and adolescents. Nineteen studies in this sample were

HIGH SCHOOL STIGMA INTERVENTION 5
evaluated among adolescents (age 12-18). Results indicated that, on average, education and contact-based
interventions were both effective for adolescents at reducing stigma (i.e., attitudes/stereotypes, negative
affect, and intended social distancing). In-person contact interventions yielded the largest effect sizes
overall and specifically for intended social distancing and behavioral intentions toward someone living
with mental illness (e.g., willingness to help). Nevertheless, many existing programs for youth still need
to be further evaluated for efficacy.
National Alliance on Mental Illness - Ending the Silence
Ending the Silence (ETS), developed by the National Alliance on Mental Illness (NAMI, the
largest grassroots mental health nonprofit in the US), is one national, standardized approach that can be
further evaluated and used as a vehicle to overcome the aforementioned limitations in stigma reduction
research (NAMI, 2015). ETS is a one-day, classroom-based presentation that lasts approximately 50
minutes. To date, however, the ETS program has not been thoroughly empirically evaluated. Analyzing
nearly 2,000 post-test surveys from middle and high school students in New York City, researchers
(DeLuca, Evans, & Yanos, 2018) found an overall 80% satisfaction rate with ETS (e.g., would
recommend the program to others; believed the presenters communicated effectively). Ninety percent or
more of students agreed that they know the early warning signs of mental illness, that they now knew how
to help themselves or a friend if they noticed mental health warning signs, and that the presenters
communicated effectively.
Only three studies have evaluated ETS beyond the standard NAMI post-test survey. Wong and
colleagues (2015) used a pre/immediate post-test, no control group design with high school students in
California. Results indicated some significant changes on individual items related to social distance,
emotional responses, attitudes, and knowledge, but no changes in help-seeking or peer support. In an
unpublished master’s thesis, Taniyama (2016) also evaluated ETS among high school students in
California using a no control group design, but included a stronger pre-, immediate post-, and 6-week
follow-up method. Results indicated significant improvements in emotions, knowledge, and attitudes at
the post-test, which maintained at 6-weeks (Taniyama, 2016). Lastly, Wahl, Rothman, Brister, and

HIGH SCHOOL STIGMA INTERVENTION 6
Thompson (2018) recently evaluated ETS in five US states, using a pre/post/follow-up (4-6 weeks) non-
randomized design, including a no intervention control group, and a 12-item outcome measure
(knowledge, stereotypes, social distance, and help-seeking). Using repeated measures analysis of
variance, results indicated positive changes at the immediate follow-up for ETS, though these changes
appeared to gradually rebound at the 4-6 week follow-up. The largest changes for students who received
ETS in this study were being able to recognize the warning signs of mental health conditions, and
knowledge of what to do to seek help if experiencing a mental health condition. Items asking about
recovery of people with mental illness (e.g., ability to get jobs) and about intended social distance (e.g.,
invite to home) returned to baseline levels at the 4-6 week follow-up.
Current Study
The current study is the first randomized controlled trial (RCT) of ETS. This study also
addresses other common limitations to youth stigma reduction research, including lack of follow-up,
failure to account for socio-demographic covariates and other predictors of stigma, a lack of standardized
and reliable stigma measures, a lack of attention paid to youth developmental processes (e.g., identity),
and no explicit linkage of findings to a conceptual model of adolescent mental health stigma (Pinto-Foltz
& Logsdon, 2009; Silke et al., 2016). We hypothesized that students who received ETS (v. active control)
would show significant improvements in mental health knowledge, negative stereotypes, intended social
distancing, negative emotional responses (affect), help-seeking attitudes, anticipated stigma, disclosure
worry, and self-stigma, from Time 1 (baseline) to Time 2 (immediate follow-up after program). Effects
were expected to be stronger at Time 2 and for primary outcome variables (i.e., mental health knowledge,
negative stereotypes, intended social distancing, negative affect, help-seeking attitudes) than secondary
outcome variables (i.e., anticipated stigma, disclosure worry, self-stigma).
Method
Participants. Two hundred and thirty-two students from one New York City public high school
were approached to take part in the study. Using a passive parent/guardian consent approach, 208 students
(90%) assented to take part, but two of those 208 students returned opt-out forms, resulting in a sample

HIGH SCHOOL STIGMA INTERVENTION 7
size of 206 students (M-age = 15.41, SD = 0.94, range: 13-18). Demographic characteristics of the sample
or presented in Table 1. The sample was predominantly female (56.2%) and participants were racially
diverse, with significant proportions of students identifying as European-American (35.0%), African-
American (20.9%), Latino/a/x (15.5%), Asian-American (13.6%). At the time of this study, the
participating high school had not yet established a formal mental health curriculum for their students,
although efforts are currently underway to do this in the state of New York (Kaufman, 2018).
[Table 1 here]
[CONSORT Figure 1 here]
Procedure. Several schools were offered the opportunity to participate in this study. Schools
were selected in a non-randomized fashion, either via personal connection or recommendation by NAMI.
One school agreed to have a representative group of students in fourteen classes participate in the study.
Institutional Board Review (IRB) approval was received from the researchers’ university IRB and the
local Department of Education IRB. Inclusion criteria included being a high school student and speaking
English well enough to complete the questionnaire. This study was completed between September 2017
and February 2018. Individual classrooms were visited at least five times (two visits to describe the study
and collect assent/consent forms; one visit for the presentation; two final visits for follow-ups). After data
collection was completed, teachers of each class were offered an in-person debriefing. Five research
assistants were trained to assist with school visits and data collection.
Study participants completed questionnaires at four time points – baseline (Time 1), immediate
post-test (Time 2), four-week follow-up (Time 3), and eight-week follow-up (Time 4). A pilot study was
conducted (DeLuca, Evans, Reyes, & Yanos, 2016) to determine survey length, identify issues with
survey implementation, and determine the appropriateness of items. The spacing out of Time 1 and Time
2 was designed to prevent any validity threats related to serial administration and to reduce the burden of
completing two questionnaires and watching a presentation in one sitting. Questionnaires were also
counterbalanced; four identical versions of the questionnaires were created (with measures randomly
ordered) to control for order effects. Intervention and control group students completed identical

HIGH SCHOOL STIGMA INTERVENTION 8
questionnaires. Participants’ questionnaires were linked across time points by anonymous identification
numbers.
This study followed the guidelines of the Consolidated Standards of Reporting Trials
(CONSORT) for cluster randomized trials (Campbell, Piaggio, Elbourne, & Altman, 2012). In order to
minimize imbalance across intervention and control groups, blocks were first stratified so that a similar
range of grade levels would be represented in each condition. Overall, fourteen rows were created
(separated into two stratified blocks of seven rows), each listing a randomized condition assignment
(intervention or control) and a sequence of dates for the five study visits. The order of condition
assignments in each block was generated by a computer algorithm. (www.randomizer.org). As teachers
responded to an online survey about availability, their classes were put into the first available slot of these
stratified blocks.
Presentations. Participants in the intervention group received ETS conducted by two speakers
from a local NAMI Affiliate. Both speakers were experienced and had given presentations for several
years. All students in the intervention group received the ETS presentation from the same pair of
speakers. For the current study, ETS presentations were shortened to account for the constraints of the
school’s 45 minute periods and the study’s pre- and post-test evaluations. Typically, presentations lasted
35-40 minutes in this study, with half of the presentation dedicated to psychoeducation and half dedicated
to a personal story from someone with lived experience. In regard to deviations from the standard ETS
50-minute program, the presenters in this study summarized some educational points/slides, omitted an
educational video, and made the in-person story more concise. Efforts were made to balance the amount
of education and contact, and to still have time for students to discuss the program.
Students in the active control group received a presentation of parallel length on “careers in
psychology” (adapted from Wood & Wahl, 2006, p. 48). This presentation was unrelated to stigma, and
included a series of videos from the American Psychological Association and facts on psychology
careers, followed by a discussion lead by the principal investigator (one control group presentation was
given by a trained research assistant, due to scheduling conflicts). As per Wood and Wahl's (2006) design,

HIGH SCHOOL STIGMA INTERVENTION 9
in order to reduce the overt demand for changed responding and minimize potential confusion for control
participants being asked to complete measures seemingly unrelated to their presentation about psychology
careers, students were informed that they were being asked to participate in two major tasks: First, they
were told that they were serving as audience members for a presentation on a psychology-related topic.
Second, they were told that they were completing some questionnaires that are being pilot tested (i.e.,
“Some of the questionnaires you will complete are being pilot tested among adolescents and thus may not
be directly relevant to the presentation you receive”), thereby framing the completion of instruments and
the presentation as separate components. The true methodological connection between these components,
however, were shared as part of the debriefing.
Measures
1
Measures were selected that aligned with multidimensional conceptualizations of stigma
(DeLuca, 2019; Evans-Lacko et al., 2010; Link & Phelan, 2001; Link, Yang, Phelan, & Collins, 2004;
Pescosolido & Martin, 2015; Silke, Swords, & Heary, 2016). Measures that had been previously used in
research with adolescents or young adults were prioritized for selection.
Primary outcome variables. The Attitudes about Mental Illness and Its Treatment Scale (AMIS;
Kobau et al., 2010) was used to assess negative stereotypes toward persons with mental health problems.
A 7-item AMIS scale was used in this study on a 5-point Likert scale. The 4-item Categorical Thinking
subscale of the Attitudes Toward Serious Mental Illness-Adolescent Version (ATSMI-AV; Watson et al.,
2005) was also used to further assess stereotypes (also a 5-point Likert scale). Internal consistency for
AMIS was poor (Cronbach’s Alpha = 0.52 at Time 1). Internal consistency for ATSMI-AV scale was
questionable (Cronbach’s Alpha = 0.68 at Time 1). The Reported and Intended Behavior Scale (RIBS;
Evans-Lacko et al., 2011) was used to assess intended social distancing behaviors. This 4-item measure is
rated on a 5-point Likert scale (1 = agree strongly, 5 = disagree strongly) and has been validated with
adolescents (Mansfield, Humphrey, & Patalay, 2019). Internal consistency for the RIBS was acceptable
1
Scales were coded so that higher scores indicated higher levels of that phenomenon (e.g., stigma, knowledge). Scales were
transformed when necessary.

HIGH SCHOOL STIGMA INTERVENTION 10
(Cronbach’s Alpha = 0.79 at Time 1). The Mental Health Knowledge Schedule (MAKS; Evans-Lacko et
al., 2010) was used to measure stigma-related mental health knowledge. The MAKS is a 12-item measure
using a 5-point Likert scale. The MAKS is meant to be used in conjunction with attitude and behavior-
related measures when assessing stigma reduction programs. Internal consistency for the MAKS was poor
(Cronbach’s Alpha = 0.27 at Time 1). The revised Attribution Questionnaire (r-AQ; Pinto, Hickman,
Logsdon, & Burant, 2012; Watson et al., 2004), a 5-item measure using a 7-point Likert scale developed
specifically for adolescents, was used to measure emotional responses toward a hypothetical student with
mental illness. Internal consistency for the r-AQ was acceptable (Cronbach’s Alpha = 0.71 at Time 1).
The Intentions to Seek Counseling Inventory (ISCI; Cash, Begley, McCown, & Weise, 1975) was used to
measure mental health help-seeking intentions. The ISCI, a 10-item measure on a 4-point Likert scale,
consists of common problems that adolescents and young adults may seek counseling for (e.g.,
relationship difficulties, depression, concerns about sexuality), and asks participants how likely they
would be to seek counseling for such problem. A two-item measure related to peer support intentions was
also used in this study (Wong et al., 2015). Internal consistency for the ISCI was good (Cronbach’s Alpha
= 0.88 at Time 1). Internal consistency for the Peer Support scale was unacceptable (Cronbach’s Alpha =
0.25 at Time 1).
Secondary outcome variables. The Perceptions of Stigmatization by Others for Seeking Help
scale (PSOSH; Vogel, Wade, & Ascheman, 2009) assesses the perceived stigma persons anticipate from
those they interact with. The PSOSH is a 5-item scale on a 5-point Likert scale. Internal consistency for
the PSOSH was good (Cronbach’s Alpha = 0.86 at Time 1). The Self-Stigma of Seeking Help scale
(SSOSH; Vogel, Wade, & Haake, 2006) is a 10-item scale on a 5-point Likert scale consisting of items
related to feelings of inadequacy and inferiority for seeking mental health treatment. Overall, SSOSH
assesses threats to one’s self-evaluation due to seeking help and internalized stigma. Internal consistency
for the SSOSH was good (Cronbach’s Alpha = 0.82 at Time 1).The Disclosure Expectations Scale (DES;
Vogel & Wester, 2003) was used to directly assess disclosure worries about confidentiality in regard to
mental health services. The DES includes eight questions using a 5-point Likert scale about the

HIGH SCHOOL STIGMA INTERVENTION 11
anticipated utility and risk of disclosing personal information to a counselor. The DES comprises two
subscales of four items each – Anticipated Risks (DES-AR) and Anticipated Benefits (DES-AB). Internal
consistency was acceptable for both subscales (DES-AR Cronbach’s Alpha = 0.78 at Time 1; DES-AB
Cronbach’s Alpha = 0.79 at Time 1).
Predictor variables. Identity development was measured via the Self-Concept Clarity Scale
(SCCS; Campbell et al., 1996), which assesses the consistency and stability of adolescents’ self-beliefs.
The SCCS is a 12-item scale, measured using a 5-item Likert scale. SCCS was only measured at Time 1.
The SCCS was coded so that higher scores indicated a stronger, more cohesive self-concept. Internal
consistency for the SCCS was good (Cronbach’s Alpha = 0.86 at Time 1). Given the relationship between
mental health knowledge and other dimensions of stigma, the MAKS was used as a predictor during some
data analyses. Other covariates included race/ethnicity, gender identity, grade level, age, and prior contact
with mental illness (“Do you have a family member who is diagnosed with a mental health problem?” and
“Do you have a close friend who is diagnosed with a mental health problem?”), consistent with past
research showing that female adolescents and adolescents with prior contact endorse less stigma (e.g.,
Dolphin & Hennessy, 2016).
Qualitative assessment. At Time 3, participants in both groups were asked to respond to two
open-ended questions: “What did you like best about the presentation?” and “What is one suggestion you
have for making this presentation better?” These questions are similar to those included in NAMI’s usual
satisfaction survey for ETS.
Data Analysis
First, descriptive analyses were conducted to provide sample characteristics and to explore
potential baseline differences between intervention and control group participants (using χ2 analyses or t-
tests). Next, analyses were completed to evaluate the longitudinal effects of a youth stigma reduction
program. Mixed-effects multilevel modelling (MLM) using the SPSS MIXED procedure (in SPSS v25)
was used to investigate main treatment effects and group by time interactions (i.e., the influence of
randomized group membership on the multiple dimensions of stigma over time, controlling for

HIGH SCHOOL STIGMA INTERVENTION 12
covariates). Mixed effects analyses were an appropriate statistical method for this repeated measure
design, because these analyses consider correlated data (e.g., as would be expected between repeated time
points) and unequal variances, accommodate for missing data (e.g., maximum likelihood estimates), and
allow for the inclusion of random effects and fixed effects . Questionnaires were also reviewed for
students who wrote in unusual responses (e.g., unusual gender identity, or an older age written down), or
who completed the survey extremely quickly (two standard deviations below the mean). Overall, no cases
were removed from the dataset based on these criteria. The apparent high quality of the data may have
partly been a function of the in-person nature of the study and the presence of research assistants and
teachers.
A power analysis was also specifically conducted for analyses of the clustered data (Campbell,
Mollison, Steen, Grimshaw, & Eccles, 2000), considering three levels: 1) between-student differences, 2)
within-student differences, and 3) between class differences. Since cluster power analysis requires a
calculation of intraclass correlation coefficients (ICC), ICCs were calculated first. The ICC was
calculated by using unconditional mean models for each outcome to estimate variance at each level (Shek
& Ma, 2011). The average ICC in this study was 0.039 across outcome measures (range: 0.01 to 0.08),
similar to prior, similarly designed stigma reduction studies (Chisholm et al., 2016; Winkler et al., 2017).
This value means that approximately 4% of the variance in outcome measures was due to classroom
effects. With this ICC and an average cluster size of 15, the power analysis indicated that a sample of 374
would be needed to detect moderate effects at an alpha level of .05 and a power level of .8. Since ICC
results indicated that classroom explained only a small percentage (4% on average) of the variation in
outcomes, and initial analyses found that the inclusion of classroom did not significantly change the
estimates of the models, classroom was not included as a random effect in the mixed models and random
effects were not used in the models presented below, for ease of presentation and interpretation.
Intent-to-treat analyses were conducted for all randomized students, regardless of “exposure” (as
long as one time point was completed). All predictors were included and analyzed within models based
on a priori hypotheses. Exposed only findings are not included here (i.e., removal of nineteen participants

HIGH SCHOOL STIGMA INTERVENTION 13
who were absent at Time 2), since the results were identical to the intent-to-treat analyses. Fixed effects
included randomized group and assessment time. Classroom was included as a random effect. Post-hoc,
Bonferroni-adjusted analyses were utilized for multiple comparisons.
Qualitative Data Analysis
Three coders (two research assistants and the principal investigator) on the research team
analyzed the text data using a consensual qualitative research framework (CQR; Hill, Thompson, & Nutt
Williams, 1997; Hill et al., 2005). All coders were trained in the CQR approach before starting this
process. First, the reviewers independently reviewed the data to develop general topic areas, then
expanded on each area with a brief summary of the domain and lastly, compared and contrasted the
categories to identify overlap between categories and the potential for merging categories or creating sub-
categories. Throughout this process, groups of text were placed into categories/domains, reviewed, and
re-grouped in subcategories as necessary. Double coding of data was allowed in some cases, but efforts
were made overall to merge categories and create specific domains (Hill et al., 1997). The coders met two
times for consensual validation. During these meetings, coders discussed areas of agreement and
disagreement, and coding differences were resolved. When differences could not be resolved, a senior
auditor and stigma expert from the research team helped to resolve the difference. Following the
consensus of all reviewers, categories and subcategories were derived and labeled with a name and
description. A tabulation of the number of unique respondents corresponding to the related category was
also provided.
Quantitative Results
Drop-out and treatment exposure. After completing at least one time point, a total of 14
students opted out of the study (11 in the control group and 3 in the intervention group). This difference
between randomized conditions was significant (p = .03, as per Fisher’s Exact Probability Test).
Participants who dropped out were more likely to have a family member with mental illness (54%, n = 7,
compared to just 25% of the non-dropped out sample) (χ2 = 4.00, df = 1, p = 045), and more likely to be
Arab/Middle-Eastern (40% dropped out, though the Arab/Middle-Eastern sample was small) (χ2 = 14.06,

HIGH SCHOOL STIGMA INTERVENTION 14
df = 5, p = 015). Participants who dropped out did not differ from other participants in regard to stigma
endorsement, self-concept clarity, or other socio-demographics. The majority of students (91%) were
considered “exposed” to their randomized condition (i.e., were present for class and completed Time 2
survey). Most missing data were due to participant absence rather than attrition.
Intent-to-Treat Outcome Analyses
The findings on the relationship between intervention assignment and change in outcomes over
time are presented in Table 2. These analyses included all participants, regardless of whether they
dropped out of the study or were absent during data collections. Various predictors were added to the
models, including gender, contact with mental illness (family and friend), school grade, race/ethnicity,
mean mental health knowledge across time points (MAKS), and baseline self-concept clarity (SCCS).
Two-way interactions (group by time) were also included in the analyses (controlling for
predictors/covariates) to determine if there were significant outcome changes over time that differed by
randomized group. Significant effects of time are presented below, in addition to between and within
group changes over time. As noted, post-hoc, Bonferroni-adjusted analyses were utilized to assess mean
differences in outcome by group at each time point. Effect size was calculated using Cohen’s d to assess
the magnitude of overall change from baseline to post-treatment in ETS versus the control group.
[Table 2 here]
Primary Outcomes
For one of the negative stereotypes scales (AMIS), there was a significant group by time
interaction (F = 3.55, df = 3, 481.12, p = .014) with all predictors in the model, indicating that
participants in the ETS group had a significant reduction in mental illness stereotypes over time in
comparison to control group participants (while controlling for other predictors; see Figure 2). This
significant change between groups was evident at Time 2 (p < .0005, 95% CI = -0.44 to -0.17) and at
Time 3 (p = .024, 95% CI = -0.29 to -0.02), but not Time 1 (p = .23) or Time 4 (p = .088). Overall higher

HIGH SCHOOL STIGMA INTERVENTION 15
mental health knowledge (MAKS) was a significant predictor
2
(p < .0005, B = -0.33, 95% CI = -0.41 to -
0.25) of lower negative stereotypes, as was family contact with mental illness
3
(no) (p = .004, B = 0.16,
95% CI = 0.05 to 0.27). There was not a significant main effect of time in this model (F = 1.89, df = 3,
487.10, p = .13). In terms of within group differences from Time 1 to Time 4, the difference between the
baseline AMIS score and final follow-up score for ETS participants was not significant (as per a paired
samples t-test). The magnitude of the differences in the AMIS means between groups across all time
points was small to medium (Cohen’s d = .44). Students who received ETS showed a 7% decrease in
negative stereotypes from pre (M = 2.15, SD = 0.46) to immediate post-test (M = 1.99, SD = 0.54),
whereas there was a nonsignificant increase for the control group from pre (M = 2.22, SD = 0.43) to
immediate post-test (M = 2.30, SD = 0.52). There was no change from pre to 2-months post-test for
students who saw ETS. There was not a significant group by time interaction for the categorical thinking-
negative stereotypes scale (ATSMI-AV), with and without predictors in the model. However, the effect of
time was significant with all predictors in the model (F = 3.67, df = 3, 489.04, p = .012), indicating a
reduction in categorical thinking in both groups over time.
[Figure 2]
Although the group by time interaction for intended social distance (RIBS) was significant (F =
4.08, df = 3, 512.39, p = .007) without any predictors in the model, the model became non-significant
with predictors added to the model (F = 2.20, df = 3, 483.89, p = .087). The effect of time was significant
in this latter model (F = 5.08, df = 3, 489.42, p = .002). MAKS (more knowledge)
4
(p < .0005, B = -0.30,
95% CI = -0.40 to -0.21), gender (female)
5
(p < .0005, B = -0.31, 95% CI = -0.46 to -0.16), and family
contact
6
(no) (p = .018, B = 0.21, 95% CI = 0.04 to 0.39) were predictors of lower social distance.
Students who received ETS showed a 12% decrease in intended social distancing from pre (M = 2.02, SD
= 0.72) to immediate post-test (M = 1.77, SD = 0.68), whereas there was a nonsignificant decrease for the
2
MAKS was a significant predictor of AMIS at all timepoints
3
Family contact was a significant predictor of AMIS at Time 1 and Time 3
4
MAKS was a significant predictor of RIBS at all timepoints
5
Gender was a significant predictor of RIBS at Time 1, Time 2, and Time 4
6
Family contact was a significant predictor of RIBS at Time 3

HIGH SCHOOL STIGMA INTERVENTION 16
control group from pre (M = 1.95, SD = 0.68) to immediate post-test (M = 1.90, SD = 0.66). Students who
received ETS also showed a 6% decrease in intended social distancing from pre to 2-months post-test.
The magnitude of the differences in the RIBS means between groups across all time points was quite
small (Cohen’s d = .06).
The group by time interaction for mental health knowledge (MAKS) was significant (F = 3.10, df
= 3, 495.13, p = .026) with all predictors in the model (see Figure 2), indicating that participants in the
ETS group had a significant increase in knowledge over time in comparison to control group participants
(while controlling for other predictors; MAKS was not included as a covariate in this model). This
significant change between groups was evident at all follow-up time points: Time 2 (p = .002, 95% CI =
0.07 to 0.33), Time 3 (p = .010, 95% CI = 0.04 to 0.30), and Time 4 (p = .034, 95% CI = 0.01 to 0.28).
Time was also a predictor (F = 11.52, df = 3, 495.08, p < .0005), and gender (female)
7
was a significant
predictor (p = .020, B = 0.11, 95% CI = 0.02 to 0.20) of higher knowledge, as was close friend contact
with mental illness (no)
8
(p = .021, B = -0.12, 95% CI = -0.23 to -0.02). Students who received ETS
showed a 9% increase in knowledge from pre (M = 3.41, SD = 0.44) to immediate post-test (M = 3.71, SD
= 0.45), whereas there was a nonsignificant increase for the control group from pre (M = 3.42, SD = 0.43)
to immediate post-test (M = 3.51, SD = 0.44). Students who received ETS also showed a 6% increase in
knowledge from pre to 2-months post-test. The magnitude of the differences in the MAKS means
between groups across all time points was small to medium (Cohen’s d = .24).
Although the group by time interaction for negative affect (r-AQ) was significant (F = 2.92, df =
3, 500.75, p = .034) without any predictors in the model, it became non-significant when predictors were
added (F = 2.48, df = 3, 480.14, p = .061). The effect of time was not significant in this latter model (F =
1.34, df = 3, 486.23, p = .260). The magnitude of the differences in the r-AQ means between groups
across all time points was small (Cohen’s d = .16). There was not a significant group by time interaction
for intentions to seek counseling (ISCI), without and with predictors, but time was significant (F = 3.22,
7
Gender was a significant predictor of MAKS at Time 2 and 4
8
Close friend contact was a significant predictor of MAKS at Time 1 and 3

HIGH SCHOOL STIGMA INTERVENTION 17
df = 3, 486.99, p = .023) with predictors in the model, indicating improvements in both groups over time.
Close friend contact (no)
9
(p = .028, B = -0.20, 95% CI = -0.38 to -0.02) and lower self-concept clarity
10
(p = .028, B = -0.12, 95% CI = -0.23 to -0.01) were predictors of help-seeking. Similarly, there was not a
significant group by time interaction for the Peer Support scale (with and without predictors), but the
effect of time was significant (F = 4.10, df = 3, 479.03, p = .007) with predictors in the model. The
magnitude of the differences in the ISCI and Peer Support means between groups across all time points
was very small (ISCI Cohen’s d = .003 and Peer Support Cohen’s d = .05). Since internal consistency was
particularly low for this scale, two additional analyses were conducted using each individual item from
the Peer Support scale as an outcome variable. Findings were still similar; with and without predictors
added, no significant interaction effects were observed.
Secondary Outcomes
The group by time interaction for perceptions of stigma for seeking help (PSOSH) was significant
(F = 2.96, df = 3, 482.91, p = .032) with predictors in the model, as was the effect of time (F = 3.29, df =
3, 490.12, p = .020), indicating that each group was changing over time in different ways. Based on the
pattern of PSOSH changes by group, no significant differences were found at each time point. Other
predictors of PSOSH included mental health knowledge
11
(p = .049, B = -0.13, 95% CI = -0.25 to -
0.001), female gender
12
(p = .002, B = -0.23, 95% CI = -0.38 to -0.09), school grade
13
(9th and 10th grade)
(p = .010, B = 0.19, 95% CI = 0.05 to 0.34), self-concept clarity
14
(p < .0005, B = -0.19, 95% CI = -0.29
to -0.08), and race/ethnicity
15
(identifying as Asian-American/Pacific Islander) (p < .0005, B = 0.41, 95%
CI = 0.21 to 0.61). Students who received ETS showed an 8% decrease in perceptions of stigma from pre
(M = 2.03, SD = 0.74) to immediate post-test (M = 1.87, SD = 0.83), and students in the control group
showed a 10% decrease in perceptions of stigma from pre (M = 2.19, SD = 0.76) to immediate post-test
9
Close friend contact was a significant predictor of ISCI at Time 4
10
Self-concept clarity was a significant predictor of ISCI at Time 1 and Time 3
11
MAKS was a significant predictor of PSOSH at Time 4
12
Gender was a significant predictor of PSOSH at Time 1, Time 2, and Time 3
13
School grade was a significant predictor of PSOSH at Time 2
14
Self-concept clarity was a significant predictor of PSOSH at Time 1, Time 2, and Time 4
15
Race/ethnicity was a significant predictor of PSOSH at Time 1, Time 2, and Time 4

HIGH SCHOOL STIGMA INTERVENTION 18
(M = 1.98, SD = 0.68). There was a 2% increase in perceptions of stigma from pre to 2-months post-test
for students who saw ETS, but an 11% decrease in such perceptions for the control group from pre to 2-
months post-test. The magnitude of the differences in the PSOSH means between groups across all time
points was very small (Cohen’s d = .02). The group by time interaction and time main effect for self-
stigma of seeking help (SSOSH) were non-significant, without and with predictors. The magnitude of the
differences in the SSOSH means between groups across all time points was small (Cohen’s d = .16).
The group by time interaction for disclosure worries: anticipated risk (DES-AR) was significant
(F = 4.68, df = 3, 481.54, p = .003) with predictors in the model, as was the effect of time (F = 3.42, df =
3, 486.18, p = .017), indicating that each group was changing over time in different ways. This change
between groups was significant at Time 4 only (p = .013, 95% CI = -0.65 to -0.08). Other predictors of
anticipated risk included gender (female)
16
(p = .027, B = 0.26, 95% CI = 0.03 to 0.48), school grade
17
(9th and 10th) (p = .017, B = 0.28, 95% CI = 0.05 to 0.50), and self-concept clarity
18
(p < .0005, B = -0.37,
95% CI = -0.52 to -0.21). Students who received ETS showed a 9% decrease in anticipated risks from pre
(M = 3.53, SD = 0.99) to immediate post-test (M = 3.17, SD = 0.97), whereas there was a nonsignificant
increase for the control group from pre (M = 3.39, SD = 1.06) to immediate post-test (M = 3.43, SD =
1.03). Students who received ETS also showed a 14% decrease in anticipated risks from pre to 2-months
post-test. The group by time interaction and time main effect for anticipated benefits (DES-AB) was non-
significant, with and without predictors in the model. The magnitude of the differences in the DES-AR
and DES-AB means between groups across all time points was small (Cohen’s d = .21 and .23,
respectively).
Qualitative Results
The 95 participants who were present for Time 2 for the ETS presentation provided written
feedback. The majority of these participants reported satisfaction with and positive feedback for the ETS
16
Gender was a significant predictor of DES-AR at Time 1 and Time 2
17
School grade was a significant predictor of DES-AR at Time 2 and Time 4
18
Self-concept clarity was a significant predictor of DES-AR at all timepoints

HIGH SCHOOL STIGMA INTERVENTION 19
presentation. Specific themes were identified within two sections: “Aspects Liked Best” and “Suggestions
for Improvement” (see Table 3 for a summary). Regarding “Aspects Liked Best,” participants
highlighted the importance of personal stories and experiences (e.g.,“The presenters really opened up to
us about their personal life and struggles”). Similarly, participants reported liking the psychoeducation,
specifically the resources handed out (e.g., “I liked the references we were given at the end for us to
use”), in addition to the video/visual images component of the presentation. Lastly, some students
specifically highlighted the presenters’ presentation style (e.g., “I felt the presentation was well organized
and spoken”) and expressed overall positive emotion toward the presentation. Regarding Suggestions for
Improvement, students reported wanting more psychoeducation and resources (e.g., “go into what
differentiates certain mental illness”) and wanting to hear more personal stories and experiences. A
smaller number of students wanted more videos and visuals, with only four students suggesting the
presentation be more concise. A larger number of students indicated that there should be more interaction
and discussion (e.g., “make it more interactive with students”)
[Table 3 here]
Discussion
Findings from the current study indicated that ETS (compared to the active control) had a small
but significant impact on negative stereotypes and mental health knowledge, consistent with results from
prior youth stigma reduction studies (Corrigan et al., 2012). Trends further indicated potential positive
effects for ETS participants in regard to reduced intentions to socially distance from people with mental
illness, reduced negative affect, and increased intentions to seek counseling. Consistent with the
hypothesis, effects were typically the strongest at Time 2 (immediate follow-up).
The current study addressed gaps in the current research and had several strong design and
methodological components. First, this study connected adolescent mental health stigma to a model and
theory of stigma. Second, researchers partnered with a national organization and evaluated a standardized
program that can be replicated, improved, and potentially dismantled in the future, in order to identify key
ingredients. Third, this study used an RCT design with an active control group and three follow-up points,

HIGH SCHOOL STIGMA INTERVENTION 20
along with controls for covariates. Studies of youth stigma reduction programs have rarely employed
randomized design (including an active control group), and even fewer have used follow-ups beyond a
pre- and post-test. Fourth, from a recruitment perspective, this study used passive parent/guardian
consent, leading to a high enrollment rate in the study. In the future, researchers may consider advocating
for passive parent/guardian consent or for mature minors’ participation without caregiver permission (see
American Psychological Association, 2018).
Contrary to hypotheses for primary outcomes, no changes were observed in regard to reduced
negative stereotypes in terms of categorical thinking, or improved intentions to help a peer with a mental
health problem, but the effects of time were significant for these models. For categorical thinking, it is
possible that content in both presentations (ETS and active control) provided students with a more
nuanced and realistic view of mental health. Additionally, it is possible that maturation, a threat to
internal validity, occurred for all students, whereby normal developmental changes led to less
dichotomous thinking about mental illness among students. In regard to intentions to provide peer
support, it is possible that the ETS presentation needs to provide students with more concrete information
on how to provide peer support. This was an area for improvement suggested in a prior post-test only
evaluation of ETS (DeLuca et al., 2018).
In regard to secondary outcomes, effects were weaker overall compared to primary outcomes,
consistent with hypotheses. It is unclear why perceptions of stigma for seeking help (PSOSH) generally
decreased in the control group over time. Research using PSOSH as an outcome variable in stigma
reduction programs is mixed (Hackler, 2011; Lopez, 2018; McGuire-Wise, 2016; Setti et al., 2019).
Given this and also that perceived and anticipated stigma are among the strongest predictors of help-
seeking (Clement et al., 2015; Gulliver et al., 2010), this is an important topic to target and evaluate in
future studies. In regard to self-stigma of seeking help (SSOSH) as an outcome, no significant interaction
was found in the current study. It is also possible that there was no effect on self-stigma, since ETS does
not explicitly focus on self-stigma. Future studies should target and measure this aspect of stigma, since

HIGH SCHOOL STIGMA INTERVENTION 21
meta-analyses and systematic reviews have found that self-stigma is one of the strongest predictors of
help-seeking behaviors (e.g., Nam et al., 2013).
A significant group by time interaction was found for anticipated risks of disclosing to a
counselor, whereby participants in the ETS group anticipated lower risk over time, which became
significant at Time 4 between groups. Based on these results, students who saw the ETS presentation felt
more comfortable—and less vulnerable—in potentially disclosing personal feelings and information to a
mental health counselor. This is a potentially important finding, because past meta analyses have found
that disclosure worries are predictors of help-seeking intentions (Nam et al., 2013). This is only the
second study to measure disclosure worries as an outcome of a stigma reduction program (Demyan &
Anderson, 2012). Male gender, being an 11th or 12th grade student, and having a stronger self-concept
were predictors of lower anticipated risk. Adolescent males generally perceive less risk and are more
willing to engage in riskier behaviors than adolescent females (Reniers, Murphy, Lin, Bartolomé, &
Wood, 2016), which may partially explain this result, however mental health stigma tends to be higher
among young males. More research is warranted in this area. More broadly, higher grade level was also a
predictor lower stigma across several outcomes (ATSMI-AV, MAKS, DES-AR) and this variable should
continue to be studied. In regard to gender, identifying as female was a predictor of lower stigma across
many outcomes as well. Some researchers (Koller & Stuart, 2016) have suggested that future
interventions may need to incorporate gender-specific stigma reduction programming. Compared to prior
ETS studies, these results confirm the benefits of ETS in terms of reducing stereotypes and negative
affect and improving knowledge (Taniyama, 2016; Wahl et al., 2018; Wong et al., 2015). Wong and
colleagues (2015) similarly found no significant impact of the ETS intervention on peer support or help-
seeking though, again, the results in regard to personal help-seeking intentions were trending. Similar to
Wahl and colleagues’ (2018) study, the effects of the intervention in the current study appeared to
generally decrease over time for some measures.
Qualitative results. Participants who were audience members for the ETS presentation overall
had positive impressions of the program. Students most enjoyed the personal story part of the

HIGH SCHOOL STIGMA INTERVENTION 22
presentation, followed by the educational information. Students also believed the presenters were credible
and competent, which is an important factor for programs (Cerully, Collins, Wong, Seelam, & Yu, 2018).
In terms of suggestions, students believed that ETS could be improved by including more personal
stories, education, and videos/visuals. Most importantly, many students suggested that future
presentations incorporate more interaction and discussion. This suggestion is consistent with prior
evaluations of ETS (DeLuca et al., 2018), which found that students wanted more encouragement from
presenters to participate in the presentation. This suggestion is also consistent with calls for adolescent
stigma reduction programs to consider cognitive and socio-emotional features of adolescent development
(Newcomb-Anjo, 2018). Other researchers have suggested interactive interventions for youth via active
learning strategies, incorporating youth stories, and promoting youth leadership (Ahmad et al., 2019;
Austin & Schwartz, 2018; Bulanda, Bruhn, Byro-Johnson, & Zentmyer, 2014).
Limitations of Current Study and Future Directions
The current study only sampled one high school in an urban area of the US and was
underpowered as per a cluster randomized controlled trial power analysis. Although the demographics of
this high school were generally reflective of NYC public high schools, it is possible that this sample was
different in some ways from other schools (e.g., potentially having more baseline interest in mental
health, given their agreement to participate in the study). In terms of design, although a randomized
design with an active control group was used, it is possible that bias was unintentionally introduced by the
researchers. To this end, a selection bias may have been present whereby teachers with mental health
contact and/or strong beliefs about mental health education were more willing to participate. Further,
given that only one school participated in this study, it is possible that students in different randomized
groups spoke about the presentations after Time 2 (i.e., contamination), though efforts were made to
conceal the true purpose of the study, and having a control and treatment group within the same school
helped to control for disentangle internal validity factors (e.g., history, maturation) within a same-school
context.

HIGH SCHOOL STIGMA INTERVENTION 23
In terms of the presentation, although the same speakers and format were used for each class,
some presentations were slightly shorter than others (e.g., due to starting late), and no presentation lasted
the full 50 minutes (but instead 35-40 minutes). Though this aspect of the study may give more weight to
the findings in regard to ecological validity, it is possible that this shortened presentation format
decreased the impact of the intervention. Corrigan and colleagues (2010) have found that another NAMI
stigma intervention (for adults) is equally effective in a 90-minute format (original design) and 30-minute
format (pared down design), but more research is needed on adapting ETS. Relatedly, the speakers in this
study primarily shared personal stories regarding depression and an eating disorder. There are specific
stigmas toward eating disorders (e.g., beliefs about fragility and attention-seeking; see Roehrig &
McLean, 2010), but also some common stigmas that are endorsed across mental health conditions (e.g.,
personal responsibility). Further, depression tends to be less stigmatized than other mental health
experiences, such as psychosis (Pescosolido et al., 2013). It is possible that personal stories about other
mental health diagnoses may have different effects on the outcome, though few studies collect this
information and researchers have called for this to be an area of future investigation (Koller & Stuart,
2016). In terms of content, ETS’ psychoeducation component is primarily psychosocial in nature (e.g.,
describing the effects of stress and environment on mental health, and how to use social support), but
there is a brief discussion of biological aspects of mental health. The effect of stigma reduction programs
may differ based on educational content and future studies should consider this (Ojio et al., 2020). The
effect of stigma reduction programs can also differ based on multiple forms of contact (e.g., Deb et al.,
2019) and other presenter factors. Lastly, to this end, although the two speakers were diverse in this study
and there was one young presenter, future research should continue to monitor the impact of speaker
demographics on outcomes, and try to match speaker demographics (e.g., age, race/ethnicity, language)
with student demographics when possible (see Chen et al., 2016). ETS and other mental health awareness
and stigma reduction presentations can also be studied and developed with elementary and middle school
students. Overall, programs must also consider intersectionality more broadly and acknowledge how

HIGH SCHOOL STIGMA INTERVENTION 24
mental health intersects with race/ethnicity, gender, age, class, and sexual orientation, in terms of both
perception and personal experience (Corrigan, Rüsch, & Scior, 2018; DuPont-Reyes et al., 2019).
In terms of measurement, some scales had low internal consistency and results should be
interpreted with this caveat. Moreover, some scale ranges in this study were relatively constricted, with
students in this study skewing toward being relatively non-stigmatizing. Ceiling effects such as these have
been noted in prior studies (Evans-Lacko et al., 2011) and multidimensional measures of stigma (as well
as measures of social desirability bias) should continue to be incorporated in future research.
Additionally, all scales were self-report and not diagnosis-specific, and no objective measures of behavior
were included in the current study. Future studies should explore the effects of stigma reduction programs
by evaluating stigma toward various mental health conditions, not just “mental health” or “mental illness”
in general. Future studies should also employ in-depth pre-test assessments of mental health knowledge
and conceptualizations, in order to determine what youth believe mental illness is.
Conclusions
The results suggest that NAMI’s Ending the Silence is well-liked by youth and has positive
effects on multiple stigma dimensions for high school youth. As a standardized program within a national
organization, ETS can be developed further to continue improving mental health knowledge, reducing
stigma, and increasing inclusion and help-seeking behaviors, and to maintain these effects over time.
Instead of acting as a solo intervention, ETS may work best with booster sessions and in tandem with
mental health school curriculum approaches (e.g., Milin et al., 2016), youth-involved community
approaches (e.g., Ramey & Rose-Krasnor, 2015), youth-led stigma reduction and mental health
promotion approaches (e.g., Bulanda et al., 2014; Eisenstein et al., 2019; Parikh et al., 2018), and with
other youth social justice initiatives (e.g., Corrigan, Watson, Byrne, & Davis, 2005; Mayberry, 2013).
Future investigations into the manifestations of stigma and ways to reduce stigma can lead to better
understandings of the stigma process and improvements in life outcomes for youth.
References

HIGH SCHOOL STIGMA INTERVENTION 25
Ahmad, S. I., Leventhal, B. L., Nielsen, B. N., & Hinshaw, S. P. (2019). Reducing mental-illness
stigma via high school clubs: A matched-pair, cluster-randomized trial. Stigma and
Health. Advance online publication. http://dx.doi.org/10.1037/sah0000193
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders
(5th ed.). Arlington, VA: American Psychiatric Publishing.
American Psychological Association (2018). APA resolution on support for the expansion of
mature minors’ ability to participate in research. Retrieved from
https://www.apa.org/about/policy.
Auerbach, R. P., Mortier, P., Bruffaerts, R., Alonso, J., Benjet, C., Cuijpers, P., ... & Murray, E.
(2018). WHO world mental health surveys international college student project:
Prevalence and distribution of mental disorders. Journal of Abnormal Psychology.
Advance online article.
Austin, L. J., & Schwartz, S. E. (2018). Addressing mental health stigma in early adolescence:
Middle school antistigma interventions. Adolescent Research Review, 1-11.
Bulanda, J. J., Bruhn, C., Byro-Johnson, T., & Zentmyer, M. (2014). Addressing mental health
stigma among young adolescents: evaluation of a youth-led approach. Health & Social
Work, 39(2), 73-80.
Campbell, M. K., Mollison, J., Steen, N., Grimshaw, J. M., & Eccles, M. (2000). Analysis of
cluster randomized trials in primary care: a practical approach. Family Practice, 17(2),
192-196. https://doi.org/10.1093/fampra/17.2.192
Campbell, M. K., Piaggio, G., Elbourne, D. R., & Altman, D. G. (2012). CONSORT 2010
statement: Extension to cluster randomised trials. BMJ, 345, 1-21.

HIGH SCHOOL STIGMA INTERVENTION 26
Campbell, J. D., Trapnell, P. D., Heine, S. J., Katz, I. M., Lavallee, L. F., & Lehman, D. R.
(1996). Self-concept clarity: Measurement, personality correlates, and cultural
boundaries. Journal of Personality and Social Psychology, 70(1), 141-156.
http://dx.doi.org/10.1037/0022-3514.70.6.1114
Cash, T. F., Begley, P. J., McCown, D. A., & Weise, B. C. (1975). When counselors are heard
and not seen: Initial impact on physical attractiveness. Journal of Counseling Psychology,
22(4), 273-279. https://doi.org/10.1037/h0076730
Cauce, A. M., Domenech-Rodríguez, M., Paradise, M., Cochran, B. N., Shea, J. M., Srebnik, D.,
& Baydar, N. (2002). Cultural and contextual influences in mental health help seeking: A
focus on ethnic minority youth. Journal of Consulting and Clinical Psychology, 70(1),
44-55. doi: 10.1037//0022-006X.70.1.44
Cerully, J. L., Collins, R. L., Wong, E., Seelam, R., & Yu, J. (2018). Differential response to
contact-based stigma reduction programs: Perceived quality and personal experience
matter. Psychiatry Research, 259, 302-309.
Chen, S. P., Koller, M., Krupa, T., & Stuart, H. (2016). Contact in the classroom: Developing a
program model for youth mental health contact-based anti-stigma education. Community
Mental Health Journal, 52(3), 281-293. doi: 10.1007/s10597-015-9944-7
Chisholm, K., Patterson, P., Torgerson, C., Turner, E., Jenkinson, D., & Birchwood, M. (2016).
Impact of contact on adolescents’ mental health literacy and stigma: the SchoolSpace
cluster randomised controlled trial. BMJ Open, 6(2), 1-11.
Clement, S., Schauman, O., Graham, T., Maggioni, F., Evans-Lacko, S., Bezborodovs, N. ... &
Thornicroft, G. (2015). What is the impact of mental health-related stigma on help-

HIGH SCHOOL STIGMA INTERVENTION 27
seeking? A systematic review of quantitative and qualitative studies. Psychological
Medicine, 45(1), 11-27. doi: 10.1017/S0033291714000129
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. New York,
NY:Routledge.
Corrigan, P. W. (2004). How stigma interferes with mental health care. American Psychologist,
59(7), 614-625. doi: 10.1037/0003-066X.59.7.614
Corrigan, P. W., & Al-Khouja, M. A. (2018). Three agendas for changing the public stigma of
mental illness. Psychiatric Rehabilitation Journal, 41(1), 1-7.
Corrigan, P. W., Druss, B. G., & Perlick, D. A. (2014). The impact of mental illness stigma on
seeking and participating in mental health care. Psychological Science in the Public
Interest, 15(2), 37–70. doi:10.1177/1529100614531398
Corrigan, P. W., Lurie, B. D., Goldman, H. H., Slopen, N., Medasani, K., & Phelan, S. (2005).
How adolescents perceive the stigma of mental illness and alcohol abuse. Psychiatric
Services, 56(5), 544–550. doi:10.1176/appi.ps.56.5.544
Corrigan, P., Michaels, P. J., & Morris, S. (2015). Do the effects of antistigma programs persist
over time? Findings from a meta-analysis. Psychiatric Services, 66(5), 543-546. doi:
10.1176/appi.ps.201400291
Corrigan, P. W., Morris, S. B., Michaels, P. J., Rafacz, J. D., & Rüsch, N. (2012). Challenging
the public stigma of mental illness: A meta-analysis of outcome studies. Psychiatric
Services, 63(10), 963-973. doi: 10.1176/appi.ps.201100529
Corrigan, P. W., Rafacz, J. D., Hautamaki, J., Walton, J., Rüsch, N., Rao, D., ... & Reeder, G.
(2010). Changing stigmatizing perceptions and recollections about mental illness: The

HIGH SCHOOL STIGMA INTERVENTION 28
effects of NAMI’s In Our Own Voice. Community Mental Health Journal, 46(5), 517-
522.
Corrigan, P. W., Rowan, D., Green, A., Lundin, R., River, P., Uphoff-Wasowski, K., ... &
Kubiak, M. A. (2002). Challenging two mental illness stigmas: Personal responsibility
and dangerousness. Schizophrenia Bulletin, 28(2), 293-309.
http://psycnet.apa.org/doi/10.1093/oxfordjournals.schbul.a006939
Corrigan, P. W., Rüsch, N., & Scior, K. (2018). Adapting disclosure programs to reduce the
stigma of mental illness, Psychiatric Services, 69(7), 826-828.
Corrigan, P. W., Watson, A. C., Byrne, P., & Davis, K. E. (2005). Mental illness stigma:
Problem of public health or social justice?. Social Work, 50(4), 363-368.
http://dx.doi.org/10.1093/sw/50.4.363
Corrigan, P. W., Watson, A. C., Otey, E., Westbrook, A. L., Gardener, A. L., Lamb, T. A., &
Fenton, W. S. (2007). How do children stigmatize people with mental illness? Journal of
Applied Social Psychology, 37(7), 1405–1417. doi:10.1111/j.1559-1816.2007.00218.x
Deb, T., Lempp, H., Bakolis, I., Vince, T., Waugh, W., Henderson, C., and the INDIGO READ
Study Group (2019). Responding to experienced and anticipated discrimination (READ):
anti-stigma training for medical students towards patients with mental illness–study
protocol for an international multisite non-randomised controlled study. BMC Medical
Education, 19, 41-49. https://doi.org/10.1186/s12909-019-1472-7
DeLuca, J. S. (2019). Conceptualizing adolescent mental illness stigma: Youth stigma
development and stigma reduction programs. Adolescent Research Review, 1-19.
https://doi.org/10.1007/s40894-018-0106-3

HIGH SCHOOL STIGMA INTERVENTION 29
DeLuca, J. S., Evans, M., Reyes, I., & Yanos, P. T. (2016). [An initial evaluation of In Our Own
Voice in a diverse high school]. Unpublished raw data.
DeLuca, J. S., Evans, M., & Yanos, P. T. (2018). [Findings from implementation of the national
alliance on mental illness’ ‘Ending the Silence’ in New York City: Methodological and
theoretical considerations for adolescent stigma reduction programs]. Unpublished raw
data.
Demyan, A. L., & Anderson, T. (2012). Effects of a brief media intervention on expectations,
attitudes, and intentions of mental health help seeking. Journal of Counseling
Psychology, 59(2), 222-229. https://psycnet.apa.org/doi/10.1037/a0026541
Dolphin, L., & Hennessy, E. (2016). Depression stigma among adolescents in Ireland. Stigma
and Health, 1(3), 185-200http://dx.doi.org/10.1037/sah0000025
DuPont-Reyes, M. J., Villatoro, A. P., Phelan, J. C., Painter, K., & Link, B. G. (2019).
Adolescent views of mental illness stigma: An intersectional lens. American Journal of
Orthopsychiatry. Advance online publication. https://doi.org/10.1037/ort0000425
Eisenstein, C., Zamperoni, V., Humphrey, N., Deighton, J., Wolpert, M., Rosan, C., ... &
Edbrooke-Childs, J. (2019). Evaluating the Peer Education Project in secondary schools.
Journal of Public Mental Health, 18(1), 58-65.
Evans-Lacko, S., Little, K., Meltzer, H., Rose, D., Rhydderch, D., Henderson, C., … &
Thornicroft, G. (2010). Development and psychometric properties of the mental health
knowledge schedule. Canadian Journal of Psychiatry, 55(7), 440-448.
https://doi.org/10.1177/070674371005500707
Evans-Lacko, S., Rose, D., Little, K., Flach, C., Rhydderch, D., Henderson, C., ... & Thornicroft,
G. (2011). Development and psychometric properties of the reported and intended

HIGH SCHOOL STIGMA INTERVENTION 30
behaviour scale (RIBS): A stigma-related behaviour measure. Epidemiology and
Psychiatric Sciences, 20(3), 263-271. doi:10.1017/S2045796011000308
Goffman, E. (1963). Stigma: Notes on the management of spoiled identity. Englewood Cliffs, NJ:
Prentice Hall.
Gulliver, A., Griffiths, K. M., & Christensen, H. (2010). Perceived barriers and facilitators to
mental health help-seeking in young people: A systematic review. BMC Psychiatry,
10(1), 113-121. http://doi.org/10.1186/1471-244X-10-113
Hackler, A. H. (2011). Contact and stigma toward mental illness: Measuring the effectiveness of
two video interventions (Doctoral dissertation, Iowa State University, Ames, United
States of America). Retrieved from https://lib.dr.iastate.edu/
Hartman, L. I., Michel, N. M., Winter, A., Young, R. E., Flett, G. L., & Goldberg, J. O. (2013).
Self-stigma of mental illness in high school youth. Canadian Journal of School
Psychology, 28(1), 28-42.
Hill, C. E., Knox, S., Thompson, B. J., Williams, E. N., Hess, S. A., & Ladany, N. (2005).
Consensual qualitative research: An update. Journal of Counseling Psychology, 52(2),
196-205.
Hill, C. E., Thompson, B. J., & Williams, E. N. (1997). A guide to conducting consensual
qualitative research. The Counseling Psychologist, 25(4), 517-572
Kaufman, R. (2018, July 2). New York, Virginia become first to require mental health education
in schools. CNN. https://www.cnn.com/2018/07/02/health/mental-health-schools-
bn/index.html

HIGH SCHOOL STIGMA INTERVENTION 31
Kobau, R., DiIorio, C., Chapman, D., & Delvecchio, P. (2010). Attitudes about mental illness
and its treatment: Validation of a generic scale for public health surveillance of mental
illness associated stigma. Community Mental Health Journal, 46(2), 164-176.
doi:10.1007/s10597-009-9191-x
Koller, M., & Stuart, H. (2016). Reducing stigma in high school youth. Acta Psychiatrica
Scandinavica, 134, 63-70.doi: 10.1111/acps.12613
Lopez, R. A. (2018). Effects of video interventions on perceptions of mental health stigma in
Latino college students (Master’s thesis, California State University, Stanislaus, Turlock,
United States of America). Retrieved from http://stanislaus-scholarworks.calstate.edu
Link, B. G., & Phelan, J. C. (2001). Conceptualizing stigma. Annual Review of Sociology, 27(1),
363-385. doi: 10.1146/annurev.soc.27.1.363
Link, B. G., Yang, L. H., Phelan, J. C., & Collins, P. Y. (2004). Measuring mental illness stigma.
Schizophrenia Bulletin, 30(3), 511–541. doi:10.1093/oxfordjournals.schbul.a007098
Mansfield, R., Humphrey, N., & Patalay, P. (2019). Psychometric validation of the Reported and
Intended Behavior Scale (RIBS) with adolescents. Stigma and Health. Advance online
publication. https://doi.org/10.1037/sah0000200
Mayberry, M. (2013). Gay-straight alliances: Youth empowerment and working toward reducing
stigma of LGBT youth. Humanity & Society, 37(1), 35-54.
McGuire Wise, S. D. (2016). The effects of anti-stigma interventions in resident advisors'
attitudes toward mental illness (Doctoral dissertation, University of Toledo, Toledo,
United States of America). Retrieved from https://etd.ohiolink.edu
Mellor, C. (2014). School-based interventions targeting stigma of mental illness: systematic
review. The Psychiatric Bulletin, 38(4), 164-171. doi: 10.1192/pb.bp.112.041723

HIGH SCHOOL STIGMA INTERVENTION 32
Merikangas, K. R., He, J. P., Burstein, M., Swendsen, J., Avenevoli, S., Case, B., ... & Olfson,
M. (2011). Service utilization for lifetime mental disorders in US adolescents: results of
the National Comorbidity Survey–Adolescent Supplement (NCS-A). Journal of the
American Academy of Child & Adolescent Psychiatry, 50(1), 32-45.
Milin, R., Kutcher, S., Lewis, S. P., Walker, S., Wei, Y., Ferrill, N., & Armstrong, M. A. (2016).
Impact of a mental health curriculum on knowledge and stigma among high school
students: A randomized controlled trial. Journal of the American Academy of Child &
Adolescent Psychiatry, 55(5), 383-391.
Nam, S. K., Choi, S. I., Lee, J. H., Lee, M. K., Kim, A. R., & Lee, S. M. (2013). Psychological
factors in college students' attitudes toward seeking professional psychological help: A
meta analysis. Professional Psychology: Research and Practice, 44(5), 37-45.
http://psycnet.apa.org/doi/10.1037/a0029562
National Alliance on Mental Illness (2015). Ending the silence overview 2015. Retrieved from
https://www.nami.org
Newcomb-Anjo, S. E. (2019). Applying what is known about adolescent development to
improve school-based mental health literacy of depression interventions: Bridging
research to practice. Adolescent Research Review, 4, 235-248.
Ojio, Y., Yamaguchi, S., Ohta, K., Ando, S., & Koike, S. (2020). Effects of biomedical messages
and expert-recommended messages on reducing mental health-related stigma: a
randomised controlled trial. Epidemiology and Psychiatric Sciences, 29, 1-9. https://
doi.org/10.1017/S2045796019000714
Parikh, S. V., Taubman, D. S., Antoun, C., Cranford, J., Foster, C. E., Grambeau, M., ... &
Salazar, S. (2018). The Michigan Peer-to-Peer Depression Awareness Program: School-
based prevention to address depression among teens. Psychiatric Services, 69(4), 487-

HIGH SCHOOL STIGMA INTERVENTION 33
491.
Perry, Y., Petrie, K., Buckley, H., Cavanagh, L., Clarke, D., Winslade, M., ... & Christensen, H.
(2014). Effects of a classroom-based educational resource on adolescent mental health
literacy: A cluster randomised controlled trial. Journal of Adolescence, 37(7), 1143-1151.
Pescosolido, B. A., & Martin, J. K. (2015). The stigma complex. Annual Review of
Sociology, 41, 87-116. https://doi.org/10.1146/annurev-soc-071312-145702
Pescosolido, B. A, Medina, T. R., Martin, J. K., & Long, J. S. (2013). The “backbone” of stigma:
identifying the global core of public prejudice associated with mental illness. American
Journal of Public Health, 103(5), 853–860. doi:10.2105/AJPH.2012.301147
Pinto, M. D., Hickman, R., Logsdon, M. C., & Burant, C. (2012). Psychometric evaluation of the
revised attribution questionnaire (r-AQ) to measure mental illness stigma in adolescents.
Journal of Nursing Measurement, 20(1), 47-58. https://doi.org/10.1891/1061-
3749.20.1.47
Pinto-Foltz, M. D., Logsdon, M. (2009). Conceptual model of research to reduce stigma related
to mental disorders in adolescents. Issues in Mental Health Nursing, 30, 788–795.
http://doi.org/10.3109/01612840903267620
Pinto-Foltz, M. D., Logsdon, M. C., & Myers, J. A. (2011). Feasibility, acceptability, and initial
efficacy of a knowledge-contact program to reduce mental illness stigma and improve
mental health literacy in adolescents. Social Science & Medicine, 72, 2011-2019.
https://doi.org/10.1016/j.socscimed.2011.04.006
Ramey, H. L., & Rose-Krasnor, L. (2015). The new mentality: Youth–adult partnerships in
community mental health promotion. Children and Youth Services Review, 50, 28-37.

HIGH SCHOOL STIGMA INTERVENTION 34
Reniers, R. L., Murphy, L., Lin, A., Bartolomé, S. P., & Wood, S. J. (2016). Risk perception and
risk-taking behaviour during adolescence: the influence of personality and gender. PloS
One, 11, 1-14.
Roehrig, J. P., & McLean, C. P. (2010). A comparison of stigma toward eating disorders versus
depression. International Journal of Eating Disorders, 43, 671-674.
Salerno, J. P. (2016). Effectiveness of universal school-based mental health awareness programs
among youth in the sUnited States: A systematic review. Journal of School Health, 86,
922–931. http://doi.org/10.1111/josh.12461
Schachter, H. M., Girardi, A., Ly, M., Lacroix, D., Lumb, A. B., van Berkom, J., & Gill, R.
(2008). Effects of school-based interventions on mental health stigmatization: A
systematic review. Child and Adolescent Psychiatry and Mental Health, 2, 1-14.
doi:10.1186/1753-2000-2-18
Setti, V. P. C., Loch, A. A., Modelli, A., de Almeida Rocca, C. C., Hungerbuehler, I., van de
Bilt, M. T., ... & Rössler, W. (2019). Disclosing the diagnosis of schizophrenia: A pilot
study of the ‘Coming Out Proud’ intervention. International Journal of Social
Psychiatry, Advance online publication.
Silke, C., Swords, L., & Heary, C. (2016). The development of an empirical model of mental
health stigma in adolescents. Psychiatry Research, 242, 262–270.
http://doi.org/10.1016/j.psychres.2016.05.033
Shek, D. T., & Ma, C. (2011). Longitudinal data analyses using linear mixed models in SPSS:
Concepts, procedures and illustrations. The Scientific World Journal, 11, 42-76.
http://dx.doi.org/10.1100/tsw.2011.2
Spencer, M. S., Chen, J., Gee, G. C., Fabian, C. G., & Takeuchi, D. T. (2010). Discrimination
and mental health–related service use in a national study of Asian Americans. American

HIGH SCHOOL STIGMA INTERVENTION 35
Journal of Public Health, 100, 2410-2417. doi: 10.2105/AJPH.2009.176321
Taniyama, S. L. (2016). An evaluation of the effectiveness of the National Alliance of Mental
Illness’s Ending the Silence program. Unpublished master’s thesis, California
Po;ytechnic State University, San Luis Obispo
Thornicroft, G., Mehta, N., Clement, S., Evans-Lacko, S., Doherty, M., Rose, D., ... &
Henderson, C. (2016). Evidence for effective interventions to reduce mental-health
related stigma and discrimination. The Lancet, 387, 1123-1132.
doi:10.1192/bjp.bp.114.151944
Vogel, D. L., Wade, N. G., & Ascheman, P. L. (2009). Measuring perceptions of stigmatization
by others for seeking psychological help: Reliability and validity of a new stigma scale
with college students. Journal of Counseling Psychology, 56, 301–308.
doi:10.1037/a0014903
Vogel, D. L., Wade, N. G., & Haake, S. (2006). Measuring the self-stigma associated with
seeking psychological help. Journal of Counseling Psychology, 53, 325–337.
doi:10.1037/0022-0167.53.3.325
Vogel, D. L., & Wester, S. R. (2003). To seek help or not to seek help: The risks of self-
disclosure. Journal of Counseling Psychology, 50, 351-361.
http://dx.doi.org/10.1037/0022-0167.50.3.351
Wahl, O., Rothman, J., Brister, T., & Thompson, C. (2018). Changing student attitudes about
mental health conditions: NAMI ending the silence. Stigma and Health. Advance online
publication.
Watson, A. C., Miller, F. E., & Lyons, J. S. (2005). Adolescent attitudes toward serious mental
illness. The Journal of Nervous and Mental Disease, 193, 769–772.
doi:10.1097/01.nmd.0000185885.04349.99

HIGH SCHOOL STIGMA INTERVENTION 36
Watson, A. C., Otey, E., Westbrook, A. L., Gardner, A. L., Lamb, T. A., Corrigan, P. W., &
Fenton, W. S. (2004). Changing middle schoolers' attitudes about mental illness through
education. Schizophrenia Bulletin, 30, 563-572.
Wei, Y., Hayden, J. A., Kutcher, S., Zygmunt, A., & McGrath, P. (2013). The effectiveness of
school mental health literacy programs to address knowledge, attitudes and help seeking
among youth. Early Intervention in Psychiatry, 7, 109-121.
Winkler, P., Janoušková, M., Kožený, J., Pasz, J., Mladá, K., Weissová, A., ... & Evans-Lacko,
S. (2017). Short video interventions to reduce mental health stigma: A multi-centre
randomised controlled trial in nursing high schools. Social Psychiatry and Psychiatric
Epidemiology, 52, 1549-1557. https://doi.org/10.1007/s00127-017-1449-y
Wood, A. L., & Wahl, O. F. (2006). Evaluating the effectiveness of a consumer-provided mental
health recovery education presentation. Psychiatric Rehabilitation Journal, 30, 46-53.
http://psycnet.apa.org/doi/10.2975/30.2006.46.53
Wong, E. C., Collins, R. L., Cerully, J. L., Roth, E., Marks, J. S., Yu, J., … Voice, O. (2015).
Effects of stigma and discrimination reduction trainings conducted under the California
mental health services authority: An evaluation of NAMI’s Ending the Silence. Retrieved
from http://www.rand.org/pubs
World Health Organization (2005). “Atlas: Child and adolescent mental health resources: Global
concerns, implications for the future” [PDF file]. Retrieved from
http://www.who.int/mental_health/resources/Child_ado_atlas.pdf
Yamaguchi, S., Mino, Y., & Uddin, S. (2011). Strategies and future attempts to reduce
stigmatization and increase awareness of mental health problems among young people: A
narrative review of educational interventions. Psychiatry and Clinical Neurosciences, 65,
405–415. http://doi.org/10.1111/j.1440-1819.2011.02239.x

HIGH SCHOOL STIGMA INTERVENTION 37

HIGH SCHOOL STIGMA INTERVENTION 38
Table 1.
Sociodemographic Characteristics of Participants at Baseline
ETS
N = 105
Control
N = 101
Total
N = 206
N (%)
N (%)
N (%)
Χ2
df
p
Gender:
Male
Female
Transgender
Gender Fluid
46 (45)
55 (54)
0 (0)
1 (1)
41 (41)
59 (58)
1 (1)
0 (0)
87 (43)
114 (56)
1 (.5)
1 (.5)
.26
1
.61
Grade
9
10
11
12
0 (0)
57 (54)
31 (30)
17 (16)
17 (17)
31 (31)
41 (41)
12 (12)
17 (8)
88 (43)
72 (35)
29 (14)
26.87
3
<.0005
Race/Ethnicity:
African-American
European-American
Latino/a/x
Asian-American
Arab/Middle-Eastern
Native American
Multiethnic/racial
Other
21 (20)
40 (38)
19 (18)
13 (12)
2 (2)
0 (0)
8 (8)
2 (2)
22 (22)
32 (32)
13 (13)
15 (15)
3 (3)
0 (0)
16 (16)
0 (0)
43 (21)
72 (35)
32 (16)
28 (14)
5 (2)
0 (0)
24 (12)
2 (1)
5.03
5
.41
Close Friend Contact
Yes
No
37 (35)
68 (65)
31 (31)
69 (69)
68 (33)
137 (67)
.25
1
.62
Family Contact:
Yes
No
28 (27)
77 (73)
26 (26)
74 (74)
54 (26)
151 (74)
<.0005
1
.99
M (SD)
M (SD)
M (SD)
t
df
p
Age
15.3
(.86)
15.5
(1.01)
15.4 (.94)
-1.37
204
.17
Note. ETS refers to Ending the Silence. Some values do not add to 206 because of data cleaning,
rounding, or missing responses. χ2 analyses for race/ethnicity omitted the group “Other” and “Native
American” since these groups violated the assumption of the analysis (< 5 cases).

HIGH SCHOOL STIGMA INTERVENTION 39
Table 2.
Estimated Marginal Means and Standard Errors for Mental Health Stigma Outcomes for Randomized
Groups
Measures
Randomized
Group
Baseline
n = 198
Immediate
Post-Test
n = 187
1-month
n = 181
2-months
n = 171
Group by Time
Interaction
M
SD
M
SD
M
SD
M
SD
ES
F
df
p
Negative Stereotypes
(AMIS)
ETS
2.15
.46
1.99
.54
2.05
.49
2.15
.54
.44
3.55
3,
481
.014
Control
2.22
.43
2.30
.52
2.21
.46
2.28
.50
Categorical Thinking
(ATSMI-AV)
ETS
1.93
.66
1.77
.62
1.84
.70
1.83
.62
.14
.207
3,
484
.892
Control
2.01
.70
1.91
.67
1.97
.64
1.97
.62
Intended Social
Distance (RIBS)
ETS
2.02
.72
1.77
.68
1.82
.68
1.89
.77
.06
2.20
3,
484
.087
Control
1.95
.68
1.90
.66
1.88
.67
2.00
.68
Knowledge (MAKS)
ETS
3.41
.44
3.71
.45
3.61
.53
3.62
.48
.24
3.10
3,
495
.026
Control
3.42
.43
3.51
.44
3.44
.44
3.47
.39
Negative Affect (r-
AQ)
ETS
2.01
.80
1.90
.75
2.12
.83
2.04
.79
.16
2.48
3,
480
.061
Control
1.99
.81
1.97
.79
1.92
.75
1.94
.80
Intentions to Seek
Counseling (ISCI)
ETS
2.11
.64
2.10
.68
2.13
.66
2.20
.68
.003
1.89
3,
483
.131
Control
2.22
.61
2.03
.71
2.02
.71
2.11
.80

HIGH SCHOOL STIGMA INTERVENTION 40
Note. Higher AMIS and ATSMI-AV = higher negative stereotypes. Higher RIBS = higher intended social
distance. Higher MAKS = higher mental health knowledge. Higher r-AQ = higher negative affect. Higher
ISCI and Peer Support = higher intentions to seek help and help a peer, respectively. Higher PSOSH and
SSOSH = higher perceptions of stigma and self-stigma, respectively. Higher DES-AR = higher
anticipated risk. Higher DES-AB = higher anticipated benefits. SD refers to standard deviation. ES refers
to effect size for overall mean score difference between ETS and control group (via Cohen’s d). Group by
time interaction significance level refers to full models (i.e., all predictors included).
Intentions to Provide
Peer Support (Peer)
ETS
4.03
.55
4.13
.66
3.92
.68
3.92
.68
.05
1.08
3,
472
.356
Control
4.12
.63
4.05
.61
3.95
.59
3.89
.71
Perceptions of
Stigma (PSOSH)
ETS
2.03
.74
1.87
.83
2.09
.77
2.07
.87
.02
2.96
3,
483
.032
Control
2.19
.76
1.98
.68
1.89
.69
1.96
.79
Self-Stigma
(SSOSH)
ETS
2.58
.68
2.53
.68
2.61
.66
2.57
.71
.16
1.51
3,
479
.211
Control
2.65
.67
2.69
.66
2.73
.72
2.81
.72
Anticipated Risks
(DES-AR)
ETS
3.53
.99
3.17
.97
3.20
.96
3.05
.99
.21
4.68
3,
482
.003
Control
3.39
1.06
3.43
1.03
3.41
1.14
3.41
1.08
Anticipated Benefits
(DES-AB)
ETS
3.15
.97
3.16
.94
3.18
.88
3.15
.81
.23
.313
3,
466
.816
Control
2.91
.86
3.02
.84
3.02
.84
3.02
.96

HIGH SCHOOL STIGMA INTERVENTION 41
Table 3.
Qualitative Feedback for Ending the Silence (N = 95)
Domain
Percentage Endorsed
n
Aspects Liked Best
Personal stories and experiences
31%
30
Psychoeducation
24%
23
Videos and visual images
11%
10
Presentation Style
6%
6
Overall positive emotions
6%
6
Suggestions for Improvement
More psychoeducation and resources
14%
13
More personal stories and experiences
9%
9
More videos and visuals
7%
7
More interaction and discussion
24%
23
More concise
4%
4
No suggestions
20%
19
Note. Percentages are derived by: total domain responses/95. Some totals do not equal 95, since some
participants did not provide qualitative feedback. For the category “No suggestions,” participants
explicitly stated they had no suggestions.

HIGH SCHOOL STIGMA INTERVENTION 42
Figure 1
Consolidated Standards of Reporting Trials (CONSORT) participant flow diagram.
14 high school classes
assessed for eligibility (n = 232)
Students who completed assent
(n = 208; 90%)
Students excluded (n = 26)
 Parent/guardian opt out (n = 3)
 Declined to participate or absent for
first two visits (n = 23)
Completed T3 (n = 94)
Completed T4 (n = 87)
7 classes allocated to Ending the Silence
intervention (n = 105)
 Completed pretest/T1 (n = 101)
 Received allocated intervention and
completed posttest/T2 (n = 95)
Completed T3 (n =87)
Completed T4 (n = 84)
7 classes allocated to control group (n = 101)
 Completed pretest/T1 (n = 97)
 Received allocated intervention and
completed posttest/T2 (n = 92)
Allocation
Follow-Ups
Randomized 14 classes
(N = 206)
Enrollment
Recruitment

HIGH SCHOOL STIGMA INTERVENTION 43
Figure 2.
Estimated marginal means between groups for negative stereotypes (AMIS) and mental health knowledge
(MAKS) over Time (Time 1 = baseline, Time 2 = immediate post-test, Time 3 = one month follow-up,
Time 4 = two month follow-up).