Clinical Psychology Review 27 (2007) 226 – 239

The functions of deliberate self-injury: A review of the evidence

E. David Klonsky ⁎
Department of Psychology, Stony Brook University, Stony Brook, NY 11794-2500,United States
Received 23 November 2005; received in revised form 9 August 2006; accepted 24 August 2006

Abstract

Deliberate self-injury is defined as the intentional, direct injuring of body tissue without suicidal intent. The present article
reviews the empirical research on the functions of self-injury. This literature includes self-reports of reasons for self-injuring,
descriptions of the phenomenology of self-injury, and laboratory studies examining the effects of self-injury proxies on affect and
physiological arousal. Results from 18 studies provide converging evidence for an affect-regulation function. Research indicates
that: (a) acute negative affect precedes self-injury, (b) decreased negative affect and relief are present after self-injury, (c) self-injury
is most often performed with intent to alleviate negative affect, and (d) negative affect and arousal are reduced by the performance
of self-injury proxies in laboratory settings. Studies also provide strong support for a self-punishment function, and modest
evidence for anti-dissociation, interpersonal-influence, anti-suicide, sensation-seeking, and interpersonal boundaries functions. The
conceptual and empirical relationships among the different functions remain unclear. Future research should address the
measurement, co-variation, clinical correlates, and treatment implications of different functions.
Published by Elsevier Ltd.

Keywords: Self-injury; Self-injurious behavior; Deliberate self-harm; Suicidal behavior; Self-mutilative behavior; Reinforcement; Phenomenology

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
2. Psychosocial characteristics, descriptive psychopathology, and treatment. . . . . . . . . . . . . . . . . . . . . . . 227
3. Empirical research on the functions of self-injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
3.1. Inclusion criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
3.2. Functions examined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
3.3. Methodological considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
3.4. Findings from studies examining self-reports of reasons . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
3.5. Findings from studies examining self-reports of phenomenology . . . . . . . . . . . . . . . . . . . . . . . 233
3.6. Findings from laboratory studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
3.7. Summary of empirical research on the functions of self-injury . . . . . . . . . . . . . . . . . . . . . . . . 235
4. Suggestions for future research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
4.1. Explore conceptual and empirical relationships among functions . . . . . . . . . . . . . . . . . . . . . . . 235

⁎ Tel.: +1 631 632 7801.

E-mail address: [email protected].

0272-7358/$ - see front matter. Published by Elsevier Ltd.

doi:10.1016/j.cpr.2006.08.002
E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239 227

4.2. Measure functions comprehensively . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

4.3. Clarify affective-changes associated with self-injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
4.4. Establish causal link between affective changes and self-injury . . . . . . . . . . . . . . . . . . . . . . . . 236
4.5. Explore mechanisms of affect-regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
4.6. Employ alternatives to retrospective self-report paradigms . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
4.7. Extend research to community samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
4.8. Investigate clinical implications of different functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

1. Introduction

Deliberate self-injury (to be referred to as self-injury for the remainder of the paper) can be defined as the intentional
and direct injuring of one's body tissue without suicidal intent (Herpertz, 1995; Muehlenkamp, 2005). Other names
have also been used to refer to this behavior, including moderate self-mutilation by Favazza and Rosenthal (1993),
deliberate self-harm by Pattison and Kahan (1983), self-wounding by Tantam and Whittaker (1992), and parasui-
cide by Ogundipe (1999). Self-injury is differentiated from stereotypic self-injurious behaviors seen in individuals with
mental retardation, and from severe forms of self-mutilation such as limb amputation seen in psychotic individuals.
Skin-cutting appears to be the most common type of self-injury, but other forms include burning, scratching, banging or
hitting body parts, and interfering with wound healing (Briere & Gil, 1998; Favazza & Conterio, 1989; Herpertz, 1995;
Langbehn & Pfohl, 1993).
The phenomenon of self-injury has concerned mental health professionals for decades. In their seminal paper, Graff
and Mallin (1967) recognized self-injury as a mainstream psychiatric problem, stating that “[nonsuicidal] wrist-slashers
have become the new chronic patients in mental hospitals, replacing the schizophrenics” (p. 36). Efforts to characterize
the scope and nature of the problem increased in the 1970s (Carr, 1977; Lester, 1972), and by the early 1980s some
were calling for self-injury to comprise its own diagnostic entity (Pattison & Kahan, 1983).
Research on self-injury has increased in recent years, and much is now known about the prevalence and risk
factors for self-injury in various populations (Skegg, 2005; Gratz, 2003). However, we continue to lack a sufficient
understanding of the functions of self-injury. This understanding would inform treatment, and provide a meaningful
context for research on the etiology, classification, prevention, and treatment of self-injury. The goal of the present
paper is to consolidate and advance knowledge about the functions of self-injury. Towards this end, I first
summarize the research on the psychosocial characteristics, descriptive psychopathology, and treatment of self-
injury. I then present a detailed review of the empirical literature on the functions of self-injury. I conclude by
describing research endeavors that would further clarify the functions of self-injury and their implications for clinical
practice.

2. Psychosocial characteristics, descriptive psychopathology, and treatment

According to the psychiatric nosology (APA, 2000), self-injury is a symptom of borderline personality disorder.
Although research has documented a strong relationship between self-injury and this disorder (Klonsky, Oltmanns, &
Turkheimer, 2003; Simeon et al., 1992; Stanley, Gameroff, Michalsen, & Mann, 2001; van der Kolk, Perry, & Herman,
1991; Zlotnick, Mattia, & Zimmerman, 1999), patients with other diagnoses may also self-injure, including those with
major depression, anxiety disorders, substance abuse, eating disorders, posttraumatic stress disorder, schizophrenia,
and several personality disorders (Haw, Hawton, Houston, & Townsend, 2001; Herpertz, Sass, & Favazza, 1997;
Klonsky et al., 2003; Zlotnick et al., 1999). Self-injury occurs in non-clinical populations as well. Studies have reported
a history of self-injury in 4% of the United States population (Briere & Gil, 1998), 4% of military recruits (Klonsky
et al., 2003), and 14% or more of college students (Favazza, DeRosear, & Conterio, 1989; Gratz, 2001; Whitlock,
Eckenrode, & Silverman, 2006). Some studies have found that more women than men self-injure (e.g., Zlotnick et al.,
1999), whereas others have not found gender differences (e.g., Briere & Gil, 1998; Stanley et al., 2001). Several studies
have reported an inverse correlation between lifetime rate of self-injury and participants' age (Briere & Gil, 1998;
228 E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239

Soloff, Lis, Kelly, Cornelius, & Ulrich, 1994; Zlotnick et al., 1999), suggesting that self-injury has become more
common in recent years.
Self-injury usually first appears between ages 14 and 24 (Favazza & Conterio, 1989; Herpertz, 1995). Research
suggests that the most common form of self-injury is skin-cutting, occurring in between 70 and 97% of individuals who
self-injure, followed by banging or hitting (21%–44%) and burning (15%–35%) (Briere & Gil, 1998; Favazza and
Conterio; Herpertz; Langbehn & Pfohl, 1993; Nijman et al., 1999; Wilkins & Coid, 1991). Many individuals who self-
injure use more than one method (Favazza and Conterio; Gratz, 2001; Herpertz). Estimates for the average number of
lifetime instances of self-injury are variable, ranging from 3.4 (Soloff et al., 1994) to 50 (Favazza and Conterio).
Treating patients who self-injure presents a number of challenges for therapists. Self-injury may cause
psychological distress to the patient and his or her loved ones, and can result in physical damage that requires
medical treatment or even leads to death. In addition, patients who self-injure may be hospitalized against their will,
even though some mental professionals fear this intervention may damage the therapeutic relationship and may not be
helpful for these patients (Huband & Tantam, 1999).
Research on Dialectical Behavior Therapy (DBT) for patients who self-injure has been encouraging. DBT is an
intensive treatment involving individual and group modalities developed by Marsha Linehan (1993). When compared
to treatment as usual in the community, DBT appears to produce better improvements in frequency of self-injury,
hospitalizations, and many other outcome variables (Linehan, 2000; Robins & Chapman, 2004). However, few studies
have compared DBT to more rigorous control conditions involving manualized treatments or treatment by experts
(Robins & Chapman, 2004; but see Linehan et al., 2006), and many patients treated with DBT continue to self-injure
even if less frequently. Understanding the functions of self-injury, or in other words, the variables that motivate and
reinforce the behavior, could greatly improve prevention and treatment.

3. Empirical research on the functions of self-injury

The primary aim of this article is to review the empirical research on the functions of self-injury. Below I describe
the inclusion criteria, functional theories examined, methods utilized to examine functions, and findings from the
empirical literature.

3.1. Inclusion criteria

Studies reporting data that directly address the functions of self-injury (i.e., motivating and reinforcing
variables) were included in the present review. Examples include studies that assess motivations or reasons for
self-injury, as well as studies that examine variables temporally associated with self-injury, such as affect-states or
events that precede, accompany, or follow self-injury, and which may be conceptualized as providing motivation
or reinforcement. Studies on psychosocial and clinical correlates of self-injury are summarized at the beginning of
this article and are not included in the empirical review of self-injury functions. In addition, studies that analyze
behaviors without ruling out suicidal intent were excluded. Literature searches and examinations of the references
sections of relevant papers yielded 18 empirical studies on the functions of self-injury that met the inclusion
criteria.

3.2. Functions examined

Aggregating results on the functions of self-injury is a difficult task because different studies examined
different functions and used different terms and items to reference and define these functions. On the basis of a
close reading of the empirical literature, I identified seven functions that were repeatedly examined. These
functions are listed and described in Table 1. I attempted to use labels and descriptions for these seven functions
that accurately reflect how they have been described and operationally defined in the empirical literature, although
I acknowledge there may be additional ways to organize, define, and label the functions that have been studied. I
refer to the seven functions as affect-regulation, anti-dissociation, anti-suicide, interpersonal boundaries,
interpersonal-influence, self-punishment, and sensation-seeking. The central theme of each model is presented
below. It is important to note that these functional models are not mutually exclusive since different functions may
co-occur and overlap conceptually.
E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239 229

Table 1
Functions of deliberate self-injury examined in the empirical literature
Function Description of function
Affect-regulation To alleviate acute negative affect or aversive affective arousal
Anti-dissociation a To end the experience of depersonalization or dissociation
Anti-suicide To replace, compromise with, or avoid the impulse to commit suicide
Interpersonal boundaries To assert one's autonomy or a distinction between self and other
Interpersonal-influence To seek help from or manipulate others
Self-punishment To derogate or express anger towards oneself
Sensation-seeking To generate exhilaration or excitement
a
“Feeling generation” is another term for the anti-dissociation model of self-injury. When individuals dissociate they may feel unreal, outside their
body, or nothing. The physical sensations produced by self-injury may help individuals feel something and thus end episodes of dissociation,
depersonalization, or derealization.

The affect-regulation model of self-injury1 suggests that self-injury is a strategy to alleviate acute negative affect or
affective arousal (Favazza, 1992; Gratz, 2003; Haines, Williams, Brain, & Wilson, 1995). Linehan (1993) theorizes
that early invalidating environments may teach poor strategies for coping with emotional distress. Individuals from
these environments and/or with biological dispositions for emotional instability are less able to manage their affect and
are therefore prone to use self-injury as a maladaptive affect-regulation strategy. How self-injuring may diminish
negative affect is unknown, although both psychological (Brown, Comtois, & Linehan, 2002; Suyemoto, 1998) and
biological (Russ, Roth, Kakuma, Harrison, & Hull, 1994) mechanisms have been proposed.
The anti-dissociation model characterizes self-injury as a response to periods of dissociation or depersonalization.
Gunderson (1984) suggests that individuals who self-injure experience prolonged dissociative episodes when loved
ones are absent. Episodes of dissociation or depersonalization may also occur as a result of the intense emotions that
self-injurers feel. Causing physical injury to oneself may shock the system – perhaps through the sight of blood
(Simpson, 1975) or the physical sensation (Gunderson, 1984) – and thereby interrupt a dissociative episode, and lead
one to regain a sense of self. Feeling generation is another term used to characterize this function. Individuals who
dissociate may describe feeling unreal or nothing at all, and self-injury may be a way to generate emotional and
physical sensations that allow individuals to feel real or alive again.
The anti-suicide model views self-injury as a coping mechanism for resisting urges to attempt suicide. From this
perspective, self-injury may be thought of as a means of expressing suicidal thoughts without risking death, and serves
as a replacement for or compromise with the desire to commit suicide (Suyemoto, 1998). For example, Himber (1994)
described a patient who felt she became suicidal after long periods of time without cutting, and that cutting prevented
suicidal ideation.
The interpersonal-influence model stipulates that self-injury is used to influence or manipulate people in the self-
injurer's environment (Chowanec, Josephson, Coleman, & Davis, 1991; Podovall, 1969). Self-injury is conceptualized
as a cry for help, a means of avoiding abandonment, or an attempt to be taken more seriously or otherwise effect
people's behavior (Allen, 1995). For example, an individual might self-injure to elicit affection from a significant other
or loved one, or to elicit reinforcing responses from authority figures or peers in correctional, clinical, or school
settings. One who self-injures for these reasons may or may not be aware of the reinforcement provided by others'
reactions to the self-injury.
The interpersonal boundaries model holds that self-injury is a way to affirm the boundaries of the self (Carroll,
Schaffer, Spensley, & Abramowitz, 1980; Podovall, 1969; Suyemoto, 1998). Proponents of this model tend to draw
upon object-relations theory. Self-injurers are thought to lack a normal sense of self due to insecure maternal
attachments and a subsequent inability to individuate from the mother (Friedman, Glasser, Laufer, Laufer, & Wohl,
1972). Marking the skin, which separates individuals from the environment and other people, is thought to affirm a
distinction between oneself and others, and assert one's identity or autonomy.
The self-punishment model suggests that self-injury is an expression of anger or derogation towards oneself.
Linehan (1993) hypothesizes that self-injurers have learned from their environments to punish or invalidate

1
Terms such as affect-regulation, emotion regulation, and mood regulation are used by different psychologists in different ways. These concepts
have become popular and efforts to reach consensus about terminology and definitions are ongoing (e.g., Larsen, 2000); discussion of this larger
issue is beyond the scope of this paper.
230 E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239

themselves. Many have reported that self-directed anger and self-derogation are prominent characteristics of those
who self-injure (Bennun, 1983; Herpertz et al., 1997; Klonsky et al., 2003; Soloff et al., 1994). Self-injury may
therefore be experienced as familiar and ego-syntonic, and become a way of self-soothing when faced with
emotional distress.
The sensation-seeking model regards self-injury as a means for generating excitement or exhilaration in a manner
similar to sky-diving or bungee jumping. This model has received less attention in the theoretical literature, perhaps
because it is not readily apparent in clinical populations. Nevertheless, this function has been examined repeatedly in
the empirical literature (Nixon, Cloutier, & Aggarwal, 2002; Osuch, Noll, & Putnam, 1999; Shearer, 1994).

3.3. Methodological considerations

Three methodological approaches have been used to examine the functions described above: self-report of reasons
for self-injuring, self-report of phenomenology, and laboratory studies. Self-report of reasons refers to studies in which
self-injurers indicate their reasons or motivations for self-injuring. For the purposes of the present review, the terms
motivations and reasons are used interchangeably and no conceptual distinction is made between them. In these
studies, participants are asked to identify their reasons for self-injuring from a list of potential reasons, or in some cases
report their reasons in response to open-ended questions. While participants' opinions about their motivations for self-
injury can offer valuable insight into its functions, this approach has important limitations. Verbal reports of mental
processes are often invalid (Nisbett & Wilson, 1977). Self-injurers may not know why they self-injure or have difficulty
verbalizing reasons and offer explanations that are not accurate. Others may fabricate explanations if they are
embarrassed by their true reasons.
Self-report of phenomenology refers to studies that ask participants about the experience of self-injuring.
Participants may be asked to describe emotions or events that precede, accompany, or follow a self-injury episode.
Answers to these questions can help illuminate the variables that prompt and reinforce self-injury. This approach has
the advantage of reducing demand characteristics because participants are not overtly asked to give reasons for, or
justify, a socially unacceptable behavior. Participants may therefore be less likely to withhold, alter, or fabricate
information. Nevertheless, self-reports of phenomenology, like self-reports of reasons, are limited by people's tendency
to misunderstand or mischaracterize their mental processes (Nisbett & Wilson, 1977).
Laboratory studies offer an alternative to self-report designs. These studies assess the effects of proxies for self-
injury and have been used to evaluate the affect-regulation function. For example, self-injurers are asked to visualize
cutting or to perform a painful task while affect and affective arousal are assessed with self-report and physiological
measures. Demonstrating decreases in negative affect and arousal when participants perform a self-injury proxy, but
not control tasks, would provide evidence that self-injuring reduces negative affect. Laboratory studies are not
susceptible to the same limitations as traditional self-report studies. In these studies, self-report of affect-states is
immediate rather than retrospective, and there is the opportunity to use physiological measures of affective arousal such
as skin conductance and heart rate that are less susceptible to conscious manipulation. An important limitation of this
approach is that proxies for self-injury may have poor external validity. Practical and ethical issues preclude the
laboratory study of actual self-injury. In addition, laboratory paradigms are not well suited for investigating some
functions.
Table 2 summarizes the methods and findings of the 18 studies on the functions of self-injury included in the present
review. Each study used either one or two of the three methods described above. Nine used only self-report of reasons,
four only self-report of phenomenology, two used self-report of reasons and phenomenology, and three were laboratory
studies of proxies for self-injury. Discussion of these studies will be organized by methodological approach.

3.4. Findings from studies examining self-reports of reasons

Eleven studies examined self-reported reasons for self-injury (see Table 2). In all 11 studies, reasons suggesting an
affect-regulation function were heavily endorsed. For example, in a sample of 75 women with borderline personality
disorder, 96% of participants endorsed reasons indicative of emotion relief such as “to stop bad feelings” and “to relieve
anxiety or terror” (Brown et al., 2002). Similarly, in another a study of women with borderline personality disorder, the
reason “to reduce anxiety and despair that I feel I can't otherwise control” received the second highest mean rating
among 17 possible reasons (Shearer, 1994).
E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239 231

Table 2
Methods and results from 18 empirical studies on the functions of deliberate self-injury
Study n of self-injurers Average age % Female Sample type a Methods b Functions studied and
supported c
Briere and Gil (1998) 93 35 96 In- and outpatient R, P AD, AR, IB, II, SP
Brain et al. (1998) 35 Not given 57 Outpatient L AR
Brown et al. (2002) 75 30 100 Outpatient BPD R AD, AR, II, SP
Coid (1993) 67 29 100 Forensic BPD P AR
Favazza and Conterio (1989) 240 28 100 Non-clinical R AD, AR, SP
Haines et al. (1995) 15 21 0 Forensic L AR
Herpertz (1995) 54 27 87 Inpatient R AD, AR, II, SP
Jones et al. (1979) 39 18–20 56 General and P AR, II
psychiatric hospital
Kemperman et al. (1997) 38 30 100 Inpatient BPD P AD, AR
Kumar et al. (2004) 50 15 62 Inpatients R AR, SP, II, SS
Nixon et al. (2002) 42 16 86 In- and outpatient R AR, AS, II, SP, SS
Laye-Gindhu and Schonert-Reichl 64 15 75 Non-clinical R,P AD, AR, AS, II, SP, SS
(2005)
Nock and Prinstein (2004) 108 15 70 Inpatient R AD, AR, II, SP
Osuch et al. (1999) 75 37 85 Inpatient R AR, SP, II, SS
Penn et al. (2003) 23 16 28 Forensic R AD, AR, SP
Russ et al. (1992) 22 24 100 Inpatient BPD L AR
Shearer (1994) 41 Not given 100 Inpatient BPD R AD, AR, AS, IB, II, SP, SS
Wilkins and Coid (1991) 74 25 100 Forensic P AR, AD
a
BPD indicates that all participants had been diagnosed with borderline personality disorder.
b
This column indicates the methods used to study functions. R = self-report of reasons; P = self-report of phenomenology; L = laboratory study of a
proxy for self-injury.
c
AR = affect-regulation, AD = anti-dissociation, AS = anti-suicide, IB = interpersonal boundaries, II = interpersonal-influence, SP = self-
punishment, and SS = sensation-seeking. The presence of an abbreviation for a function indicates that the function was examined by a particular
study. If a function received strong empirical support in a particular study (e.g., that items relevant to a particular function were endorsed by more than
half the sample or more often than items related to other functions), its abbreviation is underlined and bolded. If a function received modest empirical
support in a particular study (e.g., items relevant to a particular function were endorsed by between 5 and 49% of the sample, or that it is possible but
not certain that the items endorsed support a particular function), its abbreviation is underlined. If a function was examined but not supported by a
particular study, its abbreviation is neither underlined nor bolded.

Reasons consistent with an affect-regulation function are also prominent in non-borderline samples. Seventy-six
percent of a heterogeneous group of psychiatric inpatients selected “tension release” as a reason for self-injury, the only
reason endorsed by the majority of the sample (Herpertz, 1995). In another inpatient sample, reasons related to affect-
regulation (e.g., “to decrease feelings of rage”) received substantially higher ratings than reasons related to other
functions (Osuch et al., 1999). Likewise, in a mixed sample of psychiatric inpatient and outpatient self-injurers, 80%
endorsed the reason “distraction from painful feelings,” 77% “manage stress,” 77% “release pent-up feelings,” and
75% “reduction of tension” (Briere & Gil, 1998). In the only study of self-injury reasons in a non-clinical adult sample,
the three reasons identified most often were all related to affect-regulation: “to control their mind when it is racing”
(endorsed by 72% of the sample), “to feel relaxed” (65%), and “to feel less depressed” (58%) (Favazza & Conterio,
1989).
In adolescent samples, affect-regulation reasons are also endorsed by a majority of participants. For example, in
Penn, Esposito, Schaeffer, Fritz, and Spirito (2003), 23 incarcerated adolescents with a history of self-injury were
administered the Functional Assessment of Self-Mutilation (FASM; Lloyd, Keeley, & Hope, 1997, as cited in Penn
et al.), a self-report measure of 22 reasons for self-injury. The most common reason was “to stop bad feelings,” selected
by 65% of participants. Nock and Prinstein (2004) administered the FASM to 89 adolescent psychiatric inpatients who
had self-injured within the past year. Once again, the reason most often chosen was “to stop bad feelings,” although the
majority was smaller in this case (53% of the sample). Similar results were found in a community sample of adolescent
self-injurers, where 80% indicated they self-injured because “I felt very unhappy or depressed” and 45% endorsed the
reason “it helped me to release tension or stress and relax” (Laye-Gindhu & Schonert-Reichl, 2005). Thus reasons
related to affect-regulation were endorsed by a majority of participants in both adult and adolescent samples, although
these reasons might be slightly less common in adolescents than adults.
232 E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239

Self-reported reasons related to a self-punishment function were investigated in 11 studies. This function received
strong support in six of the studies. For example, in a sample of women with borderline personality disorder, the item
“to punish myself for being ‘bad’” received the highest mean rating of 17 possible reasons (Shearer, 1994). In another
study of women with borderline personality disorder, 63% endorsed “self-punishment” as a reason for self-injury
(Brown et al., 2002). Likewise, 83% of self-injurers in a mixed sample of psychiatric patients selected “self-
punishment” as a reason (Briere and Gil, 1998). Self-punishment reasons were also common in adolescent self-injurers.
Half of adolescents receiving psychiatric treatment endorsed the reason “punish self for being bad/bad thoughts”
(Nixon et al., 2002). Self-punishment reasons were particularly apparent in a community sample of adolescent self-
injurers, where 70% endorsed the reason “I did not like myself,” 64% selected “I felt like a failure,” and 63% endorsed
“I was angry at myself” (Laye-Gindhu & Schonert-Reichl, 2005).
However, support for self-punishment reasons was only modest in five other studies. For example, in an
inpatient sample of self-injurers, reasons related to “self-hatred and self-punishment” were endorsed by just over
10% of participants (Herpertz, 1995). In addition, in a sample of adolescent inpatient self-injurers, only 32%
selected the reason “to punish yourself” (Nock & Prinstein, 2004). In two additional studies of psychiatric
inpatients, one on adolescents (Kumar et al., 2004) and one on adults (Osuch et al., 1999), reasons indicative of a
self-punishment functions received substantially lower ratings than reasons indicative of an affect-regulation
function. Although it is unclear why self-punishment reasons are more prominent in some studies than others,
recent research suggests a possible explanation. In Klonsky (2006), young adults with a history of skin-cutting rated
potential reasons for self-injury as being primary, secondary, or not relevant. Although affect-regulation and self-
punishment reasons were both selected by a majority of participants, affect-regulation reasons were overwhelmingly
rated as primary and self-punishment reasons as secondary. The seemingly inconsistent pattern of findings in the
literature may reflect that most participants identify with self-punishment reasons but fewer consider them to be
principal reasons.
Reasons suggesting an anti-dissociation function were investigated in eight studies, with mixed results. In women
with borderline personality disorder, reasons for self-injury related to anti-dissociation/feeling generation (e.g., “to stop
feeling numb or dead,” “to feel something, even if it is pain”) were endorsed by 54% of the sample (Brown et al., 2002).
Similar results were found in a community sample of self-injurers, in which 55% of participants selected the reason “to
feel real again” (Favazza & Conterio, 1989). Likewise, 60% of a sample of incarcerated adolescent self-injurers
endorsed the reason “to feel something even if it is pain” (Penn et al., 2003).
However, five other studies examining anti-dissociation reasons for self-injury found less support for this function.
For example, in Herpertz (1995), only 9% of the psychiatric inpatient self-injurers endorsed the reason “termination of
depersonalization.” Similarly, only 7% of an inpatient sample of women with borderline personality disorder
considered “to produce feelings and a sense of being real when I feel numb and ‘out of touch’” to be one of their top
three reasons for self-injury (Shearer, 1994). In two studies of adolescent self-injurers, reasons related to an anti-
dissociation function were endorsed by approximately one third of the participants (Laye-Gindhu & Schonert-Reichl,
2005; Nock & Prinstein, 2004).
Nine studies examined reasons indicative of an interpersonal-influence function. Only one found strong evidence
for this function. In Brown et al. (2002), 61% of a sample of borderline women endorsed reasons related to
interpersonal-influence such as “to get other people to act differently or change”. Interpersonal-influence reasons were
less evident in other studies. For example, in Shearer (1994), only 17% of women with borderline personality disorder
selected “to seek support and caring from others” as one of their top three reasons, and only 5% selected “to control the
reactions and behavior of others” as one of their top three reasons. Similarly, 40% of self-injurers from a heterogeneous
psychiatric sample selected the reason “get attention, [or] ask for help” (Briere & Gil, 1998), and 24% of self-injurers in
an inpatient sample selected the reason “longing for care and attention” (Herpertz, 1995). Results were similar in
adolescent populations. Just over 30% of adolescent self-injurers from a non-clinical sample endorsed the reason “I
wanted other people to see how desperate I was” (Laye-Gindhu and Schonert-Reichl, 2005). In an inpatient adolescent
sample the reasons “to get control of a situation” and “to try to get a reaction from someone” were endorsed by less than
15% of the participants (Nock & Prinstein, 2004).
Five studies examined reasons suggesting a sensation-seeking function. All found modest evidence for this function.
For example, only 5% of women with borderline personality disorder regarded “to provide a sense of excitement or
stimulation that feels exhilarating” as one of their top three reasons for self-injury (Shearer, 1994). Likewise, in a study
examining six motivations for self-injury in psychiatric inpatients, sensation-seeking reasons (e.g., “to experience a
E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239 233

‘high’ that feels like a drug high”) were endorsed the least often (Osuch et al., 1999). Studies of adolescents found
similar results. Just over 10% of adolescent self-injurers from a non-clinical population endorsed the reason “I thought
it would be fun” (Laye-Gindhu and Schonert-Reichl, 2005), and fewer than 10% of adolescent psychiatric patients
selected the reason “for excitement” (Nixon et al., 2002).
Three studies examined reasons indicative of an anti-suicide function, with each study finding modest evidence for
this function. In a sample of women with borderline personality disorder, the reason “to prevent me from acting on
suicidal feelings” received the seventh highest mean rating among 17 reasons (Shearer, 1994). In an adolescent
psychiatric sample, just under half of the participants selected the reason “stop suicidal ideation or attempts” (Nixon
et al., 2002). Likewise, 41% of adolescent self-injurers from a non-clinical population endorsed the reason “it stopped
me from killing myself’ (Laye-Gindhu & Schonert-Reichl, 2005).
Only two studies examined reasons suggesting an interpersonal-boundaries function. Both found modest evidence
for this function. Shearer (1994) found that the reason “to do something that only I have control of and no one else can
control” was rated as one of the top-three reasons for self-injury by 22% of women with borderline personality disorder.
Similarly, Briere and Gil (1998) reported that 26% of psychiatric patients endorsed “ownership of body” as a reason for
self-injury.
Taken together, studies on self-reports of reasons for self-injury yield consistent findings. A majority of participants
in the 11 studies reviewed cited the desire to alleviate negative affect or affective arousal as a reason for self-injuring. In
addition, self-punishment reasons were strongly endorsed in six studies, and apparent to a lesser extent in five others.
Reasons associated with anti-dissociation and interpersonal-influence were identified by participants in several studies,
although usually by a minority of participants. Only five studies investigated sensation-seeking reasons, three studies
anti-suicide reasons, and two interpersonal boundaries reasons, with each study producing modest evidence for these
functions.

3.5. Findings from studies examining self-reports of phenomenology

Some researchers examining the functions of self-injury have avoided presenting participants with direct questions
about their reasons. This approach may reduce demand characteristics because participants are not asked to provide
reasons for their self-injury, which may be experienced as having to explain or justify a socially unacceptable behavior.
Instead, participants are asked questions about events or emotions that precede, accompany, or follow episodes of self-
injury. These questions help clarify variables that prompt and reinforce self-injury.
Six studies examined self-reports about the phenomenology of self-injury. Results from these studies, like
those that analyzed self-reports of reasons, provide support for an affect-regulation function of self-injury.
Specifically, findings indicate that self-injury is usually preceded by acute negative affect, and that self-injury is
associated with a substantial reduction of negative affect. For example, 93% of a sample of women with
borderline personality disorder reported that self-injury often occurred following feelings such as tension,
depression, irritability, restlessness, racing thoughts, and anxiety, and 94% of self-injurers in this sample reported
that they experience relief from these affect-states following self-injury (Coid, 1993). In another sample of self-
injurers with borderline personality disorder, approximately 90% of the participants indicated that they felt less
“angry” and “anxious” and more “peaceful” after self-injury as compared to before (Kemperman, Russ, &
Shearin, 1997).
Similar results are apparent in non-clinical populations. In a community sample of self-injurers recruited through
advertisements, the affect-states reported to be most common before self-injury were “anger at others,” “hurt,” and
“emptiness,” whereas “relief” was reported to be the most common affect-state after self-injury (Briere & Gil, 1998).
Similarly, in a community sample of adolescent self-injurers, affect-states such as “angry,” “depressed,” “lonely,” and
“frustrated” were reported to be present prior to self-injury and diminish after self-injury (Laye-Gindhu & Schonert-
Reichl, 2005).
Although studies on the phenomenology of self-injury primarily address an affect-regulation function, some
findings also bear on other functions. For example, in a prison sample of women, 26% indicated during a clinical
interview that self-injury was preceded by “depersonalization” or “derealization” (Wilkins & Coid, 1991). This result
may suggest that the anti-dissociation function is relevant for these individuals. However, in another study participants
reported that dissociative symptoms peaked during rather than before self-injury (Kemperman et al., 1997), which
seems inconsistent with the idea that self-injury serves to reduce dissociative symptoms. In addition, one study
234 E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239

provides possible support for the interpersonal-influence function because interpersonal conflict was reported to be a
frequent antecedent of self-injury (Jones, Congiu, Stevenson, Strauss, & Frei, 1979). However, the study did not assess
whether self-injury had been used to address the interpersonal conflict. Finally, evidence from one study supported a
sensation-seeking function because increases in feelings such as “happy” and “excited” were reported to be associated
with self-injury (Laye-Gindhu & Schonert-Reichl, 2005).

3.6. Findings from laboratory studies

In contrast to self-report studies, laboratory studies examine the psychological and physiological influences
of proxies for self-injury in controlled settings. This approach avoids the limitations of traditional self-report
approaches because it allows for the measurement of affect during or shortly after performance of the self-injury
proxy, and allows affective arousal to be assessed by physiological in addition to self-report measures. In addition,
laboratory studies have the potential to allow for causal inferences about the psychological and physiological effects
of proxies for self-injury. A trade-off is that generalizability of findings is limited by the external validity of the self-
injury proxy.
In the first study to apply this approach, Haines et al. (1995) hypothesized that self-injurers administered self-
injury imagery would exhibit increases in physiological arousal and negative affect during the initial stages
(leading up to the self-injury act), and decreases in arousal and negative affect when the behavior and its
aftermath were recalled. This pattern was not hypothesized to occur for control participants or for self-injurers
imagining control events such as arguing with a significant other or accidentally cutting oneself with a kitchen
knife. Participants were 38 men belonging to one of three groups: 15 prison inmates who self-injured with low
suicidal intent, 11 non-self-injuring prisoners, and 11 college student controls without a history of self-injury or
incarceration.
Following the self-injury imagery, participants indicated their subjective responses to the imagery by completing
visual analogue scales with the following bipolar dimensions: relaxed–tense, relaxed–anxious, calm–angry, unafraid–
afraid, happy–sad, normal–unreal, and relieved–uptight. In addition, as participants visualized the events they wore
electrodes measuring several indices of physiological arousal, including finger blood volume (FBV), finger pulse
amplitude (FPA), heart rate (HR), maximum cardiotachometer (CMAX), minimum cardiotachometer (CMIN),
respiration (RESP), and skin resistance level (SRL).
Six of the seven measures of physiological arousal displayed the predicted pattern for self-injurers'
responses to the self-injury script. Physiological arousal was high while participants imagined the period of
time leading up to self-injury, and reduced substantially when participants imagined self-injuring and the period
of time after self-injury. Subjective measures displayed a similar pattern. Self-injurers reported feeling more
tense, anxious, angry, sad, and uptight before self-injury, and more relaxed, calm, happy, and relieved
afterwards. This pattern differed from that observed for control events (e.g., an interpersonal argument, accident
with a kitchen knife).
Brain, Haines, and Williams (1998) used similar methodology to study 35 self-injurers from two outpatient clinics,
and 35 control participants who were university undergraduates with no history of self-injury. A pattern emerged that
was similar to that observed in Haines et al. (1995). For the self-injurers administered the self-injury imagery, arousal
was high before self-injury and decreased substantially after self-injury. This pattern was not observed when self-
injurers imagined control events, and there were no significant differences between self-injurers' and control
participants' physiological responses to control imagery. Results from both studies provide evidence that self-injuring
reduces negative affect and arousal.
In a second approach to the study of self-injury in the laboratory, Russ et al. (1992) used the cold pressor test as a
proxy for self-injury. The cold pressor test is often used in pain-threshold research and requires participants to immerse
their hands in cold (10 °C) water for up to four minutes. The cold pressor test may simulate self-injury because it
requires participants to intentionally perform an activity that causes acute physical pain, although the external validity
of this task for simulating self-injury is obviously uncertain. The cold pressor test was administered to 17 borderline
patients with a history of self-injury and six control participants without a psychiatric or self-injury history. The
borderline self-injuring group was further split into two groups, those who reported experiencing pain during self-
injury (bpd-pain; n-6) and not experiencing pain during self-injury (Bpd-no pain; n = 11). Immediately before and after
administration of the cold pressor test, all participants completed measures of mood, including depression, anxiety,
E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239 235

anger, confusion, and vigor. The entire procedure was repeated three times on three consecutive days, and mood ratings
were aggregated across the three administrations.
The borderline patients with self-injury histories exhibited reductions in self-reported depression, anxiety, and
confusion following the cold pressor test, although results were only statistically significant for the BPD-no pain group.
The BPD-no pain group also demonstrated a statistically reliable decrease in anger following the cold pressor test.
Interestingly, the control group also exhibited decreases in depression, anxiety, anger, and confusion following the cold
pressor test, although only results for anxiety were statistically reliable. To the extent the cold pressor test is a
satisfactory proxy for self-injury, results are consistent with the affect-regulation model of self-injury, and suggest that
self-injuring is associated with decreases in negative affect. It is intriguing that control participants also exhibited a
trend of reduced negative affect following the cold pressor test. It may be that pain or injury can inhibit the experience
of negative affect regardless of an individual's psychiatric history, but only certain types of people come to rely on self-
injury as a coping mechanism.

3.7. Summary of empirical research on the functions of self-injury

Eighteen studies examining seven functions of self-injury were reviewed. Affect-regulation was the only function
examined in all 18 studies. Eleven studies examined self-punishment, 10 anti-dissociation, 10 interpersonal-
influence, five sensation-seeking, three anti-suicide, and two interpersonal boundaries. Even when taking into
account that affect-regulation was examined most often, findings were strongest for this function. An integration of
results indicates that: (a) acute negative affect precedes self-injury; (b) decreased negative affect and relief are
present after self-injury; (c) most self-injurers identify the desire to alleviate negative affect as a reason for self-
injuring; and (d) the performance of proxies for self-injury in the laboratory leads to reductions in negative affect
and arousal. Several studies also provided strong evidence for a self-punishment function. The anti-dissociation,
interpersonal-influence, sensation-seeking, anti-suicide, and interpersonal boundaries functions received modest
support. Of note, the general pattern of findings regarding the seven functions tended to remain consistent,
regardless of the type of sample (e.g., non-clinical vs. clinical vs. forensic, adult vs. adolescent, outpatient vs.
inpatient, women vs. men).

4. Suggestions for future research

4.1. Explore conceptual and empirical relationships among functions

Although the affect-regulation function of self-injury was apparent in every study reviewed, results suggest that self-
injury also serves at least six other functions. Evidence of multiple functions can be interpreted in several ways, which
are not mutually exclusive: (a) different functions may distinguish different subgroups of self-injurers; (b) multiple
functions for self-injury may exist concurrently within-individuals; (c) functions of self-injury may evolve over time
within-individuals; and (d) different functional models may overlap conceptually and describe different aspects of the
same phenomenon. Future research should assess functions at multiple time points to investigate between- and within-
individual differences in functions for self-injury. In addition, studies should examine the co-variation among functions
to determine which are most related and may be better conceptualized as aspects of the same functional model. Finally,
although multiple functions may co-exist or overlap, some functions may be primary, in that they provide stronger
reinforcement and are more likely to lead to repeated self-injury. Studies should therefore examine which functions and
reinforcing variables best predict the frequency and longevity of self-injury, and therefore might be conceptualized as
providing the most reinforcement for the behavior.

4.2. Measure functions comprehensively

Researchers examining co-variation or other aspects of functions should strive to measure functions
comprehensively. Existing studies have typically relied on ad-hoc measures that assess only a few functions. For
example, in the 11 studies examining reasons for self-injury, nine different measures of reasons were used, and the
modal study examined four of the functions listed in Table 1. An ideal instrument would measure a wide variety of
possible functions, contain multiple items and types of items for each function, and have psychometric properties and a
236 E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239

factor structure that have been examined in large and diverse samples. The development of such a measure in and of
itself would be a welcome contribution to the field.
In addition, future research should utilize methodologies appropriate for evaluating interpersonal in addition to
intrapersonal functions of self-injury. Methodologies used to explore the functions of self-injury have focused more on
affective and physiological variables and less on social and interpersonal variables. Studies should more carefully and
thoroughly assess interpersonal events that tend to precede and prompt self-injury, and obtain relevant information both
from those who self-injure and family members or friends of individuals who self-injure. Informant data may be
particularly useful for clarifying interpersonal functions of self-injury, although it should also be noted that many
individuals conceal their self-injury from those around them.

4.3. Clarify affective-changes associated with self-injury

The changes in affect associated with self-injury should be examined in more detail. Some studies emphasize the
valence of affect-states, and demonstrate that negatively valenced affect-states precede self-injury and diminish
following self-injury. Other studies emphasize affective arousal, and show that arousal peaks before self-injury and
drops afterwards (e.g., Haines et al., 1995). The phenomenological studies conducted so far appear to implicate affect-
states with negative valence and high-arousal, such as anxiety, tension, and anger (Briere & Gil, 1998; Jones et al.,
1979; Kemperman et al., 1997), as opposed to lower-arousal negative affect-states such as sad, lonely, or empty. Future
research should separately examine the changes in affective valence and arousal that are associated with self-injury. In
addition, analyses should identify the changes in valence and arousal that best predict frequency and longitudinal
course of self-injury. This approach will help reveal the affective changes that provide reinforcement for self-injury and
encourage the behavior to become habitual.

4.4. Establish causal link between affective changes and self-injury

Research should address whether the relationship between self-injury and subsequent changes in affect is indeed
causal. Brain et al. (1998), Haines et al. (1995), and Russ et al. (1992) provide intriguing laboratory evidence that
proxies for self-injury cause reductions in negative affect and arousal. Additional designs should be developed that
better approximate the conditions that accompany self-injury episodes. For example, a frustrating task could be
administered before a self-injury proxy to simulate the rising negative affect that typically precedes self-injury. A
quicker return to affective baseline following the self-injury proxy as compared to control tasks would contribute
evidence that self-injury causes reductions in negative affect and arousal. Researchers should include non-self-injurers
in these paradigms as well. If self-injury proxies reduce negative affect in non-self-injurers as well as self-injurers, it
would appear that self-injury diminishes the experience or presence of negative affect and arousal via a common
mechanism (as opposed to a mechanism unique to self-injurers).

4.5. Explore mechanisms of affect-regulation

Research should investigate how self-injury reduces negative affect and arousal. There are numerous hypotheses
regarding psychological and physiological mechanisms. For example, self-injury may distract people from acute
negative affect in a manner similar to how distraction can be used effectively to cope with depression (Just and Alloy,
1997). A theory of affect-regulation based on distraction would have to explain why less damaging and more socially
acceptable methods of distraction like exercise, cold showers, or calling a friend are not sufficient. Self-injury may also
be thought of as a form of self-care. People who self-injure may be unable to soothe themselves when dealing with
intense emotions (Linehan, 1993). The experience of being helpless in the face of negative emotions can be frustrating
and overwhelming. Some individuals may be better able to care for physical than emotional injuries. Self-injuring,
then, may provide a soothing and gratifying opportunity to competently care for oneself by examining, cleaning,
bandaging, or otherwise tending to the wound.
Other theories emphasize biological effects of self-injury, such as the release of endorphins. Endorphins are
endogenous opioids that are released in response to physical injury, act as natural painkillers, and induce pleasant
feelings (Hawkes, 1992; Willer, Dehen, & Cambier, 1981). Some speculate that the endorphin-release following self-
injury may cause the mood effects that alleviate acute negative affect (Favazza & Conterio, 1988; Richardson &
E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239 237

Zaleski, 1986). Notably, at least one study has failed to support an endorphins model of self-injury. The opiate
antagonist naloxone did not diminish improvements in affect following the performance of a self-injury proxy, even
though naloxone should block the effects of endorphins (Russ et al., 1994). Nevertheless, these and other biological
explanations of self-injury should continue to be pursued because understanding the biochemistry of self-injury is
essential for understanding why people self-injure.

4.6. Employ alternatives to retrospective self-report paradigms

Real-time and laboratory methods for studying self-injury should continue to be utilized and enhanced. Most
studies on self-injury have relied on participants' retrospective self-reports. Although these studies provide
valuable information, their validity is debatable because memories about self-injury may be inaccurate,
participants may lack insight into their motivations in the first place, and some individuals may fabricate
explanations they feel are more socially acceptable than their true reasons. It is therefore important to conduct
studies of self-injury that are (a) not dependent on self-report and (b) not retrospective. Some of the laboratory
studies described above, which utilize physiological measures of affective arousal, offer valuable alternatives to
self-report studies. At the same time, the external validity of laboratory studies is limited and such approaches are
best viewed as supplementary.
Ecological momentary assessment (ECA) studies are useful alternatives to retrospective studies. Participants can
keep daily logs of self-injury and potentially related variables including self-injury ideation, suicidal ideation, affect-
states, interpersonal events, and other problem behaviors such as binging, purging, substance use, gambling, impulsive
sex, and impulsive spending. This approach would allow participants to contribute real-time descriptions of the
phenomenology and functions of self-injury, and illuminate other behaviors that may serve similar functions to self-
injury. Some intriguing ECA studies on self-injury and related problems are underway (Rafaeli, Howland, & Drejet,
2005).

4.7. Extend research to community samples

Researchers should continue to examine the functions of self-injury in non-clinical in addition to patient and
forensic samples. Most studies included in the present review examine clinical or forensic populations, which are only
subsets of the larger population of self-injurers. Since about 4% of individuals from large non-clinical populations self-
injure (Briere & Gil, 1998; Klonsky et al., 2003), and rates are even higher in college students (Favazza et al., 1989;
Gratz, 2001; Whitlock et al., 2006), it is likely that some people who self-injure are not in treatment and do not meet
criteria for a psychiatric disorder. Limiting the scope of investigations to patient and prison populations precludes a
complete understanding of self-injury and may obscure the phenomenology, functions, and treatment implications of
self-injury. In addition, research in community settings such as secondary schools may lay the groundwork for the
development of interventions that can be implemented at the community level.

4.8. Investigate clinical implications of different functions

Research should explore how assessing the functions of self-injury in patients can improve case con-
ceptualization and treatment planning. Unfortunately, the clinical lore includes misconceptions about the functions
of self-injury. For example, many clinicians emphasize interpersonal-influence functions (Bancroft & Hawton,
1983; Gough & Hawkins, 2000), even though the present review suggests that self-injury serves this function for a
minority of self-injurers. Research on functions could be used as a basis for identifying sub-types of self-injurers
with different diagnostic profiles and levels of risk for suicide and other undesirable outcomes. For example, Nock
and Prinstein (2005) found that people who self-injure to alleviate acute negative affect feel more hopeless and are
more likely to have attempted suicide than those who identify other reasons for self-injury. This finding
(especially if replicated) suggests that clinicians should assess the functions of their patients' self-injury, and
consider taking extra precautions and choosing more aggressive treatment strategies for patients who self-injure to
regulate negative affect. Additional research on the diagnostic, prognostic, and treatment implications of different
functions would substantially improve the ability of mental health professionals to meet the particular needs of
their patients.
238 E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239

Acknowledgement

I thank Thomas F. Oltmanns, Ph.D., Eric Turkheimer, Ph.D., and Alexis Black, M. A. for comments on earlier
versions of this paper. This work was supported in part by Grant MH67299 from the National Institute of Mental Health
and by the Office of the Vice President of Research at Stony Brook University.

References

Allen, C. (1995). Helping with deliberate self-harm: Some practical guidelines. Journal of Mental Health, 4, 243−250.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders, fourth edition : Washington, D.C text revision.
Bancroft, J., & Hawton, K. (1983). Why people take overdoses: A study of psychiatrists' judgements. British Journal of Medical Psychology, 56,
197−204.
Bennun, I. (1983). Depression and hostility in self-mutilation. Suicide and Life Threatening Behavior, 13, 71−84.
Brain, K. L., Haines, J., & Williams, C. L. (1998). The psychophysiology of self-mutilation: Evidence of tension reduction. Archives of Suicide
Research, 4, 227−242.
Briere, J., & Gil, E. (1998). Self-mutilation in clinical and general population samples: Prevalence, correlates, and functions. American Journal of
Orthopsychiatry, 68, 609−620.
Brown, M. Z., Comtois, K. A., & Linehan, M. M. (2002). Reasons for suicide attempts and nonsuicidal self-injury in women with borderline
personality disorder. Journal of Abnormal Psychology, 111, 198−202.
Carr, E. G. (1977). The motivation of self-injurious behavior: A review of some hypotheses. Psychological Bulletin, 98, 800−816.
Carroll, J., Schaffer, C., Spensley, J., & Abramowitz, S. I. (1980). Family experiences of self-mutilating patients. American Journal of Psychiatry,
137, 852−853.
Chowanec, G. D., Josephson, A. M., Coleman, C., & Davis, H. (1991). Self-harming behavior in incarcerated male delinquent adolescents. Journal
of the American Academy of Child and Adolescent Psychiatry, 30, 202−207.
Coid, J. W. (1993). An affective syndrome in psychopaths with borderline personality disorder? British Journal of Psychiatry, 162, 641−650.
Favazza, A. R. (1992). Repetitive self-mutilation. Psychiatric Annals, 22, 60−63.
Favazza, A. R., & Conterio, K. (1988). The plight of chronic self-mutilators. Community Mental Health Journal, 24, 22−30.
Favazza, A. R., & Conterio, K. (1989). Female habitual self-mutilators. Acta Psychiatrica Scandinavica, 79, 283−289.
Favazza, A. R., DeRosear, L., & Conterio, K. (1989). Self-mutilation and eating disorders. Suicide and Life-Threatening Behavior, 19, 352−361.
Favazza, A. R., & Rosenthal, R. J. (1993). Diagnostic issues in self-mutilation. Hospital and Community Psychiatry, 44, 134−140.
Friedman, M., Glasser, M., Laufer, E., Laufer, M., & Wohl, M. (1972). Attempted suicide and self-mutilation in adolescence: Some observations from
a psychoanalytic research project. International Journal of Psychoanalysis, 53, 179−183.
Gough, K., & Hawkins, A. (2000). Staff attitudes to self-harm and its management in a forensic psychiatric service. The British Journal of Forensic
Practice, 2, 22−28.
Graff, H., & Mallin, R. (1967). The syndrome of the wrist cutter. American Journal of Psychiatry, 124, 36−42.
Gratz, K. L. (2001). Measurement of deliberate self-harm: Preliminary data on the deliberate self-harm inventory. Journal of Psychopathology and
Behavioral Assessment, 23, 253−263.
Gratz, K. L. (2003). Risk factors for and functions of deliberate self-harm: An empirical and conceptual review. Clinical Psychology: Science and
Practice, 10, 192−205.
Haines, J., Williams, C. L., & Brain, K. L. (1995). The psychopathology of incarcerated self-mutilators. Canadian Journal of Psychiatry, 40,
514−522.
Gunderson, J. G. (1984). Borderline personality disorder. Washington, DC: American Psychiatric Press.
Haines, J., Williams, C. L., Brain, K. L., & Wilson, G. V. (1995). The psychophysiology of self-mutilation. Journal of Abnormal Psychology, 104,
471−489.
Haw, C., Hawton, K., Houston, K., & Townsend, E. (2001). Psychiatric and personality disorders in deliberate self-harm patients. British Journal of
Psychiatry, 178, 48−54.
Hawkes, C. H. (1992). Endorphins: The basis of pleasure? Journal of neurology, neurosurgery, and psychiatry, 55, 247−250.
Herpertz, S. (1995). Self-injurious behavior: Psychopathological and nosological characteristics in subtypes of self-injurers. Acta Psychiatrica
Scandinavica, 91, 57−68.
Herpertz, S., Sass, H., & Favazza, A. (1997). Impulsivity in self-mutilative behavior: Psychometric and biological findings. Journal of Psychiatric
Research, 31, 451−465.
Himber, J. (1994). Blood rituals: Self-cutting in female psychiatric inpatients. Psychotherapy, 31, 620−631.
Huband, N., & Tantam, D. (1999). Clinical management of women who self-wound: A survey of mental health professionals' preferred strategies.
Journal of Mental Health, 8, 473−487.
Jones, I. H., Congiu, L., Stevenson, J., Strauss, N., & Frei, D. Z. (1979). A biological approach to two forms of self-injury. Journal of Nervous and
Mental Disease, 167, 74−78.
Just, N., & Alloy, L. B. (1997). The response styles theory of depression: Tests and an extension of the theory. Journal of Abnormal Psychology, 106,
221−229.
Kemperman, I., Russ, J. J., & Shearin, E. (1997). Self-injurious behavior and mood regulation in borderline patients. Journal of Personality
Disorders, 11, 146−157.
Klonsky, E.D. (2006). The functions of non-suicidal skin-cutting in college students. Manuscript under editorial review.
E.D. Klonsky / Clinical Psychology Review 27 (2007) 226–239 239

Klonsky, E. D., Oltmanns, T. F., & Turkheimer, E. (2003). Deliberate self-harm in a nonclinical population. Prevalence and psychological correlates.
American Journal of Psychiatry, 160, 1501−1508.
Kumar, G., Pepe, D., & Steer, R. A. (2004). Adolescent psychiatric inpatients' self-reported reasons for cutting themselves. Journal of Nervous and
Mental Disease, 192, 830−836.
Langbehn, D. R., & Pfohl, B. (1993). Clinical correlates of self-mutilation among psychiatric inpatients. Annals of Clinical Psychiatry, 5, 45−51.
Larsen, R. J. (2000). Toward a science of mood regulation. Psychological Inquiry, 11, 129−141.
Laye-Gindhu, A., & Schonert-Reichl, K. A. (2005). Nonsuicidal self-harm among community adolescents: Understanding the “whats” and “whys” of
self-harm. Journal of Youth and Adolescence, 34, 447−457.
Lester, D. (1972). Self-mutilating behavior. Psychological Bulletin, 78, 119−128.
Linehan, M. M. (1993). Cognitive-behavioral treatment for borderline personality disorder. New York: Guildford Press.
Linehan, M. M. (2000). The empirical basis of Dialectical Behavior Therapy: Development of new treatments versus evaluation of existing
treatments. Clinical Psychology: Science and Practice, 7, 113−119.
Linehan, M. M., Comtois, K. A., Murray, A. M., et al. (2006). Two-year randomized controlled trial and follow-up of dialectical behavior therapy vs
treatment by experts for suicidal behaviors and borderline personality disorder. Archives of General Psychiatry, 63, 757−766.
Lloyd, E., Kelley, M. L., & Hope, T. (1997). Self-mutilation in a community sample of adolescents: Descriptive characteristics and provisional
prevalence rates. Paper presented at the Annual meeting of the Society for Behavioral Medicine, New Orleans, LA.
Muehlenkamp, J. J. (2005). Self-injurious behavior as a separate clinical syndrome. American Journal of Orthopsychiatry, 75, 324−333.
Nijman, H. L. I., Dautzenberg, M., Merckelbach, H. L. G. J., Jung, P., Wessel, I., & Campo, J. (1999). Self-mutilating behavior in psychiatric
inpatients. European Psychiatry, 14, 4−10.
Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84,
231−259.
Nixon, M. K., Cloutier, P. F., & Aggarwal, S. (2002). Affect regulation and addictive aspects of repetitive self-injury in hospitalized adolescents.
Journal of the American Academy of Child and Adolescent Psychiatry, 41, 1333−1341.
Nock, M. K., & Prinstein, M. J. (2004). A functional approach to the assessment of self-mutilative behavior. Journal of Consulting and Clinical
Psychology, 72, 885−890.
Nock, M. K., & Prinstein, M. J. (2005). Contextual features and behavioral functions of self-mutilation among adolescents. Journal of Abnormal
Psychology.
Ogundipe, L. O. (1999). Suicide attempts vs. deliberate self-harm. British Journal of Psychiatry, 175, 90.
Osuch, E. A., Noll, J. G., & Putnam, F. W. (1999). The motivations for self-injury in psychiatric patients. Psychiatry, 62, 334−346.
Pattison, E. M., & Kahan, J. (1983). The deliberate self-harm syndrome. American Journal of Psychiatry, 140, 867−872.
Penn, J. V., Esposito, C. L., Schaeffer, L. E., Fritz, G. K., & Spirito, A. (2003). Suicide attempts and self-mutilative behavior in a juvenile correctional
facility. Journal of the American Academy of Child and Adolescent Psychiatry, 42, 762−769.
Podovall, E. M. (1969). Self-mutilation within a hospital setting: A study of identity and social compliance. British Journal of Medical Psychology,
42, 213−221.
Rafaeli, E., Howland, M., Drejet, R. (2005). Using diary methods in the study of psychopathology. Paper presented at the 39th annual convention of
the Association for Behavioral and Cognitive Therapies, Washington, DC, November 20.
Richardson, J. S., & Zaleski, W. A. (1986). Endogenous opiates and self-mutilation. American Journal of Psychiatry, 140, 867−872.
Robins, C. J., & Chapman, A. L. (2004). Dialectical behavior therapy: Current status, recent developments, and future directions. Journal of
Personality Disorders, 18, 73−89.
Russ, M. J., et al. (1992). Pain perception in self-injurious patients with borderline personality disorder. Biological Psychiatry, 32, 501−511.
Russ, M. J., Roth, S. D., Kakuma, T., Harrison, K., & Hull, J. W. (1994). Pain perception in self-injurious borderline patients: Naloxone effects.
Biological Psychiatry, 35, 207−209.
Shearer, S. L. (1994). Phenomenology of self-injury among inpatient women with borderline personality disorder. The Journal of Nervous and
Mental Disease, 182, 524−526.
Simeon, D., Stanley, B., Frances, A., Mann, J. J., Winchel, R., & Stanley, M. (1992). Self-mutilation in personality disorders: Psychological and
biological correlates. American Journal of Psychiatry, 149, 221−226.
Simpson, M. A. (1975). The phenomenology of self-mutilation in a general hospital setting. Canadian Psychiatric Association Journal, 20, 429−434.
Skegg, K. (2005). Self-harm. The Lancet, 366, 1471−1483.
Soloff, P. H., Lis, J. A., Kelly, T., Cornelius, J., & Ulrich, R. (1994). Self-mutilation and suicidal behavior in borderline personality disorder. Journal
of Personality Disorders, 8, 257−267.
Stanley, B., Gameroff, M. J., Michalsen, V., & Mann, J. J. (2001). Are suicide attempters who self-mutilate a unique population? American Journal of
Psychiatry, 158, 427−432.
Suyemoto, K. L. (1998). The functions of self-mutilation. Clinical Psychology Review, 18, 531−554.
Tantam, D., & Whittaker, J. (1992). Personality disorder and self-wounding. British Journal of Psychiatry, 161, 451−464.
van der Kolk, B. A., Perry, J. C., & Herman, J. L. (1991). Childhood origins of self-destructive behavior. American Journal of Psychiatry, 148,
1665−1672.
Whitlock, J., Eckenrode, J., & Silverman, D. (2006). Self-injurious behavior in a college population. Pediatrics, 117, 1939−1948.
Wilkins, J., & Coid, J. (1991). Self-mutilation in female remanded prisoners: I. An indicator of severe psychopathology. Criminal Behaviour and
Mental Health, 1, 247−267.
Willer, J. D., Dehen, H., & Cambier, J. (1981). Stress induced analgesia in humans. Science, 212, 680−691.
Zlotnick, C., Mattia, J. I., & Zimmerman, M. (1999). Clinical correlates of self-mutilation in a sample of general psychiatric patients. Journal of
Nervous and Mental Disease, 187, 296−301.