Posted: August 2nd, 2022

Complete a summary paper. Be sure to limit your summary paper to three pages or less.

Tartaro2015 Richardsetal.2014 Tapiaetal.2016 Wolford-ClevengerSmith2015

 Complete a summary paper. Be sure to limit your summary paper to three pages or less. 

The Prison Journal
2015, Vol. 95(1) 23 –42

© 2014 SAGE Publications
Reprints and permissions:

sagepub.com/journalsPermissions.nav
DOI: 10.1177/0032885514563269

tpj.sagepub.com

Article

What Is Obvious?
Federal Courts’
Interpretation of the
Knowledge Requirement
in Post–Farmer v. Brennan
Custodial Suicide Cases

Christine Tartaro1

Abstract
The Supreme Court ruling on Farmer v. Brennan established the federal
court standard for determining when corrections staff members should be
held legally responsible for inmates’ injuries resulting from inmate–inmate
violence or self-harm. Legal scholars lamented that requiring plaintiffs to
prove deliberate indifference placed a heavy burden on plaintiffs. Two slightly
encouraging signs appeared to be Farmer’s end to the “individual-specific”
rule and the decision that courts had the option of ruling that the threat of
harm was too obvious for the defendants to ignore. This article involves a
review of post-Farmer custodial suicide cases and the impact that these two
aspects of the Farmer decision has had on these cases.

Keywords
Suicide litigation, Section 1983, Farmer v. Brennan, self-harm, Eighth amendment

Suicide is the leading cause of death in jails and the fifth most common cause
of death in state prisons in the United States (Noonan, 2012). Lawsuits tend

1The Richard Stockton College of New Jersey, Galloway, USA

Corresponding Author:
Christine Tartaro, Professor of Criminal Justice, The Richard Stockton College of New Jersey,
101 Vera King Farris Drive, Galloway, NJ 08205, USA.
Email: Christine.tartaro@stockton.edu

563269TPJXXX10.1177/0032885514563269The Prison JournalTartaro
research-article2014

mailto:Christine.tartaro@stockton.edu

http://crossmark.crossref.org/dialog/?doi=10.1177%2F0032885514563269&domain=pdf&date_stamp=2014-12-17

24 The Prison Journal 95(1)

to follow attempted and completed suicides that occur in custody, and plain-
tiffs typically seek relief in both federal and state courts. Schlanger (2003)
surveyed state prison systems and 44 jails with a capacity of at least 1,000
inmates and asked about the types of inmate lawsuits that have been filed
against their facilities. Approximately one third of the large jails and prison
systems had been sued for a custodial suicide in the 3 years preceding the
survey. Suicide cases ranked seventh among reasons why the large jails were
subject to litigation.

Incarcerated inmates have a right to certain protections under the constitu-
tion. People who are not in custody of a prison, jail, or law enforcement
agency have the freedom to pick up a telephone and dial 911 or travel to a
medical office for assistance with physical and mental health problems.
Placement in custody removes that ability, and the result is that the govern-
ment entity who takes away that freedom becomes responsible for prevention
of and response to serious harm. Custodial staff members who do not handle
this protection properly put their agency at risk of litigation.

Options for Relief in Court

Inmates, or families of inmates in the event of an inmate’s death, can file a
tort claim in state court or a civil rights claim in federal court. A tort is a pri-
vate wrong or injury that the court may remedy by awarding the plaintiff
compensation and damages in the form of a monetary award (Wallace &
Roberson, 2000). In the event of an inmate suicide or attempted suicide, the
most common type of tort is negligence, where the plaintiffs allege that offi-
cials committed a breach of duty that was the proximate cause of some harm.
Tort cases do not allow for plaintiffs to ask for attorney fees or any injunctive
relief (Anderson & Dyson, 2001).

Tort claims involving custodial suicide or attempted suicide usually
involve wrongful death lawsuits, and the standard for establishing wrongdo-
ing on the part of the defendants is negligence. The establishment of negli-
gence involves the question of whether the officer’s act or failure to act
created an unreasonable risk to the detainee. There are four elements required
for proving negligence: (a) the officials responsible for the inmate had a legal
duty to care for that person, (b) there was a breach of that duty, (c) the breach
was the proximate cause of injury, and (d) there was an actual injury
(Kappeler, Vaughn, & Del Carmen, 1991).

Plaintiffs who sue in the federal courts do so through title 42, Section
1983, of the U.S. Code, commonly referred to as “Section 1983.” Section
1983 lawsuits deal with the conditions of the inmates’ confinement, including
denial of medical care or proper suicide prevention. Civil rights claims

Tartaro 25

provide plaintiffs with more options for relief, including not just monetary
awards that torts offer but attorney’s fees and injunctions (Anderson &
Dyson, 2001). While the potential rewards associated with civil rights claims
are attractive to plaintiffs, the legal standards established by years of federal
court decisions has made it very difficult for plaintiffs to succeed.

Proving that law enforcement or correctional personnel violated the con-
stitutional rights of inmates, as is required in federal court, is much more
difficult than proving negligence in state court. Prior to 1964, few inmates
succeeded in convincing federal courts that they were even entitled to civil
rights protections. The case of Ruffin v. Commonwealth (1871) set the prec-
edent that inmates were to be considered “slaves of the state” and, therefore,
were not entitled to protection under the Bill of Rights. The federal courts
overturned Ruffin in the 1960s with Fulwood v. Clemmer (1962) and Cooper
v. Pate (1964). In both cases, the courts acknowledged that the Bill of Rights
applies to inmates and that meant that the Eighth Amendment’s prohibition of
cruel and unusual punishment included protections for convicted offenders.
While the Eighth Amendment does not specifically address pretrial detain-
ees, the Fourteenth Amendment’s Due Process Clause demands that the rights
of pretrial detainees be at least as great as those afforded to convicted offend-
ers (see City of Revere v. Massachusetts General Hospital, 1983).

Deliberate Indifference Pre-Farmer

The specific issue of inmates’ rights to medical care was not addressed until
the Supreme Court heard the landmark case of Estelle v. Gamble (1976).
Gamble was a Texas prison inmate who was assigned to farming duty. On
November 9, 1973, Gamble was unloading a truck when a bale of cotton fell
on him. He was seen by a doctor who diagnosed his injury as a lower back
strain. The doctor prescribed a muscle relaxer and a pain reliever and placed
Gamble on “cell-pass, cell-feed” status for 2 days, meaning that Gamble was
to remain in his cell except for meals and showers. The doctor also ordered
that Gamble be moved to a lower bunk bed, but the prison administration did
not comply with that order. After additional checks by the doctor, the cell-
pass, cell-feed status was extended a few more times until Gamble had been
out of work for 3 weeks. In early December, the doctor cleared Gamble for
light work duty. The inmate claimed that he was still too injured to work and
refused his assignment. He was placed on disciplinary segregation and
ordered to see another doctor. The second doctor prescribed medication for
high blood pressure and back pain. The prescription for his blood pressure
medication, however, was not filled for 4 days because the correctional staff
lost it. Gamble continued to receive medical care through December and

26 The Prison Journal 95(1)

January, and the Department of Corrections ordered Gamble back to work
again in January. During a January 31, 1974, disciplinary hearing, a member
of the prison medical staff testified that Gamble was in “first-class” medical
condition and could return to work. The prison staff moved Gamble to soli-
tary confinement as punishment for his refusal to work. Gamble’s medical
condition worsened while on solitary, and he was hospitalized for heart prob-
lems. Gamble filed suit the following month, claiming that the prison staff
violated his Eighth Amendment rights with their inadequate medical care.

Gamble’s complaint was eventually heard by the Supreme Court, and
while he lost, the case is remembered for setting the standard needed for
inmates to prove that their medical care, or lack thereof, violated the constitu-
tion’s prohibition of cruel and unusual punishment. The Estelle court estab-
lished deliberate indifference as the standard for judging violations of the
constitution when the issue in question is satisfactory medical care. The court
ruled that inmates had to prove that the staff’s action or inaction resulted in
“unnecessary and wanton infliction of pain.” Years later, the Supreme Court
further defined deliberate indifference as entailing “something more than
negligence, but is satisfied by something less than acts or omissions for the
very purpose of causing harm or with knowledge that harm will result. Thus,
it is the equivalent of acting recklessly” (Farmer v. Brennan, 1994, p. 818).
The loss of Gamble’s blood pressure prescription, which resulted in a 4-day
delay in his treatment, was considered to be an accident by the court, as there
was no proof that the staff intentionally withheld the medication from the
inmate (Estelle v. Gamble, 1976).

Estelle v. Gamble involved the question of when deprivation of medical care
for a physical ailment becomes a constitutional violation. Federal courts had
not yet addressed the question of inmates’ rights to mental health care. That
opportunity came when Larry Grant Bowring filed suit for violation of his
Eighth Amendment rights. Bowring was housed in a Virginia state prison and
was denied parole on three grounds, one of which was a poor outcome on a
psychological evaluation. Bowring responded that that state must provide him
with a psychological diagnosis and treatment so that he might qualify for
parole. Bowring did not prevail because he failed to establish that he did actu-
ally have a mental illness. In its ruling, the U.S. Appeals Court for the Fourth
Circuit did acknowledge that inmates had just as much a right to mental health
treatment as they did to care for physical ailments1 (Bowring v. Godwin, 1977).

The Estelle ruling established deliberate indifference as the standard for
civil rights cases involving medical care, and Bowring confirmed that mental
health treatment does constitute medical care, but there remained ambiguity
about the appropriate test to apply to deliberate indifference cases. Marschke
(2004) noted that, following the Estelle ruling, federal circuits seemed to be

Tartaro 27

split on how to apply the deliberate indifference standards to cases involving
suicides in custody. Courts differed in their use of either the objective or sub-
jective standard of proof.2 With an objective standard, the plaintiffs must sim-
ply demonstrate that the defendants should have known of the risk. The
subjective standard requires that the plaintiffs provide evidence that the
defendants had actual knowledge of the inmate’s problem (Hanser, 2002).

One example of the courts using the objective standard for determining
deliberate indifference in a suicide case is Kocienski v. City of Bayonne
(1991). Helen Catherine Garity committed suicide in the Bayonne Municipal
Jail in 1988 by using her pantyhose to hang herself. This department had a
history of dealing with Garity, so there was a documented history of mental
illness. There were, however, no apparent signs that she was suicidal at the
time of her December 1988 incarceration. When the district court considered
the defendants’ request for summary judgment, they used the objective stan-
dard. Specifically, the court noted,

In order to sustain a claim under the due process clause, Kocienski must, therefore,
show that the officers in question acted with deliberate or wanton indifference to
Garity’s serious psychological needs, needs of which they were aware or should
have been aware at the time of her incarceration. (p. 19, emphasis added)

The court did grant summary judgment, due to the plaintiff’s failure to dem-
onstrate that there was enough evidence of suicidal intent that would have
met the “should have known” standard.

When considering custodial suicide, the pre-Farmer case that provided
the most guidance to district court judges was Manarite v. Springfield (1992).
The Manarite court set a three-prong standard for determining that there was
deliberate indifference in a custodial suicide case. First, plaintiffs must show
that an unusually serious risk of harm was present. Second, there had to be
proof that the defendants actually knew or were willfully blind to the elevated
risk. Finally, there had to be proof that the defendants failed to take steps to
address that risk. The Manarite court approached suicide cases with subjec-
tive standard, allowing only for actual knowledge and “willful blindness” of
an elevated risk to be a sufficient trigger for the necessity to act.

The Farmer Decision

While the lower courts had been split between using “should have known” and
“actual knowledge” when considering corrections officials’ responsibility for
caring for a suicidal detainee since the 1976 Estelle ruling, the Supreme Court
ended that debate with the Farmer v. Brennan (1994) decision (Cohen, 2006).

28 The Prison Journal 95(1)

Farmer was a preoperative transsexual who dressed in a feminine manner and
took hormones that were smuggled into prison. He was serving time in the
federal prison system, and he did so mostly in segregation. The reasons for his
segregation were sometimes for disciplinary purposes, but he was also placed
in segregation at times for his safety. Farmer was transferred to a higher secu-
rity institution and placed in general population where he claimed to have been
beaten and raped. Farmer filed a Section 1983 claim, arguing that the correc-
tions officials should have known of the risk of placing him in general popula-
tion. The district court granted the defendants summary judgment on the
grounds that they lacked actual knowledge of the danger. After the appeals
court affirmed the summary judgment, the Supreme Court took up the question
of whether the objective standard used to define civil law recklessness or the
subjective definition of recklessness used in criminal was appropriate to define
deliberate indifference. The Farmer court mandated the subjective test. In his
opinion, Justice Souter defined the knowledge part of deliberate indifference as
having actual knowledge of an objectively substantial risk of serious harm. In
the event that there is evidence that allows fact finders to conclude that officials
were aware of the risk because the risk was obvious and that evidence is also
satisfactory in meeting the subjective test. When addressing circumstances
when correctional authorities merely should have been aware of a problem,
Souter wrote, “An official’s failure to alleviate a significant risk that he should
have perceived but did not, while no cause for commendation, cannot under
our cases be condemned as the infliction of punishment” (p. 826).

Robertson (2004) considered the impact of Farmer on federal custodial
suicide cases and suggested that the decision encouraged two types of
defenses. The first is the ostrich defense where defendants will argue that
they were simply ignorant of the facts supporting the existence of a high level
of risk. According to the Farmer ruling,

Whether the officials had the requisite knowledge is a question of fact subject
to demonstration in the usual ways, including inference from circumstantial
evidence; a fact finder may conclude that the officials knew of a substantial risk
from the very fact that the risk was obvious; if the prisoner presents evidence
showing that a substantial risk of inmate attacks was longstanding, pervasive,
well-documented, or expressly noted by the officials in the past, and if the
circumstances suggest that the officials had been exposed to information
concerning the risk and thus must have known about the risk, then such
evidence may be sufficient to permit a trier of fact to find that the officials had
actual knowledge of the risk. (HN6)

The second possible defense is the ignorance defense and involves claims
that the defendants failed to consider the facts and draw the inference that

Tartaro 29

there was an actual serious risk. Robertson acknowledged that the use of the
subjective standard does allow for ignorance of obvious risks to remain a
defense but not if “evidence showed he merely refused to verify underlying
facts that he strongly suspected to be true, or declined to confirm inferences
of risk that he strongly suspected to exist” (Farmer v. Brennan, 1994, p. 829).

The Farmer court’s ruling that the subjective standard of recklessness, and
not the civil law objective standard, was sufficient for establishing deliberate
indifference was a setback to plaintiffs. There was, however, one potential
positive aspect of the Farmer ruling that could work to the plaintiffs’ advan-
tage and that was the change to the individual-specific requirement. The
individual-specific requirement prohibited general categories for potential
suicide risk from being relevant in attempts to prove liability. Hanser (2002)
explained, “The individual-specific requirement must show a strong likeli-
hood of risk that is personal to the victim, involving previous threats or
attempts of suicide that are both recent and genuine in nature and are specifi-
cally known by prison staff” (p. 462) and acknowledged that “the individual-
specific requirement had been one of the most difficult hurdles for inmate
plaintiffs to overcome in past litigation. Thus, the Farmer decision seemed to
produce drawbacks and possible benefits for the inmate plaintiff party” (p.
464). After reviewing civil rights claims under 42 USC Section 1983 in the 6
years following the Farmer decision, Hanser expressed doubt that that the
change to the individual-specific requirement actually produced the antici-
pated opportunities for inmate plaintiffs and their decedent parties.

In instances when defendants are claiming lack of knowledge of risk to an
inmate, the Farmer ruling made it imperative that plaintiffs be able to demon-
strate that the defendants did have that knowledge or that the risk was obvious.
Inmates and their survivors could do so through providing evidence that either
the officials knew of a specific suicide risk for a particular inmate (individual-
specific knowledge) or that conditions were such that inmates should be con-
sidered part of “identifiable groups at heightened risk of suicide” (Hanser,
2002, p. 464). If plaintiffs can do that, they then have to show that the defen-
dants were able to infer, on the basis of the known or obvious facts, that the
threat was real. What follows is a review of post-Farmer 42 USC Section
1983 cases that involve defendant arguments of ignorance of suicide risk. This
discussion will highlight the challenges for plaintiffs in federal courts.

Rulings Related to the Individual-Specific Rule

While it does not involve a case of suicide, Hale v. Tallapoosa County (1995)
provides a good post-Farmer example of a case involving violence in a cor-
rectional facility where the individual-specific rule was not applied. Hale was

30 The Prison Journal 95(1)

jailed on failing to appear in court for a hearing on a marijuana charge in May
1990. He was placed in the “bullpen” with several inmates, some of whom
were being held for murder and attempted murder. Hale was concerned for
his safety but did not approach corrections officers to express his fear. After
Hale was beaten while officers were out of earshot and eyeshot of the bull-
pen, Hale filed a Section 1983 lawsuit. Hale was able to prevail in court by
showing that (a) his placement into the bullpen put him at a substantial risk
of serious harm, (b) the defendants were deliberately indifferent to that risk,
and (c) the placement in the bullpen caused him to be beaten. There was evi-
dence that serious inmate–inmate violence occurred regularly in that facility
when it was subjected to overcrowded conditions. In addition, the adminis-
trators were aware of that record of violence during overcrowded periods,
and they disregarded the risk by failing to address the problem. What is
important here is that there was no specific threat to Hale prior to the incident,
but the courts denied summary judgment to the defendants due to the evi-
dence of the generalized risk.

Despite the Farmer ruling, there are examples of suicide cases decided in
district courts, with some rulings affirmed in appeals court, after 1994 where
the individual-specific rule seems to have been applied to the detriment of the
plaintiffs. Robert Frake was arrested and placed in a police holding cell.
While some pieces of clothing commonly used to fashion ligatures in lockups
were confiscated from Frake, he was permitted to keep his jacket, which he
subsequently used to hang himself. Frake’s survivors brought a Section 1983
suit, arguing that there were 20 suicides and 163 attempted suicides by hang-
ing in the holding area, in addition to Frake’s, in the 7 years preceding Frake’s
suicide. The plaintiffs argued that the city’s knowledge of the suicide attempts
combined with its failure to remove the horizontal metal bars that were often
used in detainees’ attempts at self-harm amounted to deliberate indifference.
The district court granted the defendants summary judgment on the grounds
that the city was not aware that Frake, specifically, was suicidal. Knowledge
that the holding cells appeared to be dangerous and provided ample opportu-
nity for suicide attempts was not enough to prove deliberate indifference in
the absence of actual knowledge of a threat specific to Frake (Frake v. City of
Chicago, 2000). This ruling was upheld by the Seventh Circuit Court of
Appeals.

The District Court for the Northern District of Texas made a similar ruling
in Posey v. Southwestern Bell Telephone (2006). Brian Posey was placed in a
holdover cell with a built-in telephone following his arrest. Posey’s mother
called the police station to complain that he was harassing her by phone.
Officers reacted by transferring Posey to a cell that had a broken phone.
Posey used the phone cord to hang himself. Plaintiffs filed suit, citing

Tartaro 31

evidence that there were three suicides using a phone cord at the facility in
the decade preceding Posey’s death. The court decided that three deaths do
not constitute a pattern, and as there was no proof of knowledge that Posey
was suicidal, the defendants were entitled to summary judgment.

In Crocker v. County of Macomb (2005), the district court granted the
defendants summary judgment, and the U.S. Court of Appeals, Sixth Circuit
upheld that ruling, despite the fact that the jail was found to be violation of its
own policies. Carl William Tarzwell was arrested for an outstanding warrant,
but police and the booking staff noted that he smelled of alcohol. The jail did
have two detox cells that allow for constant supervision from the officer’s
duty station, but the booking officer lacked the authority to assign anyone to
those cells. Tarzwell was placed in a holding cell out of view from the work
station. Jail regulations prohibit the placement of any materials on the cell
windows that would obstruct the officers’ view, but the cell where Tarzwell
was located had papers on the windows. There was a phone in that cell, and
the phone cord was not encased in metal sheathing as it was supposed to be.
The plaintiff committed suicide in that cell, using the phone cord to asphyxi-
ate himself. The plaintiff’s family filed suit, claiming deliberate indifference
to Tarzwell’s medical needs. The district court ruled that, to succeed, the
plaintiff would have had to demonstrate that Tarzwell’s risk of suicide was
known or obvious to the corrections staff and that they deliberately failed to
act. Tarzwell had been drinking, but he was able to walk and talk. Without
proof of actual knowledge of the risk specific to Tarzwell, the placement in a
cell with an obstructed view and the availability of an exposed telephone cord
were not enough to establish deliberate indifference.

Rulings on the Obviousness of Risk of Suicide

In the years following Farmer, the lower courts elaborated on exactly what
constitutes an obvious risk. The district court in northern Alabama provided
criteria for the awareness prong of deliberate indifference in Greffey v.
Alabama Department of Corrections (1998). James Edward Morrison was
being transferred from a jail to the custody of the Alabama Department of
Corrections when he slashed his wrist with a razor. When the prison staff
evaluated him, he denied suicidal ideation and claimed that the incident with
the razor was in protest of poor treatment by officers. After being transferred
to another prison, he told the psychologist that he tried to commit suicide
twice in the past. Morrison was placed in segregation after reporting that
inmates were trying to rape him, and he committed suicide while in a cell by
himself. The court granted the defendants summary judgment and provided
guidance for future courts dealing with the issue of awareness. The result was

32 The Prison Journal 95(1)

a four-pronged requirement to satisfying awareness, including that the (a)
inmate previously threatened or attempted suicide, (b) the prior threat or
attempt was known to the defendants, (c) the prior threat or attempt was
somewhat recent, and (d) the prior threat or attempt appeared to be genuine.

How recent does a threat or attempt have to be for courts to decide that
officials must interpret it as a sign that the inmate needs help or extra supervi-
sion? That was the issue in Holland v. City of Atmore (2001). Holland slit his
wrists while in the community, and his family called 911. Holland told the
responding paramedics and deputies that he would continue to try to commit
suicide. Two months after that incident, he was arrested. He was put on sui-
cide watch, because he was intoxicated and remarked that life was not worth
it. While he was in the cell, he banged his head against the wall and was taken
to the hospital. Five months later, Holland was arrested again but appeared to
be calm. One of the staff members from the jail handled the original 911 call
from the wrist slashing incident, so that incident was known to the jail offi-
cials, as was the previous suicide watch and head banging incident. The judge
ruled that

An individual that threatens or attempts suicide has unequivocally expressed an
immediate desire or intent to end his life, and that desire or intent may readily
be assumed to persist only until the passions provoking it have cooled
sufficiently for reason and self-love to regain primacy. A strong likelihood of
self-annihilation may remain for periods of a few hours or even a few days after
a suicide attempt or threat, but the plaintiffs have identified no support for the
proposition that an individual may remain on the brink of suicide for months at
a time, much less that Holland did so. (p. 25)

While Holland’s suicide attempt and threats were known and were consid-
ered genuine, those incidents were considered to be too distant from the sui-
cide for the court to consider Holland’s risk to be obvious.

The case of Snow v. City of Citronelle (2005) involves the courts consider-
ing a case with a more recent and known suicide attempt. Snow was arrested
for driving while intoxicated, and she complained of having a seizure while
in her holding cell. An officer took her to the hospital for treatment, and he
overheard Snow telling the doctor that she had attempted suicide in the past.
That officer went off duty later that night without relaying that information to
any of the other officers. Due to Snow’s erratic and violent behavior while in
the cell, another officer called her parents to ask if they would take custody
of her. They refused but came to the holding facility a few hours later to drop
off some medication. Both parents testified that they informed the officer
(Chennault) that she was suicidal. Chennault called the county jail to inquire

Tartaro 33

about Snow and was told by one of the jail staff members that she tried to cut
her wrist at the jail during the past month. Snow was not put on suicide watch,
and Chennault finished her shift without sharing that information with the
next shift. Snow hanged herself in her cell later the next day. The district
court granted summary judgment to all officers involved in this case, but the
appeals court found that they erred in granting summary judgment to
Chennault, as it is clear that she knew of the significant risk to Snow’s safety
and did nothing to address that risk.

An example of courts ruling in favor of plaintiffs when it appears that
defendants are consciously avoiding obvious signs of risk can be found in
Terry v. Rice (2003). Donald Terry had a long history of mental illness and
incarcerations at the Montgomery County Jail in Indiana. He had a history
of suicide attempts at this jail, and he displayed suicidal behavior in his
1998 incarceration when he cut his wrist while in the jail on March 21. The
jail staff transferred him to a diagnostic center where he was placed on
suicide watch. Six days later, he was transferred back to the jail for finan-
cial reasons, and he was not put on suicide watch. In the month preceding
the wrist-cutting incident, the inmate’s mother called the nurse and the
sheriff to remind them that Terry needed to be on Thorazine. When he was
at the diagnostic center, the doctor adjusted his medication dosages, but the
jail staff did not review his medical paperwork when he returned to the jail,
nor did they review the paperwork when the inmate’s mother called again
to ask them if her son’s medication had been adjusted. Terry’s mother vis-
ited him on April 12, and he seemed irrational during the visit. Later that
day, Terry found out that his girlfriend was leaving him. He committed
suicide that night. The district court denied summary judgment and quoted
McGill v. Duckworth (1991) in its ruling by noting, “Going out of your way
to avoid acquiring unwelcome knowledge is a species of intent. Being an
ostrich involves a level of knowledge sufficient for conviction of crimes
requiring specific intent” (p. 44).

If a past suicide attempt is known to the authorities but not recent, a com-
bination of other factors can be used to provide enough evidence to require
corrections personnel to respond to the threat to safety. Matt Sanville had a
long history of mental illness, and his corrections file provided evidence that
the correctional staff in the Wisconsin Prison System was aware of that his-
tory. Sanville’s presentence investigation report included a note about the 30
misconduct reports that he received during confinement at the county jail,
and both the prosecutor and defense attorneys requested that the judge find a
sentence that included something other than incarceration on account of the
defendant’s mental illness. Sanville was sent to prison, and his behavior
improved during the short period of time that he was medicated. Per the

34 The Prison Journal 95(1)

doctor’s recommendation, he stopped taking his medication due to a stomach
ache which eventually led to an appendectomy. After surgery, Sanville told
the prison psychiatrist that he no longer wanted or needed medication, and
the psychiatrist removed the inmate from his patient list. Sanville’s mental
state declined to the point where he stopped eating and lost one third of his
body weight and was sending letters to his mother that included discussions
of his death. Sanville covered the window of his cell with toilet paper, and the
corrections staff did not remove it, nor did they use the video monitor to
check on his condition for 5 hours. By the time they checked, Sanville had
committed suicide. The district court granted summary judgment to the
defendants, but the appeals court reversed that ruling for the officers respon-
sible for watching Sanville. They noted that the combination of knowledge of
Sanville’s previous suicide attempts, his weight loss, and his letters to his
mother made it obvious that this was not a normally functioning individual
(Sanville v. McCaughtry, 2001).

Inmates who are suicidal do not always disclose that information, even
if they are specifically asked by staff members. In conversations with cor-
rections personnel, this author has heard numerous officers state that the
truly suicidal deny suicidal ideation because they want to be put in a posi-
tion where they can carry out a suicide without intervention from staff or
other inmates. Strickler v. McCord (2004) is an example of such a case
where an inmate with a history of suicidal ideation and self-harming behav-
ior denied his suicidal intent to corrections staff so that he would be able to
attempt suicide. Donald Strickler survived his attempt and then sued the jail
staff for a violation of his Fourteenth Amendment rights, arguing that the
staff should have seen the signs of depression that he exhibited. Strickler
was arrested for driving while intoxicated. Police noted that he had whis-
key, a loaded shotgun, and a suicide note in the car. Strickler denied suicidal
ideation when talking to the police officer and booking officer. Strickler’s
estranged wife called her brother-in-law who was a deputy, and the deputy
informed the jail staff that Strickler was suicidal. Two weeks after his arrest,
Strickler told a counselor that he was having suicidal thoughts, but the
counselor approved him for placement in the general population. A few
weeks later, he attempted suicide by overdosing on Prozac that he hoarded,
but that attempt was unsuccessful and went unnoticed by the correctional
staff, as all it did was prompt Strickler to sleep through dinner. When he
woke up from the Prozac-induced sleep, he cut himself and was found in
his cell soaked in blood. Strickler sued for a violation of his civil rights, but
the court granted summary judgment to the defendants. Strickler argued
that his depression, record of previous suicide attempts, pending divorce,
macabre drawings that he made and kept in his cell, and the fact that he

Tartaro 35

slept through dinner during the undetected Prozac suicide attempt were suf-
ficient signs of an obvious risk. The district court judge disagreed, and
Judge Sharp wrote in his opinion that “A duty to take the maximum precau-
tions for all inmates with any level of suicide risk would make life misera-
ble for any persons at a low level of risk for suicide” (p. 28).

Documentation of suicidal ideation and previous attempts combined with
a violation of the jail’s own policies does not necessarily result in a finding of
deliberate indifference to an inmate’s needs. John Stewart was arrested for
violating probation and placed in the Waldo County (Maine) Jail. At the time
of his booking, he was clearly drunk, and the officer who completed the
screening form checked “intoxicated” and “appears to be despondent/
depressed.” The officer decided to place him in the detox area rather than the
suicide prevention area. Stewart was placed in this same jail 6 weeks earlier
and was documented as being both mentally ill and having a history of sui-
cidal ideation and attempts. For his most recent incarceration, Stewart was in
custody for only 90 minutes before he committed suicide. The officer who
placed Stewart in the detox area had not reviewed his incarceration history,
so he was unaware of the history of suicide ideation and attempts. Following
the suicide, the state department of corrections conducted an investigation
and found that the jail did not meet its minimum standards and that, accord-
ing to the jail policies, the inmate should have been placed in the special
management wing where he would have been subject to close supervision.
This was not enough, however, to establish deliberate indifference in federal
court. The judge ruled the officers did not realize that Stewart was suicidal,
and the court did not consider the neglect to review the inmate’s file within
the first 90 min of incarceration to be a purposeful attempt at avoiding acquir-
ing knowledge of a risk (Stewart v. Waldo County, 2004).

There have been instances of courts finding that the risk to the inmate was
obvious to corrections official even in the absence of previous suicide
attempts. John Rosario Junior was a juvenile when he was confined for men-
tal health treatment and evaluation in a hospital following a fight with his
mother. Rosario was diagnosed with depression, and his hospital records
included information about his suicidal ideation and a statement from his
grandmother about an incident when the juvenile planned to commit suicide
by stabbing. After he was discharged from the hospital, Rosario’s mother told
his probation officer about her son’s condition and his need for medication.
One month later, Rosario was sent to a juvenile facility. He was not screened
by the mental health staff, in violation of the facility’s policies. The facility
staff was aware that Rosario was supposed to take Ritalin, but they never
arranged for him to receive his medication. Rosario committed suicide less
than 6 weeks into his stay at the facility. The District Court of Illinois, Eastern

36 The Prison Journal 95(1)

Division denied the defendants’ motion for summary judgment, ruling that
the plaintiff can demonstrate that there is a “strong likelihood” that Rosario
would commit suicide while in custody. In his opinion, Judge Shadur wrote,

It almost defies common sense to think that a 14-year-old with behavioral
problems that were sufficiently serious to merit hospitalization for over a
month, coupled with intense therapy, could be shut off cold turkey and placed
in a youth correctional facility without further mental health counseling. (Viero
v. Bufano, 1995, p. 26)

Shannon Novak did not have a history of suicide attempts. His doctor in the
community diagnosed him as a paranoid schizophrenic who was impulsive
and a possible suicide risk. When he was arrested, his mother talked to the
deputies about his mental health issues, and the jail personnel agreed to watch
him closely. When he was booked, Novak denied any suicidal ideation or past
attempts. He spent 2 weeks in the general population. During that time, an
inmate approached officers to report that Novak had been pounding on the
door and giggling to himself. The staff did not observe anything unusual. The
Appeals Court judges for the Seventh Circuit granted summary judgment to
the defendants, noting that “Strange behavior alone, without indications that
that behavior has a substantial likelihood of taking a suicidal turn, is not suf-
ficient to impute subjective knowledge of a high suicide risk to jail personnel”
(p. 9). The plaintiffs argued that the risk was obvious and that they discovered
several jail policies that differed from what was required by state statute. The
judges responded that the plaintiffs failed to show that there was a pattern of
suicides at the jail that would have made the jail administration aware that
their policies concerning treatment of mentally ill inmates was inadequate,
and “Even if we were to find that Novack’s suicide itself was a result of uncon-
stitutional conduct, a single instance of allegedly unconstitutional conduct
does not demonstrate a municipality’s deliberate indifference to the constitu-
tional rights of its inhabitants” (pp. 13-14). Judge Williams dissented, noting
that the inmate had threatened to commit suicide 9 months earlier when
arrested by the same police department. That, combined with the mother’s
discussion with the deputies and his behavior in the jail was enough to show
that the jailers knew of the risk to Novak (Turbin v. County of Wood, 2000).

Rulings on Correctional Staff Understanding That
There Is a Substantial Risk

The Farmer ruling requires plaintiffs to present evidence that the risk was
known or obvious and that the officials were able to draw the inference that

Tartaro 37

the inmate in question was at risk of serious harm. Since Farmer, the lower
courts have heard a few cases where this issue was in question.

James Boncher committed suicide while incarcerated in the Brown County
(Wisconsin) Jail. Brown had a long history of alcoholism and three prior
suicide attempts, but this information was not known to the jail staff. During
the entry screening, when he was asked if he had a mental or emotional prob-
lem, he responded affirmatively. When asked if he ever attempted suicide, he
said, “yeah, a couple days ago but I’m fine now.” The officers assumed he
was joking and considered him a “happy drunk.” Boncher was placed in a
regular cell and committed suicide 45 minutes later. The district court granted
summary judgment to the defendants after concluding that it appeared that
the jail staff members assumed that Boncher was joking about his previous
suicide attempt (Boncher v. Brown County, 2001).

In Boncher, jail staff only had Boncher’s words and no written or oral his-
tory of his substance abuse and mental health problems. The corrections staff
was aware of mental health problems and suicide threats in Wever v. Lincoln
County (2004), and the question before the court was whether corrections
personnel were deliberately indifferent to Wever when they ignored his state-
ments. Wever called 911 because he was distraught. When the police arrived,
he initially agreed to go peacefully to the hospital. For reasons that were not
clear to the district court, the police arrested Wever en route to the hospital so
that he was diverted to the jail. He repeatedly told the officers that he would
commit suicide if he was placed in jail. When he entered the jail, Wever said,
“now it’s time.” The police told the jail staff about his threats on the way to
the jail. Wever was placed in an isolation cell with a blanket. He committed
suicide within 30 min. The district court denied the defendants summary
judgment, and the Court of Appeals, Eighth Circuit affirmed that denial on
the grounds that the jail staff appeared to have enough information to realize
that Wever was suicidal.

In another case, police were transporting Brenda Clustka to civil protec-
tive custody when she became distraught. Clustka screamed that the officers
should kill her or she would kill herself. Clustka, who was not handcuffed as
required by police department policy, then wrapped the seatbelt around her
neck in an attempt to choke herself. The police testified that, while they did
not believe that she was joking about wanting to commit suicide, they did not
take the attempt seriously because she would have to let go of the seatbelt
upon losing consciousness. Once she let go of the seatbelt, she would be able
to breathe again, so the officers did not consider it a real suicide attempt.
When the officers arrived at the jail, they did not tell anyone about the inci-
dent in the car. Clustka committed suicide the next day. Attorneys for the
government argued that the police did not consider what they witnessed to be

38 The Prison Journal 95(1)

something that required action to address a serious risk. The district court
granted summary judgment to the officers, but the Appeals Court for the
Ninth Circuit overturned that ruling. Judge Reinhardt wrote the appeals court
opinion and speculated that the officers failed to report the suicidal gesture,
not because they failed to perceive the threat but because to do so would have
revealed the officers’ violation of the handcuff policy (Conn v. City of Reno,
2010).

Discussion

The Supreme Court’s ruling in Estelle v. Gamble confirmed that denial of
medical care should be considered a violation of inmates’ Eighth
Amendment rights in circumstances that involve deliberate indifference to
the inmate’s serous medical needs. The Fourteenth Amendment’s due pro-
cess clause requires that Eighth Amendment protections be extended to pre-
trial detainees. Eighteen years later, the Supreme Court clarified the
meaning of deliberate indifference in the Farmer v. Brennan ruling. The
objective standard of recklessness, which offers plaintiffs the opportunity
to establish deliberate indifference by proving that a risk was such that
officials should have been aware of it, was not sufficient to prove that cor-
rections officials violated an inmate’s constitutional rights. The subjective
standard, with the standard of recklessness used in criminal courts, was
considered by the 1994 U.S. Supreme Court to be the appropriate standard
for deliberate indifference. Only evidence that officials actually knew of
the threat or that the threat was obvious would be sufficient in federal court.
In addition, once evidence of the threat is presented, the plaintiffs need to
show clear evidence that officials drew the inference that the risk was genu-
ine and required a response.

The Farmer ruling was not considered to be an entirely negative outcome
for future plaintiffs, as the court gave inmates and their advocates two encour-
aging signs. The first was the decision that, under certain circumstances, it
was not necessary for plaintiffs to prove that officials were aware of a spe-
cific risk against the inmate. Instead, knowledge about certain categories of
people or a documented history of problems in some locations of the jail
could be enough to demonstrate that officials were aware of the risk to
inmates. Hanser (2002) studied lower court rulings in the first few years after
Farmer and expressed skepticism about whether the lower courts had, in fact,
jettisoned the individual-specific requirement. Three post-Farmer cases
involving suicides that involved inmates being placed in settings that would
increase their chances of being able to commit suicide were presented earlier
in this article, and all three provided evidence that the federal courts continue

Tartaro 39

to use the individual-specific requirements when considering deliberate
indifference in custodial suicide cases.

In Frake and Posey, plaintiffs attempted to demonstrate that there was a
generalized risk of suicide in the locations where the detainees were held, as
the police were aware that inmates held in those same cells had found ways
to commit suicide while under intermittent supervision. Despite this aware-
ness, the police continued to place people into those cells without constant
supervision and that led to additional suicides. In Crocker v. County of
Macomb (2005), the presence of papers covering Tarzwell’s cell door pro-
vided him with the opportunity to use an improperly stored phone cord to
wrap around his neck and remain that way without being detected. Papers
were not permitted to be used for privacy, and phone cords were not permit-
ted to be unsheathed because they presented inmates with the opportunity to
harm themselves or others. None of this evidence helped to produce favor-
able rulings for the plaintiffs. All three courts contended that, in the absence
of proof that the corrections personnel knew specifically that the inmates in
question were suicidal, the officials were not deliberately indifferent to the
inmates’ needs.

The federal courts’ definition of what constitutes enough information for
the risk to an inmate to be obvious to correctional staff is narrow, and it
remains difficult for plaintiffs to convince courts that the risk was obvious.
The four-pronged test presented by the Greffey v. Alabama Department of
Corrections (1998) court and subsequent rulings by other districts showed
that knowledge of one or multiple previous suicide attempts and/or threats is
not enough to present an obvious risk. The attempt must have been recent,
known to defendants, and must have appeared to have been genuine for cor-
rections and police personnel to be accountable for knowing that the inmate
is at risk. The 11th Circuit Court of Appeals echoed this in Holland v. City of
Atmore (2001), noting that suicidal crises are fleeting and are unlikely to
continue uninterrupted for weeks or months. Some pieces of evidence, com-
bined with knowledge of the inmate’s mental health or suicidal history, are
sufficient for courts to rule that there was an obvious risk. For example, being
told of the inmate’s recent suicidal intent by other corrections staff and fami-
lies of the inmates tends to generate an obvious risk (see Snow v. City of
Citronelle, 2005; Viero v. Bufano, 1995). Knowledge of an inmate’s mental
health status, combined with awareness of serious changes to the inmate and
clear negligence on the part of the medical staff to address a known mental
health problem is also sufficient to establish an obvious risk (Sanville v.
McCaughtry, 2001). Any one of the aforementioned factors alone, however,
is unlikely to be sufficient, at least in the federal courts. Even a finding that a
jail or detention facility violated some of its own policies while supervising

40 The Prison Journal 95(1)

an inmate who then committed suicide is not enough to prove deliberate
indifference to an inmate’s risk of harm unless the risk to that specific inmate
was obvious (Crocker v. Macomb, 2005; Stewart v. Waldo County, 2004). An
inmate’s admission of a very recent suicide attempt might also not be enough
to generate knowledge of a suicide risk if the inmate appears to be jovial at
the time that he or she makes that disclosure (Boncher v. Brown County,
2001).

While there are several examples of cases where defendants were denied
summary judgment and the cases were able to then either be settled out of
court or proceed to a jury trial, the standards for awareness of a suicide risk
set in Farmer v. Brennan seemed to have offered little assistance to plain-
tiffs as they try to demonstrate that police or corrections officials violated
the Eighth and/or Fourteenth Amendment rights of people who harm them-
selves while in custody. Regardless of the optimism that Farmer would lead
to lower courts discarding the individual-specific requirement when deter-
mining officials’ liability for harm to inmates, the courts continued to grant
summary judgment to defendants in cases where dangerous conditions
existed but that the plaintiffs did not prove that the defendants knew about
a specific threat to the inmate in question. The courts have also set the stan-
dards for what constitutes an “obvious” threat to an inmate so high that it
remains a bar that is often too high for plaintiffs. Several of the cases
reviewed include statements from federal judges acknowledging that there
was likely negligence associated with the inmate’s death, but negligence is
a lower standard than deliberate indifference and, therefore, insufficient for
federal courts.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publi-
cation of this article.

Notes

1. The court did place limits on what corrections agencies have to provide, stating
that corrections staff had to act when a mental health care provider determined
that the symptoms provide evidence of a serious condition, the condition is cur-
able or treatable, and delay or denial of treatment had the potential to result
in substantial harm. In addition, the treatment was to be limited to what could

Tartaro 41

be provided at a reasonable cost rather than what was merely desirable by the
inmate (Bowring v. Godwin, 1977).

2. For a discussion of how the courts were split and details of particular cases, see
Marschke (2004).

References

Anderson, J. F., & Dyson, L. (2001). Legal rights of prisoners. Lanham, MD:
University Press of America.

Cohen, F. (2006). Custodial suicide: Yet another look. Jail Suicide/Mental Health
Update, 15(1), 1-11.

Hanser, R. D. (2002). Inmate suicide in prisons: An analysis of legal liability under 42
USC Section 1983. Prison Journal, 82, 459-477.

Kappeler, V. E., Vaughn, M. S., & Del Carmen, R. V. (1991). Death in detention:
An analysis of police liability for negligent failure to prevent suicide. Journal of
Criminal Justice, 19, 381-393.

Marschke, L. A. (2004). Proving deliberate indifference: Next to impossible for men-
tally ill inmates. Valparaiso University Law Review, 39, 487-539.

Noonan, M. E. (2012). Mortality in local jails and state prisons, 2000-2010-statistical
tables. Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics.

Robertson, J. (2004). The impact of Farmer v. Brennan on jailer’s personal liability
for custodial suicide: Ten years on. Jail Suicide/Mental Health Update, 13(1),
1-5.

Schlanger, M. (2003). Inmate litigation: Results of a national survey. In National
Institute of Corrections (Ed.), Large jail network, annual exchange (pp. 1-12).
Longmont, CO: National Institute of Corrections Jails Division.

Wallace, H., & Roberson, C. (2000). Legal aspects of corrections. Incline Village,
NV: Copper House.

Court Cases

Boncher v. Brown County 272 F.3d 484 (2001).
Bowring v. E. Godwin 551 F.2d 44 (1977).
City of Revere v. Massachusetts General Hospital, 463 U.S. 239 (1983)
Cole v. Fromm 94 f.3 254 (1996).
Cooper v. Pate, 378 U.S. 546 (1964).
Conn v. City of Reno 591. F.3d 1081 (2010).
Crocker v. County of Macomb 119 Fed. Appx. 718 (2005).
Estelle v. Gamble 429 U.S. 97 (1976).
Farmer v. Brennan, 128 L. Ed. 2d 811 (1994).
Frake v. City of Chicago, 210 F. 3d 779 (2000).
Fulwood v. Clemmer 206 F. Supp 370 (1962).
Greffey v. Alabama Department of Corrections 996 F. Supp. 1368 (1998).
Hale v. Tallapoosa County, 50 F. 3d 1579 (1995).
Holland v. City of Atmore 168 F.Supp. 2d 1303 (2001).

42 The Prison Journal 95(1)

Kocienski v. City of Bayonne 757 F. Supp. 457 (1991).
Manarite v. Springfield, 957 F.2d 953 (1992).
McGill v. Duckworth 944 F.2d 344 (1991).
Posey v. Southwestern Bell Telephone 430 F. Supp 2d 616 (2006).
Ruffin v. Commonwealth, 62 Va. 790 (1871).
Sanville v. McCaughtry: 266 F. 3d 724 (7th Cir. 2001).
Snow v. City of Citronelle, et al. 420 F.3d 1262 (2005).
Stewart v. Waldo County 350 F. Supp. 2d 215 (2004).
Strickler v. McCord: 306 F. Supp 2d 8181 (2004).
Terry v. Rice WL 19221818 (2003).
Turbin v. County of Wood, 226 F.3d at 534 (2000).
Viero v. Bufano 925 F. Supp. 1374 (1995).
Wever v. Lincoln County 388 F. 3d 601 (2004).

Author Biography

Christine Tartaro is a professor of criminal justice, Richard Stockton College of
New Jersey. Her research interests include new generation jails, suicide in correc-
tional facilities, the mentally ill in the criminal justice system, and violence in prisons
and jails.

Feminist Criminology
2014, Vol 9(1) 24 –44

© The Author(s) 2013
Reprints and permissions:

sagepub.com/journalsPermissions.nav
DOI: 10.1177/1557085113501221

fcx.sagepub.com

Article

An Examination of the Media
Portrayal of Femicide–
Suicides: An Exploratory
Frame Analysis

Tara N. Richards1, Lane Kirkland Gillespie2, and
M. Dwayne Smith3

Abstract
Recent research has been focused on the portrayal of intimate partner homicides in the
news media with specific emphasis on the most commonly occurring type, femicides
(the murder of a female intimate partner by a male intimate partner). One important
finding in the analysis of intimate partner homicide is the striking number of femicides
that are followed by perpetrator suicide. Whereas homicide followed by suicide is
a rare occurrence in the context of crime generally, within the context of intimate
partner homicide, femicide–suicide is common. The present research utilized content
analysis to explore the media coverage of a near population of femicide–suicide cases
in the North Carolina from 2002 to 2009 (n = 86). An examination of the article titles
showed that the majority of titles (54%) assigned to the articles describe the crime as an
ambiguous homicide or homicide–suicide and do not indicate the relationship between
the perpetrator and victim. In comparison, results show that 78% of the articles’ text
defined the homicide–suicide as domestic violence. Specifically, in cases where the news
coverage defined the femicide–suicide as domestic violence, the authors identified 4
media frames used (1) femicide–suicide by a male perpetrator, (2) femicide–suicide due
to loss of perpetrator control, (3) femicide–suicide as a mercy killing, and (4) femicide–
suicide due to jealousy. Implications for societal perceptions of violence against women
as well as corresponding victims’ policies/services are presented and discussed.

Keywords
femicide, media issues, qualitative research, victim services, intimate partner violence

1University of Baltimore, MD, USA
2Boise State University, ID, USA
3University of South Florida, Tampa, FL, USA

Corresponding Author:
Tara N. Richards, School of Criminal Justice, University of Baltimore, 1420 North Charles Street,
AC242B, Baltimore, MD 21201, USA.
Email: trichards@ubalt.edu

501221 FCX9110.1177/1557085113501221Feminist CriminologyRichards et al.
research-article2013

http://crossmark.crossref.org/dialog/?doi=10.1177%2F1557085113501221&domain=pdf&date_stamp=2013-10-17

Richards et al. 25

Research consistently demonstrates that domestic violence is a pervasive public health
problem for women (for a review, see Belknap, 2007). In addition, femicide, the kill-
ing of woman by a male intimate partner (Dawson & Gartner, 1998), is a leading cause
of death for females (Anderson, 2002), with some cases ending in perpetrator suicide
(Liem, 2010; Marzuk, Tardiff, & Hirsch, 1992). Despite such evidence, the media
consistently portrays domestic violence as an individualized problem instead of a
widespread social issue (Bullock & Cubert, 2002; Richards, Gillespie, & Smith, 2011;
Taylor, 2009). This is problematic because the news media have the unique ability to
choose which “personal” problems are “invested with a broader meaning and made
available for public consumption” (Sacco, 1995, p. 142). According to Best (1989),
the construction of social issues begins when a private problem is identified and
invested with broader meaning for society. The interpretation of such social problems
is important at the individual and societal levels. At the personal level, naming a prob-
lem assists individuals in understanding their experiences in relation to the experi-
ences of others as well as making decisions on a course of action (Kelly, 1988). At the
public level, naming a problem is the first step to the creation of collective resources
or policy solutions (Mehrotra, 1999). Consequently, the ways in which the news media
choose to frame domestic violence, especially at its deadliest, can have important ram-
ifications, influencing how society perceives the dynamics of such violence, its vic-
tims and perpetrators, and most importantly, the public’s role in potential solutions
(Gillespie, Richards, Givens, & Smith, 2013).

Femicide–suicide is often the violent culmination of years of male perpetrated
domestic violence (Moracco, Runyan, & Butts, 1998). Although recent scholarship
has focused on media representations of femicides alone (Bullock & Cubert, 2002;
Gillespie et al., 2013; Meyers, 1994; Richards et al., 2011; Taylor, 2009), media repre-
sentations of femicide–suicides have been largely ignored. This gap in the literature is
problematic because it hides the dangers of domestic violence—batterers often devolve
to the point where they take their own lives as well as the lives of their female intimate
partners. The present study aims to expand our knowledge concerning the media’s
representation of femicide–suicide and the media’s impact on the construction of vio-
lence against women as a social problem. First, the presentation of homicide–suicide
in the news will be reviewed. Second, a brief overview of the social construction of
crime and gender will be presented. Next, existing media frames used by journalists to
tell the story of domestic violence will be examined. Finally, the present study will
build on the past literature by exploring the media frames employed in stories that
define the femicide–suicide as domestic violence compared with stories that fail to
define the femicide–suicide as domestic violence and discussing how such framing
may impact policies concerning violence against women.

Homicide–Suicide in the News

Since no study to date has been focused on how the media frames homicide–suicides,
Marzuk et al.’s (1992) typology of homicide–suicides is a useful place to begin
this examination. Marzuk et al. identified the four most common types of

26 Feminist Criminology 9(1)

homicide–suicide based on the relationship between victim and offender: (1) spousal
homicide–suicide due to jealousy, (2) spousal homicide–suicide due to declining
health, (3) filicide–suicide, and (4) familicide–suicide. As reviewed by Liem (2010),
research indicates that the killing of an intimate partner is the most common type of
domestic homicide (compared with other types such as parents killing their children)
and intimate partner homicide–suicide is the most common type of homicide–suicide.
Past studies have also found that in the case of intimate partner homicide–suicides the
perpetrator is often male, there is a history of physical abuse by the perpetrator, and
the perpetrator suffers from mental illness (primarily depression; Liem, 2010).
Furthermore, two subgroups of intimate partner homicide–suicide have been identi-
fied: (1) those that indicate jealousy by the perpetrator or, more likely, dependency on
the victim by the perpetrator and (2) those that reference old age or ill health by the
victim and/or the perpetrator (Liem, 2010; Marzuk et al., 1992).

According to existing research, homicide–suicide is one of the most newsworthy
cases of suicide and is more likely to appear in print media than independent acts of
suicide (Jamieson, Jamieson, & Romer, 2003). Jamieson et al. (2003) found that while
homicide occurs in fewer than 5% of suicides, homicide–suicide stories accounted for
more than one third of all suicide news coverage in the New York Times for the years
1990, 1995, and 1999. In regard to the newsworthiness of homicide–suicide, inter-
views with news reporters indicate that homicide–suicides are perceived as very news-
worthy, more so than suicides, and that homicide–suicides are likely to always be
covered, whereas suicides must display certain characteristics to be worthy of news
coverage (Jamieson et al., 2003). Furthermore, when the cause of a suicide can be
labeled or connected to a larger problem, the newsworthiness increases.

Research on media representation in the 1990s suggests that the problem of domes-
tic violence has been utilized by journalists to provide a compelling narrative for
homicide–suicides (Jamieson et al., 2003). However, several more recent studies have
indicated that contextualizing homicides and homicide–suicides within the frame of
intimate partner violence (IPV) as a social problem only occurs in the minority of
cases (ranging from 10% to 34%) (Bullock & Cubert, 2002; Gillespie et al., 2013;
Richards et al., 2011; Taylor, 2009). These findings suggest that further examination
of the reporting of intimate partner homicide–suicide in print journalism is necessary
to understand the presentation of homicide–suicide in the news and assess how this
particular type of social problem is represented to society.

Social Construction of Violence Against Women

The understanding of violent crime victimization and/or offending is often a “medi-
ated experience” (Surette, 2007) given that the majority of individuals have little
firsthand experience with violent crime, either as a victim or offender. Consequently,
individuals gain their understandings or “social construction” of crimes like domestic
violence via avenues such as the media, in the form of a mediated experience. More
specifically, Berns (2001) suggests that newspaper articles are one of the “public
arenas where images of domestic violence are constructed, debated, and reproduced”

Richards et al. 27

(pg. 263). Thus, the degree to which the news media provide distorted images con-
cerning domestic violence (e.g., she lets him abuse her, she could leave if she wanted,
etc.) as well as victims’ roles in such violence (e.g., if you were a better wife/mother
the abuse would stop) propagate myths that violence against women is not a serious
crime, that women are responsible (or partly responsible) for their victimization, or
both. On the other hand, accurate reporting of domestic violence, especially at its
deadliest, provides a unique opportunity to shape public opinion and mobilize com-
munity support.

Social problems can be framed in many ways, and these various frames suggest
divergent causal explanations and corresponding resolutions (Gusfield, 1989;
Schneider, 1985). The news media are an important part of this framing process due to
their power to proliferate some views and repress others. In addition, the media fre-
quently emphasize only certain kinds of criminals and their victims, while downplay-
ing or ignoring others, and disseminating powerful messages concerning who matters
most in society. Mainstream media representations also create an interpretive frame-
work for solutions to social problems that favor some types of social change over oth-
ers (Ericson, Baranek, & Chan, 1991). The media framing literature has examined
how the media utilize templates to present a structured understanding of social phe-
nomena to society. While these templates may vary in their content and accuracy they
provide the public with an avenue for understanding a broad array of events including
violent crime (Surette, 2007). In the case of domestic violence, these frames shape the
public’s perception about the occurrence, impact, and characteristics of violence
against women (Bullock & Cubert, 2002). As campaigns to prevent domestic violence
become increasingly widespread and news media attention is focused on legislative
issues relating to violence against women (e.g., reauthorization of the Violence Against
Women Act), it is prudent to examine how such violence is presented, and by exten-
sion understood, to the public. Funding and resources for victim services (including
suicide prevention), as well as support for legislation, may be influenced by how rel-
evant issues are presented by the media to society.

Framing Femicide

Media frames are based on three primary components: sources, word choice or lan-
guage, and context. The three components of media frames, sources, language, and
context, will be briefly discussed in reference to how they may shape perceptions of
domestic violence,1 including femicide–suicide, and how they operate in framing fem-
icide events. Examinations regarding the influence of sources in framing domestic
violence events have focused on three source types: official sources such as law
enforcement, individuals who personally know/knew the victim and/or offender (e.g.,
neighbors, friends, family), and qualified non-law enforcement sources (e.g., advo-
cates and shelter operators in the case of domestic violence). It has been well estab-
lished that, in regard to crime news, law enforcement are the primary contact for
information and as a result the opinions of law enforcement influence the framing of
criminal events (Chermak, 1995; Ericson, 1989; Fishman, 1981; Surette, 2007). Law

28 Feminist Criminology 9(1)

enforcement officers are often first responders and are most likely to provide news
media with a description of the crime and those involved from a legalistic standpoint
(e.g., history of police visiting the home). Police and officials are also most likely to
be quoted or paraphrased in newspaper coverage of femicide, because they have the
details of the fatal event, and thus, police appear as sources in all types of media
frames (Gillespie et al., 2013; Richards et al., 2011). Extant research has indicated that
official sources provide information that reinforce severely limited “official” defini-
tions of what should be considered criminal behavior (Gorelick, 1989; Humphries,
1981) and that the knowledge of official sources is limited to previous legal interac-
tions between the victim(s), perpetrator(s), and police. Limited definitions of crime
may be especially relevant in the study of femicide–suicides since suicide-completers
are often pathologized as “sick” and the deaths of their victims may be seen as a trag-
edy of circumstance.

In addition to law enforcement sources, the media may also attempt to cite indi-
viduals who know/have known the victim and offender. While family members and
close friends are more likely to have intimate knowledge about the victim and/or
offender, the media more routinely utilize neighbors as sources in cases of domestic
violence (Gillespie et al., 2013; Richards et al., 2011; Taylor, 2009). Neighbors might
be a more convenient source of information or they might be more willing to talk to
the media than family members, but they also are shown to be more likely to charac-
terize the event as an isolated incident when in fact there was a history of abuse
(Taylor, 2009). The framing of femicide–suicide events in particular may be influ-
enced by the use of neighbors in comparison to family or friends since the event can-
not be shaped by future coverage regarding the legal outcome of the perpetrator.
While the knowledge of friends and family may be more descriptive than the knowl-
edge of neighbors, the most important source in the framing of domestic violence
incidences as representative of a larger social issue is experts and professionals in the
field (Richards et al., 2011).

Previous studies have lamented the limited use of qualified non-law enforcement
sources such as academics, and in the case of domestic violence, victim advocates
(Bullock & Cubert, 2002; Byerly, 1994; Gillespie et al., 2013; Meyers, 1997; Richards
et al., 2011; Taylor, 2009). Given the public health concerns associated with suicide
reporting (in regard to preventing suicide contagion) citing sources that have profes-
sional knowledge about domestic violence, femicide, and suicide is imperative to
framing femicide–suicide as a prominent social problem. In cases of femicide–sui-
cides, domestic violence experts are vital to contextualizing such events as domestic
violence and dispelling myths about the true dangerousness of batterers.

The second influential framing component is language. Word selection affects the
public’s understanding and interpretation of events (Bullock & Cubert, 2002; Gillespie
et al., 2013. In reference to reporting on femicide, language has the power to portray
domestic violence as a social issue that requires community action or as a singular,
isolated violent event caused by the action or inaction of the victim. The identification
of a homicide–suicide event involving intimate partners as domestic violence, IPV, or
femicide requires the utilization of specific language. It is often the case that titles/

Richards et al. 29

headlines and even the content of an article fail to include language that distinguishes
a domestic homicide from a homicide between strangers (Bullock & Cubert, 2002;
Gillespie et al., 2013). Distinguishing domestic violence from other forms of violence
is important for accurately portraying femicide and drawing attention to the most com-
mon type of homicide–suicide event, femicide–suicide. Furthermore, language is
thought to be extremely important in shaping suicide events in a way that will mini-
mize the likelihood of contagion effects (Centers for Disease Control and Prevention
(CDC), 2001; Jamieson et al., 2003; O’Carroll et al., 1994). For example, it has been
suggested that when reporting suicide events the word suicide should not appear in the
headline; the deceased should be described has “having died by suicide” not “a sui-
cide” or “committed suicide”; and in reference to contrasting deaths and attempts,
terms associated with success should be avoided, (e.g., “the death was a result of a
successful suicide attempt”; CDC, 2001). Therefore, language used to describe femi-
cide–suicide may influence the framing of such events in ways that are distinct from
femicides not followed by perpetrator suicide.

Third, in examining how the media frames femicide–suicide, it is necessary to con-
sider context. Prior literature has revealed that domestic violence is most often cyclical
with periods of grave violence by the perpetrator followed by periods of calm and
tranquility (for a review, see Belknap, 2007). Prior history of domestic violence is the
most common risk factor of deadly domestic violence while previous arrest for IPV,
including highly supervised sanctions for offenders, continued safety planning for vic-
tims, and coordination of advocates, law enforcement, and the community are all pro-
tective factors (Campbell, Glass, Sharps, Laughon, & Bloom, 2007). Despite this,
more often than not, journalists do not effectively contextualize domestic violence as
an ongoing personal and social issue (Bullock & Cubert, 2002; Gillespie et al., 2013;
McNeill, 1992; Taylor, 2009). The recommendations for contextualizing femicide
more effectively, such as detailing the couple’s history of abuse, indicating the preva-
lence of femicide, and treating femicide as distinct from other forms of violence, may
come in conflict with some of the recommendations of suicide reporting, like empha-
sizing a decrease in the occurrence of suicide, avoiding use of the word suicide in the
title, and avoiding the prominent placement of suicide stories (CDC, 2001). Some
recommendations for contextualizing femicide and suicide events are congruent such
as not romanticizing the event or emphasizing the positive traits or aspects of the ini-
tiator, including referral information and sources for crisis intervention or shelters,
citing professionals familiar with the issue (suicide, domestic violence), and avoiding
oversimplifying the causes or presenting the causes as unknowable or unexplainable
(Bullock & Cubert, 2002; CDC, 2001; Gillespie et al., 2013; Taylor, 2009).
Contextualizing femicide–suicide events may be unique from femicide and suicide
events. The differences and commonalities between femicide, suicide, and femicide–
suicide may influence how these events are framed. Therefore, in addition to the iden-
tified typologies of homicide–suicide events, the present study is also informed from
the media frames employed in the representation of femicide events.

A recent frame analysis of newspaper articles covering femicide cases in North
Carolina for the years 2002-2007 identified media frames for articles defining the

30 Feminist Criminology 9(1)

event as domestic violence and for articles not defining the femicide event as domestic
violence (Gillespie et al., 2013). For domestic violence-defined (DV-defined) and
non-domestic violence-defined (non-DV-defined) articles, four frames emerged: (1)
normalizing the event as commonplace, (2) framing the event as an isolated incident,
(3) focusing on the actions of the victim or victim blaming, and (4) finding fault with
the criminal justice system. In addition, a fifth frame used in DV-defined articles (5)
domestic violence as a social problem contextualized the femicide as domestic vio-
lence. For articles not defined as domestic violence, two additional frames were identi-
fied, (6) minimizing the event by focusing on an alternative issue in the lives of the
victim/perpetrator (e.g., illness), and (7) personal loss of control or moral breakdown
by the perpetrator. Those articles that focused primarily on a secondary issue were
largely used in cases where the couple was elderly and the femicide resulted in perpe-
trator suicide. Given that prior homicide–suicide research (see Liem, 2010; Marzuk
et al., 1992) has identified declining health as one of the two subgroups of intimate
partner homicide–suicide, this frame may be especially pertinent in the present sample
of femicide–suicide articles.

Present Study

Existing research on news reporting of femicide incidences have indicated through
descriptive statistics that a portion of femicide cases involve perpetrator suicide (see
for example, Richards et al., 2011; Taylor, 2009); however, the portrayal of femi-
cide–suicide cases have not been specifically analyzed or discussed in these studies.
Considering the occurrence and social implications of femicide–suicide in domestic
violence, examining how these stories are framed by print media is important to
informing society’s perception of domestic violence and the prevalence of femicide–
suicide as the final outcome. While existing literature has examined the media rep-
resentation of femicide broadly (Bullock & Cubert, 2002; Gillespie et al., 2013;
Meyers, 1994; Richards et al., 2011; Taylor, 2009), the present study is the first to
use content analysis to examine the print media’s representation of femicide–suicide
specifically.

As suggested above, there are several frameworks commonly used in reporting
femicide-only and/or homicide–suicide reporting. We examine the applicability of
these existing frames to femicide–suicide, by analyzing the components of these
frames—the sources used, language and word choice, and the context of the incident—
in newspaper coverage of a near population femicide–suicides in North Carolina
from 2002-2009. As evidenced by the existing studies examining femicide in the
news, there appears to be a duality in the representation of intimate partner homicide
by the media such that the majority of stories do not define the event as domestic
violence, while a minority of stories defines the event as domestic violence. The
present study intends to contribute to the literature on the representation of domestic
violence in the news by examining the framing of femicide–suicide cases over a
period of 7 years in a single state. The following research questions will guide the
present study:

Richards et al. 31

1. How are femicide–suicide cases reported by the news media in terms of estab-
lished framing components—source, language, and context? To what extent
are femicide–suicide cases contextualized as domestic violence in news
reporting?

2. Do news articles that contextualize the femicide–suicide as domestic violence
use different frame(s) than those that do not contextualize the femicide–suicide
as domestic violence?

Method

Data Collection

The initial step in data collection was to isolate the population of femicide–suicide
cases for the state of North Carolina by obtaining a list of femicides (N = 405) from
the North Carolina Coalition Against Domestic Violence (NCCADV) for the years
2002 (the 1st year of their work) through 2009 (the last complete year). This group
is comprised of a network of service providers and law enforcement personnel who
systematically identify domestic murders across the state by conducting daily
searches of news media stories and public sources (see http://nccadv.org/homi-
cides.htm).

The next step in the process was to exclude all cases that could not be identified
specifically as cases of femicide–suicide. Femicide–suicides were operationalized as
femicides (the killing of a female by a male intimate partner) that resulted in the imme-
diate suicide of the perpetrator (n = 109). Adopting a strict interpretation of femicide–
suicide, cases where the perpetrator was arrested and/or went on the run after the
femicide and later committed suicide (n = 9) were excluded from this study in an
attempt to analyze cases in which the homicide and suicide were part of a continuous
action on the part of the perpetrator. As described further hereafter, this screening pro-
cess resulted in a list of 100 cases. Items of interest were then recorded for each case
from NCCADV records. These included the name of the victim and perpetrator, the
date of the incident, the relationship of the victim and perpetrator, the weapon used,
and the age and race of the victim and the perpetrator (when available).

Once a complete list of known femicide–suicides was compiled, the researchers
performed a news article search using the electronic newspaper databank, Access
World News. Access World News presently includes 30 newspapers serving rural and
metropolitan areas of North Carolina. News articles were identified using two search
terms for each case: (1) the victim’s name paired with the key words “murder” and
“suicide” and (2) the perpetrator’s name paired with the key words “murder” and “sui-
cide.” Following this search method, cases for which coverage could not be found
were subject to a wider internet search that focused mostly on the websites of small-
town or small-region newspapers near the setting of the femicide–suicide. Articles
authored by specific writers as well as anonymously authored “wire” stories were
included; editorials and articles on the broad issue of domestic violence that men-
tioned a case as an example were excluded.

32 Feminist Criminology 9(1)

Missing Data

Having exhausted all internet search options, newspaper coverage for a total of 14
cases (approximately 14% of femicide–suicides) could not be recovered. This missing
data can be classified into three categories: (1) 2 cases where the femicide–suicide was
mentioned in an article about domestic violence, but had no articles dedicated to the
case specifically; (2) 1 case where only an editorial about the femicide–suicide
appeared in the newspaper; and (3) 11 cases where a story about the femicide–suicide
did not appear in the newspapers accessed. Overall, the proportion of cases for which
coverage was obtained (86%) represents a substantial sample (actually, near popula-
tion) of femicide–suicides that occurred in North Carolina during the 8 years of
2002-2009.

Description of Sample

As described hereinbefore, there were 405 femicides in North Carolina recorded by
the NCCADV for the period 2002-2009; of these 109 (27%) were determined to be
femicide–suicides. After excluding cases where the suicide did not immediately fol-
low the femicide, and cases for which coverage could not be identified, 86 cases
remained for analysis. In total, 147 newspaper stories were found for the 86 cases.
Among the 86 femicide–suicides cases in the sample, 61% merited 1 newspaper arti-
cle, 24% 2 articles, and the remaining 15% 3 or more articles. The most covered
femicide–suicide case was covered in 9 different newspaper articles. Table 1 presents
the comparison of the descriptive characteristics for the missing cases and the cases
included in the sample.

Analytic Technique

Content analysis was conducted using Atlas.ti V5.0 (Muhr, 2004), a qualitative data
management package. Each of the 147 newspaper articles representing the 86 femi-
cide–suicide cases were uploaded into an Atlas.ti database and were electronically
coded by one of two researchers according to a coding scheme adapted from past
research examining newspaper coverage of femicide (see Gillespie et al., 2013). For
the first level of coding, the two researchers reviewed 25 news articles independently,
noting key framing constructs described previously (see appendix for coding scheme
employed). After this initial step in coding, both researchers discussed their individual
coding to determine that a unified coding schema had been followed. A comparison of
coding decisions yielded a Cohen kappa of .80, a level that exceeds the established
threshold of acceptability (.50; Landis & Koch, 1977). Once inter-rater reliability was
established, each coder independently coded approximately half of the remaining sam-
ple. Each article was read for content specifically relating to the three components of
media frames: sources of information, language (e.g., word choice in titles, character-
istics of the victim and perpetrator), and context (e.g., couple’s history, perpetrator’s
history of violence).

Richards et al. 33

Results

The following results describe how femicide–suicide articles are framed by the news
media. First, the articles’ titles are examined to determine how femicide–suicides are
described by their headlines. Next, the three components of print media framing—
sources, language, and context—are briefly explored. Finally, the various media
frames used by journalists in cases of femicide–suicides are identified.

Article Titles

First, each articles’ title was examined to determine whether the article was identifi-
able to readers as a femicide–suicide event or even as a crime between intimates. Titles
were collapsed into three different categories: (1) titles that described the crime as an
ambiguous homicide/homicide–suicide (n = 79), (2) titles that identified the relation-
ship between the perpetrator and victim, but were vague as to the crime (n = 30), and
(3) titles that indicated the crime was a femicide–suicide (n = 38).

The majority of titles (54%) assigned to femicide–suicide articles described the
crime as an ambiguous homicide or homicide–suicide. These titles did not indicate a
relationship between the perpetrator and the victim such as “Sheriff’s office: Deaths a
murder, suicide” (Rickert, 2007) and “Durham deaths murder–suicide” (Associated

Table 1. Characteristics of Femicide–Suicides: Missing Cases Versus Sampled Cases.

Case characteristic Missing cases (N = 14) Sampled cases (N = 86)

N (%) N (%)

Relationship
Husband 5 (36) 38 (44)
Boyfriend 5 (36) 20 (23)
Ex-boyfriend 1 (7) 8 (9)
Ex-husband 1 (7) 8 (9)
Estranged husband 2 (14) 12 (14)
Weapon used
Gun 11 (79) 79 (92)
Knife/Cutting instrument 3 (31) 5 (6)
Physical force/Blunt trauma — 1 (1)
Other (e.g., drowning,

burning, etc)
— 1 (1)

County
Urban 11 (79) 70 (81)
Rural 3 (21) 16 (19)
Age M (Range) M (Range)
Victim 39 (26-56) 39 (18-84)
Perpetrator 42 (25-61) 42 (19-80)

34 Feminist Criminology 9(1)

Press, 2008). Other times, the title did not even reveal that the crime was a murder–
suicide such as, “Deaths shock friends” (Dorell & Weigl, 2003) or “Two shot to death
near beach town” (Associated Press, 2002b).

The next category of titles (20%) assigned to femicide–suicide events indicated that
there was an intimate relationship between the two victims, but was vague on the cir-
cumstances of the crime. Many times these titles did not distinguish which individual
was the perpetrator and which was the victim. Examples of such titles include, “Couple
dead in apparent murder–suicide” (Associated Press, 2003a) and “Husband, wife
found dead in home—Couple had recently separated, police suspect murder–suicide”
(Mitchell, 2008). In other cases, titles did not indicate that a murder and a suicide had
occurred, “Love tangle ends in two deaths” (LaGrone, 2007).

A third group of titles (26%) did describe the crime as a femicide–suicide incident.
These titles distinguished that the perpetrator was male and the victim was female and
described the nature of their present or past intimate relationship. Examples of this
type of title included: “Woman feared for life—Wife in murder–suicide case took out
restraining order against husband” (Boone & Hewlett, 2007); “2 die in murder–sui-
cide, Police: Husband shot wife at workplace, then turned gun on himself” (Ferreri &
Velliquette, 2004); and “Woman murdered in street—Shooter turns weapon on him-
self in apparent domestic dispute” (Wetherington, 2009).

Framing Components

As a first step in exploring the ways in which the media frames femicide–suicides, the
framing components of the femicide–suicide articles were examined. Findings are pre-
sented in Table 2. Beginning with sources, the majority of articles (88%) used the
police as a primary source of information about femicide–suicide cases while only a
minority of articles included the input of domestic violence experts (14%). Findings
also revealed that neighbors were used as sources in fewer articles than friends and
family, 29% versus 52% of articles, respectively. Next, the language used to describe
the perpetrator and/or relationship between the perpetrator and the victim was explored.
Results indicated that perpetrator substance use and/or mental health issues were men-
tioned in only a minority of news coverage, 3% and 6% of articles, respectively.
Similarly, other potential triggers revealed in past homicide–suicide literature such as
infidelity (3%), marital discord (7%), separation (4%), and money problems (1%)
were also rarely mentioned in femicide–suicide news articles. Finally, in regard to
contextualizing the femicide–suicide as domestic violence, only 13% of articles indi-
cated prior domestic violence by the perpetrator; 3% of articles described past mutual
domestic violence between the victim and perpetrator; and 14% of articles indicated
ambiguous domestic violence between the couple without distinguishing the perpetra-
tor from the victim. It was also found that 16% of articles discussed different mecha-
nisms the victim used to protect herself from domestic violence by the perpetrator
such as taking out a protective order against the perpetrator or using a domestic vio-
lence shelter; however, 9% of those articles argued that the victim did not take enough
protective action by, for example, not showing up for court dates or returning to the

Richards et al. 35

abuser. Finally, 14% of articles mentioned other state and local incidences of deadly
domestic violence, while 5% provided information for local marches and/or vigils to
raise awareness concerning the problem of domestic violence.

Media Frames

Finally, the different media frames used for femicide–suicide articles were identified.
These results are presented in Table 3. Similar to past research on media framing of
femicide-only cases (Gillespie et al., 2013), the first distinction made by the research-
ers regarding how the media frames femicide–suicides was to distinguish articles that
identified the crime as femicide–suicide (i.e., domestic violence committed by a male
perpetrator against a female victim) from those that did not. Articles that failed to
define incidents as femicide–suicides were framed in one of two ways, as an ambigu-
ous murder–suicide (n = 25) or as a mystery (n = 6).

Twenty-five articles (17%) framed the femicide–suicide as an ambiguous homicide
meaning news coverage failed to indicate who the victim was and who was the

Table 2. Components of Femicide–Suicide News Articles.

Total articles (N = 147)

Article components N (%)

Sources
Official/Public 129 (88)
Neighbors 42 (29)
Friends/Family 76 (52)
DV experts/Advocates 21 (14)
Language
Perpetrator substance use 4 (3)
Perpetrator mental health issues 9 (6)
Infidelity 3 (2)
Marital discord 10 (7)
Perpetrator/Victim separated 5 (4)
Money problems 2 (1)
Context—Prior IPV
History of DV by the perpetrator 13 (9)
History of mutual DV 3 (2)
History of ambiguous DV 14 (10)
Victim protective actions 24 (16)
Not enough victim protective actions 13 (9)
DV as a social problem
Mentions other cases of DV 20 (14)
Mentions marches or vigils for DV 7 (5)

Note. DV = domestic violence.

36 Feminist Criminology 9(1)

perpetrator and, in some cases, that the incident was a murder–suicide. In one such
article, the scene was described in the following way, “‘Their bodies lay opposite each
other across the Lee Road home’s living room,’ said law enforcement. Law enforce-
ment said it was too early to determine who shot whom or what led to the shooting”
(Mitchell, 2008). Another article described a femicide–suicide where the father,
mother, and two children were all found dead, “the lead investigator wouldn’t say
which parent was suspected, saying investigators would wait for initial autopsy reports
this week before releasing details” (Kirkpatrick, 2009).

Another 6 articles (4%) framed the femicide–suicide as a mystery, for example one
article described,

“There were no signs of any other criminal or suspicious activity at the residence,” police said in
a news release. They added that they had not previously responded to calls for service at the West
Davis Street house. “It was not an ongoing domestic situation,”(the police captain) said.
(Brevorka, 2005)

In another femicide–suicide article, law enforcement stated that they were “leaning”
toward murder–suicide:

“That’s the general consensus by looking at the bodies and the information gathered at the
scene,” an officer said. “We are pretty much still leaning toward marriage rumblings (as the
possible motive). But we may never know. The two people who could really shed light on it
aren’t here.” (Rickert, 2007)

Of particular interest to the present study was the identification of frames used in
cases where the news coverage defined the femicide–suicide as domestic violence.
As shown in Table 3, these frames include (1) femicide–suicide by a male perpetra-
tor (n = 85), (2) femicide–suicide due to loss of perpetrator control (n = 20),
(3) femicide–suicide as a mercy killing (n = 6), and (4) femicide–suicide due to
jealousy (n = 5).

First, 85 articles (58%) clearly identified the crime as a femicide–suicide commit-
ted by a male intimate partner. These articles were more likely to talk about past

Table 3. Media Frames Used in Femicide–Suicide Articles.

Frames Total articles (N = 147)

Frames appearing in articles indicating femicide–suicide
Male perpetrated femicide–suicide 85 (58)
Loss of perpetrator control 20 (14)
Mercy killing 6 (4)
Perpetrator jealousy 5 (3)
Frames appearing in articles not indicating femicide–suicide
Ambiguous murder–suicide mystery deaths 25 (17)
Mystery deaths 7 (5)

Richards et al. 37

domestic violence between the couple and cite domestic violence experts and/or
resources for domestic violence. In one such article, an attorney lamented, “What are
you supposed to do? We’ve got the protection order, but it’s only a piece of paper. It’s
not a bulletproof vest” (Ferreri & Velliquette, 2004). In another article, a domestic
violence expert warned,

Folks need to take domestic-violence incidences very seriously. If someone is in danger and
we’re believing the best about their abuser, then we’re not giving them the help they need . .
. It’s very difficult for a victim to leave an abusive relationship. It takes a great bit of strength,
faith and courage. (Boone & Hewlett, 2007)

Among the articles defined as femicide–suicides, a minority of articles contained
frames used to assign a “cause” as to why the perpetrator committed the deadly domes-
tic violence. The first frame, perpetrator loss of control, was used to frame 20 articles
(14%). The articles described femicide–suicide events that occurred after the female
intimate partner had left the perpetrator. For example, in one case law enforcement
expressed the following after a woman was killed after leaving her abuser, “Victims
need to take every threat seriously. We’ve learned a person becomes more dangerous
when they feel they have nothing left to lose” (Hess, 2002). In another case, after sepa-
rating from her abuser, family members stated this about why a woman’s husband
committed the femicide–suicide, “He couldn’t stand the fact that she was leaving him
and taking the kids” (Manware & White, 2002). In an additional article a witness
described that the perpetrator allowed everyone to flee the scene except the victim and
then, “He then took (the victim) behind a small shed in the woods nearby, where the
two exchanged heated words. He said, ‘If I can’t have you, nobody can’ . . . we heard
two shots” (Kenna, 2003).

In another 5 articles (3%), perpetrator jealousy was identified as the cause of the
femicide–suicide. In these cases, the victim had not only left her abuser but had begun
a new relationship. In one such case, a neighbor described that the perpetrator had
“cracked” and that the perpetrator “wanted her for himself . . . He just couldn’t stand
for her to have someone for herself” (Associated Press, 2003b). In another case, a man
shot his wife and then himself after finding out a day earlier that his estranged wife had
begun seeing a co-worker (McDonald, 2004).

Finally, 6 articles (4%) described the femicide–suicides as “mercy killings.” While
it may have been expected that more articles using this frame would be present in a
sample of femicide–suicides, this number is fairly similar to prior research findings
with articles representing femicide (not followed by suicide) events (i.e., Gillespie
et al. [2013] found 6 (2.7%) articles used this frame).In these cases, the victim was
suffering from a serious illness. Articles described as mercy killings did not assign any
blame to the perpetrator but instead described the act in sympathetic terms. For exam-
ple, one article stated, “Authorities believe that a Wilkes County man shot his extremely
ill wife before shooting and killing himself this past weekend” (Associated Press,
2002a), while another declared that “after more than 55 years of marriage, an elderly
couple’s life together came to an end in an apparent murder–suicide” (Welch, 2004).

38 Feminist Criminology 9(1)

Discussion

The present study is the first research to date to explore the ways femicide–suicides are
framed by the news media. The research examined newspaper coverage for a near
population of femicide–suicide cases in North Carolina over an 8-year period. The
present study began by exploring the language used to describe the femicide–suicide
event in the title/headline. Word choice in headlines is important because it conveys
the main idea of the article and allows readers to choose stories of interest or impor-
tance to them. In addition, many news consumers may “skim” the headlines instead of
reading entire articles to gain information about news. Thus, it is vital that femicide–
suicides be correctly titled. First, findings demonstrate that the majority of femicide–
suicide articles are titled in ways that indicate that they are ambiguous homicides.
Therefore, the majority of femicide–suicides articles may be easily misidentified by
readers as stories about stranger homicides instead of deadly domestic violence con-
cluding in perpetrator suicide. Next, a group of stories included titles that identify a
relationship between the perpetrator and the victim but are vague on the crime. Readers
may believe that these articles describe the murders of husbands by wives or even the
murders of a couple by a stranger. Finally, a minority of titles indicated that the stories
describe femicide–suicides by specifically designating the victim in contrast to the
perpetrator as well as the relationship between both individuals. As a result, news
consumers may accurately identify only a select few of the total articles sampled as
femicide–suicide. These results are similar to findings by Bullock and Cubert (2002)
that headlines fail to include language that distinguished an intimate partner homicide
from a homicide between strangers. Making this type of distinction about the relation-
ship between the victim and offender is vital for the transmission of information con-
cerning the frequency and nature of femicide–suicides to readers.

Next, we examined the framing components including the sources, language, and
context of the femicide–suicide articles. Beginning with sources, similar to prior
research concerning domestic violence news coverage (see Gillespie et al., 2013;
Richards et al., 2011; Taylor, 2009), the majority of femicide–suicide articles used law
enforcement as the main source of case information while domestic violence experts
and/or advocates were rarely cited. Contrary to the prior literature focused on femicide
cases generally, the present findings revealed that for femicide–suicide cases friends
and family were used as sources more often than neighbors. The increased use of fam-
ily and friends could be indicative of a difference in journalistic treatment of femicide–
suicide (or homicide–suicide) cases compared with femicide cases, (i.e., journalists
may approach family members more frequently in these cases, because there are mul-
tiple deceased individuals or because there is no longer a perpetrator to explain the
event). Family and friends may also be more willing to talk to journalists in cases
where the perpetrator is no longer living. Regardless of the reason, this finding is par-
ticularly intriguing and may indicate a need for further examination in future studies
of suicide and femicide reporting.

In addition to the articles’sources, the present study also examined the language
and resulting contextualization of femicide–suicide events in comparison to

Richards et al. 39

previous studies of femicide representations. Contrary to the previous literature on
the media presentation of suicide and/or homicide–suicide, but similar to the prior
research on femicide, the present research revealed that perpetrator mental health
and/or substance use issues were rarely discussed by journalists. Suicide research
has indicated that the vast majority of suicide victims have psychiatric problems,
most commonly mood disorders and substance abuse, and that these issues and treat-
ment options should be contained in articles reporting suicide events (CDC, 2001).
It is important to note that this research and reporting suggestions are in reference to
suicide victims specifically and there may be differences between suicide victims
and the perpetrators of murder–suicide. Despite these potential variances, assessing
the role of mental illness is important for accurately portraying problems and solu-
tions associated with domestic violence and suicide. In addition, divergent from
established information concerning the evolution of domestic violence, perpetrator/
victim separation was also rarely mentioned in femicide–suicide articles. While spe-
cific contextual elements were less present in the sample of femicide–suicides com-
pared with femicides generally, overall, a slightly higher percentage (14%) of
articles contextualized femicide–suicide within the greater problem of domestic vio-
lence compared with the 12% of femicide-only articles identified by Gillespie et al.
(2013) and the 10% of articles indentified by Bullock and Cubert (2002; but see
Taylor, 2009).

Finally, the present research explored the use of different types of media frames
within femicide–suicide news coverage. As mentioned previously, the present find-
ings revealed two distinct ways the media contextualized femicide–suicides (1) those
that did not identify the crime as domestic violence committed by a male perpetrator
against a female victim (not a femicide–suicide) and (2) those that did identify the
crime as domestic violence (as a femicide–suicide). Although the titles rarely reflected
as much, the text of the majority of femicide–suicide articles described the event as
domestic violence. Specifically, 78% of the articles in the present research defined
femicide–suicides as domestic violence compared with only 38% of the femicide
articles analyzed by Gillespie et al. (2013). Results demonstrate that four media
frames represented articles that identified femicide–suicides as domestic violence. In
line with previous studies identifying an explanatory typology of homicide–suicide
events, the present study found a spousal murder–suicide due to jealousy frame and
spousal murder–suicide due to declining health (a mercy killing) frame. The present
study found that articles emphasizing the perpetrator’s loss of control or those that
did not suggest a specific reason/motive for the femicide–suicide outnumbered these
two frames. To our interest, divergent from prior research on femicide cases (Gillespie
et al., 2013), a faulty criminal justice frame was not identified in the present research.
Articles that did not identify the crime as femicide–suicides were framed as either
ambiguous homicide–suicides or as a mystery. These articles failed to discuss prior
domestic violence in the lives of the couple or even identify the present murder of the
female victim by the male perpetrator as domestic violence. Likewise, these articles
did not utilize the expertise of domestic violence advocates and/or provide informa-
tion on the local resources for domestic violence. These findings indicate that

40 Feminist Criminology 9(1)

femicide–suicide cases are distinguishable from femicide cases in their circum-
stances, outcome, and journalistic representations.

Limitations

The present study is not unique in its research limitations. First, while the sample of
articles represents a near population of identified cases from a singular state over a
period of 8 years, the identified frames may not be generalizable to other geographic
areas. As discussed in previous studies using the present data (Gillespie et al., 2013;
Richards et al., 2011), the NCCADV is active in the state and has developed relation-
ships with the local media—this relationship may influence the prevalence of domes-
tic violence language and contextualization in North Carolina’s news coverage.
Comparing media frames used in states without an active coalition to states with more
active advocacy (such as the case with the present study) may be useful in identifying
the influence coalitions have on the media. In addition, while the present results are
comparable to the findings from recent femicide news representation examinations,
this study did not include a comparison sample of non-femicide homicide–suicides.
Examining the framing of articles on homicide–suicide between non-intimates may be
useful in bridging the suicide and domestic homicide media representation literatures.
Finally, frame analysis of femicide and femcide–suicide in other forms of media is
needed. In an increasingly digital age, print journalism may be becoming less influen-
tial than electronic news sources such as internet sites, television news programs, and
social networking outlets. Addressing the evolving methods for the dissemination of
news and the influence these various types of media have on the collective social con-
struction of the problem of domestic violence is necessary to this line of inquiry.

Policy Implications and Future Research

As previously mentioned, interpretation and understanding of violent crime com-
monly occurs through “mediated experience” (Surette, 2007) as opposed to personal
experience. At first glance this notion may appear inconsequential; however, the
potential implications should not be ignored. Specifically, societal reactions, including
those of policymakers, based on inaccurate media portrayals may result in misguided
strategies to address a form of violence. Therefore, it is requisite that the media por-
trayal of violence against women represents the true nature of such violence and that
the media embrace the opportunity to assist in the dissemination of accurate informa-
tion concerning social problems and the resources available to victims. Furthermore,
assessing media representations of intimate partner abuse, femicide, and femicide–
suicide can prove useful in developing publishing recommendations for reporters.

The CDC has been active in publishing recommendations for the reporting of sui-
cide events primarily due to concerns about contagion effects. These recommenda-
tions are often not implemented in the reporting of femicide–suicide, and, as
mentioned previously, may at times conflict with the reporting suggestions for
increasing domestic violence understanding and awareness. Assessments of the

Richards et al. 41

reporting of murder–suicide and femicide–suicide are important in terms of framing
femicide as a social problem and in terms of creating responsible reporting efforts
that are in agreement with the goals of the suicide and domestic violence advocacy
networks. Even within the literature addressing the reporting on suicide, murder–
suicide is treated as a separate category and has received less attention in terms of
explicit reporting suggestions. While the present analysis was intended as an initial
exploratory examination of media representations of the most common type of mur-
der–suicide, femicide–suicide, it also lays the groundwork for assessing responsible
reporting recommendations. Future research should examine the degree to which
homicide–suicide recommendations are followed in femicide–suicide articles, how
suicide and homicide–suicide recommendations may adhere to or conflict with the
goals of domestic violence advocacy, and develop reporting recommendations for
femicide–suicide cases specifically. Although homicide–suicide is rare compared
with suicide, this future research agenda is important because (1) homicide–suicide is
more likely to be reported on than suicide events, (2) femicide–suicide is the most
common type of homicide–suicide event, and (3) femicide followed by perpetrator
suicide is prevalent among domestic homicides—occurring in approximately one
quarter to one third of all domestic homicide cases (Campbell et al., 2007; Richards
et al., 2011;Van Wormer, 2008).

Conclusion

The media have the unique ability to harness the attention of news consumers and
shape public opinion concerning important social problems such as domestic violence
as well as society’s role in prevention and intervention efforts. In this vein, Surette
(2007) suggests that the power of the media is an obstacle and an opportunity—the
present research demonstrates the veracity of Surrette’s assertion. While there is opti-
mism in finding that the majority of the sampled femicide–suicides are framed as
domestic violence, the fact remains that many femicide–suicides are explained away
in overly simplistic terms, as isolated instances of perpetrator jealously or loss of con-
trol. In addition, in several cases the media failed to portray femicide–suicides as
domestic violence at all, instead presenting it as an ambiguous homicide–suicide or a
mystery. Such articles are missed opportunities to channel the media’s power to edu-
cate and advocate for victims by providing news consumers accurate depictions con-
cerning deadly domestic violence followed by perpetrator suicide. It is our hope that
research in this area will continue to encourage meaningful partnerships between
researchers, advocates, and the media and that similar to recommendations for the
reporting of suicide, best case practices for reporting on femicide–suicides will be
developed, disseminated, and implemented on a national level.

Authors’

Note

Portions of this research were presented at the 2011 meeting of the American Society of
Criminology in Washington, D.C.

42 Feminist Criminology 9(1)

Acknowledgments

The authors acknowledge the North Carolina Coalition Against Domestic Violence, which con-
tinually compiles data on intimate partner homicides making this research possible.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship,
and/or publication of this article.

Funding

The author(s) received no Financial support for the research, authorship, and/or publication of
this article.

Note

1. A more detailed description of the role of sources, language, and context in frame develop-
ment can be found in Gillespie, Richards, Givens, and Smith (2013).

References

Anderson, R. N. (2002). Deaths: Leading causes for 2000. National Vital Statistics Reports,
50, 1-88.

Associated Press. (2002a, October 8). Man shot himself, authorities say. Winston-Salem
Journal, p. B2.

Associated Press. (2002b, May 30). Two shot to death in near beach town. The News &
Observer, p. B4.

Associated Press. (2003a, July, 5). Couple dead in apparent murder-suicide. The News &
Observer, p. B7.

Associated Press. (2003b, December 16). Threats, jealousy led to murder-suicide of four:
Neighbors say family had a history of domestic violence. The Charlotte Observer, p. 2B.

Associated Press. (2008, October 21). Durham deaths murder-suicide. The News & Observer. p. B2.
Belknap, J. (2007). The invisible woman: Gender, crime, and justice. Belmont, CA: Thompson-

Wadsworth.
Berns, N. (2001). Degendering the problem and gendering the blame: Political discourse on

women and violence. Gender & Society, 15, 262-281.
Best, J. (1989). Images of issues: Typifying social problems. Hawthorne, NY: Aldine de Gruyter.
Boone, L. R., & Hewlett, M. (2007, March 15). Woman feared for life: Wife in murder-suicide

case took out restraining order against husband. Winston-Salem Journal. p. B1.
Brevorka, J. (2005, September 30). Couple’s autopsies pending. The News & Observer. p. B3.
Bullock, C., & Cubert, J. (2002). Coverage of domestic violence fatalities by newspapers in

Washington State. Journal of Interpersonal Violence, 17, 475-499.
Byerly, C. (1994). An agenda for teaching news coverage of rape. Journalism Educator, 49, 59-69.
Campbell, J. C., Glass, N., Sharps, P. W., Laughon, K., & Bloom, T. (2007). Intimate partner

homicide: Review and implications of research and policy. Trauma, Violence, & Abuse, 8,
246-269.

Centers for Disease Control and Prevention (CDC). (2001). National Institute of Mental Health,
Substance Abuse and Mental Health Services Administration, Office of the Surgeon
General, American Foundation for Suicide Prevention, American Society of Suicidology,

Richards et al. 43

Annenberg Public Policy Center. Reporting on suicide: Recommendations for the media.
Retrieved from http://www.sprc.org/library/sreporting

Chermak, S. (1995). Victims in the news: Crime and the American news media. Boulder, CO:
Westview Press.

Dawson, M., & Gartner, R. (1998). Differences in the characteristics of intimate femicides: The
role of relationship state and status. Homicide Studies, 2, 378-399.

Dorell, O., & Weigl, A. (2003, August 29). Death shock friends. The News & Observer. p. B1.
Ericson, R. V. (1989). Patrolling the facts: Secrecy and publicity in police work. British Journal

of Sociology, 40, 205-226.
Ericson, R. V., Baranek, P. M., & Chan, J. B. L. (1991). Visualizing deviance: A study of news

organization. Toronto, Ontario, Canada: University of Toronto Press.
Ferreri, E., & Velliquette, B. (2004, November 30). 2 die in murder–suicide, Police: Husband

shot wife at workplace, then turned gun on himself. Chapel Hill Herald. p. 1.
Fishman, M. (1981). Police news: Constructing an image of crime. Urban Life, 9, 371-394.
Gillespie, L. K., Richards, T. N., Givens, E., & Smith, M. D. (2013). Framing intimate partner

victimization: Why the media’s spin matters in newspaper coverage of femicide. Violence
Against Women, 19, 222-244.

Gorelick, S. (1989). “Join our war”: The construction of ideology in a newspaper crimefighting
campaign. Crime & Delinquency, 35, 421-436.

Gusfield, J. (1989). Constructing the ownership of social problems: Fun and profit in the welfare
state. Social Problems, 36, 431-441.

Hess, C. (2002, June 20). Four murder-suicides, one year. Kinston Free Press. Retrieved from
www.kinston.com

Humphries, D. (1981). Serious crime, news coverage, and ideology: A content analysis of crime
coverage in metropolitan newspaper. Crime & Delinquency, 27, 191-205.

Jamieson, P., Jamieson, K. H., & Romer, D. (2003). The responsible reporting of suicide in print
journalism. American Behavioral Scientist, 46, 1643-1660.

Kelly, L. (1988). How women define their experiences of violence. In K. Yllo & M. Bograd
(Eds.), Feminist perspectives on wife abuse (pp. 114-132). Newbury Park, CA: Sage.

Kenna, A. (2003, July 7). Fishing lake site of murder-suicide. Shelby Star. Retrieved from www.
shelbystar.com

Kirkpatrick, C. D. (2009, October 13). Father in killings had lost his job − 2 sons hospitalized
ex-UNCC employee, wife, dead in apparent murder-suicide. The Charlotte Observer. p. 1B.

LaGrone, S. (2007, August 28). Love tangle ends in two deaths. The News & Observer. p. B1.
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical

data. Biometrics, 33, 159-174.
Liem, M. (2010). Homicide followed by suicide. Aggression and Violent Behavior, 15, 153-161.
Manware, M., & White, S. (2002, November 10). 4 Lexington deaths may be murder-suicide:

Family finds bodies of man, estranged wife, 2 children in house. The Charlotte Observer, p. 1B.
Marzuk, P. M., Tardiff, K., & Hirsch, C. S. (1992). The epidemiology of murder-suicide.

Journal of the American Medical Association, 267, 3179-3183.
McDonald, T. (2004, November). Couple’s kin recall violence. The News & Observer, p. B1.
McNeill, S. (1992). Woman killer as tragic hero. In J. Radford & D. E. H. Russell (Eds.),

Femicide: The politics of woman-killing (pp. 178-183). New York, NY: Twayne.
Mehrotra, M. (1999). The social construction of wife abuse: Experiences of Asian Indian

women in the United States. Violence Against Women, 5, 619-640.
Meyers, M. (1994). News of battering. Journal of Communications, 44, 47-63.
Meyers, M. (1997). News coverage of violence against women: Engendering blame. Thousand

Oaks, CA: Sage.

44 Feminist Criminology 9(1)

Mitchell, H. (2008, January 17). Husband, wife found dead in home: Couple had recently sepa-
rated police separated police suspect murder-suicide. The Charlotte Observer. p. 2B.

Moracco, K. E., Runyan, C. W., & Butts, J. (1998). Femicide in North Carolina. Homicide
Studies, 2, 422-446.

Muhr, T. (2004). Atlas.ti 5.0 software. Berlin, Germany: Scientific Software Development.
O’Carroll, P. W., Potter, L. B., Aronowitz, E., Westchester, P. K., Linskey, D., Moscicki, E., .

. .Jarvis, J. (1994). Suicide contagion and the reporting of suicide: Recommendations from
a national workshop. Atlanta, GA: US Department of Health and Human Services, Centers
for Disease Control and Prevention.

Richards, T. N., Gillespie, L. K., & Smith, M. D. (2011). Exploring news coverage of femicide:
Does reporting the news add insult to injury. Feminist Criminology, 6, 178-202.

Rickert, E. (2007, June 28). Sheriff’s office: Deaths a murder, suicide. The Daily Reflector.
Retrieved from www.reflector.com

Sacco, V. (1995). Media constructions of crime. Annals of the American Academy of Political
and Social Science, 539, 141-154.

Schneider, J. (1985). Social problems theory: The constructionist view. Annual Review of
Sociology, 11, 209-229.

Surette, R. (2007). Media, crime, and criminal justice: Images, realities and policies (3rd ed.).
Belmont, CE: Thomson/Wadsworth Publications.

Taylor, R. (2009). Slain and slandered: A content analysis of the portrayal of femicide in the
news. Homicide Studies, 13, 21-49.

Van Wormer, K. (2008). The dynamics of murder-suicide in domestic situations. Brief Treatment
and Crisis Intervention, 8, 274-282.

Welch, C. (2004, July 7). Ailing husband, wife found dead in home: Victims of apparent mur-
der-suicide. Star-News, pp. 1A, 6A.

Wetherington, T. (2009, April 26). Woman murdered in street: Shooter turns weapon on himself
in apparent domestic dispute. The Daily Herald. Retrieved from www.rrdailyherald.com

Author Biographies

Tara N. Richards, PhD, is an assistant professor in the School of Criminal Justice at the
University on Baltimore. Her major research interests include violence against women; mental
health, substance abuse, and trauma/violence; and evaluation research. Some of her most recent
published work appears in Crime & Delinquency, Journal of Interpersonal Violence, and
Violence & Victims.

Lane Kirkland Gillespie, PhD, is an assistant professor in the Department of Criminal Justice
at Boise State University. Her research focuses on violence and victimization, including the
relationship between gender and crime, intimate partner violence, corporate victimization, and
the role of the victim in criminal justice processes. Her publications can be found in Feminist
Criminology, Violence Against Women, and Homicide Studies.

M. Dwayne Smith, PhD, is a professor in the Department of Criminology at the University of
South Florida where he serves as the institution’s senior vice provost and dean, Office of
Graduate Studies. He has published numerous articles and book chapters on crime and criminal
justice topics, focusing especially on homicide and capital punishment research. He is the
founding editor of Homicide Studies: An International and Interdisciplinary Journal and has
coedited (with Margaret A. Zahn) two books, Homicide: A Sourcebook of Social Research and
Studying and Preventing Homicide: Issues and Challenges.

Article

Suicidal Ideation in Juvenile
Arrestees: Exploring Legal
and Temporal Factors

Mike Tapia
1
, Henrika McCoy

2
, and Lynsey Tucker

3

Abstract
Mental health screening data (Massachusetts Youth Screening Instrument version 2 [MAYSI-2]) and
offense history were used to study levels of suicidal ideation in a sample of juvenile arrestees held in a
large, urban detention center located in a predominately Hispanic Southwestern U.S. city. We used
t-tests and multinomial logistic regression to examine the relationships with particular attention to
temporal issues. Results indicated that offense history, the timing of arrests, and demographics did
influence levels of suicide ideation. We discuss these findings, strengths and limitations, and directions
for future research.

Keywords
juvenile offenders, suicidal ideation, MAYSI-2, Hispanic youth

Introduction

Adolescent suicide is a complex and multidimensional problem with a high incidence in the United

States. Suicide rates among adolescents and young adults nearly tripled between 1952 and 1996; and

by 1999, the U.S. Surgeon General, David Satcher, deemed suicide a major public health concern (U.S.

Department of Health and Human Services, 1999). In 2001, rates of suicide had risen so high that the

U.S. Department of Health and Human Services reported that a greater number of adolescents and

young adults were dying each year from suicide than from AIDS, birth defects, cancer, chronic lung

disease, heart disease, influenza, pneumonia, and stroke combined. By 2011, suicide was the second

leading cause of death among 15- to 24-year-olds with over 4,800 lives lost that year (Centers for Dis-

ease Control and Prevention [CDC], 2014a). Among that age-group, there is one successful comple-

tion for every 100–200 attempts made with 20% of all deaths attributable to suicide (CDC, 2012).
Suicidal behavior varies by gender. For example, when compared to boys, girls are at a higher

risk for and have higher rates of suicidal behavior, attempts, and ideation (CDC, 2013; Lewinsohn,

1
New Mexico State University, Las Cruces, NM, USA

2
University of Illinois at Chicago, Chicago, IL, USA

3 University of Texas San Antonio and The Center for Health Care Services, San Antonio, TX, USA

Corresponding Author:

Mike Tapia, New Mexico State University, Las Cruces, NM, 88003-8001, USA.

Email: mtapia@nmsu.edu

Youth Violence and Juvenile Justice
2016, Vol. 14(4) 468-483
ª The Author(s) 2015
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1541204015579522
yvj.sagepub.com

http://www.sagepub.com/journalsPermissions.nav

http://yvj.sagepub.com

http://crossmark.crossref.org/dialog/?doi=10.1177%2F1541204015579522&domain=pdf&date_stamp=2015-04-14

Rohde, Seeley, & Baldwin, 2001; Wagner, 2009). Boys however, have a higher rate of completion

(CDC, 2014b), likely due to the use of more aggressive and lethal means during their attempts

(Wagner, 2009). Differences also exist in suicidal behavior by race and ethnicity (Goldston et al.,

2008). In 2012, for African American youth of age 12–18, suicide rates were 3.01 per 100,000, for

Hispanic youth rates were 3.62 per 100,000, and for White youth rates were 5.86 per 100,000

(National Center for Injury Prevention and Control, CDC, 2014). Despite White youth having the

highest rates of completion, Hispanic youth have the highest rates of suicidal ideation, plans, and

attempts among all youth (CDC, 2013).

Suicide and Juvenile Offenders

It has been well documented that incarcerated youth in particular are at high risk for attempting and

dying by suicide (Casiano, Katz, Globerman, & Sareen, 2013; CDC, 2014a; Gallagher & Dobrin,

2006; Mallett, DeRigne, Quinn, & Stoddard-Dare, 2012). Juvenile offenders are approximately

3 times more likely to die by suicide than youth in the general population (Gallagher & Dobrin,

2006). In fact, suicide is the leading reason juvenile offenders die when confined (Bureau of Justice

Statistics, 2005). This is highly concerning because in 2011, there were almost 62,000 juvenile

offenders in residential placement (Hockenberry, 2014). One of the most striking reasons for these

higher rates for incarcerated youth are partially due to the violent means they often use during their

suicide attempts (Rohde, Mace, & Seeley, 1997).

During the 15 years since suicide was deemed a public health risk for youth, there has been

increased awareness about its occurrence, especially related to juvenile offenders. In 2006, Gallagher

and Dobrin reported that 95% of facilities who had experienced a juvenile offender dying by suicide
had instituted mental health screening. The need for such screening became even further highlighted

in 2013, when the National Action Alliance for Suicide Prevention [NAASP] recommended that all

youth who were arrested or detained be screened for suicidal ideation. Thus, juvenile justice systems

in 47 states, and 44 statewide, accomplished that goal by implementing the Massachusetts Youth

Screening Instrument version 2 (MAYSI-2; Grisso et al., 2012).

Research about instruments used to screen for suicide ideation (SI) in this population, such as the

MAYSI-2, the Suicidal Ideation Questionnaire, the Suicidal Behaviors Questionnaire–Revised, and the

Global Appraisal of Individual Needs–Short Screener, has mostly focused on the clinical application,

validity, and generalizability of the measures across diagnoses, races/ethnicities, and gender (Archer,

Simonds-Bisbee, Spiegel, Handel, & Elkins, 2010; Cauffman & MacIntosh, 2006; Ford, Hartman,

Hawke, & Chapman, 2008; McCoy, 2014; McCoy, Vaughn, Maynard, & Salas-Wright, 2014; Vincent,

Grisso, Terry, & Banks, 2008a). In addition, the focus in the literature about suicide and juvenile detai-

nees has been primarily on clinical predictors such as the presence of psychiatric disorders, suicidal

ideation and/or attempts, substance use, and trauma exposure (Abram et al., 2008; Archer, Stredny,

Mason, & Arnau, 2004; Cauffman, 2004; Chapman & Ford, 2008; Nolen et al., 2008; Wasserman &

McReynolds, 2006). Understudied is the relationship between suicidal ideation and the legal char-

acteristics of juvenile detainees such as crime type and recidivism. This study aims to help fill that

gap by paying special attention to some of the temporal issues in the relationship between offense

history and suicidal ideation using MAYSI-2 and offense data from a sample of juveniles.

The Massachusetts Youth Screening Instrument, Version 2

As previously noted, the MAYSI-2 is often used to screen for the presence of SI. It is a brief, self-

report, mental health screening tool designed to be administered to juvenile offenders, age 12–17,

within 24–48 hr of entry into any decision-making point in the juvenile justice system (Grisso,

Barnum, Fletcher, Cauffman, & Peuschold, 2001). It identifies current feelings, thoughts, and

Tapia et al. 469

behaviors indicative of mental or emotional needs that necessitate immediate response or further

assessment (Grisso et al., 2001; Grisso & Quinlan, 2005). Furthermore, it is used as a triage tool to

identify youth whom report symptoms of distress (e.g., depressed mood) or who manifest feelings

or behaviors (e.g., suicide potential) that may require immediate intervention or further assess-

ment (Grisso et al., 2001). However, it is not designed to identify disorders from the Diagnostic

and Statistical Manual of Mental Disorders (Grisso et al., 2001).

The MAYSI-2 has 52 yes or no questions and takes approximately 10 min to administer. All

items are worded at the fifth-grade reading level and it contains the following 7 scales: Alcohol/

Drug Use, Angry-Irritable, Depressed-Anxious, Somatic Complaints, SI, Thought Disturbance

(boys only), and Traumatic Experiences. The number of items on each scale ranges from 5 to

9 and respondents answer each item using the time frame ‘‘in the past few months.’’ Six of the

scales use the cutoffs Caution and Warning. Caution indicates clinical significance and a score of

Warning means that the youth has scored in the top 15% of youth in the juvenile justice system
(Grisso & Barnum, 2006). The Traumatic Events scale however uses a simple count where youth

respond to whether they have ever experienced one of the identified events during their lifetime.

The SI scale contains 5 items with potential scores ranging from 0 to 5. A score of 2 is desig-

nated as Caution while a score between 3 and 5 is designated as Warning. The 5 items are (1)

‘‘Have you wished you were dead?’’ (2) ‘‘Have you felt life was not worth living?’’ (3) ‘‘Have

you felt hopeless?’’ (4) ‘‘Have you felt like hurting yourself?’’ (5) ‘‘Have you felt like killing

yourself?’’

A number of studies have examined the psychometric properties of the MAYSI-2’s scales,

including comparing it with measures that were conceptually similar by using Pearson’s r and

receiver operating characteristics (ROC) curves. From that research, the SI scale was shown to

have strong correlations and good to excellent accuracy including the Children’s Depressive

Inventory r ¼ .51 (Tille & Rose, 2007); the Millon Adolescent Clinical Inventory Suicide Ten-
dency scale r ¼ .57 (Butler et al., 2007), r ¼ .61 (Grisso et al., 2001), and when using the Caution
cutoff, area under the curve (AUC) of .94 for boys and .91 for girls (Grisso et al., 2001); the Neg-

ative Suicide Ideation Index from the Positive and Negative Suicide Ideation Inventory r ¼ .68
(Tille & Rose, 2007); the Suicide Ideation Questionnaire r ¼ .58 (Chapman & Ford, 2008); and
the Youth Self-Report Self-Destructive Scale r ¼ .44 (Grisso et al., 2001).

When Archer, Stredny, Mason, and Arnau (2004) compared the respondent’s current suicidal

ideation to the Caution cutoff on the SI scale, there was good specificity (.81) and sensitivity

(.90). When the respondent’s history of suicidal attempts was compared to the Caution cutoff, sen-

sitivity (.66) was weaker while specificity (.85) increased slightly. When compared to the Warning

cutoff, results were similar, that is, current suicidal ideation, sensitivity ¼ .90 and specificity ¼ 0.88;
and a history of suicidal ideation, sensitivity ¼ .48 and specificity ¼ .90. Archer, Simonds-Bisbee,
Spiegel, Handel, and Elkins (2010) used data from detained adolescents (n ¼ 1,192) to further exam-
ine the MAYSI-2’s validity. They used a standard of r ¼ .15 to determine correlation at a meaningful
level, and the SI scale was found to have convergent validity with prior suicidal ideation (boys

r ¼ .30 and girls r ¼ .24), prior suicidal behaviors without medical attention (boys r ¼ .27 and girls
r ¼ .28), prior suicide attempts with medical intervention (boys r ¼ .23 and girls ¼ .28), and a his-
tory of self-mutilating behavior (boys r ¼ .28 and girls r ¼ .30).

A study by Wasserman et al. (2004) of n ¼ 325 detained youth examined how well MAYSI-2
scales mapped onto the Diagnostic Interview Schedule for Children-IV. Results showed that when

prior suicide attempts were compared to SI, youth who scored Caution on the SI scale were more

likely (OR [odds ratio] ¼ 6.12, p < .001) to report lifetime suicide attempts. When SI and suicide attempts were examined on a continuum, youth who reported a recent suicide attempt were more

likely (OR ¼ 3.46, p < .001) to show an increase on the SI scale. ROC analyses were also used to determine whether a score above Caution on the SI scale could accurately identify recent suicide

470 Youth Violence and Juvenile Justice 14(4)

attempts, lifetime attempts, and no attempt both with and without a plan, results were AUC ¼ 0.96,
0.73, and 0.55, respectively.

Past MAYSI-2 SI Scale Research Findings

A recent, and the only, meta-analysis of MAYSI-2 data was examined by Vincent, Grisso, Terry, and

Banks (2008a) of sex, race, and ethnicity differences in a national sample of system-involved youth.

The data contained 70,423 intake records with MAYSI-2 scores from 283 juvenile justice probation,

detention, or corrections sites from 19 states across the country. On at least one of the five MAYSI-2

scales, 72% of girls and 63% of boys scored above Caution. On the SI scale, girls (29%) had a much
higher prevalence of scoring above Caution than boys (15%). The pattern of girls consistently scor-
ing higher than boys on the SI scale repeated in the numerous other studies where gender was exam-

ined (e.g., Archer et al., 2010; Archer, Stredny, Mason, & Arnau, 2004; Cauffman, 2004; Gretton &

Clift, 2011; Hayes, McReynolds, & Wasserman, 2005; Kerig, Moeddel, & Becker, 2011; Nordness

et al., 2002; Stathis et al., 2008; Wasserman et al., 2004).

Differences by race and ethnicity have also been examined. Vincent et al. (2008a) used a national

sample of White (39%), Black (34%), and Hispanic (24%) youth and found that most youth scored
below Caution on the SI scale. The differences by race or ethnicity were small to virtually nonexis-

tent and the percentages of youth scoring above Caution on the SI scale were as follows: White

(22%), Black (15%), and Hispanic (17%). In other studies (e.g., Archer, Stredny, & Arnau, 2004;
Cauffman, 2004; Dalton, Evans, Cruise, Feinstein, & Kendrick, 2009; Grisso et al., 2001) when

White youth were compared to Black youth on the SI scale, White youth had higher means or scored

Caution at higher percentages. When Black and Hispanic youth were compared, Hispanic youth

were more likely than Black youth to score higher on the SI scale (Cauffman, 2004; Grisso et al.,

2001). However, results on the SI scale were varied across studies when Hispanic youth were com-

pared to White youth. For example, Grisso, Barnum, Fletcher, Cauffman, and Peuschold (2001)

found that White and Hispanic males had equal means, while Cauffman (2004) found that White

males had lower means than Hispanic males. Grisso et al. (2001) also found that White females had

lower means than Hispanic females while Cauffman (2004) found the opposite.

The Current Study

The current study will fill two important gaps in the literature about SI among youth involved in the

juvenile justice system. First, where past research has uncovered many of the social and psychological

correlates of SI, this study explores some of the legal and temporal correlates. To accomplish this,

average scores on the MAYSI-2 SI scale are compared by crime type, crime severity, arrest frequency,

and demographics. In addition, SI scores for repeat arrestees at two points within a 6-month time frame

are examined to determine whether outcomes vary when a youth quickly recidivates. Second, while

various recent studies conducted with MAYSI-2 data do include Hispanic youth (e.g., Aalsma,

Schwartz, & Perkins, 2014; Cauffman & MacIntosh, 2006; Coker et al., 2014; Grisso et al., 2001;

Vincent et al., 2008a), many others exclude them (e.g., Cruise, Marsee, Dandreaux, & DePrato,

2007), have a small sample size (e.g., Maney, 2011), or do not delineate the findings for Hispanics

very well (McCoy et al., 2014). This study however includes youth residing in a large, Hispanic dom-

inated urban county and will allow a unique and needed focus in Hispanic juvenile offenders.

Hypotheses

Our first hypothesis (Hypothesis 1): Offenders who commit serious crime (felonies or violent

crimes) will have higher SI scores than less serious offenders who commit misdemeanors or status

Tapia et al. 471

offenses. Engaging in more serious crimes reflects a higher level of dysfunction in a youth’s

psychological or social background (Lynam et al., 2000; Valdez, Kaplan, & Codina, 2000). Thus,

we posit there will be higher levels of distress, such as suicidal ideation, for youth entering the

juvenile justice system for such crimes.

Our second hypothesis (Hypothesis 2): Having prior offenses will be negatively associated with

SI scores due to a desensitization effect. The tests for this hypothesis are 3-fold, focusing on the

difference in SI scores between (1) youth with priors versus those without (Prevalence test) and

(2) Time 1 and Time 2 arrests for repeat arrestees (Frequency test) and then on examining the

effect of the number of arrests on SI (Multivariate test). For the Prevalence test, we expect the

mean SI score for youth with priors to be lower than that of first offenders. For the Frequency test,

we used only the subset of youth arrested more than once during a specific 6-month time period

(n ¼ 222). We expect for scores at Arrest Time 1 to be higher than at Arrest Time 2. In the Multi-
variate test, we control for other legal and demographic items and expect for the number of arrests

(ranging from 1 to 4) to be negatively associated with SI.

We posit the negative relationship between the number of detentions and level of suicidal idea-

tion will be either an artificial or genuine institutionalization effect. Due to prior detention experi-

ences, youth might expect that reporting certain symptoms could lead to further screening, unwanted

intervention, and possibly a lengthier detention stay. That knowledge might cause youth to be decep-

tive in their responses in order to mask symptoms of psychological distress (McCoy, 2014; NAASP,

2013) or because they believe that services would be ineffective so it would be fruitless to report

them (Coker et al., 2014). Regardless, this outcome would be categorized as an artificial effect.

A genuine effect would occur when a youth’s psychological distress symptoms are genuinely lower

and ultimately there is less fear and lower feelings of suicidal ideation related to potential long-term

system involvement.

Our third hypothesis (Hypothesis 3): Younger offenders will score higher than older offenders on

the SI scale. This hypothesis is supported from findings in a study conducted by Morris et al. (1995)

of youth in 39 juvenile correctional facilities, where being younger was associated with higher levels

of SI and more suicide attempts. Because younger offenders are less likely to commit acts leading to

arrest and detention, their psychological distress levels are likely to be more acute than that of older,

more experienced offenders. We posit that youth with less experience being processed (i.e.,

younger) are likely to view such an event as direr than older youth, who on average should be more

desensitized to the process. In regression, we control for prior offenses, in order to avoid confound-

ing age effects in Hypothesis 3 with the hypothesized desensitization effects in Hypothesis 2.

Our fourth hypothesis (Hypothesis 4): Race-ethnicity and gender will influence SI scores. Spe-

cifically, Hispanic youth will have higher SI means than White and Black youth and females will

have higher SI means than boys. According to Zayas and Pilat (2008), Hispanic females report some

of the highest levels of suicidal ideation of any demographic group. Based on prior research exam-

ining race-ethnicity and gender in the MAYSI-2, coupled with the identified stressors in Bexar

County, we posit our results will also show that Hispanic youth, and girls in particular, will demon-

strate higher levels of SI.

Method

Study Setting

Bexar County (i.e., San Antonio), Texas, is one of the nation’s largest urban hubs for U.S. Hispanics,

who comprise 59% of its population (U.S. Census Bureau, 2014b). It has high rates of childhood
poverty (U.S. Census Bureau, 2014a), teen pregnancy (City of San Antonio, 2012), and the state

of child welfare there has been characterized as dismal (Stoeltje, 2014; The Stephen Group,

472 Youth Violence and Juvenile Justice 14(4)

2014). This is particularly concerning because many youth who are involved in the child welfare

system are often also involved in the juvenile justice system (McCoy & Bowen, 2014) and Bexar

County has high rates of involvement in both systems (Bexar County Juvenile Probation Depart-

ment, n.d.a.). The multitude of stressors and risks to youth that this entails likely increase levels

of psychological and emotional distress for this group. Thus, the current family and social conditions

experienced by Hispanic juvenile offenders in Bexar County are possibly more toxic than in many

other contexts. Therefore, it might be expected that youth in Bexar County may show higher levels

of distress, and SI in particular, than a national sample.

Sample Description

This study sample includes juveniles (N ¼ 763), aged 10–18, who were apprehended, arrested, and
detained in the Bexar County Detention Center for incidents occurring between December 1, 2009,

and May 31, 2010. The race and ethnic composition of the sample is Hispanic (n ¼ 569, 75%), Black
(n ¼ 99, 13%), White (n ¼ 90, 12%), and Other (n ¼ 5, 0.01%); the gender breakdown is male
(n ¼ 560, 73%) and female (n ¼ 203, 27%). Regarding age, 56% of the sample is below age 15.
Table 1 contains the complete descriptive data for each variable.

Notably, fewer youth scored Caution (5%) versus Warning (10%) on the SI scale. Most youth
were 1 Time arrestees during the 6-month period (71%) and committed either misdemeanors or sta-
tus offenses (77%). Table 2 contains the average SI scores of various subgroup pairs, which are com-
pared with t-tests to determine whether subgroup differences are statistically significant. Table 3

contains the average SI scores of the subset of repeat arrestees at Arrest Time 1 and Arrest Time

2. Table 4 contains results of the multinomial logistic regression.

We examined two types of data from the Bexar County Juvenile Probation Department as fol-

lows: (1) scores from each MAYSI-2 administered to arrestees as part of their detention intake and

(2) related offense data for each juvenile offender. All noncustodial arrests, such as ‘‘paper arrest’’

cases where a youth was detained on a warrant for missing a court date or for violating probation,

were dropped. In addition, cases with missing data on current and prior offenses were dropped. The

study’s hypotheses were tested using t-tests and multinomial logistic regression; our analyses were

conducted using SPSS 21.

Table 1. Descriptives.

Variable n %

SI caution (Time 1) 36 4.7
SI warning (Time 1) 76 10.0
No SI cautions/warnings (Time 1) 642 84.1
Single arrest 541 71.0
Two arrests 175 22.9
Three or more arrests 47 6.1
Felonies (Time 1) 175 22.9
Violent offenses (Time 1) 223 29.2
Misdemeanors and status offenses (Time 1) 588 77.1
Prior offenses (at Time 1) 530 69.5
10–15 age-group (at Time 1) 429 56.2
Males 560 73.4
Hispanics 569 74.6
Blacks 99 13.0
Whites 90 11.8

Note. SI ¼ suicide ideation.

Tapia et al. 473

For each hypothesis, we compare the mean scores of youth on the SI scale and use t-tests assum-

ing equal variances on items, to determine whether any observed differences are statistically signifi-

cant for different groups according to demographic (e.g., race, gender, and age) and legal categories

Table 2. SI by Paired Groups.

Group Mean SD Cohen’s d t Value

Felony 0.51 1.21 –0.116 –1.32
Misdemeanors 0.66 1.38
Violent 1.02 1.68 0.383 5.21**
Status offenders 0.47 1.14
Prior record 0.44 1.08 –0.430 –6.03**
No priors 1.06 1.73
Single arrest 0.69 1.40 0.154 1.84
Repeat arrests 0.49 1.19
Ages 10–15 0.73 1.43 0.174 2.32*
Ages 16–18 0.50 1.21
Male 0.44 1.14 –0.487 –6.47**
Female 1.14 1.69
Hispanic 0.54 1.24 –0.253 –3.21**
Blacks and Whites combined 0.90 1.58
Black 0.83 1.58 0.159 1.58
Whites and Hispanics combined 0.60 1.30
White 1.00 1.63 0.286 2.78**
Blacks and Hispanics combined 0.58 1.29

Note. SI ¼ suicide ideation.
*p < .05. **p < .01.

Table 4. Multinomial Logistic Regression of SI Cutoffs on Legal Items, Age, and Controls.

Caution Warning

OR 95% CI OR 95% CI

Intercept — —
Felony 0.999 [0.44, 2.28] 0.74 [0.38, 1.44]
Violent 1.02 [0.46, 2.29] 1.93 [1.11, 3.34]
Priors (yes/no) 1.32 [0.58, 2.98] 0.41 [0.23, 0.70]
Number of arrests, past 6 months 0.74 [0.40, 1.37] 0.88 [0.56, 1.37]
Ages 10–15 2.21 [1.02, 4.80] 1.00 [0.58, 1.74]
Male 0.53 [0.26, 1.09] 0.37 [0.22, 0.62]
Hispanic 0.73 [0.27, 1.99] 0.40 [0.21, 0.76]
Black 0.85 [0.23, 3.14] 0.88 [0.39, 1.98]

Note. N ¼ 753. SI ¼ suicide ideation; OR ¼ odds ratio; CI ¼ confidence interval.

Table 3. Repeat Arrestees, SI Means at Time 1 and Time 2.

Group Mean SD Cohen’s d t Value

Time 1 0.49 1.19 –0.073 6.32*
Time 2 0.58 1.27

Note. n ¼ 222. SI ¼ suicide ideation.
*p < .01.

474 Youth Violence and Juvenile Justice 14(4)

(e.g., offense type, severity, frequency, and history). Multinomial logistic regression is conducted

with the independent variables and controls in a model predicting SI scores of Caution or Warning,

with No Caution/No Warning as the reference category in order to evaluate hypotheses while con-

trolling for potential confounds. The results will show which variables increase the odds that youth

SI scores exceed the Caution and Warning cutoffs. For these analyses, the independent variables,

namely, felony, violent, priors, and age 10–15 are measured as dummy variables, coded 0 and 1; the

remaining independent variable, the number of arrests during the study’s 6-month period, has a

range from 1 to 4. The control items, sex, and race-ethnicity are dummy coded 0 and 1 with female

as the reference category for gender and White as the reference category for race-ethnicity.

Additionally, for the second hypothesis, three analyses were conducted. First, we examine the

difference in SI scores between youth with priors and those without using a t-test. This analysis uses

a left-censored time frame because youth in this study range in age from 10 to 18 with some criminal

histories extending retrospectively 8 years so the date of first offense is not measured. For the second

analysis, the average SI scores at Arrest Time 1 and Arrest Time 2 were compared for a subset

of repeat arrestees (n ¼ 222), also using a t-test. This group had two or more arrests during a
6-month period, enabling a test of whether SI scores decrease when a youth reenters the system and

retakes the MAYSI-2 within a short time frame (Table 3). Finally, the impact of the number of

arrests on the SI thresholds of Caution and Warning is further reexamined using multinomial logistic

regression (Table 4).

Results

Our first hypothesis was that offenders who committed serious crime (felonies or violent crimes)

would have higher SI scores than less serious offenders who commit misdemeanors or status

offenses. In partial support, youth who committed violent crimes had significantly higher SI scores

than youth who committed status offenses (t ¼ 5.21, p < .01). In addition, committing a violent offense significantly increased the odds of scoring Warning on the SI scale (OR ¼ 1.93).

Second, we hypothesized that having prior offenses would be negatively associated with SI

scores due to a desensitization effect. The first test conducted for this hypothesis (arrest Prevalence)

was proven true. Table 2 shows that first offenders had significantly higher SI means ( ¼ 1.06) than
youth with priors (M ¼ 0.44) and results were statistically significant (t ¼ 6.03, p < .01). Further- more, in regression, having prior offenses decreased the odds of a Warning score (OR ¼ 0.40). On the second test (arrest Frequency), results were mixed. First, note that in Table 2, the multiple arrests

category had a significantly lower SI mean than the single arrest category, indicative of a desensi-

tization effect. However, when comparing SI scores at arrest time 1 and arrest time 2 only (Table 3),

the average score increases significantly during this brief period, opposite of the desensitization pre-

diction in Hypothesis 2. Finally, the multivariate test shows that the odds of scoring a Caution or

Warning on the SI scale decrease with the number of arrests, which, although not significant, is indi-

cative of desensitization. For this final test, there is no significant effect and the hypothesis is ulti-

mately not supported.

Results supported the third hypothesis: Younger offenders would score higher than older offen-

ders on the SI scale. The SI mean for the younger age-group (10–15) is significantly higher (t ¼ 2.32,
p < .05) than that of the older youth (16–18). Further, the only independent variable that significantly

predicts the SI Caution cutoff is the young age-group (OR ¼ 2.21).
The final hypothesis was that race-ethnicity and gender would influence SI scores. As hypothe-

sized, females (M ¼ 1.14) have a much higher SI average than that of males (M ¼ 0.44) and results
were statistically significant (t ¼ 6.47, p < .01). Being male also decreased the odds of scoring a Warning on SI (OR ¼ 0.37). Contrary to the second portion of the hypothesis, Hispanics have a sig- nificantly lower SI score than Blacks and Whites combined (t ¼ 3.21, p < .01), Blacks had higher SI

Tapia et al. 475

scores than Whites and Hispanics combined (t ¼ 1.58, n.s.), and Whites had significantly higher
scores than Blacks and Hispanics combined (t ¼ 2.78, p < .01). Finally, when race-ethnicity was examined in regression, the only significant results were that being Hispanic decreased the odds

of scoring a Warning (OR ¼ 0.40) when compared to Whites.

Discussion

Most research about the MAYSI-2 has not focused on the offense characteristics of the detained

youth to which it is administered. This study sought to highlight the relationship between several

legal items and SI, with particular attention to some of the temporal issues therein. Ultimately,

there was full or partial support for each of the four hypotheses. Where analyses of MAYSI-2

data on Hispanics are sparse in the literature, this study makes a unique contribution in that arena

as well.

Contrary to our expectation, youth charged with felonies had lower SI scale means than those

with misdemeanors and status offenses. Although the difference was not significant, prior research

does show that youth who engage in more serious crimes have higher levels of psychological dis-

tress. However, felonies are a broad category and include numerous crimes that may not be per-

ceived to be serious by a juvenile and that do not lead to self-injurious behaviors. For example,

in Texas, many property crimes are designated as felonies, but there is no compelling reason to

expect for such behavior to be associated with high levels of anxiety in young offenders.

As hypothesized, the SI scale mean for violent offenders was significantly higher than for non-

violent offenders. This finding has implications for direct service staff, particularly because it would

not be unusual for these youth to be managed differently from other youth, such as being housed

separately or engaged in more intensive services, which are targeted to youth with a history of vio-

lence. The training of these staff regarding how they assess and screen for suicidal ideation youth

who commit violent offenses should account for the increased potential SI with this group. In addi-

tion, the ultimate disposition for a youth who has committed a violent felony versus a status offense

is vastly different. It is the norm that youth adjudicated to have committed a violent felony can be

sent to the Texas Juvenile Justice Department (TJJD) or a post-adjudication secure correctional

facility (Texas Attorney General, 2014).

The second hypothesis addresses the following two competing assumptions: (1) youth with

prior arrest experiences would be desensitized to the negative aspects of detention and therefore

score lower on an SI scale and (2) anytime a youth is arrested, it is a negative experience, thus SI

scores will show no difference or even be higher for those with prior arrests. It is important to note

that only about 4–5% of national, representative samples of the youth population will report ever
having been arrested (Tapia, 2012), making it a rather rare, and likely traumatic event. It would

follow that regardless of the amount of time between arrests, a youth will experience distress about

being rearrested and detained, particularly if they had been able to ‘‘stay out of trouble’’ for an

extended period of time. Support for either of these ideas in this study was contingent on the tim-

ing and number of arrests.

The comparison of SI means for those with prior offenses and those without was a basic, general

test in support of the second hypothesis. Indeed, it was a rather large effect, where the average SI

score is over twice as large for first offenders as for those with priors. Then, in regression, having

prior arrests significantly decreases the odds of scoring a Warning on the SI scale, further indicating

that first time arrestees experience higher levels of SI because they fear the unknown about their

detention experience. It may also indicate an awareness by repeat offenders of the possible limits

of the sanctions that can be imposed by the system when they recidivate over a lengthy time frame.

Ultimately, a youth might express feelings of failure or concern, but perhaps not to the extent of

producing suicidal ideation.

476 Youth Violence and Juvenile Justice 14(4)

Among the subset of repeat arrestees within a 6-month period (n ¼ 222), however, there was no
desensitization effect. In fact, the average SI scores at the second arrest were significantly higher

than at Arrest Time 1. This challenges our ideas about institutionalization effects and emphasizes

the importance of the temporal dimension in our study. Given the high prevalence of SI, thoughts,

behaviors, and completions in the juvenile justice population, this may be genuine versus being an

artificial survey effect. The higher scores are possibly attributed to feelings of failure about being

rearrested so quickly or be due to concerns about legal consequences, particularly if the youth had

multiple prior arrests and the latest, quickly occurring arrest could result in a more severe disposi-

tion. Psychological distress symptoms are also likely genuine since youth with a recent history of

detention will have a fresh recall of its many negative aspects (see Mendel, 2009), especially in a

large, urban county facility.

Interestingly, two arrests may be a threshold for the effect of multiple arrests on the SI score and

the relationship seems to be curvilinear. When the number of arrests (ranging 1–4) within a 6-month

time frame measure was used in regression, it did not increase the odds of a Caution or Warning

reading. In fact, the number of arrests was negatively associated with SI scores (n/s), which is a

reversal of the effect found with only two arrests as described earlier. These results again support

our ideas about institutionalization effects where youth having frequent, repeat arrests are ‘‘experi-

enced’’ and seem to have lower levels of mental crisis about this circumstance, yet we recognize the

possibility of mis-specification. While we do not employ a squared term on this exploratory article,

clearly there is a need to further test the relationship in that manner in future research.

Ultimately, we must also keep in mind that some part of the results may reflect youth manip-

ulation of their responses to the SI items. As noted previously, some youth attempt to mask psy-

chological distress symptoms with deceptive responses to avoid further screening, unwanted

interventions, and possibly a lengthier stay in detention. The scores may reflect such savviness,

thus more research, perhaps of the qualitative type is needed to further explore this issue with

Hispanic youth in the way that it has been done for Black youth (McCoy, 2014).

As hypothesized, younger offenders did have significantly higher SI scores than older youth and

being younger significantly predicted scoring Caution, although not Warning. Despite our effort to

isolate the age effect in regression, these results may indicate that age serves as a proxy for desen-

sitization due to a prior offense history; the older a youth is the less psychologically distressed they

feel. It is also possible that younger offenders are more impacted by being separated from their sup-

port system and have concerns about what might happen to them during and after their detention. In

a normative framework, the odds of having contact with the juvenile system can only increase with

age and thus the age effect may ultimately be a function of past detention experiences.

As expected, females had much higher SI means than males, over twice as high. Those results

support the fourth hypothesis and reflect the results from the detention portion of the 2003 national

sample (n ¼ 29,817) analyzed by Vincent, Grisso, Terry, and Banks (2008b). The females in the
current study (M ¼ 1.14) had almost equal means to those in Vincent et al.’s national sample
(M ¼ 1.15) whereas the males in the current study had lower SI means (M ¼ 0.44) than males
in that U.S. sample (M ¼ 0.62). The results from our study continue to reinforce prior research that
females in the juvenile justice system, and in this case who are detained, experience higher levels

of suicidal ideation than their male counterparts. This is possibly because they are in a system that

is primarily designed to respond to male needs (Veysey, 2003), leading to much higher levels of psy-

chological distress. It may also be because girls are less likely to be detained and therefore those that

are have higher levels of psychological distress (Veysey, 2003). Regardless, it is clear that female

juvenile offenders may require a particular focus on feelings of suicidal ideation that may be present.

Finally, consistent with other recent research using MAYSI-2 data (Aalsma et al., 2014),

Hispanics had the lowest scores among race groups. They had a significantly lower SI score than

Blacks and Whites combined. Whites, meanwhile, had significantly higher scores than Blacks and

Tapia et al. 477

Hispanics combined. When compared to the national-level data of Vincent et al. (2008b), the most

noticeable difference was that Hispanic youth in our study (M ¼ 0.54) scored far lower on the SI
scale than they did nationally (M ¼ 0.84). The patterns for White youth, on the other hand, were
the same between the two studies (White M ¼ 1.00 in Bexar County vs. M ¼ 0.95 in United
States). Finally, Black youth in Bexar scored much higher (M ¼ 0.83) than their counterparts in
the United States (M ¼ 0.53). The scores for Hispanic youth stand out as being far lower than
expected in Bexar County. These results may reflect that Hispanic youth in this particular context

are resilient in the face of typical mental health stressors. Hispanic youth in this sample are nested

in a predominately Hispanic community, perhaps echoing research by Locke and Newcomb

(2005) showing that the sense of familial support can serve as a protective factor against suicid-

ality. The centrality of family (i.e., familism) which is particularly important for Hispanics

(Tienda & Mitchell, 2006) is likely reinforced by their community context. This and the possibility

of strong cultural norms against this ideation and practice may help these youth to eclipse their

feelings of suicidal ideation. This is all despite often experiencing a high rate of common stressors

for depression and mental illness (Gallo, Penedo, Espinoza, & Arguelles, 2009; Umaña-Taylor &

Updegraff, 2007), which are certainly present in Bexar County.

Strengths and Limitations

This study used self-report data from the MAYSI-2, which may be unreliable due to social desirabil-

ity dynamics. Respondents can often answer self-report items in ways that make them appear more

positive even when it is untruthful (Holtgraves, 2004). Regarding the MAYSI-2, prior research has

shown that juvenile offenders have specifically provided untrue responses to items (McCoy, 2014).

Our sample also does not reflect the typical population of the juvenile justice system nor does Bexar

County reflect most communities in the United States. These unique features mean that our findings

are not generalizable beyond this study or perhaps only to other large Hispanic-dominated counties.

Finally, the longitudinal portion of our analyses was restricted to 6 months, which left the extent of

repeat offending over a lengthier period unmeasured.

This study also had a number of strengths. We explored the underresearched influence of the level

of SI experienced by youth actively engaged in a detention setting and the impact of age, race, eth-

nicity, gender, and legal offense history on those feelings. We also make a unique contribution by

comparing the average scores on the SI scale by crime type, severity, frequency, and timing of deten-

tion in a large, Hispanic dominated urban county. The composition of Hispanic youth as the majority

in the study sample and in a setting that is predominately Hispanic provided an opportunity to exam-

ine a rarely focused on group, particularly in studies about or using the MAYSI-2.

There is a paucity of research looking at the temporal relationship between offending behaviors

and the presence of suicidal ideation. This study provides exploratory insight into that relationship as

well as on gender and race-ethnicity differences. Finally, although the analysis of the 6-month time

frame may have limited the findings, our comparison of SI scores for repeat arrestees at Arrest Time

1 and Arrest Time 2 to determine whether outcomes vary when a youth reenters the system fre-

quently is certainly a unique contribution to the literature in this subarea of juvenile research. We

now have some insight into the role that first offense and repeat arrests may have on suicidal idea-

tion. This information can be instrumental in informing how direct service staff monitors such youth

when they return to detention.

Future Research

This study’s population was youth detained in a county detention facility versus in a state correc-

tional facility. This fact is important when considering the implications of our findings because

478 Youth Violence and Juvenile Justice 14(4)

in Texas, as in many other places, state corrections facilities are the most serious placement setting

with only adult facilities classified as more serious (TJJD, 2013). State facilities range broadly in

sentence severity; and in Texas, there are youth with indeterminate sentences who are placed in low,

medium, or high security risk facilities and youth with determinate sentences that include an even-

tual transfer to an adult facility (TJJD, 2013). In addition, youth in a state correctional facility expe-

rience minimum length of stay requirements ranging from 9 to 24 months, dependent on severity of

offense and risk to public safety. Conversely, the county detention facility, used during the pre-trial

phase of juvenile processing, is where youth are taken by law enforcement when they have been

arrested for a delinquent offense (Bexar County Juvenile Probation Department, n.d.b). Further-

more, after being seen by an intake probation officer, a decision is made regarding whether to release

or detain the youth (n.d.b.).

The findings from this study begin to fill an important gap in the literature, but their applicability

should be viewed in context—focused on youth who have not been adjudicated and sentenced. If we

studied youth in state correctional facilities, we might expect for the impact of incarceration on SI to

be even more evident than it was here. The difference between temporary confinement in one’s com-

munity versus removal from the community, and long-term incarceration causing separation from

family and friends, should weigh more heavily on a youths’ state of mind and likely raise the odds

of mental health crisis. Thus, it is important to acknowledge that this study captures a different con-

textual dynamic related to suicidal ideation for juvenile offenders. Future research should build on

this study and explore how the temporal relationship between SI and offending behavior manifest for

youth who have been adjudicated and are housed in state correctional facilities.

Future research on the county detention phase of incarceration should also further explore the

relationship between offending behavior and SI using a squared term for the number of arrests to

capture the nonlinearity of the effect that we seem to have detected here. More research is also

needed to understand how detention may impact the suicidal ideation of youth arrested for violent

offenses, younger offenders, and girls. At a minimum, for most youth, arrest and detention are seri-

ous events that may reflect engaging in other forms of risky, problematic behavior. Clearly, part of

this dynamic may include careless, impulsive, and self-destructive tendencies. This possibility

coupled with the potential outcome for failing to accurately identify a youth’s SI suggests the need

for additional, or at minimum more in-depth, screening.

The relationship of race-ethnicity to SI also requires more study. The literature we reviewed on

national-level estimates of SI by race-ethnicity contains conflicting estimates regarding which group

has the highest levels of SI. Our results showing Hispanic ethnicity to be a protective factor against

SI in detained youth appears to be the first such finding with MAYSI-2 data, albeit from a particular

context and locale. We are nonetheless encouraged by this and call for continued research in this

area that expands the model specification in the way we propose earlier and that replicates the model

in other places with sizeable Hispanic populations.

Acknowledgment

The authors wish to thank Erin Wit for her early work on this project.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or

publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Tapia et al. 479

References

Abram, K. M., Choe, J. Y., Washburn, J. J., Teplin, L. A., King, D. C., & Dulcan, M. K. (2008). Suicidal

ideation and behaviors among youths in juvenile detention. Journal of the American Academy of Child

& Adolescent Psychiatry, 47, 291–300.

Aalsma, M. C., Schwartz, K., & Perkins, A. J. (2014). A statewide collaboration to initiate mental health

screening and assess service for detained youths in Indiana. American Journal of Public Health, 104,

e82–e88.

Archer, R. P., Simonds-Bisbee, E., Spiegel, D., Handel, R., & Elkins, D. (2010). Validity of the Massachusetts

Youth Screening Instrument-2 (MAYSI-2) in juvenile justice settings. Journal of Personality Assessment,

92, 337–348.

Archer, R. P., Stredny, R. V., Mason, J. A., & Arnau, R. C. (2004). An examination and replication of the psy-

chometric properties of the Massachusetts Youth Screening Instrument—Second Edition (MAYSI-2)

among adolescents in detention settings. Assessment, 11, 1–13.

Bexar County Juvenile Probation Department. (n.d.a.). Juvenile probation. Retrieved from http://gov.bexar.

org/jpd/stats.html

Bexar County Juvenile Probation Department. (n.d.b.). Probation services. Retrieved from http://home.bexar.

org/jpd/probservices.html

Bureau of Justice Statistics. (2005). Deaths in custody statistical tables: State juvenile correctional facility

deaths, 2002–2005. Retrieved from http://bjs.ojp.usdoj.gov/content/dcrp/tables/juvtab2.cfm

Butler, M., Loney, B., & Kistner, J. (2007). The Massachusetts Youth Screening Instrument as a predictor of

institutional maladjustment in severe male juvenile offenders. Criminal Justice and Behavior, 34, 476–492.

Casiano, H., Katz, L. Y., Globerman, D., & Sareen, J. (2013). Suicide and deliberate self-injurious behavior in

juvenile correctional facilities: a review. Journal of the Canadian Academy of Child and Adolescent Psy-

chiatry, 22, 118–124.

Cauffman, E. (2004). A statewide screening of mental health symptoms among juvenile offenders in detention.

Journal of the American Academy of Child & Adolescent Psychiatry, 43, 430–439. doi:10.1097/00004583-

200404000-00009

Cauffman, E., & MacIntosh, R. (2006). A Rasch differential item functioning analysis of the Massachusetts

youth screening instrument. Educational and Psychological Measurement, 66, 502–521.

Centers for Disease Control and Prevention. (2012). Suicide: Facts at a glance. Atlanta, GA: Author. Retrieved

from http://www.cdc.gov/violenceprevention/pdf/Suicide-DataSheet-a

Centers for Disease Control and Prevention. (2013). 1991–2013 high school youth risk behavior survey data.

[Online Database] Retrieved from http://nccd.cdc.gov*9/-/youthonline/

Centers for Disease Control and Prevention. (2014a). Suicide prevention. Atlanta, GA: Author Retrieved from

http://www.cdc.gov/violenceprevention/pub/youth_suicide.html

Centers for Disease Control and Prevention. (2014b). Ten leading causes of death and injury. Atlanta, GA:

Author. Retrieved from http://www.cdc.gov/injury/wisqars/leadingcauses.html

Chapman, J., & Ford, J. (2008). Relationships between suicide risk, traumatic experiences, and substance use

among juvenile detainees. Archives of Suicide Research, 12, 50–61. doi:10.1080/13811110701800830

City of San Antonio. (2012). Teen pregnancy in Bexar County. Retrieved from http://www.sanantonio.gov/Por-

tals/0/Files/health/HealthyLiving/TeenHealth/FactSheet-Teen

Coker, K. L., Wernsman, J., Ikpe, U. N., Brooks, J. S., Bushell, L. L., & Kahn, B. A. (2014). Using the

Massachusetts youth screening instrument-2 on a community sample of African American and Latino/a

juvenile offenders to identify mental health and substance abuse and treatment needs. Criminal Justice and

Behavior, 41, 492–511.

Cruise, K., Marsee, M., Dandreaux, D., & DePrato, D. (2007). Mental health screening of female juvenile

offenders: Replication of a subtyping strategy. Journal of Child and Family Studies, 16, 615–625. doi:10.

1007/s10826-006-9111-4

480 Youth Violence and Juvenile Justice 14(4)

http://gov.bexar.org/jpd/stats.html

http://gov.bexar.org/jpd/stats.html

http://home.bexar.org/jpd/probservices.html

http://home.bexar.org/jpd/probservices.html

http://bjs.ojp.usdoj.gov/content/dcrp/tables/juvtab2.cfm

http://www.cdc.gov/violenceprevention/pdf/Suicide-DataSheet-a

http://nccd.cdc.gov*9/-/youthonline/

http://www.cdc.gov/violenceprevention/pub/youth_suicide.html

http://www.cdc.gov/injury/wisqars/leadingcauses.html

http://www.sanantonio.gov/Portals/0/Files/health/HealthyLiving/TeenHealth/FactSheet-Teen

http://www.sanantonio.gov/Portals/0/Files/health/HealthyLiving/TeenHealth/FactSheet-Teen

Dalton, R. F., Evans, L. J., Cruise, K. R., Feinstein, R. A., & Kendrick, R. F. (2009). Race differences in

mental health service access in a secure male juvenile justice facility. Journal of Offender Rehabilita-

tion, 48, 194–209.

Ford, J. D., Hartman, J. K., Hawke, J., & Chapman, J. F. (2008). Traumatic victimization, posttraumatic stress

disorder, suicidal ideation, and substance abuse risk among juvenile justice-involved youths. Journal of

Child and Adolescent Trauma, 1, 75–92. doi:10.1080/19361520801934456

Gallagher, C. A., & Dobrin, A. (2006). Deaths in juvenile justice residential facilities. Journal of Adolescent

Health, 38, 662–668.

Gallo, L., Penedo, F., Espinoza, K., & Arguelles, W. (2009). Resiliency in the face of disadvantage? Do

Hispanic cultural characteristics protect health outcomes? Journal of Personality, 77, 1707–1746.

Goldston, D. B., Molock, S. D., Whitbeck, L. B., Murakami, J. L., Zayas, L. H., & Hall, G. C. N. (2008). Cul-

tural considerations in adolescent suicide prevention and psychosocial treatment. American Psychologist,

62, 14–31.

Gretton, H., & Clift, R. (2011). The mental health needs of incarcerated youth in British Columbia, Canada.

International Journal of Law and Psychiatry, 34, 109–115.

Grisso, T., & Barnum, R. (2006). Massachusetts youth screening instrument version 2: User’s manual and tech-

nical report 2006 revised edition. Sarasota, FL: Professional Resource Press.

Grisso, T., Barnum, B., Fletcher, K., Cauffman, E., & Peuschold, D. (2001). Massachusetts youth screening

instrument for mental health needs of juvenile justice youths. Journal of the American Academy of Child

and Adolescent Psychiatry, 40, 541–548.

Grisso, T., Fusco, S., Paiva-Salisbury, M., Perrauot, R., Williams, V., & Barnum, R. (2012). The Massa-

chusetts Youth Screening Instrument-Version 2 (MAYSI-2): Comprehensive research review. Worcester,

MA: University of Massachusetts Medical School. Retrieved from http://www.nysap.us/MAYSI-

2%20Review

Grisso, T., & Quinlan, J. (2005). Massachusetts youth screening instrument-revised. In T. Grisso, G. Vincent, &

D. Seagrave (Eds.), Mental health screening and assessment in juvenile justice (pp. 99–111). New York,

NY: Guilford Press.

Hayes, M. A., McReynolds, L. S., & Wasserman, G. A. (2005). Paper and voice MAYSI-2: Format compar-

ability and concordance with the voice DISC-IV. Assessment, 12, 395–403.

Hockenberry, S. (2014). Juveniles in residential placement, 2011. Washington, DC: US Department of Justice,

Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention.

Holtgraves, T. (2004). Social desirability and self-reports: Testing models of socially desirable responding. Per-

sonality and Social Psychology Bulletin, 30, 161.

Kerig, P. K., Moeddel, M. A., & Becker, S. P. (2011). Assessing sensitivity and specificity of the MAYSI-2 for

detecting trauma among youth in juvenile detention. Child & Youth Care Forum, 40, 345–362.

Lewinsohn, P. M., Rohde, P., Seeley, J. R., & Baldwin, C. L. (2001). Gender differences in suicide attempts

from adolescence to young adulthood. Journal of the American Academy of Child & Adolescent Psychiatry,

40, 427–434. doi:10.1097/00004583-200104000-00011

Locke, T., & Newcomb, M. (2005). Psychosocial predictors and correlates of suicidality in teenage Latino

males. Hispanic Journal of Behavioral Sciences, 27, 319–326.

Lynam, D. R., Caspi, A., Moffitt, T., Wikstrom, P. O., Loeber, R., & Novak, S. (2000). The interaction between

impulsivity and neighborhood context on offending: The effects of impulsivity are stronger in poor neigh-

borhoods. Journal of Abnormal Psychology, 109, 563–574.

Mallett, C., DeRigne, L. A., Quinn, L., & Stoddard-Dare, P. (2012). Discerning reported suicide attempts within

a youthful offender population. Suicide and Life Threatening Behavior, 42, 67–77. doi:10.1111/j.1943-

278X.2011.00071.x

Maney, S. (2011). Violent and nonviolent juvenile offenders: Comparisons concerning self report of mental

health symptoms and response styles. Unpublished doctoral dissertation, Massachusetts School of Profes-

sional Psychology, Boston, MA.

Tapia et al. 481

http://www.nysap.us/MAYSI-2%20Review

http://www.nysap.us/MAYSI-2%20Review

http://www.nysap.us/MAYSI-2%20Review

McCoy, H. (2014). Using cognitive interviewing to explore causes for racial differences in the MAYSI-2.

Crime and Delinquency, 60, 647–666. doi:10.1177/0011128710388922

McCoy, H., & Bowen, E. A. (2014). Disproportionality and disparities in the juvenile justice system and the

courts. In R. Fong, A. Dettlaff, J. James, & C. Rodriguez (Eds.), Eliminating racial disproportionality and

disparities: Multi systems culturally competent approaches (pp. 208–237). New York, NY: Columbia

University.

McCoy, H., Vaughn, M. G., Maynard, B. R., & Salas-Wright, C. P. (2014). Caution or warning? A validity

study of the MAYSI-2 with juvenile offenders. Behavioral Sciences and the Law, 32, 508–526. doi:10.

1002/bsl.2128

Mendel, R. A. (2009). Two decades of JDAI. Baltimore, MD: Annie E. Casey Foundation.

Morris, R. E., Harrison, E. A., Knox, G. W., Tromanhauser, E., Marquis, D. K., & Watts, L. L. (1995). Health

risk behavioral survey from 39 juvenile correctional facilities in the United States. Journal of Adolescent

Health, 17, 334–344.

National Action Alliance for Suicide Prevention. (2013). Guide to developing and revising suicide prevention

protocols for youth in contact with the juvenile justice system. Washington, DC: Author.

National Center for Injury Prevention and Control, Centers for Disease Control. (2014). 2012, United

States suicide injury deaths and rates per 100,000. Retrieved from http://webappa.cdc.gov/cgi-bin/bro-

ker.exe

Nolen, S., McReynolds, L. S., DeComo, R. E., John, R., Keating, J. M., & Wasserman, G. A. (2008). Lifetime

suicide attempts in juvenile assessment center youth. Archives of Suicide Research, 12, 111–123. doi:10.

1080/13811110701857087

Nordness, P., Grummert, M., Banks, D., Schindler, M., Moss, M., Gallagher, K., . . . Epstein, M. (2002).

Screening the mental health needs of youths in juvenile detention. Juvenile and Family Court Journal,

52, 43–50.

Rohde, P., Mace, D. E., & Seeley, J. R. (1997). The association of psychiatric disorders with suicide attempts in

a juvenile delinquent sample. Criminal Behaviour and Mental Health, 7, 187–200. doi:10.1002/cbm.172

Stathis, S., Letters, P., Doolan, I., Fleming, R., Heath, K., Arnett, A., . . . Cory, S. (2008). Use of the Massachu-

setts youth screening instrument to assess mental health problems in young people within an Australian

youth detention centre. Journal of Pediatrics and Child Health, 44, 438–443.

Stoeltje, M. F. (2014, February 11). CPS stats: Bexar County No. 2 in abuse, neglect. San Antonio Express.

Retrieved from http://www.mysanantonio.com/news/local/article/CPS-stats-Bexar-County-No-2-in-abuse-

neglect-5224476.php.

Tapia, M. (2012). Juvenile arrest in America: Race, social class, and gang membership. New York, NY: LFB

Scholarly.

Texas Attorney General. (2014). 2014 Juvenile justice handbook. Retrieved from http://www.texasattorneygen-

eral.gov/AG_Publications/pdfs/juvenile_justice

Texas Juvenile Justice Department. (2013). How offenders move through TJJD. Retrieved from http://www.

tjjd.texas.gov/about/how_movethru.aspx

The Stephen Group. (2014). DFPS CPS operational review executive summary: Assessment. Manchester, NH:

Texas Department of Family and Protective Services. Retrieved from http://www.dfps.state.tx.us/docu-

ments/about/Other_Reports/stephen_group_report/Stephens_Grp_Findings_exec_overview

Tienda, M., & Mitchell, F. (Eds.). (2006). Hispanics and the future of America. Washington, DC: National

Research Council, Panel on Hispanics in the U.S. National Academies Press.

Tille, J., & Rose, J. (2007). Emotional and behavioral problems of 13-to-18-year-old incarcerated female

first-time offenders and recidivists. Youth Violence and Juvenile Justice, 5, 426–435. doi:10.1177/

1541204007300355

Umaña-Taylor, A. J., & Updegraff, K. A. (2007). Latino adolescents’ mental health: Exploring the interrela-

tions among discrimination, ethnic identity, cultural orientation, self-esteem, and depressive symptoms.

Journal of Adolescence, 30, 549–567.

482 Youth Violence and Juvenile Justice 14(4)

http://webappa.cdc.gov/cgi-bin/broker.exe

http://webappa.cdc.gov/cgi-bin/broker.exe

http://www.mysanantonio.com/news/local/article/CPS-stats-Bexar-County-No-2-in-abuse-neglect-5224476.php

http://www.mysanantonio.com/news/local/article/CPS-stats-Bexar-County-No-2-in-abuse-neglect-5224476.php

http://www.texasattorneygeneral.gov/AG_Publications/pdfs/juvenile_justice

http://www.texasattorneygeneral.gov/AG_Publications/pdfs/juvenile_justice

http://www.tjjd.texas.gov/about/how_movethru.aspx

http://www.tjjd.texas.gov/about/how_movethru.aspx

http://www.dfps.state.tx.us/documents/about/Other_Reports/stephen_group_report/Stephens_Grp_Findings_exec_overview

http://www.dfps.state.tx.us/documents/about/Other_Reports/stephen_group_report/Stephens_Grp_Findings_exec_overview

U.S. Census Bureau. (2014a). Selected economic characteristics 2008–2012 American community survey

5-year estimates. Retrieved from http://factfinder2.census.gov/faces/tableservices/jsf/pages/productview.

xhtml?src¼bkmk
U.S. Census Bureau. (2014b). State & County QuickFacts. Bexar County, Texas. Retrieved from http://quick-

facts.census.gov/qfd/states/48/48029.html

U.S. Department of Health and Human Services. (1999). The surgeon general’s call to action to prevent

suicide. Washington, DC: Author. Retrieved from http://www.surgeongeneral.gov/library/calltoaction/

fact3.htm

U.S. Department of Health and Human Services. (2001). National strategy for suicide prevention: Goals and

objectives for action. Rockville, MD: Public Health Service. Retrieved from http://www.surgeongeneral.

gov/library/reports/national-strategy-suicide-prevention/

Valdez, A., Kaplan, C. D., & Codina, E. (2000). Psychopathy among Mexican American gang members: A

comparative study. International Journal of Offender Therapy and Comparative Criminology, 44, 46–58.

Veysey, B. M. (2003). Adolescent girls with mental health disorders involved with the juvenile justice system.

Delmar, NY: National Center for Mental health and Juvenile Justice.

Vincent, G., Grisso, T., Terry, A., & Banks, S. (2008a). Sex and race differences in mental health symptoms

in juvenile justice: The MAYSI-2 national meta-analysis. Journal of the American Academy of Child and

Adolescent Psychiatry, 57, 282–290.

Vincent, G., Grisso, T., Terry, A., & Banks, S. (2008b). [Sex and race differences in mental health symptoms in

juvenile justice: The MAYSI-2 national meta-analysis]. Unpublished raw data.

Wagner, B. M. (2009). Suicidal behavior in children & adolescents. New Haven, CT: Yale University.

Wasserman, G. A., & McReynolds, L. S. (2006). Suicide risk at juvenile justice intake. Suicide and Life Threa-

tening Behavior, 36, 239–249.

Wasserman, G. A., McReynolds, L. S., Ko, S. J., Katz, L. M., Cauffman, E., Haxton, W., & Lucas, C. P. (2004).

Screening for emergent risk and service needs among incarcerated youth: Comparing MAYSI-2 and Voice

DISC-IV. Journal of the American Academy of Child and Adolescent Psychiatry, 43, 629–639.

Zayas, L., & Pilat, A. (2008). Suicidal behavior in Latinas: Explanatory cultural factors and implications for

intervention. Suicide and Life Threatening Behavior, 38, 334–342.

Author Biographies

Mike Tapia is an assistant professor of Criminal Justice at New Mexico State University. His

research interests include juvenile corrections, race-ethnicity, and Latino gangs. He holds a PhD

in Sociology from the Ohio State University and was on the faculty at UT San Antonio’s Department

of Criminal Justice from 2003–2014.

Henrika McCoy is an assistant professor at the Jane Addams College of Social Work at the Uni-

versity of Illinois at Chicago. She has a PhD from the George Warren Brown School of Social Work

at Washington University in St. Louis. Her work, which has been funded by the Robert Wood John-

son Foundation, focuses on the intersection between juvenile delinquency, mental health disorders,

and health disparities for youth of color.

Lynsey Tucker is a master of Social Work clinician working with people experiencing homeless-

ness and serious mental illness. She received her MSW degree from UT San Antonio. Her past work

includes multicultural engagement with orphaned and vulnerable children with HIV/AIDS.

Tapia et al. 483

http://factfinder2.census.gov/faces/tableservices/jsf/pages/productview.xhtml?src=bkmk

http://factfinder2.census.gov/faces/tableservices/jsf/pages/productview.xhtml?src=bkmk

http://factfinder2.census.gov/faces/tableservices/jsf/pages/productview.xhtml?src=bkmk

http://quickfacts.census.gov/qfd/states/48/48029.html

http://quickfacts.census.gov/qfd/states/48/48029.html

http://www.surgeongeneral.gov/library/calltoaction/fact3.htm

http://www.surgeongeneral.gov/library/calltoaction/fact3.htm

http://www.surgeongeneral.gov/library/reports/national-strategy-suicide-prevention/

http://www.surgeongeneral.gov/library/reports/national-strategy-suicide-prevention/

<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile (Gray Gamma 2.2) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Warning /CompatibilityLevel 1.3 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages false /CreateJDFFile false /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.1000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails false /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams true /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness false /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Apply /UCRandBGInfo /Remove /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages false /ColorImageMinResolution 266 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Average /ColorImageResolution 175 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50286 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages true /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.40 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >>
/ColorImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >>
/JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/AntiAliasGrayImages false
/CropGrayImages false
/GrayImageMinResolution 266
/GrayImageMinResolutionPolicy /OK
/DownsampleGrayImages true
/GrayImageDownsampleType /Average
/GrayImageResolution 175
/GrayImageDepth -1
/GrayImageMinDownsampleDepth 2
/GrayImageDownsampleThreshold 1.50286
/EncodeGrayImages true
/GrayImageFilter /DCTEncode
/AutoFilterGrayImages true
/GrayImageAutoFilterStrategy /JPEG
/GrayACSImageDict << /QFactor 0.40 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >>
/GrayImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >>
/JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/AntiAliasMonoImages false
/CropMonoImages false
/MonoImageMinResolution 900
/MonoImageMinResolutionPolicy /OK
/DownsampleMonoImages true
/MonoImageDownsampleType /Average
/MonoImageResolution 175
/MonoImageDepth -1
/MonoImageDownsampleThreshold 1.50286
/EncodeMonoImages true
/MonoImageFilter /CCITTFaxEncode
/MonoImageDict << /K -1 >>
/AllowPSXObjects false
/CheckCompliance [
/None
]
/PDFX1aCheck false
/PDFX3Check false
/PDFXCompliantPDFOnly false
/PDFXNoTrimBoxError true
/PDFXTrimBoxToMediaBoxOffset [
0.00000
0.00000
0.00000
0.00000
]
/PDFXSetBleedBoxToMediaBox false
/PDFXBleedBoxToTrimBoxOffset [
0.00000
0.00000
0.00000
0.00000
]
/PDFXOutputIntentProfile (U.S. Web Coated \050SWOP\051 v2)
/PDFXOutputConditionIdentifier (CGATS TR 001)
/PDFXOutputCondition ()
/PDFXRegistryName (http://www.color.org)
/PDFXTrapped /Unknown
/Description << /ENU
>>
/Namespace [
(Adobe)
(Common)
(1.0)
]
/OtherNamespaces [
<< /AsReaderSpreads false /CropImagesToFrames true /ErrorControl /WarnAndContinue /FlattenerIgnoreSpreadOverrides false /IncludeGuidesGrids false /IncludeNonPrinting false /IncludeSlug false /Namespace [ (Adobe) (InDesign) (4.0) ] /OmitPlacedBitmaps false /OmitPlacedEPS false /OmitPlacedPDF false /SimulateOverprint /Legacy >>
<< /AllowImageBreaks true /AllowTableBreaks true /ExpandPage false /HonorBaseURL true /HonorRolloverEffect false /IgnoreHTMLPageBreaks false /IncludeHeaderFooter false /MarginOffset [ 0 0 0 0 ] /MetadataAuthor () /MetadataKeywords () /MetadataSubject () /MetadataTitle () /MetricPageSize [ 0 0 ] /MetricUnit /inch /MobileCompatible 0 /Namespace [ (Adobe) (GoLive) (8.0) ] /OpenZoomToHTMLFontSize false /PageOrientation /Portrait /RemoveBackground false /ShrinkContent true /TreatColorsAs /MainMonitorColors /UseEmbeddedProfiles false /UseHTMLTitleAsMetadata true >>
<< /AddBleedMarks false /AddColorBars false /AddCropMarks false /AddPageInfo false /AddRegMarks false /BleedOffset [ 9 9 9 9 ] /ConvertColors /ConvertToRGB /DestinationProfileName (sRGB IEC61966-2.1) /DestinationProfileSelector /UseName /Downsample16BitImages true /FlattenerPreset << /ClipComplexRegions true /ConvertStrokesToOutlines false /ConvertTextToOutlines false /GradientResolution 300 /LineArtTextResolution 1200 /PresetName ([High Resolution]) /PresetSelector /HighResolution /RasterVectorBalance 1 >>
/FormElements true
/GenerateStructure false
/IncludeBookmarks false
/IncludeHyperlinks false
/IncludeInteractive false
/IncludeLayers false
/IncludeProfiles true
/MarksOffset 9
/MarksWeight 0.125000
/MultimediaHandling /UseObjectSettings
/Namespace [
(Adobe)
(CreativeSuite)
(2.0)
]
/PDFXOutputIntentProfileSelector /DocumentCMYK
/PageMarksFile /RomanDefault
/PreserveEditing true
/UntaggedCMYKHandling /UseDocumentProfile
/UntaggedRGBHandling /UseDocumentProfile
/UseDocumentBleed false
>>
]
/SyntheticBoldness 1.000000
>> setdistillerparams
<< /HWResolution [288 288] /PageSize [612.000 792.000] >> setpagedevice

Article

A Theory-Based Approach to Understanding
Suicide Risk in Shelter-Seeking Women

Caitlin Wolford-Clevenger1 and Phillip N. Smith1

Abstract
Women seeking shelter from intimate partner violence are at an increased risk for suicide ideation and attempts compared to
women in the general population. Control-based violence, which is common among shelter-seeking women, may play a pivotal
role in the development of suicide ideation and attempts. Current risk assessment and management practices for shelter-seeking
women are limited by the lack of an empirically grounded understanding of increased risk in this population. We argue that in
order to more effectively promote risk assessment and management, an empirically supported theory that is sensitive to the
experiences of shelter-seeking women is needed. Such a theory-driven approach has the benefits of identifying and prioritizing
targetable areas for intervention. Here, we review the evidence for the link between coercive control and suicide ideation and
attempts from the perspective of Baumeister’s escape theory of suicide. This theory has the potential to explain the role of coer-
cive control in the development of suicide ideation and eventual attempts in shelter-seeking women. Implications for suicide risk
assessment and prevention in domestic violence shelters are discussed.

Keywords
coercive control, intimate partner violence, suicide risk, shelter seeking, and escape theory

Intimate partner violence (IPV) is a serious public health concern,

affecting one in three women in their lifetime in the United States

(Black et al., 2011). Research has implicated IPV for several neg-

ative mental health outcomes (Campbell, 2002), including

suicide-related thoughts and behaviors (Bergman & Brismar,

1991; Golding, 1999; Langhinrichsen-Rohling, Snarr, Smith

Slep, Heyman, & Foran, 2011; Simon, Anderson, Thompson,

Crosby, & Sacks, 2002). In particular, abused women who seek

safety at a domestic violence shelter may be at greatest risk, with

one third of shelter-seeking women reporting suicide ideation or

suicide attempts (Golding, 1999; Kirkwood, 1993). However, sui-

cide risk in shelter-seeking women may go unrecognized and

untreated as no documented, uniform procedure for suicide risk

assessment and intervention exists in shelters across the United

States. Moreover, a majority of shelters do not provide psycholo-

gical treatment, likely due to limited resources (Roberts & Lewis,

2000). Therefore, shelter-seeking women may endure long waits

for treatment, experience increased psychological distress, sui-

cide ideation, and resultant risk for death by suicide. We argue

that an empirically based theory that identifies critical compo-

nents underlying the process by which shelter-seeking women

come to be at risk for suicide is essential for guiding suicide risk

assessment and management practices in shelter settings.

The Trajectory to Death by Suicide

Suicide risk is not static and several predisposing biopsycho-

social factors converge to result in death by suicide.

Understanding the overtime variations in risk along the trajec-

tory to suicide is essential for effective suicide risk assess-

ment and management (Joiner, 2002; Silverman, Berman,

Sanddal, O’Carroll, & Joiner, 2007a). Furthermore, it is

essential to adopt clear and consistent definitions of suicide-

related thoughts and behaviors to improve the validity and

reliability of research findings and translation of findings to

clinical practice (Silverman et al., 2007a).

In many clinical settings, the earliest indicator of potential

suicide risk is evidence of suicide ideation, defined as thoughts

of suicide that can vary in frequency, duration, intensity, and

content (Silverman, Berman, Sanddal, O’Carroll, & Joiner,

2007b). However, the term suicide ideation is often used to

describe a myriad of cognitions, introducing problems with

specificity and accurate risk designation. For example, it is

important to distinguish suicide ideation from death ideation,

which involves thoughts of death, being dead, or typically not

existing (Smith et al., 2012). Although some who experience

death ideation will go on to develop suicide ideation, this expe-

rience is common (up to 48.6%) for women who have

1 Department of Psychology, University of South Alabama, Mobile, AL, USA

Corresponding Author:

Phillip N. Smith, Psychology Department, University of South Alabama, 75 S.

University Blvd. (UCOM 1000), Mobile, AL 36688, USA.

Email: pnsmith@southalabama.edu

TRAUMA, VIOLENCE, & ABUSE
2015, Vol. 16(2) 169-178
ª The Author(s) 2014
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1524838013517562
tva.sagepub.com

http://www.sagepub.com/journalsPermissions.nav

http://tva.sagepub.com

http://crossmark.crossref.org/dialog/?doi=10.1177%2F1524838013517562&domain=pdf&date_stamp=2014-01-10

experienced significant trauma histories and ongoing interperso-

nal distress (Smith et al., 2012). However, particular variations

of suicide ideation are indicative of greater risk than others. Pas-
sive suicide ideation involves thoughts of or wishes for suicide
but does not involve thoughts of a plan or method. Active suicide
ideation, on the other hand, involves thoughts of suicide, poten-
tial methods for suicide, and, when most severe, plans and pre-

parations for suicide (Joiner, Rudd, & Rajab, 1997; Silverman

et al., 2007b). Although passive suicide ideation should warrant

attention, active ideation that includes resolved plans and pre-

parations denotes greater acute suicide risk (Joiner, 2002; Joiner

et al., 1997; Joiner, Walker, Rudd, & Jobes, 1999).

In addition to the varying active and specific nature of sui-

cide ideation discussed, suicide ideation varies in intentionality

toward behavior. Given that patients’ predictions of attempts

are more accurate than some providers (Peterson, Skeem, &

Manchak, 2011), examining and understanding the nature of

the expressed or implied intent is critical (Silverman et al.,

2007b). Such intent may be subtle, remain unexpressed, or,

conversely, involve specific indications, or suicide threats, that

an individual will engage in suicide-related behaviors. These

behaviors can include self-inflicted body tissue damage with-

out intending to die (i.e., nonsuicidal self-injury [NSSI]), which

often serves specific functions, such as relieving emotional dis-

tress (Klonsky, 2007). In the presence of any nonzero level of

intent to die, such self-inflicted injury is considered a suicide

attempt or—if death results—suicide (Silverman et al.,

2007b). One of the greatest difficulties in shelter settings may

be discerning NSSI from suicide attempts, as both are prevalent

given the incidence of complex trauma, borderline personality

disorder (BPD), and posttraumatic stress disorder (PTSD) in

this population (Weaver & Clum, 1996). Complex trauma

symptomology is specifically related to early, chronic interper-

sonal trauma, and includes emotional dysregulation, dissocia-

tion, NSSI, and suicide attempts (van der Kolk, Roth,

Pelcovitz, Sunday, & Spinazzola, 2005). Such a prevalence

of complex trauma, BPD, and PTSD would make it difficult

to distinguish whether self-injury is driven by intent to die

rather than other functions (e.g., emotion regulation; Klonsky,

2007). Therefore, assessing the intent underlying suicide

threats and related behaviors will allow providers to accurately

determine level of suicide risk and intervene appropriately.

Finally, providers involved in assessing and managing

suicide risk must consider that the trajectory to suicide is not

a linear progression from passive to active suicide ideation,

from ideation to suicide attempt and, ultimately, to death

by suicide. In fact, the trajectory begins arguably prena-

tally—with biological diatheses such as variations in serotonin

and testosterone exposure and development (Arango et al.,

2001; Arango, Underwood, Gubbi, & Mann, 1995; Witte &

Smith, 2008). Environmental stressors, such as interpersonal

discord, bring about suicide ideation that—when expressed—

indicates potential risk (Yen et al., 2005). Suicide-related

thoughts often serve a function other than to prepare for suicide,

most notably the intent to escape the state of psychological pain

or psychache (Shneidman, 1993). Some argue that even the

intent to die is simply an extreme expression of the wish to

escape psychache (Silverman et al., 2007a). However, as suicide

ideation fails to serve its intended function of escape, suicide-

related thoughts and the related psychache may yield to beha-

vior. Thus, understanding the intended function underlying

expressed suicide-related thoughts and behaviors is critical. At

the beginning of the trajectory, suicide ideation may be intended

as a temporary escape from a stressor, whereas near the end of

the trajectory such thoughts may cease to provide an escape, ren-

dering death by suicide to be perceived as a more sustainable

solution (Joiner, 2002; Joiner & Rudd, 2000).

IPV

Like suicide-related thoughts and behaviors, IPV is heteroge-

neous; ranging in frequency, severity, and chronicity of vio-

lence (Langhinrichsen-Rohling, 2010). In order to better

understand and prevent IPV, researchers have attempted to

construct typologies based on presumed important aspects of

IPV. For example, Holtzworth-Munroe and Stuart (1994) clas-

sified perpetrators based on their documented differences in

physiological reactivity (Gottman, Jacobson, Rushe, & Shortt,

1995), psychopathology, and frequency, severity, and general-

ity of physical abuse (Gondolf, 1988; Hamberger & Hastings,

1986; Holtzworth-Munroe & Rehman, 2000). Such classifica-

tions were developed to specify treatment planning for each

subtype, for example, by focusing less resources on perpetra-

tors high in psychopathy (Gondolf, 1988). Unfortunately,

typologies based on perpetrators’ characteristics or forms of

abuse (i.e., physical, sexual, and psychological) have proven

limited as forms often co-occur (O’Leary, Malone, & Tyree,

1994), and characteristics fail to differentiate groups as pre-

dicted (Huss & Langhinrichsen-Rohling, 2006; Waltz, Bab-

cock, Jacobson, & Gottman, 2000). Most importantly,

however, such typologies have not been effectively employed

in clinical settings, demonstrating limited practical utility

(Langhinrichsen-Rohling, Huss, & Ramsey, 2000; Lohr,

Bonge, Witte, Hamberger, & Langhinrichsen-Rohling, 2005).

Missing from these typologies has been the consideration of

motivations for violence or the function the violence is pre-

sumed to serve. Discerning the functions of violence will iden-

tify targetable areas for violence prevention as well as refine

treatment of mental health sequelae (e.g., suicide) associated

with the violence. Johnson (1995) proposed a typology based

on the function (i.e., coercion vs. conflict) and the direction

of the violence (i.e., unidirectional vs. bidirectional). He pos-

ited two bidirectional violent relationships: one in which one

partner is physically violent in response to the other’s control-

ling violence (violent resistance) and one in which both part-

ners fight to exert control over the other (mutual violent

control; Johnson, 2006; Johnson & Ferraro, 2000). Two unidir-

ectional violent relationships were proposed: one in which the

agent’s physical violence occurs in response to escalating con-

flicts (situational couple violence) and one in which one partner

perpetrates a pattern of violence to gain control over the other

(intimate terrorism).

170 TRAUMA, VIOLENCE, & ABUSE 16(2)

Johnson’s typology is particularly useful compared to previ-

ous models in considering the application of theory to the assess-

ment and prevention of suicide in shelter-seeking women. First,

his framework is able to explain the gender symmetry debate

present in IPV research. Findings suggest that the majority of

IPV assessed in agency samples (e.g., shelters) is intimate terror-

ism perpetrated by males; whereas IPV assessed in community

samples is predominately situational violence perpetrated

equally by males and females (Frye, Manganello, Campbell,

Walton-Moss, & Wilt, 2006; Graham-Kevan & Archer, 2003;

Johnson, 2006; Johnson & Leone, 2005; Langhinrichsen-

Rohling, 2006; Laroche, 2005; Leone, 2011; Leone, Johnson,

& Cohan, 2007; Rosen, Stith, Few, Daly, & Tritt, 2005; Tanha,

Beck, Figueredo, & Raghavan, 2010). Second, evidence sug-

gests Johnson’s typology improves upon past classifications by

ascribing the heterogeneity of IPV to both partners’ use and

function of violence. He accomplishes this by placing less

emphasis on severity of violence and more on both partners’ use

of violence to either resolve conflict or exert control. Finally,

distinguishing IPV by these functions will improve our under-

standing and prevention of suicide risk in shelter-seeking women

by targeting coercive control as a risk factor (Leone, 2011).

Control-Based Violence, Suicide Ideation,
and Attempts

Control-based violence increases risk for suicide ideation,

threats, and attempts independent of physical violence severity,

substance use, and depressive symptoms, but not PTSD symp-

toms (Leone, 2011; Pico-Alfonso et al., 2006). This finding sug-

gests that PTSD symptoms may play a specific role in the

relationship between coercive control and suicide risk. Indirect

support for the relationship between coercive control and suicide

ideation and attempts has emerged from samples assumed to be

predominantly intimate terrorism targets (e.g., shelters and

emergency departments; Johnson, 1995). Golding’s (1999)

meta-analysis reported higher rates of suicide ideation and

attempts in women seeking help from IPV in shelter (34%) and
emergency department samples (20%) compared to women in
general population samples (6.6%). Additionally, women identi-
fied to be at risk for intimate partner homicide through the Dan-

ger Assessment (Campbell, Webster, & Glass, 2009) are likely

targets of coercive control, as measures of intimate partner homi-

cide risk typically tap controlling behaviors (e.g., threats, isola-

tion, and surveillance). As would be expected, women at risk for

intimate partner homicide are significantly likely to have a his-

tory of threatened or attempted suicide (Cavanaugh, Messing,

Del-Colle, O’Sullivan, & Campbell, 2011; Sato-DiLorenzo &

Sharps, 2007). Finally, psychological abuse, consisting of con-

trolling behaviors, is predictive of both suicide ideation and

attempts in women presenting to emergency departments for a

suicide attempt (Kaslow et al., 1998) or other medical problems

(Houry, Kemball, Rhodes, & Kaslow, 2006).

Based on these findings, we argue that coercive control is

most functionally relevant for the development of suicide risk

in IPV targets. Further, given that coercive control is prevalent

in shelter-seeking women, models that explicate these relation-

ships will be of importance for assessing and managing risk in

these settings. Although theoretical models of suicide are avail-

able, no discussions of the implications and applications of

such theories to this population have been offered. Such a

framework is necessary to contextualize risk and identify targe-

table areas for prevention and treatment of suicide risk in

women seeking shelter from control-based violence.

Suicide as Escape From the Self

Baumeister’s (1990) escape theory of suicide is a promising

framework for understanding suicide risk in shelter-seeking

women. Escape theory contextualizes how specific life events,

such as seeking shelter and the interpersonal discord that typi-

cally accompanies it, promote suicide risk by describing specific

psychological processes. In contrast to theories of suicide that

center on risk factors for suicide rather than specific contextual

events, such as hopelessness (Beck, Brown, Berchick, Stewart,

& Steer, 1990), social integration (Durkheim, 1897), psychache

(Shneidman, 1993), and the interpersonal theory (Joiner, 2005;

Van Orden et al., 2010), escape theory will guide identification

of targetable points for suicide risk assessment and intervention

in shelter settings. Broadly, escape theory posits that self-blame

for difficult life events increases painful self-awareness and

motivations to escape such awareness. Motivations to escape

aversive self-awareness generate a state of restricted cognitive

flexibility termed cognitive deconstruction, which ultimately

facilitates suicide ideation and attempts to die by suicide (Bau-

meister, 1990). Coercive control plays a pivotal role in the devel-

opment and maintenance of suicide risk in shelter-seeking

women in the context of escape theory.

Women seeking shelter from a violent relationship face an

acute major life event that is very difficult or discrepant with

expectations (see Figure 1). Indeed, negative life events such

as interpersonal discord (Yen et al., 2005) and rejection (Bau-

meister, 1990; Twenge, Catanese, & Baumeister, 2003) are sig-

nificant predictors of acute episodes of suicidal desire, suicide

attempts, and deaths by suicide (Conner et al., 2012; Hill, Pet-

tit, Green, Morgan, & Schatte, 2012; Smith et al., 2012). As

such, the acute interpersonal discord and act of seeking safety

from a violent relationship create a selected vulnerability for

suicide risk for these women. Seeking shelter also comes with

practical and social difficulties that increase distress such as

finding employment, housing, and engaging in legal processes.

Women exposed to greater coercive control are more likely to

experience these difficulties due to coercive tactics such as eco-

nomic control and stalking behaviors (Dutton & Goodman,

2005). Critical to the development of suicide ideation, women

exposed to greater coercive control are likely to blame them-

selves for the abuse and conflict, the need to seek shelter, and

related difficulties due to the psychological warfare of blame

and guilt perpetrated against them (Kirkwood, 1993; Peterson

& Seligman, 1983; Reviere et al., 2007). Targets of coercive

control are also more likely to make such internal attributions

due to unsuccessful past help-seeking attempts and feeling

Wolford-Clevenger and Smith 171

incapable or at fault for failing to cease the violence (Leone

et al., 2007). According to escape theory, this self-blame brings

about painful self-awareness, psychache, and motivations to elim-

inate or escape such pain (Baumeister, 1990). Of course, not all

shelter-seeking women are suicidal. Women seeking shelter from

non-controlling violence may attribute any related distress to exter-

nal sources (e.g., the perpetrator) and experience less painful self-

awareness. Such external attributions would be predicted to pro-

mote increased self-efficacy and greater relief about reaching

safety from the abuse (Peterson & Seligman, 1983). This interpre-

tation is in line with reports of greater hopefulness and less distress

during shelter seeking, despite the aforementioned stressors that

accompany shelter seeking (Clements, Sabourin, & Spiby, 2004;

Clements & Sawhney, 2000).

Thus, coercive control is the key variable underlying the pro-

cess by which shelter-seeking women become vulnerable for

painful self-awareness and resultant escape motivations. Women

exposed to greater coercive control attempt to escape such

awareness by cognitive deconstruction (Baumeister, 1990). Cog-

nitive deconstruction is a temporary state in which higher level

thinking and associated emotions are eliminated by a constricted

focus to immediate goals, physical acts, and sensory input (Bau-

meister, 1990). Such a state is akin to emotional numbing and

dissociative responses to victimization common of complex

trauma victims and individuals experiencing BPD and PTSD

symptomology (Frieze, Hymer, & Greenberg, 1987). As such,

individuals with a histories of complex trauma, BPD, and/or

PTSD may be at greater risk for suicide ideation and attempts

(Foote, Smolin, Neft, & Lipschitz, 2008; Guerra, Calhoun,

Mid-Atlantic Mental Illness Research, & Education and Clinical

Center Workgroup, 2011; Wedig et al., 2012) due in part

because they are especially predisposed to this deconstructed

state during heightened painful awareness and distress.

The narrowed focus to the present inherent to cognitive

deconstruction breeds disinhibition, irrational thought, and

problem-solving difficulties (Dixon, Heppner, & Rudd,

1994; Howat & Davidson, 2002), including deficits that

characterize suicide attempters such as Black and White,

rigid thinking (Neuringer, 1961, 1964), and impulsivity

(Patsiokas, Clum, & Luscomb, 1979). Such a cognitive–

emotional state is effective, at least initially, in coping with

the aversive self-awareness generated by internal attribu-

tions. This state, however, is only temporarily effective and

cannot be sustained, causing women to alternate between

states of painful self-awareness and the emotionless state

of cognitive deconstruction. As women begin to recognize

the temporary nature of cognitive deconstruction, hopeless-

ness about permanent escape increases. Problem-solving

deficits, irrational thought, and disinhibition make more

severe and risky thoughts and behaviors—such as suicide

ideation—more acceptable to reduce aversive self-

awareness. Women who exhibit complex trauma histories,

BPD, and PTSD symptoms may resort to NSSI as a means

for emotional relief that likely worked during previous

stressors (Brown, Comtois, & Linehan, 2002; van der Kolk

et al., 2005), which—while unintended—can be lethal

(Klonsky, 2007). As wishes for death and NSSI decrease

in effectiveness in relieving the psychache, women must

progress and attempt suicide to achieve similar escape (Bau-

meister, 1990; Joiner, 2002). The lethality of the attempt is

partially contingent on women’s ability to engage in lethal

self-injury, which is typically expressed through availability

and familiarity with lethal means (Smith & Cukrowicz,

2010). Thus, the degree to which escape is not achieved

interacts with the ability to lethally self-injure, resulting in

death by suicide (Baumeister, 1990).

Figure 1. Suicide ideation and attempts in shelter-seeking women within the framework of escape theory.

172 TRAUMA, VIOLENCE, & ABUSE 16(2)

Clinical Implications

Given the high rates of suicide ideation and attempts in shelter-

seeking women (Golding, 1999), a standard practice for identify-

ing and managing suicide risk in shelter settings is sorely needed.

Most shelters do not have adequate financial and staff resources to

provide extensive psychological treatments; making traditional

therapies for suicide and related behaviors, such as dialectical

behavior therapy, unsuitable. Therefore, we restrict our discus-

sion to recommendations for assessing and targeting immediate

suicide risk factors that are manageable within shelter settings

without prolonged therapy. We provide basic information with

references to detailed procedures from evidence-based recom-

mendations for outpatient practice (Bryan & Rudd, 2006; Joiner

et al., 1999). For individuals who have received formal training,

we emphasize areas of risk novel to the escape model presented.

This model will be especially useful for guiding identification of

risk and specific points for intervention in these settings.

Assessment

According to escape theory, events preceding and associated with

shelter-seeking are triggers for the process leading to suicide idea-

tion. Therefore, we recommend that suicide risk assessment

should be conducted at each intake. When assessing suicide risk,

practitioners should assess both chronic and acute suicide risk.

Chronic risk is described as a ‘‘baseline’’ level of risk for suicide

and is generally informed by an individual’s suicide attempt his-

tory, with two or more attempts indicating at least moderate

chronic risk (see Bryan & Rudd, 2006; Joiner et al., 1999, for a

detailed review). Individuals with elevated chronic risk have a

lower threshold for experiencing an acute crisis—even in the

absence of an external event—and tend to have more enduring

crises compared to nonattempters and single attempters (Joiner

& Rudd, 2000). Interpreted through the lens of escape theory,

multiple suicide attempters are prone to experiencing enduring

aversive self-awareness and also have frequently experienced the

futility of cognitive deconstruction in relieving psychache. There-

fore, shelter-seeking women with multiple suicide attempts are

vulnerable for experiencing their situation as an enduring crisis

that is not effectively managed by cognitive deconstruction,

which would result in suicide ideation. Practitioners should con-

tinually monitor and assess women with histories of multiple sui-

cide attempts, given that they will more readily enter distress

states and shift from cognitive deconstruction to suicide ideation.

Practitioners could identify chronic risk by examining items

on homicide risk assessments typically given at intake (e.g., dan-

ger assessment) that inquire about women’s histories of threaten-

ing or attempting suicide. Positive endorsements on these items

should direct further assessment of chronic risk by evaluating the

women’s suicide attempt histories. We recommend using a com-

prehensive interview that assesses the presence of the intent to

die in self-injuring events (e.g., The Columbia-Suicide Severity

Rating Scale; Posner et al., 2011). Given the incidence of BPD,

PTSD, and complex trauma in shelter settings, symptoms of

affect dysregulation, impulse control, and dissociation should

be noted as indicators of chronic risk given these symptoms pre-

dispose individuals for entering cognitive deconstruction and

progressing to attempt suicide in the face of internal or external

stressors (Brown et al., 2002; Joiner & Rudd, 2000).

Although chronic suicide risk is relatively static, acute risk

varies according to the precipitating stressor (e.g., interpersonal

loss), present symptomology (e.g., hopelessness), impulsivity,

and lack of protective factors (e.g., social support; Bryan &

Rudd, 2006). Standard risk assessments indicate collecting

information about the frequency, intensity, and duration of sui-

cide ideation, the presence of intent (i.e., implicit and explicit),

plans and methods considered, and preparation for suicide

(e.g., collecting pills). Additionally, methods and plans consid-

ered should be assessed until the individual denies other meth-

ods, as many suicidal individuals do not initially disclose the

most lethal method considered (Bryan & Rudd, 2006).

Given escape theory’s orientation to specific contextual events

and subsequent psychological processes, providers should use

escape theory as a guide for identifying acute risk by observing

women’s responses to the stressor of entering shelter. Responses

that would serve as critical indicators of acute risk include psy-

chache, as indicated by depressed affect, self-blame, hopeless-

ness, and evidence of the deconstructed state such as blunted or

flat affect, problem-solving deficits, disinhibition, and dissocia-

tion (Baumeister, 1990). According to escape theory, PTSD

symptoms of emotional numbing or experiential avoidance are

concerning indicators that suicide risk has increased as the

individual is attempting to manage a state of painful self-

awareness. However, as discussed, cognitive deconstruction only

provides temporary relief. Upon unsuccessful resolution of the

state of painful self-awareness, agitation and perturbation, which

are indicators of acute suicide risk, are likely to result (Busch,

Fawcett, & Jacobs, 2003; Shneidman, 1993). Careful attention

should be paid to individuals with histories of additional interper-

sonal traumas (e.g., childhood sexual abuse and adult sexual

assault), as complex symptomology predisposes them to dissocia-

tive responses to distress (van der Kolk et al., 2005).

After a thorough assessment of factors involved in chronic

and acute risk, practitioners should be well equipped to desig-

nate suicide risk and initiate safety planning and risk manage-

ment. According to escape theory and recommendations for

designating suicide risk in outpatient settings, women seeking

shelter are presenting with a noteworthy risk factor: the inter-

personal loss and violence preceding shelter. Thus, women

with chronic risk (i.e., multiple suicide attempts) are at moder-

ate risk, given their vulnerability to acute crises (Joiner et al.,

1999). Women with chronic risk reporting passive suicide idea-

tion are at severe risk, while those reporting active ideation

(e.g., identified time and method) are at extreme risk.

Nonattempters or single attempters with low acute risk and

no additional risk factors can be deemed at low or mild risk,

because they do not exhibit chronic risk but are experiencing

interpersonal stressors that are documented to increase suicide

risk (Yen et al., 2005). Endorsement of passive suicide ideation

in addition to a risk factor consistent with the theory (e.g.,

hopelessness and self-blame), nonattempters or single

Wolford-Clevenger and Smith 173

attempters should be placed at moderate suicide risk. Evidence

of active suicide ideation (e.g., plans and preparation), how-

ever, should warrant at minimum moderate risk designation.

If nonattempters or single attempters endorse active ideation

paired with an additional risk factor, severe to extreme risk

should be designated (Joiner et al., 1999). In sum, practitioners

designating suicide risk should consider the chronic and acute

risk factors consistent with escape theory including the stres-

sors of seeking safety at a shelter.

Management

Following designation of acute suicide risk, practitioners’ most

immediate decision is whether to manage the women’s suicide

risk through outpatient or inpatient referral. For women identi-

fied to be at severe to extreme acute risk for suicide, referrals

for psychiatric inpatient treatment are addressed through emer-

gency medicine. It is important to adopt a collaborative

approach when discussing seeking emergency services. Doing

so will reduce the need to break confidentiality and will reduce

the possibility of women perceiving the situation as coercive

(Rudd, Madrusiak, & Jobes, 2006).

Unfortunately, women who are not at high acute risk often

endure long waits for outpatient treatment and most shelters are

not equipped to provide in-house treatment (Roberts & Lewis,

2000). Domestic violence shelters are also limited in that the

length of stay is quite short and does not lend itself to prolonged

treatments, such as cognitive therapy. Also important, there is lit-

tle support for the notion that suicide risk abates upon treatment of

the primary psychiatric condition (Linehan, 2008). This under-

lines the usefulness of this model in shelter settings for guiding

practitioners’ management of suicide risk beyond targeting spe-

cific psychiatric symptoms, information that may not be readily

available or amendable compared to contextual risk factors in

shelters (i.e., stress resulting from recent assault). However, cli-

ents presenting with BPD and complex PTSD symptoms should

be referred to clinics specializing in intensive treatments for BPD

and complex trauma such as dialectical behavior therapy and

phase-oriented therapy; though such clinics are often sparse (van

der Kolk, 2002). Phase-oriented therapy in particular may be use-

ful for reducing suicide risk, as it aims to teach the individual to

label emotional responses to distressing stimuli to reduce avoid-

ance and emotional numbing responses to internal and external

cues of distress (van der Kolk, 2002). In short, phase-oriented

therapy would directly treat the propensity for cognitive decon-

struction during an aversive state, thus reducing risk for suicide

ideation and eventual attempts.

While awaiting outpatient treatment, practitioners should

manage clients with mild to moderate risk by first collaborating

with the women by discussing a ‘‘commitment to treatment’’

agreement, developing a safety plan, and increasing contact with

social supports (Joiner et al., 1999; Rudd et al., 2006). According

to escape theory, practitioners should aim to prevent feelings of

unmet expectations by promoting hope during the intake process

and throughout the shelter. For example, during intake, staff

should emphasize successes of shelter services in aiding women

achieve independence and freedom from their abusers. Display-

ing encouraging information throughout the shelter about surviv-

ing abuse and regaining control through available resources will

increase women’s perceptions of control, hopefulness, relief, and

reduce vulnerability to self-blame (Nurius et al., 2003). For

women who experience self-blame for failed expectations,

encouraging self-forgiveness about any perceived failures may

prevent or reduce painful self-awareness and resultant cognitive

deconstruction. Additionally, shelters may provide brief mind-

fulness intervention groups for teaching distress toleration

through body scan, meditation, and practicing self-compassion.

While IPV victims have reported difficulties with mindfulness-

based techniques given the heightened focus on painful emo-

tions, with practice they report experiencing reduced distress

(Bermudez et al., 2013). Practitioners should note that individu-

als with PTSD, complex trauma, or BPD have greater difficulties

tolerating distress and will struggle most with reducing aversive

self-awareness using mindfulness-based techniques (Sass,

Berenbaum, & Abrams, 2013).

Shelters could reduce the likelihood of cognitive deconstruc-

tion by decreasing avoidance coping and increasing active coping

skills through group-formatted problem-solving therapy (Mal-

ouff, Thorsteinsson, & Schutte, 2007; Nezu, 2004). However,

women’s use of avoidance coping strategies largely depends on

contextual influences such as cumulative trauma, violence sever-

ity, relationship length, and prior unsuccessful active coping (e.g.,

calling police and talking to a friend) (Waldrop & Resnick, 2004).

Therefore, increasing active coping strategies will only be useful

when paired with increasing resources that will aid them in the

context of a violent, isolating relationship such as the develop-

ment of social support, career skills, and accessing social services

(Moos, 1995; Sullivan, Basta, Tan, & Davidson, 1992). Given

that women with complex trauma, PTSD, or BPD have a propen-

sity toward states of avoidance and dissociation, referrals to out-

patient treatment to resolve these symptoms must be made (Amir,

Kaplan, Efroni, & Kotler, 1999). Furthermore, given that the

alternation between the deconstructed state and painful self-

awareness breeds greater hopelessness and a greater propensity

for suicide ideation, special attention may be given to women who

have been in shelter for a longer duration, especially if the alter-

nating symptoms of both negative affect and emotional numbness

are observed. Inquiring about and restricting means when possi-

ble such as potentially lethal medication and previous partners’

firearms is necessary to prevent women’s access to lethal means

for suicide. Finally, continual monitoring and awareness of

increases in acute risk is critical to ensure women’s safety (Joiner

et al., 1999).

Limitations and Future Directions

The extant literature supporting the notion that coercive con-

trol increases risk for suicide in targets of IPV suffers many

limitations (see Tables 1 and 2 for a summary of critical find-

ings and implications). First, studies either dichotomize coer-

cive control or use measures that are not designed to measure

coercive control (i.e., the danger assessment). Studies should

174 TRAUMA, VIOLENCE, & ABUSE 16(2)

measure coercive control dimensionally using a validated

measure that is designed to assess coercive control (Dutton

& Goodman, 2005) instead of comparing groups of high and

low coercive control victimization, which reduces power,

effect size, individual differences, and increases risk for Type

I error (MacCallum, Zhang, Preacher, & Rucker, 2002). Sec-

ond, no study to date investigating the relationship between

coercive control and suicide risk assesses the intent to die in

measuring a suicide attempt—an element that is critical for

valid measurement of suicide attempts. Relatedly, many stud-

ies combine suicide ideation, threats, and attempts into a sin-

gle outcome variable, making it impossible to establish the

level of risk coercive control confers for distinct suicidal out-

comes. Finally, longitudinal examination of the development

of suicide risk in targets of IPV is needed to establish a tem-

poral relationship between control-based violence experi-

enced in relationships and increased suicide risk.

In addition to these methodological limitations, the existing

literature is incomplete due to a lack of theoretical guidance.

As such, this literature would benefit from studies developing

or testing theories of suicide in explaining suicide risk in

shelter-seeking women. For example, future studies should

examine the relationships among self-blame, negative affect,

and characteristics of the deconstructed state in shelter-seeking

women who experience varying levels of control-based vio-

lence. Additionally, the literature is severely lacking in

examining whether women who are prone to suicide-related

thoughts and behaviors in addition to coercive control are

women who exhibit BPD traits. Research must examine this pos-

sible relationship, yet should not exclude the potential usefulness

of a theory in identifying functionally relevant areas for interven-

tion in women with BPD symptoms. Most importantly, though,

is that research must aim to identify what factors drive shelter-

seeking women to make an actual attempt versus engage in sui-

cide ideation. This limited area of research would benefit from

theoretical guidance. Perhaps examining the extent to which

women’s current situations are discrepant to their relationship

expectancies could delineate how some targets come to think

about suicide while others do not. Furthermore, examining past

help-seeking behaviors and related hopelessness may elucidate

why some women attempt suicide while others think about it.

Future research should explore these areas and identify areas for

reducing suicide risk in targets of IPV.

Conclusions

We have discussed how escape theory can be used to explain

the role coercive control victimization plays in the develop-

ment of suicide risk in shelter-seeking women. This model

shows promising utility for use in shelter settings, where stan-

dard suicide risk assessment and management are lacking.

Additional research is required to test the usefulness of this

model for understanding, assessing, and managing suicide risk

in shelter-seeking women.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to

the research, author

ship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, author-

ship, and/or publication of this article.

References

Amir, M. M., Kaplan, Z. Z., Efroni, R. R., & Kotler, M. M. (1999).

Suicide risk and coping styles in posttraumatic stress disorder

patients. Psychotherapy and Psychosomatics, 68, 76–81.

Arango, V., Underwood, M. D., Boldrini, M., Tamir, H., Kassir, S. A.,

Hsiung, S., . . . Mann, J. J. (2001). Serotonin 1a receptors, seroto-

nin transporter binding and serotonin transporter mRNA expres-

sion in the brainstem of depressed suicide victims.

Neuropsychopharmacology, 25, 892–903.

Arango, V., Underwood, M. D., Gubbi, A. V., & Mann, J. J. (1995).

Localized alterations in pre- and postsynaptic serotonin binding

sites in the ventrolateral prefrontal cortex of suicide victims. Brain

Research, 688, 121–133.

Baumeister, R. F. (1990). Suicide as escape from self. Psychological

Review, 97, 90–113.

Beck, A. T., Brown, G., Berchick, R. J., Stewart, B. L., & Steer, R. A.

(1990). Relationship between hopelessness and ultimate suicide: A

replication with psychiatric outpatients. American Journal of Psy-

chiatry, 147, 190–195.

Table 1. Critical Findings.

� There are few recommendations for an approach to understanding,
assessing, and managing suicide risk in shelter settings
� There has been no application of existing theories to understanding

the development of suicide risk in women seeking shelter from
intimate partner violence
� There is direct and indirect support that coercive control plays a

critical role in the development of suicide risk in shelter-seeking
women
� There is preliminary support suggesting that Baumeister’s escape

theory of suicide is a suitable framework for understanding
increased suicide risk in shelter-seeking women

Table 2. Practice and Research Implications.

� A standard practice for identifying and managing suicide risk in
shelter settings is needed
� Suicide risk assessment should be conducted for women presenting

to shelter settings
� This conceptual model based in Baumeister’s escape theory of

suicide will be useful in identifying points for assessing and managing
suicide risk in shelter settings
� Future research should be devoted to developing or testing theories

of suicide to explain increased suicide risk in shelter-seeking women
� Research examining the relationships among coercive control, self-

blame, negative affect, cognitive deconstruction, and suicide risk is
needed to test the usefulness of this theory for this population
� Future research must aim to distinguish what factors place shelter-

seeking women at risk for suicide attempts rather than death
ideation or suicide ideation

Wolford-Clevenger and Smith 175

Bergman, B., & Brismar, B. (1991). A 5-year follow-up study of 117 bat-

tered women. American Journal of Public Health, 81, 1486–1489.

Bermudez, D., Benjamin, M. T., Porter, S. E., Saunders, P. A., Myers, N.

A., & Dutton, M. (2013). A qualitative analysis of beginning mindful-

ness experiences for women with post-traumatic stress disorder and a

history of intimate partner violence. Complementary Therapies in

Clinical Practice, 19, 104–108. doi:10.1016/j.ctcp.2013.02.004

Black, M. C., Basile, K. C., Breiding, M. J., Smith, S. G., Walters, M.

L., Merrick, M. T., . . . Stevens, M. R. (2011). The national inti-

mate partner and sexual violence survey (NISVS): 2010 Summary

report. Atlanta, GA: National Center for Injury Prevention and

Control, Centers for Disease Control and Prevention.

Brown, M. Z., Comtois, K. A., & Linehan, M. M. (2002). Reasons for

suicide attempts and nonsuicidal self-injury in women with border-

line personality disorder. Journal of Abnormal Psychology, 111,

198–202. doi:10.1037/0021-843x.111.1.198

Bryan,C.J.,&Rudd,M.D.(2006).Advancesintheassessmentofsuiciderisk.

Journal of Clinical Psychology, 62, 185–200. doi:10.1002/Jclp.20222

Busch, K., Fawcett, J., & Jacobs, D. (2003). Clinical correlates of

inpatient suicide. The Journal of Clinical Psychiatry, 64, 14–19.

Campbell, J. C. (2002). Health consequences of intimate partner violence.

Lancet, 359, 1331–1336. doi:10.1016/S0140-6736(02)08336-8

Campbell, J. C., Webster, D. W., & Glass, N. (2009). The danger

assessment: Validation of a lethality risk assessment instrument for

intimate partner femicide. Journal of Interpersonal Violence, 24,

653–674. doi:10.1177/0886260508317180

Cavanaugh,C.E.,Messing,J.T.,Del-Colle,M.,O’Sullivan,C.,&Campbell,

J. C. (2011). Prevalence and correlates of suicidal behavior among adult

female victims of intimate partner violence. Suicide & Life-Threatening

Behavior, 41, 372–383. doi:10.1111/j.1943-278X.2011.00035.x

Clements, C. M., Sabourin, C. M., & Spiby, L. (2004). Dysphoria and

hopelessness following battering: The role of perceived control,

coping, and self-esteem. Journal of Family Violence, 19, 25–36.

Clements, C. M., & Sawhney, D. K. (2000). Coping with domestic

violence: Control attributions, dysphoria, and hopelessness. Journal

of Traumatic Stress, 13, 219–240.

Conner, K. R., Houston, R. J., Swogger, M. T., Conwell, Y., You, S., He,

H., . . . Duberstein, P. R. (2012). Stressful life events and suicidal

behavior in adults with alcohol use disorders: Role of event severity,

timing, and type. Drug and Alcohol Dependence, 120, 155–161.

Dixon, W. A., Heppner, P. P., & Rudd, M. D. (1994). Problem-solving

appraisal, hopelessness, and suicide ideation: Evidence for a med-

iational model. Journal of Counseling Psychology, 41, 91–98. doi:

10.1037/0022-0167.41.1.91

Durkheim, E. (1897). Le Suicide: Etude de socologie. Paris, France: F.

Alcan.

Dutton, M. A., & Goodman, L. A. (2005). Coercion in intimate partner

violence: Toward a new conceptualization. Sex Roles, 52,

743–756. doi:10.1007/s11199-005-4196-6

Foote, B., Smolin, Y., Neft, D., & Lipschitz, D. (2008). Dissociative

disorders and suicidality in psychiatric outpatients. Journal of Ner-

vous & Mental Disease, 196, 29–36.

Frieze, I. H., Hymer, S., & Greenberg, M. S. (1987). Describing the

crime victim: Psychological reactions to victimization. Profes-

sional Psychology: Research and Practice, 18, 299–315. doi:10.

1037/0735-7028.18.4.299

Frye, V., Manganello, J., Campbell, J. C., Walton-Moss, B., & Wilt, S.

(2006). The distribution of and factors associated with intimate ter-

rorism and situational couple violence among a population-based

sample of urban women in the United States. Journal of Interper-

sonal Violence, 21, 1286–1313. doi:10.1177/0886260506291658

Golding, J. M. (1999). Intimate partner violence as a risk factor for

mental disorders: A meta-analysis. Journal of Family Violence,

14, 99–132. doi:10.1023/a:1022079418229

Gondolf, E. W. (1988). Who are those guys? Toward a behavioral

typology of batterers. Violence and Victims, 3, 187–203.

Gottman, J. M., Jacobson, N. S., Rushe, R. H., & Shortt, J. W. (1995).

The relationship between heart rate reactivity, emotionally aggres-

sive behavior, and general violence in batterers. Journal of Family

Psychology, 9, 227–248. doi:10.1037/0893-3200.9.3.227

Graham-Kevan, N., & Archer, J. (2003). Intimate terrorism and com-

mon couple violence: A test of Johnson’s predictions in four Brit-

ish samples. Journal of Interpersonal Violence, 18, 1247–1270.

Guerra, V. S., & Calhoun, P. S., Mid-Atlantic Mental Illness Research,

& Education and Clinical Center Workgroup. (2011). Examining

the relation between posttraumatic stress disorder and suicidal

ideation in an OEF/OIF veteran sample. Journal of Anxiety Disor-

ders, 25, 12–18.

Hamberger, L. K., & Hastings, J. E. (1986). Personality correlates of

men who abuse their partners: A cross-validation study. Journal of

Family Violence, 1, 323–341.

Hill, R. M., Pettit, J. W., Green, K. L., Morgan, S. T., & Schatte, D. J.

(2012). Precipitating events in adolescent suicidal crises: Explor-

ing stress-reactive and nonreactive risk profiles. Suicide and

Life-Threatening Behavior, 42, 11–21. doi:10.1111/j.1943-278X.

2011.00067.x

Holtzworth-Munroe, A., & Rehman, U. (2000). General and spouse-

specific anger and hostility in subtypes of maritally violent men

and nonviolent men. Behavior Therapy, 31, 603–630.

Holtzworth-Munroe, A., & Stuart, G. L. (1994). Typologies of male

batterers: Three subtypes and the differences among them. Psycho-

logical Bulletin, 116, 476–497. doi:10.1037/0033-2909.116.3.476

Houry, D., Kemball, R., Rhodes, K. V., & Kaslow, N. J. (2006). Inti-

mate partner violence and mental health symptoms in African

American female ED patients. The American Journal of Emer-

gency Medicine, 24, 444–450.

Howat, S., & Davidson, K. (2002). Parasuicidal behaviour and interperso-

nal problem solving performance in older adults. The British Journal of

Clinical Psychology/The British Psychological Society, 41, 375–386.

Huss, M. T., & Langhinrichsen-Rohling, J. (2006). Assessing the gen-

eralization of psychopathy in a clinical sample of domestic vio-

lence perpetrators. Law and Human Behavior, 30, 571–586. doi:

10.1007/s10979-006-9052-x

Johnson, M. P. (1995). Patriarchal terrorism and common couple vio-

lence: Two forms of violence against women. Journal of Marriage

and the Family, 57, 283–294.

Johnson, M. P. (2006). Conflict and control: Gender symmetry and

asymmetry in domestic violence. Violence Against Women, 12,

1003–1018.

Johnson, M. P., & Ferraro, K. J. (2000). Research on domestic vio-

lence in the 1990s: Making distinctions. Journal of Marriage and

the Family, 62, 948–963. doi:10.1111/j.1741-3737.2000.00948.x

176 TRAUMA, VIOLENCE, & ABUSE 16(2)

Johnson, M. P., & Leone, J. M. (2005). The differential effects of intimate

terrorism and situational couple violence: Findings from the national

violence against women survey. Journal of Family Issues, 26, 322–349.

Joiner, T. E., Jr. (2002). The trajectory of suicidal behavior over time.

Suicide and Life-Threatening Behavior, 32, 33–41. doi:10.1521/

suli.32.1.33.22187

Joiner, T. E., Jr. (2005). Why people die by suicide. Cambridge, MA:

Harvard University Press.

Joiner, T. E., Jr. & Rudd, M. D. (2000). Intensity and duration of suicidal

crises vary as a function of previous suicide attempts and negative life

events. Journal of Consulting and Clinical Psychology, 68, 909–916.

Joiner, T. E., Jr. Rudd, M. D., & Rajab, M. H. (1997). The modified

scale for suicidal ideation: Factors of suicidality and their relation

to clinical and diagnostic variables. Journal of Abnormal Psychol-

ogy, 106, 260–265.

Joiner, T. E., Jr. Walker, R. L., Rudd, M. D., & Jobes, D. A. (1999).

Scientizing and routinizing the assessment of suicidality in outpa-

tient practice. Professional Psychology: Research and Practice,

30, 447–453. doi:10.1037/0735-7028.30.5.447

Kaslow, N. J., Thompson, M. P., Meadows, L. A., Jacobs, D.,

Chance, S., Gibb, B., . . . Phillips, K. (1998). Factors that mediate

and moderate the link between partner abuse and suicidal behavior

in African American women. Journal of Consulting and Clinical

Psychology, 66, 533–540. doi:10.1037/0022-006x.66.3.533

Kirkwood, C. (1993). Leaving abusive partners: From the scars of

survival to the wisdom for change. Thousand Oaks, CA: Sage.

Klonsky, E. D. (2007). The functions of deliberate self-injury: A review

of the evidence. Clinical Psychology Review, 27, 226–239. doi:10.

1016/j.cpr.2006.08.002

Langhinrichsen-Rohling, J. (2006). An examination of sheltered bat-

tered women’s perpetration of stalking and other unwanted pursuit

behaviors. Violence and Victims, 21, 579–595.

Langhinrichsen-Rohling, J. (2010). Controversies involving gender

and intimate partner violence in the United States. Sex Roles, 62,

179–193. doi:10.1007/s11199-009-9628-2

Langhinrichsen-Rohling, J., Huss, M. T., & Ramsey, S. (2000). The

clinical utility of batterer typologies. Journal of Family Violence,

15, 37–54. doi:10.1023/a:1007597319826

Langhinrichsen-Rohling, J., Snarr, J. D., Smith Slep, A. M., Heyman,

R. E., & Foran, H. M. (2011). Risk for suicidal ideation in the U.S.

Air force: An ecological perspective. Journal of Consulting and

Clinical Psychology, 79, 600–612. doi:10.1037/a0024631

Laroche, D. (2005). Aspects of the context and consequences of

domestic violence—situational couple violence and intimate ter-

rorism in Canada in 1999. Quebec City: Government of Quebec.

Leone, J. M. (2011). Suicidal behavior among low-income, African

American female victims of intimate terrorism and situa-

tional couple violence. Journal of Interpersonal Violence, 26,

2568–2591.

Leone, J. M., Johnson, M. P., & Cohan, C. L. (2007). Victim help

seeking: Differences between intimate terrorism and situational

couple violence. Family Relations, 56, 427–439.

Linehan, M. M. (2008). Suicide intervention research: A field in des-

perate need of development. Suicide and Life-Threatening Beha-

vior, 38, 483–485.

Lohr, J. M., Bonge, D., Witte, T. H., Hamberger, K. L., & Langhin-

richsen-Rohling, J. (2005). Consistency and accuracy of batterer

typology identification. Journal of Family Violence, 20,

253–258. doi:10.1007/s10896-005-5989-7

MacCallum, R. C., Zhang, S., Preacher, K. J., & Rucker, D. D. (2002).

On the practice of dichotomization of quantitative variables.

Psychological Methods, 7, 19–40. doi:10.1037//1082-989x.7.1.19

Malouff, J. M., Thorsteinsson, E. B., & Schutte, N. S. (2007). The effi-

cacy of problem solving therapy in reducing mental and physical

health problems: A meta-analysis. Clinical Psychology Review,

27, 46–57. doi:10.1016/j.cpr.2005.12.005

Moos, R. H. (1995). Development and applications of new measures

of life stressors, social resources, and coping responses. European

Journal of Psychological Assessment, 11, 1–13.

Neuringer, C. (1961). Dichotomous evaluations in suicidal individu-

als. Journal of Consulting Psychology, 25, 445–449. doi:10.

1037/h0046460

Neuringer, C. (1964). Rigid thinking in suicidal individuals. Journal

of Consulting Psychology, 28, 54–58. doi:10.1037/h0045809

Nezu, A. M. (2004). Problem solving and behavior therapy revisited.

Behavior Therapy, 35, 1–33.

Nurius, P., Macy, R., Bhuyan, R., Holt, V., Kernic, M., & Rivara, F.

(2003). Contextualizing depression and physical functioning in bat-

tered women: Adding vulnerability and resources to the analysis.

Journal of Interpersonal Violence, 18, 1411–1431.

O’Leary, K. D., Malone, J., & Tyree, A. (1994). Physical aggression in

early marriage: Prerelationship and relationship effects. Journal of

Consulting and Clinical Psychology, 62, 594–602. doi:10.1037/

0022-006x.62.3.594

Patsiokas, A. T., Clum, G. A., & Luscomb, R. L. (1979). Cognitive

characteristics of suicide attempters. Journal of Consulting and

Clinical Psychology, 47, 478–484. doi:10.1037/0022-006x.47.3.478

Peterson, C., & Seligman, M. E. (1983). Learned helplessness and vic-

timization. Journal of Social Issues, 39, 103–116.

Peterson, J., Skeem, J., & Manchak, S. (2011). If you want to know, con-

sider asking: How likely is it that patients will hurt themselves in the

future? Psychological Assessment, 23, 626–634. doi:10.1037/a0022971

Pico-Alfonso, M. A., Garcia-Linares, M. I., Celda-Navarro, N.,

Blasco-Ros, C., Echeburua, E., & Martinez, M. (2006). The impact

of physical, psychological, and sexual intimate male partner vio-

lence on women’s mental health: Depressive symptoms, posttrau-

matic stress disorder, state anxiety, and suicide. Journal of

Women’s Health, 15, 599–611. doi:10.1089/jwh.2006.15.599

Posner, K., Brown, G. K., Stanley, B., Brent, D. A., Yershova, K. V.,

Oquendo, M. A., . . . Mann, J. (2011). The Columbia-suicide

severity rating scale: Initial validity and internal consistency find-

ings from three multisite studies with adolescents and adults.

American Journal of Psychiatry, 168, 1266–1277.

Reviere, S. L., Farber, E. W., Twomey, H., Okun, A., Jackson, E., Zanville,

H., & Kaslow, N. J. (2007). Intimate partner violence and suicidality in

low-income African American women: A multimethod assessment of

coping factors. Violence Against Women, 13, 1113–1129.

Roberts, A. R., & Lewis, S. J. (2000). Giving them shelter: National

organizational survey of shelters for battered women and their chil-

dren. Journal of Community Psychology, 28, 669–681.

Wolford-Clevenger and Smith 177

Rosen, K. H., Stith, S. M., Few, A. L., Daly, K. L., & Tritt, D. R.

(2005). A qualitative investigation of Johnson’s typology. Violence

and Victims, 20, 319–334.

Rudd, M. D., Mandrusiak, M., & Joiner, T. E., Jr. (2006). The case against

no-suicide contracts: The commitment to treatment statement as a

practice alternative. Journal of Clinical Psychology, 62, 243–251.

Sass, S. M., Berenbaum, H., & Abrams, E. M. (2013). Discomfort with emo-

tion moderates distress reduction in a brief mindfulness intervention.

InternationalJournalofBehavioralConsultationandTherapy,7,24–27.

Sato-DiLorenzo, A., & Sharps, P. W. (2007). Dangerous intimate part-

ner relationships and women’s mental health and health behaviors.

Issues in Mental Health Nursing, 28, 837–848. doi:10.1080/

01612840701493535

Shneidman, E. S. (1993). Suicide as psychache. The Journal of Ner-

vous and Mental Disease, 181, 145–147.

Silverman, M. M., Berman, A. L., Sanddal, N. D., O’Carroll, P. W., &

Joiner, T. E., Jr. (2007a). Rebuilding the tower of Babel: A revised

nomenclature for the study of suicide and suicidal behaviors. Part

1: Background, rationale, and methodology. Suicide and Life-

Threatening Behavior, 37, 248–263.

Silverman, M. M., Berman, A. L., Sanddal, N. D., O’Carroll, P. W., &

Joiner, T. E., Jr. (2007b). Rebuilding the tower of Babel: A revised

nomenclature for the study of suicide and suicidal behaviors. Part

2: Suicide-related ideations, communications, and behaviors. Sui-

cide and Life-Threatening Behavior, 37, 264–277.

Simon, T. R., Anderson, M., Thompson, M. P., Crosby, A., & Sacks, J.

(2002). Assault victimization and suicidal ideation or behavior

within a national sample of U.S. adults. Suicide and Life-

Threatening Behavior, 32, 42–50.

Smith, P. N., & Cukrowicz, K. C. (2010). Capable of suicide: A func-

tional model of the acquired capability component of the

interpersonal-psychological theory of suicide. Suicide and Life-

Threatening Behavior, 40, 266–275.

Smith, P. N., Gamble, S. A., Cort, N. A., Ward, E. A., Conwell, Y., &

Talbot, N. L. (2012). The relationships of attachment style and

social maladjustment to death ideation in depressed women with

a history of childhood sexual abuse. Journal of Clinical Psychol-

ogy, 68, 78–87. doi:10.1002/jclp.20852

Sullivan, C. M., Basta, J., Tan, C., & Davidson, W. S. (1992). After the

crisis: A needs assessment of women leaving a domestic violence

shelter. Violence and Victims, 7, 269–275.

Tanha, M., Beck, C. J. A., Figueredo, A. J., & Raghavan, C. (2010).

Sex differences in intimate partner violence and the use of coercive

control as a motivational factor for intimate partner violence. Jour-

nal of Interpersonal Violence, 25, 1836–1854.

Twenge, J. M., Catanese, K. R., & Baumeister, R. F. (2003). Social

exclusion and the deconstructed state: Time perception, meaning-

lessness, lethargy, lack of emotion, and self-awareness. Journal of

Personality and Social Psychology, 85, 409–423. doi:10.1037/

0022-3514.85.3.409

van der Kolk, B. A. (2002). The assessment and treatment of complex

PTSD. In R. Yehuda (Ed.), Treating trauma survivors with PTSD

(pp. 127–156). Washington, DC: American Psychiatric Press.

van der Kolk, B. A., Roth, S., Pelcovitz, D., Sunday, S., & Spinazzola,

J. (2005). Disorders of extreme stress: The empirical foundation of

a complex adaptation to trauma. Journal of Traumatic Stress, 18,

389–399.

Van Orden, K. A., Witte, T. K., Cukrowicz, K. C., Braithwaite, S. R.,

Selby, E. A., & Joiner, T. E., Jr. (2010). The interpersonal theory of

suicide. Psychological Review, 117, 575–600. doi:10.1037/

a0018697

Waldrop, A. E., & Resick, P. A. (2004). Coping among adult female

victims of domestic violence. Journal of Family Violence, 19,

291–302.

Waltz, J., Babcock, J. C., Jacobson, N. S., & Gottman, J. M. (2000).

Testing a typology of batterers. Journal of Consulting and Clinical

Psychology, 68, 658–669. doi:10.1037/0022-006x.68.4.658

Weaver, T. L., & Clum, G. A. (1996). Interpersonal violence: Expand-

ing the search for long-term sequelae within a sample of battered

women. Journal of Traumatic Stress, 9, 783–803.

Wedig, M. M., Silverman, M. H., Frankenburg, F. R., Reich, D., Fitz-

maurice, G. G., & Zanarini, M. C. (2012). Predictors of suicide

attempts in patients with borderline personality disorder over 16

years of prospective follow-up. Psychological Medicine, 42,

2395–2404. doi:10.1017/S0033291712000517

Witte, T. K., & Smith, A. R. (2008). Affective lability and prenatal tes-

tosterone exposure: Risk factors for the acquired capability for sui-

cide. Paper presented at the meeting of the Association for

Behavioral and Cognitive Therapies, Orlando, FL.

Yen, S., Pagano, M. E., Shea, T. M., Grilo, C. M., Gunderson, J. G.,

Skodol, A. E., . . . Zanarini, M.C. (2005). Recent life events pre-

ceding suicide attempts in a personality disorder sample: Findings

from the collaborative longitudinal personality disorders study.

Journal of Consulting and Clinical Psychology, 73, 99–105. doi:

10.1037/0022-006x.73.1.99

Author Biographies

Caitlin Wolford-Clevenger is pursuing her MS in Psychology at the

University of South Alabama where she received her BA in psychol-

ogy. Her research interests focus on understanding the development of

suicide ideation, suicide attempts, and nonsuicidal self-injury in survi-

vors of interpersonal violence.

Phillip N. Smith received his PhD in clinical psychology from Texas

Tech University and completed a 2-year postdoctoral fellowship at the

Center for the Study and Prevention of Suicide in the Department of

Psychiatry at the University of Rochester Medical Center. His

research interests center on the understanding, assessment, and treat-

ment of suicidal patients.

178 TRAUMA, VIOLENCE, & ABUSE 16(2)

<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile (Gray Gamma 2.2) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Warning /CompatibilityLevel 1.3 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages false /CreateJDFFile false /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.1000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails false /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams true /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness false /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Apply /UCRandBGInfo /Remove /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages false /ColorImageMinResolution 266 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Average /ColorImageResolution 175 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50286 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages true /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.40 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >>
/ColorImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >>
/JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/AntiAliasGrayImages false
/CropGrayImages false
/GrayImageMinResolution 266
/GrayImageMinResolutionPolicy /OK
/DownsampleGrayImages true
/GrayImageDownsampleType /Average
/GrayImageResolution 175
/GrayImageDepth -1
/GrayImageMinDownsampleDepth 2
/GrayImageDownsampleThreshold 1.50286
/EncodeGrayImages true
/GrayImageFilter /DCTEncode
/AutoFilterGrayImages true
/GrayImageAutoFilterStrategy /JPEG
/GrayACSImageDict << /QFactor 0.40 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >>
/GrayImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >>
/JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/AntiAliasMonoImages false
/CropMonoImages false
/MonoImageMinResolution 900
/MonoImageMinResolutionPolicy /OK
/DownsampleMonoImages true
/MonoImageDownsampleType /Average
/MonoImageResolution 175
/MonoImageDepth -1
/MonoImageDownsampleThreshold 1.50286
/EncodeMonoImages true
/MonoImageFilter /CCITTFaxEncode
/MonoImageDict << /K -1 >>
/AllowPSXObjects false
/CheckCompliance [
/None
]
/PDFX1aCheck false
/PDFX3Check false
/PDFXCompliantPDFOnly false
/PDFXNoTrimBoxError true
/PDFXTrimBoxToMediaBoxOffset [
0.00000
0.00000
0.00000
0.00000
]
/PDFXSetBleedBoxToMediaBox false
/PDFXBleedBoxToTrimBoxOffset [
0.00000
0.00000
0.00000
0.00000
]
/PDFXOutputIntentProfile (U.S. Web Coated \050SWOP\051 v2)
/PDFXOutputConditionIdentifier (CGATS TR 001)
/PDFXOutputCondition ()
/PDFXRegistryName (http://www.color.org)
/PDFXTrapped /Unknown
/Description << /ENU
>>
/Namespace [
(Adobe)
(Common)
(1.0)
]
/OtherNamespaces [
<< /AsReaderSpreads false /CropImagesToFrames true /ErrorControl /WarnAndContinue /FlattenerIgnoreSpreadOverrides false /IncludeGuidesGrids false /IncludeNonPrinting false /IncludeSlug false /Namespace [ (Adobe) (InDesign) (4.0) ] /OmitPlacedBitmaps false /OmitPlacedEPS false /OmitPlacedPDF false /SimulateOverprint /Legacy >>
<< /AllowImageBreaks true /AllowTableBreaks true /ExpandPage false /HonorBaseURL true /HonorRolloverEffect false /IgnoreHTMLPageBreaks false /IncludeHeaderFooter false /MarginOffset [ 0 0 0 0 ] /MetadataAuthor () /MetadataKeywords () /MetadataSubject () /MetadataTitle () /MetricPageSize [ 0 0 ] /MetricUnit /inch /MobileCompatible 0 /Namespace [ (Adobe) (GoLive) (8.0) ] /OpenZoomToHTMLFontSize false /PageOrientation /Portrait /RemoveBackground false /ShrinkContent true /TreatColorsAs /MainMonitorColors /UseEmbeddedProfiles false /UseHTMLTitleAsMetadata true >>
<< /AddBleedMarks false /AddColorBars false /AddCropMarks false /AddPageInfo false /AddRegMarks false /BleedOffset [ 9 9 9 9 ] /ConvertColors /ConvertToRGB /DestinationProfileName (sRGB IEC61966-2.1) /DestinationProfileSelector /UseName /Downsample16BitImages true /FlattenerPreset << /ClipComplexRegions true /ConvertStrokesToOutlines false /ConvertTextToOutlines false /GradientResolution 300 /LineArtTextResolution 1200 /PresetName ([High Resolution]) /PresetSelector /HighResolution /RasterVectorBalance 1 >>
/FormElements true
/GenerateStructure false
/IncludeBookmarks false
/IncludeHyperlinks false
/IncludeInteractive false
/IncludeLayers false
/IncludeProfiles true
/MarksOffset 9
/MarksWeight 0.125000
/MultimediaHandling /UseObjectSettings
/Namespace [
(Adobe)
(CreativeSuite)
(2.0)
]
/PDFXOutputIntentProfileSelector /DocumentCMYK
/PageMarksFile /RomanDefault
/PreserveEditing true
/UntaggedCMYKHandling /UseDocumentProfile
/UntaggedRGBHandling /UseDocumentProfile
/UseDocumentBleed false
>>
]
/SyntheticBoldness 1.000000
>> setdistillerparams
<< /HWResolution [288 288] /PageSize [612.000 792.000] >> setpagedevice

Expert paper writers are just a few clicks away

Place an order in 3 easy steps. Takes less than 5 mins.

Calculate the price of your order

You will get a personal manager and a discount.
We'll send you the first draft for approval by at
Total price:
$0.00