Grief and Attitudes Toward Suicide in Peers Affected by a Cluster of Suicides as Adolescents

Abstract

Eighty-five young adults exposed to a cluster of peer suicides as adolescents completed measures of attitudes toward suicide, grief, and social support. Closeness to the peers lost to suicide was positively correlated with grief and the belief that suicide is not preventable, with grief further elevated in close individuals with high social support from friends. Overall, social support was related to healthy attitudes about suicide including preventability, yet it was also related to some stigmatizing beliefs. Compared with 67 young adults who had not been exposed to a suicide cluster, the exposed sample was more likely to think that suicide is normal but more likely to think of it as incomprehensible.

Full text

Grief and Attitudes Toward Suicide in Peers
Affected by a Cluster of Suicides as
Adolescents
CAROLINE H. ABBOTT,BA,AND AUDREY L. ZAKRISKI,PHD
Eighty-five young adults exposed to a cluster of peer suicides as adoles-
cents completed measures of attitudes toward suicide, grief, and social support.
Closeness to the peers lost to suicide was positively correlated with grief and
the belief that suicide is not preventable, with grief further elevated in close
individuals with high social support from friends. Overall, social support was
related to healthy attitudes about suicide including preventability, yet it was also
related to some stigmatizing beliefs. Compared with 67 young adults who had
not been exposed to a suicide cluster, the exposed sample was more likely to
think that suicide is normal but more likely to think of it as incomprehensible.
Suicide has been a long-standing problem
in young Americans, and is consistently the
second or third leading cause of death
among adolescents ages 13 to 18 (Bondora
& Goodwin, 2005). A loss to suicide can
be devastating to the survivors, and despite
evidence that a loss to suicide can be harm-
ful to peer survivors and even trigger subse-
quent suicides in adolescent peers, there is
relatively little research on adolescent peer
survivors of suicide (Bondora & Goodwin,
2005). There is even less research on ado-
lescent peer survivors of cluster suicides
due to their rarity (Gould, Wallenstein,
Kleinman, O’Carroll, & Mercy, 1990).
General findings on adjustment of
suicide survivors suggest that those bereaved
by suicide experience more feelings of
shame, rejection, stigma, and blaming than
those bereaved by other types of death
(Sveen & Walby, 2008). Adolescents who
have lost a peer to suicide are at an
increased risk of major depression, posttrau-
matic stress disorder, and suicidal ideation
compared with adolescents who have not
lost a peer to suicide, and those having clo-
ser relationships with those lost to suicide
being at even greater risk (Brent et al.,
1993). In a recent study of adults, those
who were psychologically closer to those
who died by suicide were more likely to
self-identify as “suicide survivors” (Cerel,
Maple, Aldrich, & van de Venne, 2013). In
addition, the relational attachment between
survivors and individuals who died by either
suicide or accident has been found to be
more important in explaining grief than
relationship status (sibling or child vs. par-
ent) in sudden death bereavement (Reed &
Greenwald, 1991). These findings illumi-
nate the distress faced by suicide survivors
and the importance of closeness to the
deceased. Additionally, those who had expo-
sure to suicidal behavior and had experi-
enced their own suicidal ideations had more
accepting views toward suicide (Stein,
CAROLINE H. ABBOTT and AUDREY L.
ZAKRISKI, Connecticut College, New London,
CT, USA.
Address correspondence to Caroline
Abbott, Center for Psychosocial Epidemiology
and Outcomes Research, Dana-Farber Cancer
Institute, 450 Brookline Ave., Boston MA 02215;
E-mail: CarolineH_Abbott@dfci.harvard.edu
Suicide and Life-Threatening Behavior 1
©2014 The American Association of Suicidology
DOI: 10.1111/sltb.12100

Witzum, Brom, DeNour, & Elizur, 1992).
This shows that there is a relationship
between attitudes toward suicide and expo-
sure to suicidal behavior; however, the rela-
tionship between attitudes toward suicide
and grief in suicide survivors has not yet, to
our knowledge, been examined.
In adolescents, the risk of suicide fol-
lowing exposure to another suicide is two to
four times higher than in any other age
group (Gould et al., 1990), which can lead to
suicide “clusters.” The Centers for Disease
Control define a suicide cluster as “a group of
suicides or suicide attempts, or both, that
occurs closer together in time and space
than would normally be expected in a given
community” (O’Carroll, Mercy, & Steward,
1988, p. 1). It has been estimated that
between 1% and 2% of adolescent suicides
occur in clusters, and they are very rarely
seen in age groups over 24 (Gould et al.,
1990). Thus, it is apparent that suicide clus-
ters are a rare but serious problem within
adolescents and young adults. Given the doc-
umented effects of single suicides on adoles-
cent adjustment and attitudes toward suicide,
it is likely that exposure to a cluster of sui-
cides would have even more profound and
enduring effects, yet very little is known due
to the low base rate of cluster suicides.
Another important topic to consider
when studying responses to suicide is stigma.
In Joiner’s (2010) Myths about Suicide, the
author explored different stigmatizing myths,
including the ideas that suicide is an easy
escape, cowardly, an act of anger or revenge,
selfish, and done “on a whim.” This stigma
can complicate the grieving process for
survivors and lead to less social support for
suicide survivors than is experienced by
survivors of other deaths (Lukas & Seiden,
2007). It is likely that suicide stigma can
make survivors fear disclosure and make oth-
ers uncomfortable with providing support.
Wagner and Calhoun (1991) found that sui-
cide survivors believed that those who had
not experienced suicide could not fully
understand them and therefore could not
provide sufficient social support. Most young
suicide survivors report wanting help from
mental health professionals but, unfortu-
nately, of those who received this help, most
were generally unhappy with the professional
services rendered (Dyregrov, 2009).
In adolescents and young adults in
general, social support from friends has been
found to reduce the risk of suicidal behavior
(Hirsch & Barton, 2011), and this may also
be true in groups of adolescent suicide survi-
vors (Joiner, 1999). Understanding social
support in peer groups affected by suicides is
important because adolescent suicide survi-
vors might choose to reach out within the
affected peer group, anticipating poor social
support from others who have not been
directly exposed to suicide, and following the
normative developmental inclination to
reach out for peer rather than adult support
(Gould et al., 2004). This use of peer social
support could be helpful as Wagner and
Calhoun (1991) suggested and may even pre-
vent subsequent suicides (Joiner, 1999), but
it also could have the potential for negative
consequences on the grieving process. One
possible negative consequence could be
“corumination,” which has been defined as a
repeated focus on negative emotions that can
happen within friend groups (Rose, 2002).
Corumination is related to positive friend-
ship quality and closeness, yet it can also pre-
dict an increase in depressive symptoms
(Davila et al., 2012; Rose, 2002). We already
know that adolescents are more likely to tell
a peer about suicidal thoughts than an adult
(Gould et al., 2004) and that those adoles-
cents are not likely to report the friend’s sui-
cidal ideation (Kalafat & Elias, 1992). If
suicide survivors seek out support from other
survivors and struggling adolescents already
reach out to friends first, understanding the
quality of peer social support for peer suicide
loss is very important, especially in the case
of peer cluster suicides.
In this research we examined the
adjustment of adolescent cluster suicide sur-
vivors 2 years after experiencing the third
death due to suicide by a member of their
high school graduating class (the three sui-
cides occurred between 2005 and 2010). All
participants were classmates of the three
2P
EER SUICIDE SURVIVORS

adolescents lost to suicide, although they
varied on how close they were with those
who died. Building on past research, close-
ness to the peers lost to suicide and current
social support were used to understand both
current grief and attitudes toward suicide.
We predicted that higher levels of closeness
would be associated with lingering present
grief. Because exposure to suicidal behavior
has been found to be related to more
accepting views toward suicide, we pre-
dicted that those who were more affected
by the suicides through exposure and close-
ness would have less stigmatizing, yet more
hopeless, attitudes toward suicide. We also
examined the role of social support in peer
cluster suicide bereavement to determine
whether it would be positively or negatively
related to long-term adjustment and atti-
tudes. Lastly, we compared this exposed
group to a group with similar demographics
that was not exposed to multiple peer sui-
cides, predicting that the exposed group
would show less stigmatizing, yet more
hopeless, attitudes toward suicide.
METHOD
Participants
Participants were 152 young adults
who graduated from public high schools in
the northeast. There were two samples: a
high exposure group and a comparison
group. The high exposure group was com-
prised of 85 graduates from a public, New
England high school that experienced mul-
tiple suicides starting in 2005, when partici-
pants were freshmen. Participants reported
having lost an average of 3.39 (SD =1.10)
peers to suicide. None of the participants
had witnessed the physical acts of suicide;
therefore, in this article, “exposure to sui-
cide” is used to describe the loss of class-
mates to suicide. There were 65 women
(76.5%) and 20 men with a mean age of
21.2 years old (SD =.44). The majority
(90.5%) identified as being White.
The comparison sample was com-
prised of 67 graduates of other public high
schools in the northeast who had not been
exposed to multiple suicides. Eighty-two
individuals completed the survey, but 15 had
experienced multiple suicides and were not
included in the final sample. Ten participants
had lost one peer to suicide, but they did not
significantly differ from those with no losses
to suicide on any of the study measures, so
they were retained. There were 38 (56.7%)
women, 25 men, and 4 who did not report
gender, with a mean age of 21.5 years
(SD =.44). The majority (87.3%) identified
as White. The comparison sample was
slightly older (M=21.54, SD =.77) than
the high exposure sample, t(94.70) =3.11,
p=.002, and had a more even distribution
of gender, v
2
(1, n=147) =4.27, p=.039.
Therefore, comparisons between samples
were controlled for age and gender.
Measures
Texas Revised Inventory of Grief
(TRIG). The TRIG was used to measure
unresolved grief (Faschingbauer, Devaul, &
Zisook, 1977), both past and present.
Three TRIG present grief subscales were
formed based on Futterman, Holland, Brown,
Thompson, and Gallagher-Thompson (2010):
emotional response, nonacceptance, and
thoughts. Participants filled out a TRIG for
each of up to three peers lost, with item
responses on a 1–5 Likert scale (1 =
completely false;3=both true and false;
5=completely true). The TRIG was found
to be reliable in this sample with Cron-
bach’s alphas of .93 for past grief and .94
for present grief.
Closeness Questionnaire. This measure
asked participants to identify how many
peers they lost to suicide and to rate their
relationship with each of these peers (up to
three) by indicating how close they were on
a 5-point Likert scale (1 =not close at all;
5=very close).
Attitudes Toward Suicide Scale
(ATTS). The ATTS (Renberg & Jacobsson,
ABBOTT AND ZAKRISKI 3

2003) is a 37-item and 10-factor scale that
measures a person’s feelings about and atti-
tudes toward suicide. The factors are (1) sui-
cide as a right, (2) incomprehensibility, (3)
noncommunication, (4) preventability, (5)
tabooing, (6) normal or common, (7) suicidal
process, (8) caused by relational problems,
(9) preparedness to prevent, and (10) resigna-
tion. Participants rated each item on a
5-point Likert scale (1 =strongly agree;5=
strong disagree). Due to internal review board
concerns about exposure of a vulnerable sam-
ple to multiple items expressing potentially
distressing beliefs about suicide, only the two
highest loading items from each factor were
used to reduce scale length and the potential
for participant distress. This reduction in
items per factor may have contributed to this
study’s slightly lower alpha scores for some
of the factors, compared with those reported
in the original article. Four factors (prevent-
ability, tabooing, suicidal process, and resig-
nation) had alphas for this study (.62, .40,
.42, and .62, respectively) that were compara-
ble to the original article (.63, .45, .51, and
.54). Four factors (suicide as a right, incom-
prehensibility, noncommunication, and nor-
mal or common) had lower alphas for this
study (.54, .57, .39, and .32, respectively)
than the original article (.86, .72, .64, and
.51). This drop in reliability can be explained
by the loss of up to five additional items per
factor. Two factors (caused by relational
problems and preparedness to prevent) had
much lower alphas (.169 and .170) than the
original article (.38 and .54) and were there-
fore not included in our analyses.
Stigma Toward Suicide. This measure
was created to assess how much each partic-
ipant agreed with Joiner’s (2010) stigmatiz-
ing myths about suicide. Items were as
follows: (1) “Suicide is an easy escape, one
that cowards use”; (2) “Suicide is an act of
anger, aggression, or revenge”; (3) “Suicide
is selfish”; (4) “People often die by suicide
‘on a whim’”; (5) “If people want to die
by suicide, we can’t stop them”; and (6)
“Suicide is just a cry for help.” Participants
rated their degree of agreement with each
statement on a 5-point Likert scale (1 =
strongly agree;5=strongly disagree). The
Cronbach’s alpha was .61; however, these
six myths were analyzed separately so spe-
cific stigmatizing themes could be examined
in this sample. Individual myths were mod-
estly related to relevant ATTS subscales:
Myths 1–4 (escape, anger, selfish, whim)
and 6 (cry for help) were positively corre-
lated with ATTS 2 (Incomprehensibility;
N=152; rs=.20 to .40, all ps<.05 for
reported correlations, with Myth 4 strong-
est) and negatively correlated with ATTS 7
(suicidal process; rs=.19 to .50, with
Myth 3 strongest); Myths 1–2 (escape,
anger) and 4–5 (whim, can’t stop) were
positively correlated with ATTS 3 (non-
communication; rs=.17–.40, with Myth 5
strongest); Myths 1–2 (escape, anger) and 5
(can’t stop) were positively correlated with
ATTS 5 (tabooing; rs=.17–.26, with Myth
2 strongest); Myths 5 (can’t stop) and 6 (cry
for help) were correlated with ATTS 4
(preventability; rs=.40 and .19); Myths 3
(selfish) and 5 (can’t stop) were correlated
with ATTS 10 (resignation; rs=.17 and
.26); and Myth 5 (can’t stop) was correlated
with ATTS 1 (suicide as a right; r=.18).
No myth was correlated with ATTS 6
(normal or common).
Multidimensional Scale of Perceived Social
Support (MSPSS). The MSPSS (Zimet,
Dahlem, Zimet, & Farley, 1988) measures pres-
ent perceived social support. Participants rated
family, friends, and a significant other on several
social support statements using a 7-point Likert
agreement scale (1 =very strongly disagree;
7=very strongly agree). This scale was found to
have strong internal consistency in this sample
with Cronbach’s alpha of .94 for the entire
scale, .95 for the special person subscale, .87 for
the family subscale, and .94 for the friends sub-
scale.
Procedure
All participants were recruited
through e-mail and/or social networking
websites and were given a link to complete
an online survey. For the high exposure
group, participants were told that they were
4P
EER SUICIDE SURVIVORS

being invited to participate in the study
because they were members of a high
school graduating class that had been
exposed to peer suicide, but that they did
not need to personally know any of the
deceased classmates to participate. They
were not told how many classmates died by
suicide, but they were asked how many
peers they lost to suicide. Only one partici-
pant reported not having lost any peers to
suicide, and this person was not included in
the analyses. Of the 423 students in this
high school class, 342 (80.9%) could be
contacted through social networking or
e-mail. A total of 131 students (38.3% of
those contacted) started the survey and 92
(26.9% of those contacted) completed it.
Compared with those with complete data
(85), those who dropped out early (46) had
higher averaged current grief (M=3.03,
SD =1.18) than those who completed the
survey (M=2.44, SD =.90), t(101) =2.39,
p=.019. However, they did not differ on
past averaged grief scores, highest individ-
ual past and present grief scores, highest
present grief dimension scores, closeness
ratings, or number of peers lost to suicide,
nor did they differ on grief or closeness for
the peer they reported being closest to.
The comparison sample was recruited
through online snowball sampling through
social networking contacts of the first
author to parallel the recruitment of the
exposed sample. The criteria for participa-
tion were graduation from a public high
school in the northeast and being between
the ages of 21 and 23 at the time of the
study. These individuals received the same
online survey without the TRIG. The
TRIG was excluded because it would be
irrelevant for anyone who had not experi-
enced the loss of a loved one, and difficult
to compare with the exposed sample if it
was completed on someone who had not
died by suicide. All participants in both
samples were asked whether they had lost
family members to suicide. The two
samples did not differ in exposure to family
suicides, t(148) =0.951, p=.34.
RESULTS
Analyses of the exposed sample
showed a range of grief and closeness scores
over participants and also a range in the
number of peers they reported losing to sui-
cide (see Table 1 for descriptive statistics).
Repeated measures analyses of the 77 partic-
ipants who reported grief for two peers, and
of the 61 participants who reported grief for
three peers, showed no systematic differ-
ences in past or present grief over the peers,
with Fs ranging from .38 to 1.06, and pval-
ues ranging from .79 to .31. Participants
TABLE 1
Descriptive Statistics for Numbers of Peers Lost
to Suicide, Closeness, and Grief Scores
Min./Max. M(SD)
Number of peers
lost (n=85)
1.00/5.00 3.39 (1.10)
Peer 1 (n =85)
Closeness 1.00/5.00 2.16 (1.22)
TRIG past 1.00/4.25 2.19 (0.95)
TRIG present 1.08/4.69 2.44 (0.90)
Peer 2 (n =77)
Closeness 0.00/5.00 2.24 (1.44)
TRIG past 1.00/5.00 2.29 (1.18)
TRIG present 1.00/4.85 2.58 (1.06)
Peer 3 (n =61)
Closeness 0.00/5.00 1.69 (1.55)
TRIG past 1.00/5.00 2.23 (1.23)
TRIG present 1.00/4.92 2.51 (1.09)
Closest peer (n =85)
Closeness 1.00/5.00 2.79 (1.36)
TRIG past 1.00/5.00 2.47 (1.21)
TRIG present 1.08/4.92 2.75 (1.05)
Nonacceptance 1.00/5.00 2.72 (1.05)
Thoughts 1.00/5.00 2.86 (1.20)
Emotional 1.00/5.00 2.66 (1.08)
Averaged over peers (n =85)
Closeness 1.00/4.67 2.22 (1.10)
TRIG past 1.00/4.17 2.17 (1.00)
TRIG present 1.05/4.62 2.46 (0.91)
Nonacceptance 1.00/5.00 2.56 (0.90)
Thoughts 1.00/4.60 2.49 (1.06)
Emotional 1.00/4.40 2.38 (0.95)
TRIG, Texas Revised Inventory of Grief;
Min., minimum; Max., maximum.
ABBOTT AND ZAKRISKI 5

were not explicitly asked to report on the
suicides in chronological order; instead they
were asked to report on up to three peers
who had died by suicide and were prompted
with the headers Peer 1, Peer 2, and Peer 3.
Additionally, participants who reported hav-
ing experienced more than the three peer
suicides, n=36 (42%), were asked to report
on the “three that had the most significant
impact on your life,” in no specified order.
There is some evidence that closeness may
have influenced order of reporting: Close-
ness did not differ over peers in the subsam-
ple reporting three, F(2, 60) =1.87,
p=.163, g
2
=.06, Wilks’s lambda =.94, but
trended toward a difference in the subsample
reporting only two peers, F(1, 76) =3.27,
p=.075, g
2
=.04, Wilks’s lambda =.96,
with higher closeness to the peer rated sec-
ond (M=2.47) than the peer rated first
(M=2.23).
To capture both the cumulative effect
of multiple suicides and the possibly stron-
ger impact of the suicide of the closest peer,
we conducted our analyses in two ways,
using averaged scores over all peers and
then using the scores for the peer each par-
ticipant was closest to. Cronbach’s alphas
for combined closeness, past grief, present
grief were .83, .87, and .86, respectively,
providing support for the aggregation
method. For 56.5% of the participants, the
closest peer was Peer 1; for 31.8%, it was
Peer 2; and for 11.8%, it was Peer 3. When
there was a tie for closeness ratings, we
selected the peer who was reported on first.
All means presented in Table 2 for closest
peer were significantly higher than the
means averaged over peers, with all paired-
sample ts(85) >3.45, and ps<.01.
As predicted, both number of peers
lost to suicide and closeness to the peers lost
to suicide were related to past and present
grief, including all present grief factors, with
stronger relationships seen for closeness
than for number of peers lost (see Table 2;
TABLE 2
Correlations Between Grief and Closeness to Peers Lost to Suicide (Overall and for Closest Peer),
Numbers of Peers Lost to Suicide, and Perceived Social Support from Family (N =85)
Variable
Closeness to Peers or
to Closest Peer
Number of Peers
Lost
PSS from
Family
rrr
Overall
Grief scores
TRIG past .80*** .39*** .27*
TRIG present .74*** .40*** .07
Present grief factors
Emotion .71*** .40*** .06
Nonacceptance .52*** .27* .09
Thoughts .80*** .40*** .15
Closest peer
Grief scores
TRIG past .75*** .40*** .24*
TRIG present .76 *** .40*** .07
Present grief factors
Emotion .72*** .39*** .02
Nonacceptance .55*** .26* .01
Thoughts .76*** .42*** .15
TRIG, Texas Revised Inventory of Grief; PSS, perceived social support.
*p<.05, ***p<.001 (two-tailed).
6PEER SUICIDE SURVIVORS

only subscales with significant correlations
are presented). This was seen both for grief
and closeness averaged over peers and for
grief and closeness for the closest peer. With
regard to social support, there was only a
significant negative correlation between past
grief and perceived social support (PSS) from
family (for both averaged grief and for closest
peer grief); neither social support from
friends nor significant others was related to
grief.
Univariate relationships for closeness
to peers lost to suicide, grief, and social
support with attitudes toward suicide and
suicide myths were examined next (see
Table 3). Number of peers lost to suicide
was not significantly correlated with any of
the myths or attitudes, so it was not
included in this table. Most notably, for
closeness and grief, were relations with
beliefs about preventability: ATTS 4 (pre-
ventability) scores were lower in those with
higher closeness, higher past and present
grief scores, higher present emotional grief,
and higher present thoughts about grief
(overall and for the closest peer). Similarly,
endorsement of Myth 5 (“If people want to
die by suicide, we can’t stop them”) was
higher in those with higher closeness and
past grief overall and for the closest peer.
Additionally, lower scores on ATTS 2
(incomprehensibility) were related to higher
past grief for the closest peer, and lower
scores on ATTS 1 (suicide as a right) were
related to higher overall closeness.
Perceived social support was also
related to myths and attitudes about prevent-
ability. Those with higher social support
from a special person were less likely to
endorse the myth “If people want to die by
suicide, we can’t stop them” (Myth 5), and
similarly those with higher social support
from family and friends had higher scores for
ATTS 4 (preventability). In addition, those
with higher social support from friends, fam-
ily, and a special person had lower scores for
ATTS 6 (normal or common), ATTS 10
(resignation), and ATTS 7 (suicidal process).
Those with higher social support from a
special person had lower scores for ATTS 5
(tabooing), but was more likely to endorse
the myth “Suicide is an act of anger, aggres-
sion, or revenge.” Those with higher social
support from family had lower scores for
ATTS 3 (noncommunication), but were
more likely to endorse Myth 3 (“Suicide is
selfish.”).
To examine the roles of closeness and
social support from friends in predicting
dimensions of current grief, a 2 (high or low
closeness) by 2 (high or low friend social
support) MANOVA controlling for gender
was conducted on the three present grief
dimensions. The small number of males in
the sample limited our ability to test for
interactions with gender, so gender was
entered as a covariate. Preliminary gender
comparisons for grief, social support, and
closeness revealed that women expressed
more grief than men both past and present
and on all factors, ts(82) >2.04, ps<.05,
with perhaps the exception of averaged non-
acceptance, t(82) =1.78, p=.079. Men and
women did not differ significantly on close-
ness, ts(82) <1.64, ps>.10 overall or for
their closest peer. Results were essentially
the same for all remaining analyses using
averaged closeness and grief as when using
scores for the closest peer, with stronger
effects for the former. To reduce redun-
dancy in reporting, only the results for aver-
aged scores are presented.
There was a significant multivariate
effect for closeness, Wilks’s lambda =.48, F
(3, 77) =28.28, p=.000, g
2
=.52. Main
effects showed that closer peers were
higher on all grief dimensions [thoughts:
F(1, 79) =86.00, p=.000; nonacceptance:
F(1, 79) =24.88, p=.000; emotional response:
F(1, 79) =57.99, p=.000] than peers who
were less close to those who died by
suicide. There was also a significant multi-
variate interaction between closeness and
friend social support [Wilks’s lambda =.87,
F(3, 77) =3.85, p=.013, g
2
=.13]. Inter-
action effects revealed a consistent overall
pattern showing that for those closer to the
peers who died by suicide, having high
social support from friends was related to
more grief, but for those who were not as
ABBOTT AND ZAKRISKI 7

TABLE 3
Correlations Between Closeness, Grief (Overall and with Closest Peer), and Perceived Social Support with Myths and ATTS Factors
Myth 2
Angry Act
Myth 3
Selfish
Myth 5
Can’t Stop
ATTS 1
Right
ATTS 2
Incomprehensibility
ATTS 3
Noncommunication
ATTS 4
Preventability
ATTS 5
Tabooing
ATTS 6
Normal
ATTS 7
Process
ATTS 10
Resignation
rr r r r r r r rr r
Overall
Closeness .05 .10 .26* .22* .14 .08 .32** .08 .02 .04 .07
Past grief .01 .13 .13 .15 .05 .05 .21* .03 .15 .05 .01
Present grief .04 .15 .29** .09 .18 .11 .29** .09 .19 .12 .20
Present grief factors
Emotional
response
.03 .14 .15 .12 .08 .04 .24* .01 .11 .04 .04
Nonacceptance .03 .03 .04 .14 .03 .10 .04 .04 .14 .07 .05
Thoughts .03 .17 .14 .17 .05 .04 .25* .07 .17 .12 .01
Closest peer
Closeness .05 .08 .22* .13 .16 .04 .33** .10 .02 .02 .07
Past grief .06 .08 .28* .06 .22* .09 .30** .04 .15 .10 .16
Present grief .02 .05 .12 .11 .08 .08 .26* .07 .11 .03 .01
Present grief factors
Emotional
Response
.01 .05 .16 .10 .12 .08 .27* .07 .06 .00 .04
Nonacceptance .00 .02 .03 .14 .04 .09 .12 .06 .14 .03 .01
Thoughts .04 .09 .11 .17 .06 .06 .28** .11 .12 .09 .00
Perceived social support
Friends PSS .14 .17 .13 .17 .04 .19 .26* .17 .35** .31** .34**
Special Person
PSS
.22* .13 .21 .10 .14 .19 .15 .27* .30** .33** .36**
Family PSS .11 .24* .17 .04 .17 .22* .28** .17 .43** .32** .29**
ATTS, Attitudes toward Suicide Scale; Right, suicide as a right; Process, suicidal process; PSS, perceived social support. Bold values are signifi-
cant at the .05 or .01 level.
N=85; **p<.01, *p<.05 (two-tailed).
8PEER SUICIDE SURVIVORS

close, high social support was related to less
grief. This overall pattern was significant
for all three present grief factors [thoughts:
F(1, 79) =9.79, p=.002; nonacceptance: F
(1, 79) =7.42, p=.008; and emotional
response, F(1, 79) =5.57, p=.036; see Fig-
ure 1]. Pairwise differences for friend social
support within closeness group were signifi-
cant for nonacceptance within participants
who were closest to the peers lost to suicide
(higher grief for those with more social sup-
port) and for thoughts within participants
who were not as close to the peers lost to sui-
cide (higher grief for those with less social
support).
Next we conducted a 2 (high or low
closeness) by 2 (high or low friend social
support) MANOVA with gender as a covar-
iate to examine the roles of closeness and
social support from friends in predicting
attitudes toward suicide. Although the
interaction was not significant, there was a
significant multivariate effect for friend
social support, Wilks’s lambda =.78, F(8,
72) =2.53, p=.018, g
2
=.22. Main effects
showed that those with higher social sup-
port from friends had higher scores for
ATTS 4 (preventability; M=4.04 vs. 3.55),
F(1, 79) =6.28, p=.014, and lower scores
for ATTS 5 (tabooing; M=1.94 vs. 2.26),
F(1, 79) =4.95, p=.029; ATTS 6 (normal;
M=3.48 vs. 3.85), F(1, 79) =4.31, p=
.041); ATTS 7 (suicidal process; M=2.41
vs. 2.82), F(1, 79) =7.45, p=.008; and
ATTS 10 (resignation; M =1.82 vs. 2.40),
F(1, 79) =6.13, p=.015. These results
largely echo the univariate correlational
findings for friend social support (and special
person social support) presented in Table 3.
A parallel analysis on myths revealed a
significant main effect for level of social sup-
port from friends, Wilks’s lambda =.822, F
(6, 72) =2.60, p=.025, g
2
=.18. There
was a significant univariate effect of friend
social support for Myth 5, F(1, 77) =4.44,
p=.038, g
2
=.055, with those who had less
social support from friends (M=2.21)
endorsing Myth 5 (“If people want to die by
suicide, we can’t stop them”) more than
those with greater social support from
friends (M=1.82). This positive effect of
friend social support was not evident in the
univariate correlational analyses presented in
Table 3.
Lastly, we examined differences
between the high suicide exposure group
and the low exposure comparison group on
ATTS and myths. In a two-way MANOVA,
controlling for age and gender there was a
significant multivariate main effect for
0
0.5
1
1.5
2
2.5
3
3.5
4
Low High Low High Low High
Current TRIG Scores
Averaged Closeness to Peer(s) Who Died by Suicide
Low Social Support - Friends
High Social Support - Friends
Nonacceptance Thoughts
Emotional Response
Figure 1. Current Texas Revised Inventory of Grief (TRIG) factor scores as a function of closeness to the peer(s)
who died by suicide and current social support from friends. Closeness by social support interactions for all three
TRIG factors were significant; significant pairwise comparisons within closeness level for each TRIG factor are
noted with an asterisk (*).
ABBOTT AND ZAKRISKI 9

exposure group on ATTS, Wilks’s
lambda =.873, F(8, 136) =2.48, p=.015,
g
2
=.13 with two significant univariate
effects. For ATTS 2 (incomprehensibility),
those with higher exposure to suicide
(M=3.10) were more likely to think that
suicide is incomprehensible than those with
lower exposure (M=2.71), F(1, 143) =6.89,
p=.010, g
2
=.05. For ATTS 6 (normal or
common), those with higher exposure to sui-
cide (M=3.67) were more likely to think
that suicide or suicidal ideation is normal
than those with lower exposure to suicide
(M=3.38), F(1, 143) =4.73, p=.031,
g
2
=.03. There were no significant differ-
ences on myth endorsement by exposure
group and no significant differences in social
support.
DISCUSSION
In this study we sought to understand
how exposure to a cluster of suicides as
adolescents affects the peer survivors several
years later. Closeness to the peers lost to
suicide and current social support were used
to understand grief and beliefs about sui-
cide. Results revealed both risk and resil-
ience in this sample. Present grief showed a
large range and was higher for those closest
to the peers who died by suicide and higher
for the closest peer lost to suicide. Close-
ness and grief were also linked to weaker
beliefs in the preventability of suicide.
Social support seemed to serve as a healthy
influence on both suicide attitudes and
grief, but there were some associations
between social support and stigmatizing
attitudes, and there was evidence that peer
social support might prolong grief in those
who were closest to the deceased. Finally,
compared with a group of same-aged peers
without suicide cluster exposure, these par-
ticipants reported that suicide was more
normal, yet more incomprehensible.
Seven years after the first suicide and
2 years after the last suicide, there were
many associations with grief, both past and
present, in this exposed sample. The
hypothesis that grief would be higher in
those closer to the peers lost to suicide was
clearly supported, both for overall grief and
closeness to the peers lost to suicide and for
the closest peer lost to suicide. Grief was
also higher in those who lost more peers to
suicide, but this relationship was somewhat
weaker. Social support from family was
associated with less past grief, yet support
from significant others and friends was not
(this latter finding is discussed further
below). We hypothesized that those who
were more affected by the suicides would
have less stigmatizing, yet more hopeless,
attitudes toward suicide. We did not find
any support for the hypothesis that those
more affected (i.e., those who were closer)
would have less stigmatizing thoughts about
suicide. In fact, closer individuals were
more likely to endorse the stigmatizing
myth that suicide is not preventable, as
measured by both the myth and attitude
scales. We found this same pattern for past
and present grief. One possible explanation
is that this belief could help a survivor cope
with guilt, one of the most difficult emo-
tions for suicide survivors, especially in
those who were close to the deceased. By
believing that there is nothing one can do
to prevent a suicide, the suicide survivors
may be able to avoid blaming themselves
for not having prevented it. Alternatively,
this could be a sign of hopelessness. These
young adults had repeated exposure to
suicides in their peer group and may have
witnessed multiple failed attempts at pre-
vention.
Social support in this sample was
generally related to more hopeful attitudes.
The belief in preventability was higher in
those with better social support, no matter
what the source (i.e., friends, family, signifi-
cant other). Those with higher social
support from a special person were less
likely to think suicide should be a taboo
topic, and those with higher social support
from family were less likely to think that
people who are legitimately suicidal will not
talk about it or make threats. Being able to
talk to suicidal individuals about their
10 PEER SUICIDE SURVIVORS

thoughts without taboo and taking their
threats seriously can be important steps to
getting help for them. These results suggest
that social support may have a positive role
in promoting prevention and help-seeking
behaviors.
Social support was also related to
some negative and less accepting attitudes
toward the act of suicide. Those with
higher social support (from friends, family,
and a special person) were less likely to
think that suicide or suicidal ideation is
normal or common, that suicide is a
thoughtful process (vs. an impulsive act),
and that in some cases it can be acceptable
or a relief for those involved. Additionally,
those with higher social support from a spe-
cial person were more likely to think that
suicide is an act of anger, aggression, or
revenge, and those with higher social sup-
port from family were more likely to think
that suicide is selfish. These beliefs may be
fostered in affected adolescents by families
and significant others to help them avoid
self-blame, come to terms with difficult
losses, and prevent future suicides in a clus-
ter-affected community, but they may also
promote stigma. The question of how to
best help adolescents cope in these circum-
stances without stigmatizing either the
actions of those who have already been lost
to suicide or resulting suicidal ideations
within the affected population clearly war-
rants further investigating.
Social support was not strongly
related to grief overall, and social support
from friends showed an overall pattern of
being related to higher grief in those who
were closest to the peers who died by sui-
cide, but lower grief in those who were not
as close. This pattern was seen for all
dimensions of present grief, including emo-
tional response to, thoughts about, and
nonacceptance of the losses, with the effect
on close peers being especially strong for
nonacceptance. In this exposed sample,
especially those who were close to the
deceased, the friends had all experienced
the same loss, thus “social support from
friends” may have resembled “corumina-
tion” more than productive social support
that might facilitate bereavement narrative
processing. Corumination has been defined
as a repeated focus on negative emotions
within friend groups and is related to posi-
tive friendship quality and closeness, yet it
can also predict an increase in depressive
symptoms (Davila et al., 2012; Rose, 2002).
Interactions between closeness and social
support from friends were only seen for
grief, and not for attitudes or myth
endorsement, suggesting that corumination
may shape emotional responses to specific
losses more than general attitudes.
When compared with the low expo-
sure comparison group, the high exposure
group was more likely to think that suicide
or suicidal ideation is normal, but also
incomprehensible. These findings point to
an interesting paradox. Repeated exposure
to suicide seems to have influenced percep-
tions of how common, typical, or “normal”
suicide and suicidal ideation are in the high
exposure group, but has not necessarily led
to a greater understanding of why it occurs.
In fact, individuals in the cluster-exposed
sample found suicide more incomprehensi-
ble than those with no cluster exposure. If
one believes that suicide is normal, one may
not take suicidal threats or ideation as seri-
ously as when suicide is viewed as atypical.
The influence of incomprehensibility on
prevention requires more probing.
Limitations and Future Directions
We recognize that there are limita-
tions to this study. One limitation is that all
of the participants from the high exposure
group were from the same high school.
Therefore, the findings are not necessarily
generalizable to clusters of suicides in other
communities. This study used online recruit-
ment; for that reason, members of this grad-
uating class who do not use social
networking websites, although a small
percentage, were not represented. The first
author was a member of the graduating class
studied; therefore, personal relationships
may have influenced participation (e.g., more
ABBOTT AND ZAKRISKI 11

female than male participants) or skewed the
results by shaping reporting patterns. The
survey was designed intentionally to make it
easy to drop out due to the sensitive nature
of the questions, and there was a 35% drop-
out rate. The participants who dropped out
had significantly higher present averaged
grief scores. However, they did not differ
on past averaged grief scores, closest indi-
vidual past and present grief scores, highest
present grief dimension scores, closeness
ratings, or number of peers lost to suicide.
Also, some myths (easy escape, done on a
whim) were not related to any study vari-
ables, and two ATTS factors (caused by
relational problems, preparedness to pre-
vent) could not be used due to low reliabil-
ity. This suggests the need for more
research using the full ATTS subscales or a
smaller set of relevant attitudes or myths.
The ATTS scale was shortened for this
study due to IRB concerns, which was a
limitation for analyses using this scale.
Additionally, the large number of analyses
involving these scales suggests caution is
needed when interpreting reported findings
due to the possibility of Type I error. The
comparison sample was similar to the main
sample in race or ethnicity, geographical
location of school (the northeast), type of
high school (public), exposure to family sui-
cide, and social support, but may not have
been comparable in other ways. Of the vari-
ables measures, the samples only differed
significantly in age and gender, and in the
attitudes previously noted. Gender and age
were used as control variables in compari-
sons between samples, but there may have
been other differences we did not measure
and could not control for, which is a limi-
tation. Additionally, comparisons on a
broader range of variables, including emo-
tional adjustment indices relevant to both
samples, would have been useful.
For future research, it would be help-
ful to directly measure overall grief rather
than grief for up to three peers to avoid the
inconsistencies we had in the number of
peers contributing to the overall score. Our
analyses of the closest peer provided a
partial solution to this problem, but it may
not be the best way to capture the influence
of the suicide cluster as a whole. It would
also be helpful to use a grief scale with
identical past and present items so scores
could be compared over time and a change
in grief could be assessed. With this, it
would be clearer whether grief had lessened
and for whom. A topic for future studies
would be to more directly examine how the
stigma of suicide may affect the grieving
process in peer survivors and may diminish
disclosure of being a suicide survivor. For
example, fear of stigmatizing responses may
limit disclosure and restrict the pool of
individuals survivors can seek support from.
This could increase the chances of isolation
and counterproductive corumination with
affected peers. Future research is needed to
identify the quality of social support that is
being accessed in these peer groups affected
by suicide clusters. This study has also
identified that young people who were clo-
ser to those lost by suicide are not very
hopeful about prevention. Future research
could focus on interventions to educate
about suicide prevention to promote both
hopefulness and help-seeking behaviors.
These interventions may need to target
some of the stigmatizing beliefs we have
identified that appear to be fostered in
socially supportive relationships and might
help individuals cope with peer suicides in
the short run, but be stigmatizing and
counterproductive to prevention efforts in
the long run. How to do this without
undermining community efforts to contain
the cluster is an important question for
future research.
These results can provide helpful and
tangible insight for postvention with
schools, peer groups, and community health
personnel in this under-studied population.
Because there was a negative correlation
between exposure to suicide and belief in
suicide prevention, it might be helpful to
specifically target this belief in suicide pre-
vention training and psychoeducation with
adolescents affected by cluster suicides.
Because adolescents exposed to suicides are
12 PEER SUICIDE SURVIVORS

at a greater risk of suicidal ideations, aware-
ness training and mindfulness interventions
may be important to include in postvention
work, especially because those in the
exposed sample with social support believe
that suicidal ideation is not normal. The
finding that grief was higher in those with
high exposure and high social support from
friends identifies a certain subpopulation
that may be at heightened risk of lingering
grief. Close friend groups affected by sui-
cide clusters may benefit from group ther-
apy and training on how friend groups can
avoid corumination while still helping each
other process their experiences and being
otherwise supportive.
REFERENCES
BONDORA, J. T., & GOODWIN,J.L.
(2005). The impact of suicidal content in popu-
lar media on the attitudes and behaviors of ado-
lescents. Praxis,5,5–12.
BRENT, D. A., PERPER, J. A., MORITZ, G.,
ALLMAN, C., LIOTUS, L., SCHWEERS, J., ET AL.
(1993). Bereavement or depression? The impact
of the loss of a friend to suicide. Journal of
the American Academy of Child and Adolescent Psy-
chiatry,32, 1189–1197. doi:10.1097/00004583-
199311000-00012.
CEREL, J., MAPLE, M., ALDRICH, R., &
VAN DE VENNE, J. (2013). Exposure to suicide
and identification as survivor. Crisis,34, 413–
419. doi:10.1027/0227-5910/a000220.
DAVILA, J., HERSHENBERG, R., FEINSTEIN,
B. A., GORMAN, K., BHATIA, V., & STARR,L.R.
(2012). Frequency and quality of social network-
ing among young adults: Associations with
depressive symptoms, rumination, and co-rumi-
nation. Psychology of Popular Media,1,72–86,
doi:10.1037/a0027512.
DYREGROV, K. (2009). How do the young
suicide survivors wish to be met by psycholo-
gists? A user study. Omega,59, 221–238.
FASCHINGBAUER, T. R., DEVAUL, R. A., &
ZISOOK, S. (1977). Development of the Texas
Inventory of Grief. American Journal of Psychiatry,
134, 696–698.
FUTTERMAN,A.,HOLLAND,J.M.,BROWN,P.
J., THOMPSON,L.W.,&GALLAGHER-THOMPSON,D.
(2010). Factorial validity of the Texas Revised
Inventory of Grief–Present scale among bereaved
older adults. Psychological Assessment,22, 675–687.
doi:10.1037/a0019914.
GOULD, M. S., VELTING, D., KLIENMAN,
M., LUCAS, C., GRAHAM THOMAS, J., & CHUNG,
M. (2004). Teenagers’ attitudes about coping
strategies and help-seeking behavior for suicidali-
ty. Journal of the American Academy of Child and
Adolescent Psychiatry,43, 1124–1133. doi:10.1097/
01.chi.0000132811.06547.31.
GOULD, M. S., WALLENSTEIN, S., KLEIN-
MAN, M. H., O’CARROLL, P., & MERCY, J. (1990).
Suicide clusters: An examination of age-specific
effects. American Journal of Public Health,80,
211–212.
HIRSCH, J. K., & BARTON, A. L. (2011).
Positive social support, negative social exchanges,
and suicidal behavior. Journal of American College
Health,59, 393–398. doi:10.1080/07448481.2010.
515635.
JOINER, T. E. (1999). The clustering and
contagion of suicide. Current Directions in Psycho-
logical Science,8,89–92. doi:10.1111/1467-
8721.00021
JOINER, T. E. (2010). Myths about suicide.
Cambridge: Harvard University Press.
KALAFAT, J., & ELIAS, M. (1992). Adoles-
cents’ experience with and response to suicidal
peers. Suicide and Life-Threatening Behavior,23,
315–321. doi:10.1111/j.1943278X.1992.tb00736.x.
LUKAS, C., & SEIDEN, H. M. (2007). Silent
grief: Living in the wake of suicide. Philadelphia,
PA: Jessica Kingsley.
O’CARROLL, P. W., MERCY, J. A., &
STEWARD, J. A. (1988). CDC recommendations
for a community plan for the prevention and
containment of suicide clusters. Morbidity and
Mortality Weekly Report,6,1–12.
REED, M. D., & GREENWALD, J. Y. (1991).
Survivor-victim status, attachment, and sudden
death bereavement. Suicide and Life-Threatening
Behavior,21, 385–401.
RENBERG, E. S., & JACOBSSON, L. (2003).
Development of a questionnaire on attitudes
towards suicide (ATTS) and its application in a
Swedish population. Suicide and Life-Threatening
Behavior,33,52–64.
ROSE, A. J. (2002). Co-rumination in the
friendships of girls and boys. Child Development,
73, 1830–1843. doi:10.1111/1467-8624.00509.
STEIN, D., WITZUM, E., BROM, D.,
DENOUR, A., & ELIZUR, A. (1992). The associa-
tion between adolescents’ attitudes toward
suicide and their psychosocial background and
suicidal tendencies. Adolescence,27, 949–959.
SVEEN, C. A., & WALBY, F. A. (2008).
Suicide survivors’ mental health and grief reac-
tions: A systematic review of controlled studies.
ABBOTT AND ZAKRISKI 13

Suicide and Life-Threatening Behavior,38,13–29.
doi:10.1521/suli.2008.38.1.13.
WAGNER, K. G., & CALHOUN,L.G.
(1991). Perceptions of social support by suicide
survivors and their social networks. Omega,24,
61–73.
ZIMET, G. D., DAHLEM, N. W., ZIMET,S.
G., & FARLEY, G. K. (1988). The Multidimen-
sional Scale of Perceived Social Support. Journal
of Personality Assessment,52,30–41.
Manuscript Received: August 29, 2013
Revision Accepted: February 11, 2014
14 PEER SUICIDE SURVIVORS