Content uploaded by Clara Strauss
Author content
All content in this area was uploaded by Clara Strauss on Jun 06, 2016
Content may be subject to copyright.
Review
What is compassion and how can we measure it? A review of definitions
and measures
Clara Strauss
a,b,
⁎, Billie Lever Taylor
b,c
,JennyGu
a
, Willem Kuyken
d
, Ruth Baer
d,e
,
Fergal Jones
f
, &, Kate Cavanagh
a
a
School of Psychology, University of Sussex, Falmer, East Sussex BN1 9QH, UK
b
Sussex Partnership NHS Foundation Trust, Hove BN3 7HZ, UK
c
Division of Psychiatry, University College London, 149 Tottenham Court Road, London W1T 7NF, UK
d
Department of Psychiatry, University of Oxford, Warneford Hospital, Headington, Oxford OX3 7JX, UK
e
Department of Psychology, University of Kentucky, Lexington, KY 40506-0044, US
f
Salomons Centre for Applied Psychology, Canterbury Christ Church University, Runcie Court, Broomhill Road, TunbridgeWells TN3 0TF, UK
HIGHLIGHTS
•Compassion is recognized as important across many sectors of society.
•There is lack of consensus on definition and few self/observer-rated measures exist.
•Five elements of compassion are proposed after consolidating existing definitions.
•The psychometric properties of measures could be improved, limiting their utility.
•A new measure of compassion with robust psychometric properties is needed.
abstractarticle info
Article history:
Received 19 January 2016
Received in revised form 27 April 2016
Accepted 24 May 2016
Available online 26 May 2016
The importance of compassion is widely recognized and it is receiving increasing research attention. Yet, there is
lack of consensus on definition and a paucity of psychometrically robust measures of this construct. Without an
agreed definition and adequate measures, we cannotstudy compassion,measure compassionor evaluate wheth-
er interventions designed to enhance compassion are effective. In response, this paper proposes a definition of
compassion and offers a systematic review of self- and observer-rated measures. Following consolidation of
existing definitions, we propose that compassion consists of five elements: recognizing suffering, understanding
the universality of human suffering, feeling for the person suffering, tolerating uncomfortable feelings, and mo-
tivation to act/acting to alleviate suffering. Three databases were searched (Web of Science, PsycInfo, and
Medline) and nine measures included and rated for quality. Quality ratings ranged from 2 to 7 out of 14 with
low ratings due to poor internal consistency for subscales, insufficient evidence for factor structure and/or failure
to examine floor/ceiling effects, test–retest reliability, and discriminant validity. We call our five-element defini-
tion, and if supported, the development of a measure of compassion based on this operational definition, and
which demonstrates adequate psychometric properties.
© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Keywords:
Compassion
Self-compassion
Measure
Systematic review
Definition
Contents
1. Introduction............................................................... 16
1.1. Conceptualizations of compassion: towards a definition........................................ 16
1.2. Relatedconstructs......................................................... 18
1.3. Compassion: a proposed definition ................................................. 19
Clinical Psychology Review 47 (2016) 15–27
⁎Corresponding author at: School of Psychology, University of Sussex, Falmer, Brighton BN1 9QJ, UK.
E-mail address: c.y.strauss@sussex.ac.uk (C. Strauss).
http://dx.doi.org/10.1016/j.cpr.2016.05.004
0272-7358/© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Contents lists available at ScienceDirect
Clinical Psychology Review
journal homepage: www.elsevier.com/locate/clinpsychrev
2. Method ................................................................. 19
2.1. Inclusionandexclusioncriteria ................................................... 19
2.2. Informationsources......................................................... 19
2.3. Searchstrategy...........................................................19
2.4. Assessmentofquality........................................................19
3. Results.................................................................. 20
3.1. Review of identifiedmeasures....................................................20
3.1.1. Compassionatelovescale(CLS;Sprecher&Fehr,2005) .................................... 21
3.1.2. SantaClarabriefcompassionscale(SCBCS;Hwang,Plante,&Lackey,2008)........................... 22
3.1.3. Thecompassionscale(CS-M;Martins,Nicholas,Shaheen,Jones,&Norris,2013).........................23
3.1.4. Self-compassionscale(SCS;Neff,2003b)...........................................23
3.1.5. Self-compassionscale:shortform(SCS-SF;Raes,Pommier,Neff,&VanGucht,2011).......................23
3.1.6. Thecompassionscale(CS-P;Pommier,2010)......................................... 24
3.1.7. Relationalcompassionscale(RCS;Hacker,2008)....................................... 24
3.1.8. Compassionatecareassessmenttool(CCAT;Burnell&Agan,2013)...............................24
3.1.9. TheSchwartzcentercompassionatecarescale(SCCCS;Lown,Muncer,&Chadwick,2015)..................... 25
4. Discussion................................................................25
4.1. Strengthsandlimitations......................................................26
4.2. Futureresearch...........................................................26
4.3. Conclusion.............................................................26
References................................................................... 27
1. Introduction
The importance of compassion is recognized in many segments of
society. Most of the world's religious traditions place compassion at
the center of their belief systems. International professional bodies in
healthcare, education and the justice system also emphasize the impor-
tance of compassion. In the US, compassion is enshrined in the
American Medical Association's (AMA) Principles of Medical Ethics,
with Item 1 stating that “A physician shall be dedicated to providing
competent medical services with compassion and respect for human
dignity”(AMA, 1981). In the UK, compassion is one of the six core values
in the NHS constitution (Department of Health; DoH, 2013), and calls
for a greater focus on compassion have been driven in part by high pro-
file exposés of serious failings in compassionate care at some hospitals
and care homes. The international ‘Compassion in Education’founda-
tion (CoED, 2014) offers a range of services to educational professionals
in order to promote compassion in the education system. It has also
been argued that compassion should lie at thecore of the ethical frame-
work guiding our justice systems (Norko, 2005).
An evolutionary perspective on compassion can be traced to
Darwin (1871), who stated that “those communities which included
the greatest number of the most sympathetic members would flourish
best, and rear the greatest number of offspring”(p. 130). Current the-
orists also note that compassion is reproductively advantageous,
being part of the care-giving system that has evolved to nurture and
protect the young (e.g. Gilbert, 2005; Goetz, Keltner, & Simon-
Thomas, 2010). Compassion can be seen as having evolved from an
adaptive focus on protecting oneself and one's offspring to a broader
focus on protecting others including and beyond one's immediate kin-
ship group (de Waal, 2009). Compassion may also have evolved in pri-
mates because it is a desirable criterion in mate selection and
facilitates cooperative relationships with non-kin (e.g. de Waal,
2009; Keltner, 2009).
Within the healthcare domain, compassion is believed to have nu-
merous practical advantages. It has been argued that treating patients
compassionately has wide-ranging benefits, including improving clini-
cal outcomes, increasing patient satisfaction with services, and enhanc-
ing the quality of information gathered from patients (Epstein et al.,
2005; Rendelmeir et al., 1995; Sanghavi, 2006). Conversely, compassion
fatigue may contribute to poorquality of care (Najjar, Davis, Beck-Coon,
& Carney Doebbeling, 2009). Treating oneself and others with compas-
sion is also believed to promote individual wellbeing and improve
mental health (e.g. Cosley, McCoy, Saslow, & Epel, 2010; Feldman &
Kuyken, 2011; MacBeth & Gumley, 2012). Accordingly, some re-
searchers have called for the implementation of interventions that
seek to enhance people's ability to give and receive compassion (e.g.
Gilbert,2005, 2010), arguing that compassion buffers reactivity to stress
and is central to the process of recovery from psychopathology. Other
research has focused on the developmental trajectory of compassion
and has found relationships between parenting styles and children's
levels of sympathy and caring (Eisenberg, VanSchyndel, & Hofer,
2015) and between attachment security in childhood and capacity for
compassion in adulthood (see Gillath, Shaver, & Mikulincer, 2005, for
areview).
Despite the importance of compassion and increasing interest from
researchers, clinicians, teachers, and other professionals, there is lack
of consensus on its definition and a paucity of psychometrically robust
measurement tools.Without these, scientificenquiry is greatlyimpeded
—we need consensus on a definition and valid and reliable measure-
ment tools in order to assess compassion in empirical research. This
paper has two aims: first, to suggest a definition of compassion based
on a consolidation of conceptualizations and definitions in the field
and second, to systematically review self-and observer-rated measures
of compassion.
1.1. Conceptualizations of compassion: towards a definition
According to the Oxford English Dictionary, the word “compassion”
stems from the Latin “compati”,meaning“to suffer with”. In the litera-
ture, there appears to be a broad consensus that compassion involves
feeling for a person who is suffering and being motivated to act to
help them (e.g. Goetz et al., 2010; Lazarus, 1991). For example, in his
seminal work on human emotions Lazarus defines compassion as:
“Being moved by another's suffering and wanting to help”(p. 289).
Similarly, in a major systematic review of compassion and its evolution-
ary origins, Goetz et al. define it as: “the feeling that arises in witnessing
another's suffering and that motivates a subsequent desire to help”
(p. 351). These definitions havein common the suggestion that compas-
sion is not only about feeling touched by a person's suffering, but also
about wanting to act to help them. Compassion is a fundamental tenet
of Buddhist philosophy (it is, in fact, emphasized by all the main
world religions but Buddhist perspectives on compassion have been
given greater prominence in the psychological literature) and the
Dalai Lama (1995) defines compassion in comparable terms as: “An
16 C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
openness to the suffering of others with a commitment to relieve it”.
However, within Buddhism, compassion is seen not only as an emotion-
al response but also as a responsefounded on reason and wisdom which
is embedded in an ethical framework concerned with the selfless
intention of freeing others from suffering.
More specifically, in their review of compassion within organiza-
tions, Kanov et al. (2004) argue that compassion consists of three facets:
noticing, feeling, and responding. ‘Noticing’involves being aware of a
person's suffering, either by cognitively recognizing this suffering or
by experiencing an unconscious physical or affective reaction to it. ‘Feel-
ing’is defined as responding emotionally to that suffering and
experiencing ‘empathic concern’through adopting the person's per-
spective and imagining or feeling their condition. Finally, ‘responding’
involves having a desire to act to alleviate the person's suffering. As in
Buddhist conceptualizations, this definition suggests that compassion
does not purely consist of affective and behavioral elements, but also
may have cognitive components insofar as it involves being able to
imagine and reason about a person's experiences.
Gilbert (2010) conceptualizes compassion in evolutionary terms,
arguing that compassion is an evolved motivational system designed
to regulate negative affect, where compassion is seen to have originated
from the same capacities that primates evolved to form attachment
bonds and engage in affiliative and cooperative behaviors for group
survival. He defines compassion as: “A deep awareness of the suffering
of another coupled with the wish to relieve it”(Gilbert, 2009, p. 13)
and, like Kanov et al. (2004), suggests it has cognitive, affective and
behavioral elements. Gilbert (2010) sees compassion as consisting of
six ‘attributes’: sensitivity, sympathy, empathy, motivation/caring,
distress tolerance and non-judgement. ‘Sensitivity’involves being
responsive to other people's emotions and perceiving when they need
help, which appears to correspond to Kanov et al.'s ‘noticing’facet.
‘Sympathy’(defined as showing concern for the other person's suffer-
ing) and ‘empathy’(defined as putting yourself in their shoes) together
appear to correspond to Kanov et al.'s ‘feeling’facet. Finally, ‘motivation’
to act is akin to Kanov et al.'s ‘responding’facet.
The final two components in Gilbert's (2010) model –‘distress toler-
ance’and ‘non-judgement’–are not included in Kanov et al.'s (2004)
model. Distress tolerance is defined as the ability to tolerate difficult
emotions in oneself when confronted with someone else's suffering
without becoming overwhelmed by them. Gilbert argues that this is im-
portant because if we over-identify with a person's suffering we may
feel a need to get away from them or reduce our awareness of their
distress, preventing a compassionate response. This suggests that,
although compassion is about ‘suffering with’another person, if we
feel such extreme personal distress in the face of another's suffering
that we become too focused on our own discomfort, this may hinder
our ability to help. The final element of Gilbert's model –‘non-
judgement’–is defined as the ability to remain acceptingof and tolerant
towards another person even when their condition, or response to it,
gives rise to difficult feelings in oneself, such as frustration, anger, fear
or disgust. The idea that compassion means approaching those who
are suffering with non-judgement and tolerance –even if they are in
some sense disagreeable to us –is also central to Buddhist conceptuali-
zations. For example, the Dalai Lama (2002) contends that: “for a
practitioner of love and compassion, an enemy is one of the most
important teachers. Without an enemy you cannot practice tolerance,
and without tolerance you cannot build a sound basis of compassion”
(p. 75).
Both Gilbert (2005, 2010) and the Dalai Lama are also clear that
compassion is not only felt for close others (where attachment comes
into play as well), but also for those we do not know. Similarly, Gilbert
(2003, cited in Wang, 2005) notes: “One can feel compassion for those
we might never meet (the starving children in Africa)”(p. 99–100).
The idea that compassion can be experienced towards close others
and those we do not know is also emphasized by Sprecher and Fehr
(2005) who developed a measure of ‘compassionate love’which
includes separate versions relating to close others and strangers or hu-
mankind at large.
Like Gilbert (2010);Wispe (1991) conceptualizes compassion for
others not only as being aware of and moved by suffering and wanting
to help, but also as including the ability to adopt a non-judgmental
stance towards others and to tolerate one's own distress when faced
with other people's suffering. Neff (2003a) developed this definition
of compassion for others into a model of self-compassion, arguing that
self-compassion can be viewed as compassion directed inward towards
the self. She concludes that self-compassion consists of three principal
components: kindness (being kind and non-judgmental towards the
self rather than self-critical), mindfulness (which, like ‘distress toler-
ance’, involves holding painful feelings in mindful awareness rather
than over-identifying with them), and common humanity (seeing
one's suffering as part of the human condition rather than as isolating).
It is debatable whether compassion for others and self-compassion
are in fact part of the same overarching construct. While Buddhistthink-
ing argues that differentiating compassion for others from self-
compassion means drawing a false distinction between the self and
others, and moreover that self-compassion is a prerequisite for showing
‘true’compassion towards others, recent research has found that associ-
ations between self-compassion and other-focused compassion may be
weak, or even non-existent for some populations. For example, Neff and
Pommier (2013) explored the relationship between self-compassion
and compassion for others and found that they were not correlated in
a sample of undergraduates (r= 0.00), and only weakly correlated in
a community sample and a sample of practicing meditators (r= 0.15
and 0.28 respectively). Similarly, Pommier (2010) found no association
between self-compassion and compassion for others in a sample of
undergraduates (r= 0.07). It is unclear whether the lack of association
between self-compassion and compassion for others reflects a genuine
independence between these two constructs or whether it reflects
definitional problems, weaknesses of correlational study designs or lim-
itations with current measures (e.g. Williams, Dalgleish, Karl, & Kuyken,
2014). This is an area for further empirical research.
While acknowledging some of the difficulties with equating self-
compassion with compassion for others, Pommier (2010) has applied
Neff's (2003a) model of self-compassion to a model of compassion for
others suggesting that, like self-compassion, compassion for others
can be seen as involving kindness, mindfulnessand common humanity.
In Pommier's model, ‘kindness’is defined as being understanding to-
wards others who are suffering insteadof being critical or indifferent to-
wards them. ‘Mindfulness’is seen as the ability to notice another
person's suffering and remain open to it without feeling so distressed
that you disengage from that person. ‘Common humanity’is conceptu-
alized as realizing that all humans suffer and that one could find oneself
in the position of the sufferer if one was less fortunate —a sense that
“There but for the grace of God, go I”.
This emphasis on seeing a ‘common humanity’with the person who
is suffering is also evident in Buddhist definitions of compassion, with
the Dalai Lama (2005) arguing that: “Genuine compassion must have
both wisdom and loving kindness. That is to say, one must understand
the nature of the suffering from which we wish to free others (this is
wisdom), and one must experience deep intimacy and empathy with
other sentient beings (this is loving kindness)”(p. 49). Within such
Buddhist conceptualizations, understanding the nature of suffering
(‘wisdom’) is to understand that suffering is part of what it is to be
human; suffering is a shared human experience. Similarly, in their re-
view of the role of compassion in mindfulness-based therapies,
Feldman and Kuyken (2011) describe compassion as: “an orientation
of mind that recognizes pain and the universality of pain in human ex-
perience and the capacity to meet that pain with kindness, empathy,
equanimity and patience”(p. 145).
In summary, in all these definitions compassion is seen as awareness
of someone's suffering, being moved by it (emotionally and, according
to some definitions, cognitively), and acting or feeling motivated to
17C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
help. Several definitions emphasize that, although one is moved by suf-
fering, compassion also involves being able to tolerate uncomfortable
feelings that arise in oneself as a result of seeing suffering, including
tolerating feelings of distaste, frustration or angerthat might be elicited
by that suffering. There is also a suggestion in several definitions that
compassion involves recognizing a commonality with the sufferer,
acknowledging that as a fellow being we too could find ourselves in a
similar position. Table 1 contains summaries of the major definitions
of compassion discussed in this section.
1.2. Related constructs
In definitions of compassion, reference is commonly made to related
terms such as empathy and in turn these words are often used to define
each other. The similarities between compassion and constructs such as
kindness, pity and altruism have also been noted (Goetz et al., 2010). It
is instructive to consider the overlaps and distinctions between these
terms.
Accordingto the Oxford English Dictionary, the word empathy is de-
fined as: “the power of mentally identifying oneself with (and so fully
comprehending) a person or object of contemplation”. Like compassion,
empathy has been described as a multidimensional construct,
consisting of cognitive and affective components (Davis, 1983).
Cognitive empathy can be defined as intellectually understanding
another person's emotions and perspective (Hogan, 1969), whereas
affective empathy refers to being affected by and sharing another's
emotions (Mehrabian & Epstein, 1972). Gilbert (2010);Kanov et al.
(2006) and the Dalai Lama (2005) all explicitly define compassion as
requiring empathy and therefore appear to see empathy as an essential
element ofcompassion. Even so, they suggest that compassion has addi-
tional components over and above empathy. In particular, a desire to act
or acting to alleviatesuffering is seen as a corefeature of compassion but
not empathy (see Table 1).
A second distinction between compassion and empathy is that,
whereas compassion is felt specifically in response to suffering, empa-
thy may apply to a broader range of situations, for example one could
feel empathy with someone else's anger, fear or joy (Pommier, 2010).
Moreover, Goetz et al. (2010) argue that compassion is an emotion in
its own right, whereas empathy is the vicarious experience of another's
emotions, while Sprecher and Fehr (2005) contend that compassion is
broader than empathybecause it can be felt for humanity at large, rath-
er than only in relation tospecific interpersonal encounters. In addition,
recent neuroscientificfindings suggest that different brain regions are
activated in response to compassion and empathy training (Klimecki,
Leiberg, Ricard, & Singer, 2014).
The same is true of pity, which, despite also having similarities to
compassion, does not require an inclination to help. On the contrary,
some writers have argued that pity implies that one sees someone as
unworthy of help (Lazarus, 1991), or at least involves showing conde-
scension towards them (e.g. Cassell, 2002). At the other end of the spec-
trum, altruism has a greater focus than compassion on behavioral acts
that may be at a great personal cost to the person. Also, altruistic acts
can have a broad range of motivations, that do not necessarily involve
the same elements as compassion.
Finally, compassion is frequently linked to kindness (defined by the
Oxford English Dictionary as “the quality of being friendly, generous and
considerate”). For example, Neff (2003a) and Pommier (2010) include
‘kindness’as a component of compassion and compassion has even
been defined as “intelligent kindness”(e.g. DoH, 2013). However,
these two terms have distinctions too: for example, as outlined compas-
sion includes elements beyond kindness (e.g. recognizing and being
touched bysuffering); and likewise, kindness includes elements beyond
Table 1
Major definitions of compassion in the literature in relation to the five-element definition of compassion.
Definition of compassion
Recognizing
suffering
a
Understanding
the
universality
of suffering
Emotional
resonance
Tolerating
uncomfortable
feelings
Motivation
to
act/acting
to
alleviate
suffering
1. “Being moved by another's suffering and wanting to help”(Lazarus, 1991, p. 289). ✓(implied) ✓✓
2. An openness to the suffering of others with a commitment to relieve it (Dalai Lama, 1995).
Buddhist conceptualizations also highlight cognitive components (e.g. the ability to imagine
and reason about a person's experiences) and approaching those who are suffering with
tolerance and non-judgement.
✓(implied) ✓✓✓ ✓
3. “Being touched by the suffering of others, opening one's awareness to others' pain and
not avoiding or disconnecting from it, so that feelings of kindness towards others and the desire to
alleviate their suffering emerge. It also involves offering non-judgmental understanding to those who
fail or do wrong”(Neff, 2003a,p.86–87).
✓
(explicitly
stated)
✓✓ ✓
4. Compassion consists of three facets: Noticing, feeling, and responding (Kanov et al.,
2006).
✓
(explicitly
stated)
✓✓
5. “A deep awareness of the suffering of another coupled with the wish to relieve it”
(Gilbert, 2009, p. 1 3). Compassion consists of six ‘attributes’: Sensitivity, Sympathy,
Empathy, Motivation/Caring,
Distress Tolerance, and Non-Judgement.
✓
(explicitly
stated)
✓✓ ✓
6. “The feeling that arises in witnessing another's suffering and that motivates a subsequent
desire to help”(Goetz et al., 2010, p. 351).
✓
(explicitly
stated)
✓✓
7. “An orientation of mind that recognises pain and the universality of pain in human
experience and the capacity to meet that pain with kindness, empathy, equanimity and
patience”(Feldman & Kuyken, 2011, p. 145).
✓
(explicitly
stated)
✓✓✓
8. Compassion involves three elements: Kindness, mindfulness, and common humanity
(Pommier, 2010).
✓(implied) ✓✓✓ ✓
a
Some definitions of compassion explicitly include an element of ‘recognizing suffering’, whereas in others, this is implied. We have indicated whether ‘recognizing suffering’is
explicitly stated or implied in the following way: ✓(explicitly stated) and ✓(implied).
18 C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
compassion, as kindness is not only linked to suffering (e.g. remember-
ing someone's birthday is kind but not compassionate). Additionally,
compassion may not always involve kindness in the moment (e.g. tak-
ing a ‘tough love’approach may be compassionate but not kind).
1.3. Compassion: a proposed definition
To bring together the various definitions and considerations above
and to aid the review of existing measures of compassion, we propose
anewdefinition of compassion as a cognitive, affective, and behavioral
process consisting of the following five elements that refer to both self-
and other-compassion: 1) Recognizing suffering; 2) Understanding the
universality of suffering in human experience; 3) Feeling empathy for
the person suffering and connecting with the distress (emotional reso-
nance); 4) Tolerating uncomfortable feelings aroused in response to
the suffering person (e.g. distress, anger, fear) so remaining open to
and accepting of the person suffering; and 5) Motivation to act/acting
to alleviate suffering.
We use this proposed new definition of compassion to organize the
remainder of this paper, which provides a systematic review of self- and
observer-rated measures of compassion. The psychometric properties
of identified measures are reported and rated for quality, including the
extent to which they measure each ofthe five elements outlined above.
2. Method
2.1. Inclusion and exclusion criteria
To be included in the main review, measures had to: be available in
English; include a scale explicitly defined by its authors as measuring
compassion; include a psychometric paper outlining the development
of the scale; and be obtainable either within a published article or
from the author (two attempts were made to contact the relevant au-
thors to obtain measures where necessary). Measures were excluded
if they did not assess participants' levels of compassion per se (e.g. mea-
sures of barriers to feeling compassion, fear of compassion, and empa-
thy were excluded); used non-questionnaire measures of compassion,
or included only a subscale on compassion. Because we do not yet
know the relationship between compassion for others and self-
compassion, measures of self-compassion were included because
many conceptualizations and definitions of compassion do not
distinguish between other- and self-compassion.
2.2. Information sources
The databases searched for relevant measures included Web of Sci-
ence (Thomson Reuters), PsycInfo, and Medline, from inception to 23
September 2015. Dissertations and theses that met the inclusion criteria
were reviewed along with papers published in peer-reviewed journals.
Where relevant, the most recent versions of measures were reviewed.
2.3. Search strategy
All articles including the word “compassion*”in combination with
“measure*”,“scale*”,“instrument*”or “questionnaire*”in either the
title or abstract or key words were identified. Where identified papers
referredto additional scales, reference lists were searched and any addi-
tional relevant papers retrieved. Experts in the field were also consulted
to ensure that no measures were missed.
2.4. Assessment of quality
The psychometric properties of each measure were reviewed and
measures were rated for quality based largely on Terwee et al.'s
(2007) quality criteria for health status measures. These criteria were
used because they include explicit criteria for what constitutes good
measurement properties. However, since these criteria relate to mea-
sures of health status, Barker, Pistrang, and Elliott's (2002) ‘rules of
thumb’for evaluating psychological measures were also drawn on
where these seemed more appropriate.
1
Terwee et al. award measures
a positive (+), intermediate (?), or negative (−) rating, or a rating of
0 where no information regarding the relevant criteria is provided. In
this review, in order to make scores easier to interpret, measures were
given a score of 2 if there was evidence for the criterion being fully
met, 1 if the criterion was partially met, and 0 if the criterion was not
met, or if no relevant data were reported. Scores were aggregated to
provide an overall rating. Two researchers independently scored the
measures using these criteria, and any discrepancies in scoring were re-
solved collectively. Specifically, measures were rated across the follow-
ing domains:
1. Content validity (the extent to which the domain of interest was
comprehensively sampled by the items in the questionnaire). In
this case the domain of interest was considered to be compassion
as defined in this review, rather than as defined by the scale's au-
thors. Under this criterion, Terwee et al. (2007) also emphasize the
importance of both members of the target population and experts
being involved in item development. For this criterion to be fully
met all five elements of compassion must be captured by the items
and items must have been generated in consultation with experts
and members of the intended population.
2. Factor structure (whether or not the factor structure for the measure
has been examined and supported). This criterion was included in
addition to those proposed by Terwee et al. (2007). This criterion
was scored as follows. A score of 2 was given where exploratory fac-
tor analysis (EFA) followed by confirmatory factor analysis (CFA)
have been conducted in independent samples OR where CFA has
been conducted if the factor structure has been previously proposed
theoretically (a score of 2 was only given if the factor analyses sup-
port the proposed factor structure). A score of 1 was given if only
EFA has been conducted (without CFA) and if the EFA supports the
factor structure. A score of 0 was given where either factor analysis
has not been conducted OR where EFA and/or CFA have been con-
ducted that do not support the proposed factor structure.
3. Internal consistency (the extent to which items in a (sub)scale are
inter-correlated and thus measuring thesame construct).For this cri-
terion to be fully met –in line with Terwee et al.'s (2007) criteria –
factor analyses had to have been performed on an adequate sample
size (7 ∗number of items and NN100) and Cronbach's alpha for
each identified factor had to be between 0.70 and 0.95.
4. Test–retest reliability. Based on Barker et al.'s (2002) ‘rules of thumb’
test–retest reliabilities had to be at least r= 0.70 for this criterion to
be fully met.
5. Convergent and discriminant validity (the extent to which scores on
a particular scale relate to other measures in a manner consistent
with theoretically derived hypotheses). For this criterion to be met,
Terwee et al. (2007) require that (i) specific hypotheses are formu-
lated by the scale's authors about expected correlations and (ii) at
least three quarters of results are in line withexpectations. AsTerwee
et al. do not take into account the strength of these correlations, we
1
Terwee et al. (2007) proposed the following eight quality criteria to evaluate health
status measures: 1) Content validity, 2) internal consistency, 3) criterion validity, 4) con-
struct validity (convergent and discriminant validity), 5) reproducibility (test-retest reli-
ability), 6) responsiveness, 7) floor and ceil ing effects, and 8) interpretability. We did
not include criterion validity and responsiveness as criteria, for the reasons stated in the
paper. Terwee et al. did not provide rules of thumb in terms of the size of correlation co-
efficients for the test-retest reliability criterion. They also did not account for the size of
correlations for the convergent and discriminant validity criterion. Therefore, for these
two criteria, we drew on Barker et al.'s (2002) general recommendations when evaluating
the reliability and validity of psychological measures. We also includedfactor structure as
acriterion.
19C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
also drew on Barker et al. (2002), and required that at least two cor-
relations with theoretically related constructs were at least r=0.50
to demonstrate convergent validity.
6. Floor and ceiling effects (the number of respondents achieving the
highest or lowest possible scores). This was rated based on Terwee
et al.'s (2007) criterion that no more than 15% of the sample should
receive the top or bottom score on a scale.
7. Interpretability (how differences in scores on the measure can be
interpreted, or the degree to which qualitative meaning can be at-
tached to quantitative scores). Terwee et al. (2007) require means
and SDs of scores from at least four relevant subgroups of partici-
pants to be reported (e.g. compassion scores in males vs. females,
meditators vs. non-meditators) and minimal important change de-
fined. However, as minimal important change was arguably not en-
tirely relevant to the measures in this review, consideration was
instead given to whether the authors indicated how scale scores
might be interpreted.
Terwee et al.'s (2007) quality ratings also include ‘criterion validity’
(the extent to which scores on a particular scale relate to a ‘gold
standard’)and‘responsiveness’(the ability of a scale to detect change
over time). However, these two criteria were not rated. In the case of
‘criterion validity’this was because there is no gold standard compas-
sion measure to rate scales against. In the case of ‘responsiveness’it
was because such data were not typically available and, as this criterion
relates to clinically meaningful change, arguably the majority of the
scales were not primarily designed to measure this. Therefore, the
total possible score for any measure was 14.
3. Results
3.1. Review of identified measures
Fig. 1 shows a flow diagram illustrating the search process. After re-
moving duplicates, 2146 papers were identified, with only nine mea-
sures included after screening titles, abstracts, and full texts. Table 2
provides the quality ratings of the reviewed measures and Tables 3
and 4 outline the psychometric properties of the measures. Floor and
ceiling effects are not included in Tables 3 and 4 because no studies
reported them. Similarly, although most studies included measures of
Fig. 1. PRISMA flow diagram of search strategy.
Table 2
Quality ratings of measures of compassion.
Measure Content validity Factor structure Internal consistency Test retest reliability Convergent/discriminant validity Floor/ceiling effects Interpretability Total/14
SCS 1 0 2 2 1 0 1 7
RCS 1 2 2 0 1 0 1 7
CLS 1 1 2 0 1 0 1 6
SCBCS 1 1 2 0 1 0 1 6
CS-P 1 2 1 0 1 0 1 6
CS-M 1 0 2 0 0 0 1 4
CCAT 1 0 2 0 0 0 1 4
SCS-SF 1 2 1 0 0 0 0 4
SCCCS 1 0 0 0 1 0 0 2
Note. Rating: 0 = criterion not met/insufficient data to rate criterion; 1 = criterion partially met; 2 = criterion fully met.
CCAT = Compassionate Care Assessment Tool; CLS = Compassionate Love Scale; CS-P = Pommier Compassion Scale; CS-M = Martins et al. Compassion Scale; RCS=RelationalCom-
passion Scale; SCBCS = Santa Clara Brief Compassion Scale; SCCCS = Schwartz Center Compassionate Care Scale; SCS = Self-Compassion Scale; SCS-SF = Self-Compassion Scale —
Short Form.
20 C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
related constructs to test convergent validity, none included measures
of theoretically unrelated constructs; therefore, discriminant validity is
not included in Table 4.
3.1.1. Compassionate love scale (CLS; Sprecher & Fehr, 2005)
The CLS consists of 21 self-report items, rated on a Likert scale from 1
(not at all true of me) to 7 (very true of me). The CLS is intended for the
general population and consists of two forms: one relating to significant
others (including family members and friends) and one focusing on
strangers and humanity at large.
3.1.1.1. Content validity. The scale was rated as partially satisfactory for
content validity. Items were generated by the investigators based on a
review of the literature on love and altruism and also based on a proto-
type analysis with laypeople around their concept of compassionate
love. In line with our definition of compassion, the scale includes
items related to four of our five elements of compassion identified ear-
lier: Feeling moved by other people's suffering (emotional resonance),
understanding or imagining somethingabout their condition as a fellow
being, accepting and not judging others (which implies tolerance), and
being motivated to help them. However, the CLS did not appear to con-
tain items explicitly related to recognizing suffering.
Three items include the word ‘compassion’or ‘compassionate love’,
which requires respondents to define these concepts themselves. How-
ever, it seems uncertain whether they will know what is meant by ‘com-
passionate love’or define it uniformly. Additionally, not all items on the
scale relate to those who are suffering, and it is questionable whether
items such as: “I feel happy when I see that [loved ones/others
(strangers)] are happy”and “I very much wish to be kind and good to
[my friends and family members/fellow human beings]”assess com-
passion or in fact more broadly assess empathy and kindness respec-
tively. Finally, the scale refers explicitly to either close others or
strangers, but does not allow respondents to consider people who
may not fall into either of these categories (e.g. patients responding
in relation to healthcare professionals), potentially limiting its use in
some contexts.
3.1.1.2. Factor structure and reliability. Exploratory factor analysis (EFA)
yielded a single factor structure for each version of the scale. Sprecher
and Fehr (2005) did not explicitly propose a factor structure for the
Table 3
Psychometric properties of measures of compassion (content validity, factor structure, internal consistency, and test–retest reliability).
Measure
Content validity:
Factors
captured
a
/5
Content validity:
item generation
(recipient and
expert groups
consulted?)
Proposed factor
structure
Support for factor
structure: type of
analysis conducted
(factor structure
found)b
Internal consistency:
adequate sample
size for factor
analyses?
Internal consistency:
Cronbach alpha
(for total
scale and subscales)
Test retest reliability:
r(time between
testing)
CLS 4 (U, ER, T, & A) Recipients = yes
Experts = yes
Not reported EFA (single factor for
both versions)
Yes
(N= 354)
α= 0.95 for both
close others and
strangers-humanity
versions
Not reported
SCBCS 2 (ER & A) Recipients = yes
Experts = yes
Not reported EFA (single factor) Yes
(N= 223)
α= 0.90 Not reported
CS-M 1 (A) Recipients = yes
Experts = yes
Five factors EFA (two-factor
structure found but
rejected in favour of a
single-factor model)
Yes
(N= 310)
α= 0.82 Not reported
SCS 4 (U, ER, T, & A) Recipients = no
Experts = yes
Six factors
represented under a
single overarching
construct
CFA (six factors) Yes
(N= 391)
Total α= 0.92,
Subscales = 0.75 to
0.81.
Total scale:
r= 0.93,
Subscales:
r= 0.80 to 0.88
(3 weeks)
SCS-SF 4 (U, ER, T, & A) Recipients = no
Experts = yes
Six factors
represented under a
single overarching
construct
CFA (six factors
represented under a
single overarching
construct)
Yes
(N= 415)
Total α= 0.86,
Subscales = 0.54 to
0.75.
Not reported
CS-P 4 (R, U, ER, & A) Recipients = no
Experts = yes
Six factors
represented under a
single overarching
construct
CFA (six factors
represented under a
single overarching
construct)
Yes
(sample 1: N= 439,
sample 2: N= 510)
Total α=0.90(sample
1) and 0.87 (sample 2).
Subscale αsb0.70 for
4/6 subscales in
sample 1 and 1/6
subscales in sample 2.
Not reported
RCS
(compassion for
others and
self-compassion
subscales)
4 (R, ER, T, & A) Recipients = no
Experts = yes
Four factors CFA (four factors) Yes
(N= 231)
Subscales = 0.74 to
0.84
Not reported
CCAT 3 (ER, T, & A) Recipients = yes
Experts = yes
Not reported EFA (four factors) Yes
(N= 250)
Total αN0.70 (exact
value not given),
Subscales = 0.77 to
0.87.
Not reported
SCCCS 3 (R, ER, A) Recipients = yes
Experts = yes
Not reported EFA (single factor but
analysis was not
conducted on all items)
CFA (single factor but
analysis was not
conducted on all items)
Yes
(N= 801)
Subscales = 0.97 and
0.95. Single scale but
total αmissing.
Not reported
CCAT = Compassionate Care Assessment Tool; CFA = confirmatory factor analysis; CLS = Compassionate Love Scale; CS-P = Pommier Compassion Scale; CS-M = Martins et al. Com-
passion Scale; EFA = exploratoryfactor analysis; RCS = Relational Compassion Scale; SCBCS = Santa ClaraBrief Compassion Scale;SCCCS = Schwartz Center Compassionate Care Scale;
SCS = Self-Compassion Scale; SCS-SF = Self-Compassion Scale —Short Form.
a
Five elements of compassion: R = recognizing suffering; U = understanding the universality of suffering; ER = emotional resonance; T = tolerating uncomfortable feelings; A = acting or
motivation to act to alleviate suffering.
b
For details of the factors identified, refer to the results section.
21C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
CLS prior to analysis and did not conduct CFA. Internal consistency was
high for both versions. Test–retest reliability was not reported.
3.1.1.3. Convergent validity and interpretability. Convergent validity was
supported by significant correlations in the expected directions with
measures of empathy, helpfulness, volunteerism, and spiritual experi-
ences. Limited subgroup analyses were undertaken by Sprecher and
Fehr (2005), showing that women obtained significantly higher com-
passion scores than men on both versions.
3.1.2. Santa Clara brief compassion scale (SCBCS; Hwang, Plante, & Lackey,
2008)
The SCBCS is a shortened version of Sprecher and Fehr's (2005) CLS,
consisting of five items from the original scale (the correlation between
the two scales is r= 0.95). Unlike the CLS, this scale refers exclusively to
Table 4
Psychometric properties of measures of compassion (convergent validity and interpretability).
Measure
Convergent validity: correlation (Pearson's r) of compassion measure with
measures of related constructs
a
Interpretability: subgroups tested for differences
CLS PSP other-oriented empathy subscale and empathy items from Schieman & Van
Gundy (2000):r= 0.50 to 0.68;
PSP helpfulness subscale: r= 0.23 (close others), r= 0.32 (strangers);
Frequency of church attendance: r= 0.22 (close others), r= 0.43 (strangers);
Volunteerism items from Mikulincer et al. (2005):r= 0.18 (close others), r=
0.35 (strangers);
Social support (developed by authors): r= 0.27 (strangers), r= 0.51 (close
others);
DSES: r= 0.39 (close others), r= 0.44 (strangers).
Gender (women scored significantly higher than men on both
versions)
SCBCS IRI empathic concern subscale: r= 0.65; VIQ: r= 0.48; SCSORF: r= 0.27. Gender (women scored significantly higher than men)
CS-M CLS: r= 0.66 Gender (women scored significantly higher than men),
Education (significantly higher for those with University
education versus High School or less),
Income (significantly higher for those with annual income
$40,000+ versus $10,000 or less),
Age, race, and marital status (no differences)
SCS DEQ self-criticism subscale: r=−0.65; SOC: r= 0.41; TMMS attention
subscale: r= 0.11; TMMS clarity subscale: r= 0.43; TMMS repair subscale: r=
0.55; RSES: r= 0.59.
Gender (women scored significantly lower than men)
SCS-SF Not reported Not reported
CS-P SCS: r= 0.01; SOC: r= 0.41; 3D-WS reflective subscale: r= 0.26; 3D-WS
cognitive subscale: r= 0.39; 3D-WS affective subscale: r= 0.56; QMEE: r=
0.58; IRI empathic concern subscale: r= 0.65;
IRI perspective taking subscale: r=0.35; CLS close others version: r= 0.54;
CLS strangers version: r= 0.27; SMQ: r=−0.12.
Gender (women scored significantly higher than men)
RCS (compassion for others
and self-compassion
subscales)
RCS compassion for others subscale & SCS: r= 0.24;
RCS self-compassion subscale & SCS: r= 0.65;
RCS compassion for others subscale & EACS emotional expression & processing
subscales: r= 0.41 and 0.42;
RCS self-compassion subscale & EACS emotional expression & processing
subscales: r= 0.51 and 0.46;
RCS compassion for others subscale & SCSRS inadequate & hated self subscales:
r= 0.03 and 0.12;
RCS self-compassion subscale & SCRS inadequate & hated self subscales: r=
−0.29 and −0.41;
RCS compassion for others subscale & SCSRS reassured self subscale: r= 0.01;
RCS self-compassion subscale & SCRS reassured self subscale: r= 0.43;
RCS compassion for others subscale and RSQ secure attachment: r= 0.34;
RCS self-compassion subscale and RSQ secure attachment: r= 0.31;
RCS compassion for others subscale and RSQ insecure attachment styles
(fearful, preoccupied, dismissing, anxious, and avoidant):
r=−0.23, −0.06, −0.15, −0.19, and −0.22, respectively;
RCS self-compassion subscale and RSQ insecure attachment styles (fearful,
preoccupied, dismissing, anxious, and avoidant): r=−0.22, −0.15, −0.05,
−0.03, and −0.07, respectively.
Significant differences in RCS scores between Arts and
Engineering students. The direction of the results for each
subscale was not specified.
CCAT Not reported Gender (women scored carers significantly higher than men),
Marital status and reason for hospitalisation (no differences)
SCCCS Overall satisfaction with recent hospitalisation (item set 1): r= 0.54;
Overall satisfaction with recent hospitalisation (item set 2): r= 0.60;
Satisfaction with communication and emotional support (item set 1): r= 0.72;
Satisfaction with communication and emotional support (item set 2): r= 0.64.
Not reported
3D-WS = 3-Dimensional Wisdom Scale (Ardelt, 2003); CCAT = Compassionate Care Assessment Tool; CLS = Compassionate Love Scale; CS-P = Pommier Compassion Scale; CS-M =
Martins et al. Compassion Scale; DEQ = Depressive Experiences Questionnaire (Blatt, D'Afflitti, & Quinlan, 1976 ); DSES = Daily Spiritual Experience Scale (Underwood & Teresi,
2002); EACS = Emotional Approach Coping Scale (Stanton et al., 2000); IRI = Interpersonal Reactivity Index (Davis, 1980); PSP = Prosocial Personality Battery (Penner et al., 1995);
QMEE = Questionnaire Measure of Emotional Empathy (Mehrabian & Epstein, 1972); RCS = Relational Compassion Scale; RSES = Rosenberg Self-Esteem Scale (Rosenberg, 1965);
RSQ = Relationship Scales Questionnaire (Griffin & Bartholomew, 1994); SCBCS = Santa Clara Brief Compassion Scale; SCSORF = Santa Clara Strength of Religious Faith Questionnaire
(Plante & Boccaccini, 1997); SCCCS = Schwartz Center Compassionate Care Scale; SCS = Self-Compassion Scale; SCS-SF = Self-Compassion Scale —Short Form; SCSRS = Self-
Criticising/Attacking and Self-Reassuring Scale (Gilbert et al., 2004); SMQ = Southampton Mindfulness Questionnaire (Chadwick et al., 2008); SOC = Social Connectedness Scale (Lee
& Robbins, 1995); TMMS = Trait Meta-Mood Scale (Salovey et al., 1995); VIQ = Vocational Identity Questionnaire (Dreher et al., 2007).
a
Five elements of compassion: R = recognizing suffering; U = understanding the universality of suffering; ER = emotional resonance; T = tolerating uncomfortable feelings; A = ac ting or
motivation to act to alleviate suffering.
22 C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
strangers rather than to close others. The items of the SCBCS were se-
lected because they had moderate means, high standard deviations,
and high correlations with the overall score from the CLS.
3.1.2.1. Content validity. The SCBCS was rated partially satisfactory for
content validity. The scale includes items related to two of our five ele-
ments of compassion: Emotionally connecting with other people's suf-
fering and acting to help them. However, unlike the CLS, the SCBCS
did not appear to contain items explicitly related to understanding the
universality of suffering and tolerating uncomfortable feelings, and
also did not include items explicitly related to recognizing suffering.
Two items contain the word “compassion”, again relying on respon-
dents to define these terms rather than tapping into their underlying
elements.
3.1.2.2. Factor structure and reliability. EFA yielded a single factor struc-
ture for the SCBCS and CFA was not conducted. Internal consistency
was high. Test–retest reliability was not reported.
3.1.2.3. Convergent validity and interpretability. The SCBCS was strongly
correlated with empathic concern, moderately correlated with voca-
tional identity, and showed a small correlation with strengthof religious
faith. Examination of group differences was limited to gender and
showed that women scored significantly higher than men.
3.1.3. The compassion scale (CS-M; Martins, Nicholas, Shaheen, Jones, &
Norris, 2013)
Martins et al.'s CS-M is a 10-item self-report scale developed to mea-
sure five domains of compassion: generosity, hospitality, objectivity,
sensitivity, and tolerance across social networks and relationships
(strangers, friends, and family) using a 1 (none) to 7 (all) response
scale. The aim of the scale was to provide a measure of compassion
across domains that could be enhanced through training, as the authors
argue that scales like the CLS do not lend themselves well to measuring
compassion in a way that can be targeted for education. Items were gen-
erated and evaluated by a panel of academic and community experts.
3.1.3.1. Content validity. Martins et al.'s (2013) scale was rated partially
satisfactory for content validity. The CS-M focuses exclusively on practi-
cal acts of compassion including giving financial help to others, using
your free time to help others, and doing things for others at a cost or
risk to yourself or your family and friends. Thus, only the acting to alle-
viate suffering factor of our five-factor definition is captured by the
items of theCS-M; items related to recognizing suffering, understanding
the universality of suffering, emotional resonance, and tolerating un-
comfortable feelings were not included. Additionally, it could be argued
that the scale's items measure only a limited range of acts of compassion
(giving away money, using free time to help others, sharing personal
space, or doing something for others at a cost to oneself) and if the
scale were applied to certain contexts (e.g. a healthcare context), the
items may not assess the types of actions that might be expected in
those contexts (e.g. considering ways to make those who are suffering
more comfortable).Indeed, it is not altogether clear for whatpopulation
the scale is intended. Furthermore, items such as “How many times
would you do the right thing if it puts your family at risk”do not appear
to fit well with a response ranging from “none”to “all”.
3.1.3.2. Factor structure and reliability. EFA did not support the proposed
five factor structure; the analysis identified a two-factor solution. How-
ever, the two-factor structure was rejected by the authors in favour of a
single factor model, arguing that, as all items beginning “How much of
your…?”loaded onto one factor and all items beginning “How many
times would you…?”loaded onto the second factor, the factors ap-
peared to reflect methodological differences between items rather
than substantively different constructs. Cronbach's alpha for the total
scale was acceptable. Test–retest reliability was not tested.
3.1.3.3. Convergent validity and interpretability. The authors only com-
pared their scale with the CLS (rN0.50). In terms of interpretability,
the authors provided mean scores for a range of subgroups but, though
they argue that the scale should help measure change in compassion
after training, they do not provide any indication of what level of change
on the scale would be needed to show that such training had been of
value.
3.1.4. Self-compassion scale (SCS; Neff, 2003b)
The SCS is a 26-item scalewith a 5-point response scale from “almost
never”to “almost always”.
3.1.4.1. Content validity. The scale was rated partially satisfactory overall
for content validity. Although items were selected after extensive
piloting, it is notable that this was only carried out with experts and un-
dergraduate students, even though the scale's target population includ-
ed community and clinical samples. The scale includes items related to
four of the five elements in the definition of compassion used in this re-
view: understanding the universality of suffering, emotional resonance,
the ability to tolerate distressing feelings, and feeling motivated to act or
acting to help ameliorate one's suffering. However, the scale does not
include items specifically relating to being attentive to how one is
feeling.
3.1.4.2. Factor structure and reliability. CFA of the SCS supported the six
factor model, with each of the three components of self-compassion
split into two sub-factors —one comprising ‘positively’worded and
one ‘negatively’worded items. This resulted in the following factors:
kindness versus self-judgement; mindfulness versus over-
identification; and common humanity versus isolation. However, Neff
(2003b) found only a marginal fit with a single higher-order factor,
questioning whether the six factors can be explained by a single over-
arching construct of self-compassion. Other studies have also
questioned the higher-order factor structure and the non-hierarchical
six-factor model across a range of populations, student, clinical and
meditating/non-meditating (e.g. Costa, Marôco, Pinto-Gouveia, Ferreira,
& Castilho, 2015; López et al., 2015; Williams et al., 2014). Several stud-
ies have suggested a two-factor model of self-compassion, with the fac-
tors representing the positive and negative dimensions of self-
compassion and self-criticism, respectively (e.g. Costa et al., 2015;
López et al., 2015).
Recently, Neff (2015) argued that the two-factor conceptualization
of the SCS is problematic in that it does not capture the relative balance
between the three proposed broad components of self-compassion
(self-kindness vs. self-judgement, common humanity vs. isolation, and
mindfulness vs. over-identification). Instead, Neff proposed a bifactor
model of self-compassion, where each item loads directly on to a gener-
al factor as well as their respective subscale, and suggests that re-
searchers can select whether to analyse subscale scores separately or
use a total SCS score depending on their interests.
Cronbach's alpha values for total SCS scale and subscale scores and
test–retest reliability were adequate.
3.1.4.3. Convergent validity and interpretability. Convergent validity was
supported by significant correlations in the expected direction between
the SCS and other related measures, severalof which were ≥0.50. A par-
tially satisfactory score was achieved for interpretability, as only gender
differences were reported.
3.1.5. Self-compassion scale: short form (SCS-SF; Raes, Pommier, Neff, &
Van Gucht, 2011)
Raes et al. developed a 12-item version of the SCS by selecting two
items from each of the SCS's six subscales, based on their high correla-
tions with the SCS and intended subscales, and high intercorrelations.
The SCS-SF is rated in the same way as the SCS.
23C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
3.1.5.1. Content validity. The scale was rated ‘partially satisfactory’for
content validity for the same reasons as the long form.
3.1.5.2. Factor structure and reliability. CFA supported the proposed six-
factor hierarchical structure of the measure. Internal consistency was
acceptable for the total score, but was variable for the individual sub-
scales. Test–retest reliability was not reported.
3.1.5.3. Convergent validity and interpretability. Relevant data were not
reported for convergent validity and interpretability.
3.1.6. The compassion scale (CS-P; Pommier, 2010)
The CS-P is a 24-item self-report scale targeted at the general popu-
lation and based on the argument (outlined earlier) that compassion
consists ofsix elements: Kindness (in contrastto indifference), mindful-
ness (in contrast to disengagement), and common humanity (in con-
trast to separation). Responses are given on a five-point Likert scale,
ranging from 1 (almost never) to 5 (almost always).
3.1.6.1. Content validity. The CS-P was rated partially satisfactory for
content validity. Items were devised by the author, based on theory
and research, and reviewed by a panel of experts. The scale includes
items consistent with four of our five elements of compassion: Rec-
ognizing suffering, feeling moved by suffering, understanding or
imagining something about another person's condition as a fellow
being, and motivation to act/acting to alleviate suffering. Although
in her development of the scale Pommier (2010) notes that compas-
sion requires the ability to tolerate uncomfortable feelings in the
face of suffering so that one can remain tolerant and accepting of
others, the scale appears not to directly assess this, other than asking
whether respondents “try to keep a balanced perspective on the sit-
uation”when people tell them about their problems, or whether
they tend to avoid those who suffer.
Additionally, several of the scale's items include the words ‘some-
times’,‘often’or ‘usually’which conflict with the response scale used
(‘almost never’to ‘almost always’) and makes responses difficult to in-
terpret. The response scale is perhaps also unintuitive for negatively
worded items —for example, a response of “I almost never don't feel
emotionally connected to people in pain”may be difficult for some peo-
ple to rate accurately. Similarly, items such as “Suffering is just a part of
the common human experience”cannot be answered accurately using
the scale from ‘almost never’to ‘almost always’and do not sit well
with the scale's instructions to “indicate how often you behave in the
stated manner”.
3.1.6.2. Factor structure and reliability. CFA supported the proposed six-
factor structure of the measure, and that a single higher order factor of
compassion explained the inter-correlations between the six factors.
EFA was not conducted because Neff's (2003b) Self-Compassion Scale
(SCS) had already demonstrated these six factors. However, as noted
earlier compassion for others and self-compassion were not significant-
ly correlated in Pommier's (2010) research, suggesting that the factor
structure for each measure cannot be assumed to be identical. Internal
consistency was high for the total score but mixed and inconsistent
across samples for the subscales. Test–retest reliability was not
reported.
3.1.6.3. Convergent validity and interpretability. Convergent validity was
supported by significant correlations in the expected direction between
the CS-P and other measures of compassion, empathy, perspective-
taking, and wisdom; several of these were ≥0.50. However, the CS-P
was not significantly correlated with the SCS (Neff, 2003b), a problemat-
ic finding for the scale's construct validity, given that the CS-P was devel-
oped based on the factor structure of the SCS. Additionally, while the
scale was positively correlated with the CLS, this correlation was only
small for the strangers-humanity version (r= 0.27; r= 0.54 for the
close others version). This suggests that the CS-P and the CLS may not
be measuring the same construct. Another unexpected finding was
that the Southampton Mindfulness Questionnaire (SMQ; Chadwick
et al., 2008) had a small negative correlation with the CS-P.The only sub-
group analyzed was gender, again showing that women scored higher
than men.
3.1.7. Relational compassion scale (RCS; Hacker, 2008)
The RCS consists of 16 items rated on a four-point scale (from ‘do not
agree’to ‘agree strongly’). The scale consists of four subscales which
measure respondents' compassion for others and self-compassion,
along with their beliefs about how compassionate other people are to
each other, and their beliefs about how compassionate other people
are to them. The latter two subscales extend beyond simply measuring
respondents' own levels of compassion, but the scale was nevertheless
included because it defines itself as a comprehensive measure of com-
passion and also because the subscales were psychometrically tested
individually.
3.1.7.1. Content validity. The RCS was rated ‘partially satisfactory’for con-
tent validity. The scale's items that comprise the ‘compassion for others’
subscale assess people's capacity to recognize and understand suffering
and accept and not judge others (which implies tolerance), just two of
the five elements in our definition of compassion. Additionally, some
items relate to other people's ‘experiences’in general, rather than spe-
cifically to their suffering. The items comprising the self-compassion
subscale assess emotional resonance and acting to alleviate suffering,
two of the five elements in our definition. Items related to understand-
ing the universality of suffering were not included in either subscale.
3.1.7.2. Factor structure and reliability. CFA using the final version of the
RCS supported its proposed four-factor structure. Internal consistency
was acceptable for all four subscales. Test–retest reliability was not
tested.
3.1.7.3. Convergent validity and interpretability. Although several correla-
tions with related measures were ≥0.50, specific hypotheses appear not
to have been set out in advance aboutthe expected direction of correla-
tions and, in the discussion, the author highlights some unexpected
findings. For example, the ‘compassion for others’subscale did not cor-
relate significantly with a measure of self-criticism/self-attack and self-
reassurance (Gilbert, Clarke, Hempel, Miles, & Irons, 2004) and although
the self-compassion subscale was positively correlated with the SCS
(r= 0.65), this is arguably weaker than might be expected given they
allegedly measure the sameconstruct. Only one subgroup was analyzed
for interpretability (Arts versus Engineering students), however the au-
thors stated no predictions about differences between these groups.
3.1.8. Compassionate care assessment tool (CCAT; Burnell & Agan, 2013)
The 28-item CCAT was developed to measure levels of compassion
demonstrated by individual nurses providing care for patients in acute
hospital settings. In contrast to the other scales reviewed so far, this
scale is completed by patients in relation to their carers. Respondents
rate compassionate care from two perspectives —the importance of
each item to them personally, and the degree to which their individual
nurses demonstrated these qualities. Ratings range from 1 to 4. A selec-
tion of possible items for the scale were derived from the Spiritual
Needs Survey (Galek, Flannelly, Vane, & Galek, 2005)andtheCaringBe-
haviors Inventory (Wu, Larrabee, & Putman, 2006), and refined after
consulting with hospital staff involved in implementing national criteria
for compassionate care, nurses, and patients. The CCAT focuses on four
domains: the ability of carers to establish meaningful connection (e.g.
having a sense of humor), to meet patient expectations (e.g. giving
timely treatments), display caring attributes (e.g. considering personal
needs), and exhibit capable practitioner qualities (e.g. appearing
competent).
24 C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
3.1.8.1. Content validity. Overall, the scale was considered partially satis-
factory for content validity. It includes items relating to three of the five
elements in our definition of compassion: whether patients thought
carers felt for them (emotional resonance), acted to help relieve their
suffering, and could tolerate distress (e.g. asking if they ‘remained
calm at all times’, treated them non-judgmentally, and excused their
shortcomings). Items related to recognizing suffering and understand-
ing the universality of suffering were not included. Additionally, some
of the items arerather ambiguous—for example, one item askswhether
nurses “addressed difficult issues”, which could relate to their ability to
tolerate distress, or to their ability to resolve more practical matters. It is
also questionable whether the scale is actually measuring levels of com-
passion of nurses; factor analyses appear to have been carried out based
on asking patients to rate how important each item wasto them, rather
than on asking them to rate the extent to which their carers behaved in
this way.
Furthermore, as a number of items were derived from the Spiritual
Needs Survey, thereis a fairly strong emphasis on whether spiritual sup-
port was offered to patients, which is not necessarily relevant to the
measurement of compassion for all patients. Similarly, several items
taken from the Caring Behaviors Inventory ask about whether nurses
gave timely treatments to patients, showed skill with equipment and
helped control pain; however, while competence may be important in
order to provide compassionate care, such abilities in themselves do
not necessarily equate to showing compassion. It could also be argued
that some of the areas tapped, such as providing timely treatments, con-
trolling pain, and providing access to spiritual support, depend on vari-
ables outside of nurses' power (i.e. managerial or organizational level
factors), and this raises a wider issue around the extent to which com-
passion can and should be measured at an individual or organizational
level.
3.1.8.2. Factor structure and reliability. EFA supported a four-factor struc-
ture, with the four aforementioned domains. However, as previously
noted, analyses appear to have been carried out based on asking pa-
tients to rate how important each item was to them, rather than on ask-
ing them to rate the extent to which their carers behaved in this way.
This means that it is not clear whether the scale is measuring actual
levels of compassion of their nurses per se. Additionally, the authors re-
port that only 20 of the 28 items fit into the four factors identified, but
they nonetheless appear to have retained all 28 items. Therefore, the
CCAT was given a rating of 0 for factor structure. Cronbach's alpha
values were adequate for the total scale and subscales. Test–retest
reliability was not reported.
3.1.8.3. Convergent validity and interpretability. Convergent validity was
not reported. Limited subgroup analyses were conducted for
interpretability.
3.1.9. The Schwartz center compassionate care scale (SCCCS; Lown,
Muncer, & Chadwick, 2015)
The 12-item SCCCS was developed to measure patients' ratings of
compassionate inpatient care received from physicians' during a recent
hospitalisation. Patients complete items using a ten-point scale from 1
(not at all successful) to 10 (very successful). Itemswere initially devel-
oped by a committee consisting of patients, family members of patients,
and individuals working in healthcare policy and advocacy, and were
fine-tuned in five focus groups with patients, physicians, and nurses.
3.1.9.1. Content validity. Overall, the SCCCS was considered partially sat-
isfactory for content validity. It includes items which could be
interpreted to relate to three of the five elements in our definition of
compassion: Whether patients thought physicians expressed sensitivi-
ty, care, and compassion for them (emotional resonance/acting to alle-
viate suffering), listened attentively (recognizing suffering), and acted
in ways to relieve their suffering. The SCCCS did not appear to contain
items related to understanding the universality of suffering and tolerat-
ing uncomfortable feelings. Additionally, a couple of items refer to com-
petence in caring (whether physicians spend enough time with
patients, whether physicians communicate test results in a timely man-
ner) which does not necessarily equate to showing compassion and
could be dependent on factors outside of physicians' power (i.e. mana-
gerial or organizational level factors).
3.1.9.2. Factor structure and reliability. The SCCCS originally consisted of
16 items which were split into two item sets and administered to 801
recently hospitalized patients; half were asked item set one and half
item set two. The authors conducted an EFA and CFA for each set of
items and concluded that items within each set were unidimensional.
However, they did not conduct analyses on all of the items, making it
impossible to determine whether the scale as a whole is unidimension-
al, or whether the measure consisted of two separate scales or subscales.
Despite this, the SCCCS was presented as a single scale. Although
Cronbach's alpha values were adequate for both sets, these values
were based on there being eight items in each set; the final 12-item
scale consisted of seven items from the first set and five items from
the second set after the removal of problematic items (e.g. items with
lowest item-total correlations). The alpha value for all of the scale
items was also missing. Test–retest reliability was not tested.
3.1.9.3. Convergent validity and interpretability. The authors found mod-
erate to large, positive correlations between scores on both sets of
items from the 12-item SCCCS and related constructs. Interpretability
was not tested.
4. Discussion
The first aim of this paper was to synthesize existing conceptualiza-
tions of compassion and to propose a new definition that integrates
common elements. A range of definitions from Buddhist and Western
psychological perspectives were considered and five components of
compassion were identified: recognition of suffering; understanding
its universality; feeling sympathy, empathy, or concern for those who
are suffering (which we describe as emotional resonance); tolerating
the distress associated with the witnessing of suffering; and motivation
to act or acting to alleviate the suffering. Each of these components has
been articulated by several published definitions of compassion, al-
though no single existing definition explicitly includes all five of them.
We do not claim that these five elements constitute statistically distinct
factors of an overarching construct of compassion; this possibility must
be empirically tested. However, we argue that our definition provides a
useful foundation for the development of a comprehensive new mea-
sure of compassion.
The need for a new measure is supported by the findings of our
review of existing measures of compassion. The maximum quality
rating of any measure was seven out of a possible fourteen, suggest-
ing that no scale exists that comprehensively measures compassion
and provides scores with acceptable levels of reliability and validity.
In other words, we cannot be confident that existing measures of
compassion are measuring this construct accurately and this raises
significant barriers to scientificprogressinthefield —how can we
assess compassion and evaluate the effectiveness of interventions
intended to enhance compassion if we cannot measure the construct
accurately?
Quality ratings were low both because of poor ratings for content va-
lidity (the extent to which items appeared to fit our definition of com-
passion) and because of poor or untested psychometric properties.
Internal consistency was strong for total scores but weak for many sub-
scales. Evidence for the proposed factor structure of some scales was
weak or absent. The presence of floor or ceiling effects was not exam-
ined for any scale, and test–retest reliability was examined for only
one. Convergent correlations were generally significant and in the
25C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
expected directions, but discriminant validity was not assessed. Low
quality ratings could also be attributed to measures being in their
early stages of development and initial papers being unlikely to include
a thorough test of psychometric properties. Quality ratings for compas-
sion measures may improve over time with additional research includ-
ing psychometric research.
The strongest measures identified were Neff's (2003b) Self-
Compassion Scale and Hacker's (2008) Relational Compassion Scale,
but neither of these measures capture each of the five elements in our
definition. As Neff's measure focuses on self-compassion rather than
compassion more generally or compassion for others, it is in any case
not entirely suitable as a measure that can be used to determine levels
of compassion in populations for whom compassion towards others is
of interest(e.g., healthcareprofessionals).Given the current enthusiasm
for compassion across different contexts, it is critical for future research
to develop a psychometrically robust measure of the proposed
definition of compassion as well as to explore more fully the relation-
ship between self- and other-compassion.
4.1. Strengths and limitations
A strength of this review is its contribution to greater clarity in
the conceptualization of compassion and its components, which
have previously been described in a variety of ways. The five ele-
ments of compassion extracted from our synthesis of definitions sug-
gests that compassion is a complex construct that includes emotion
but is more than an emotion, as it also includes perceptiveness or
sensitivity to suffering, understanding of its universality, acceptance,
nonjudgment, and distress tolerance, and intentions to act in helpful
ways. This conceptualization suggests that compassion can be state-
like and trait-like. Sensitivity to one's own or others'suffering, emotion-
al responsiveness, acceptance and nonjudgment in the face of suffering,
and motivation to be helpful are all likely to fluctuate across time and
situations. On the other hand, Goetz et al. (2010) present evidence sug-
gesting that compassion can be seen as a trait-like quality that endures
over time (e.g., Eisenberg et al., 2002). An implicit assumption of
compassion-focused interventions seems to be that a trait-like general
tendency to be compassionate towards oneself or others can be devel-
oped through repeated practice of skills that cultivate compassionate
states, attitudes, or behaviors. Additionally, although many of the ques-
tionnaires reviewed treat compassion as a disposition that is fairly con-
sistent across contexts, some measures conceptualize compassion as
operating within a particular context or social interaction (e.g., the
CCAT).
This review assumes that compassion can indeed be measured with
questionnaire methods. Some authors have suggested that subtle but ob-
servable behaviors, such as using a soft tone of voice, may also be valid in-
dicators of compassion (Cameron, Mazer, DeLuca, Mohile, & Epstein,
2015), while Pearson (2006) notes that acts of compassion are often ‘in-
visible’,being“simple not clever; basic not exquisite; peripheral not cen-
tral”(p. 22). This means that, as Dewar, Pullin, and Tocheris (2011) note,
“there is a danger, therefore, of measuring what is easy to quantify, rather
than what is important”(p.32).Dewaretal.alsopointoutthatcompas-
sion can be seen as something that is negotiated between individuals in
their interactions. These points suggest that, as with many psychological
variables, questionnaire measures may only provide a partial picture of
compassion. Furthermore, while questionnaire measures benefit from
being simple to administer and complete, and helpful for tapping people's
underlying attitudes where these are not directly observable, it may be
difficult for people to complete such measures accurately in some con-
texts, for example in situations where healthcare staff feel under threat
to be seen to be compassionate.
A further limitation of this review is the approach taken to iden-
tifying the definitions of compassion in Table 1. A systematic search
was attempted but the way the field has evolved does not easily
lend itself to a systematic review. Defining compassion was very
rarely the primary purpose of papers; definitions were typically em-
bedded as secondary to addressing the primary purpose of the paper.
An early search generated an unfeasibly large number of results. We
therefore relied on the expertise of the authors to identify key theo-
rists and sources in the field.
The review also assumed that individual levels of compassion should
be measured. However, it has been argued that measuring compassion
at the individual level opens people to accusations that they are not suffi-
ciently compassionate. For example, in a healthcare context, this may re-
sult in a tendency to blame healthcare professionals for failings that in fact
relate to external factors such as resourcing pressures or organizational
restructuring (Crawford, Brown, Kvangarsnes, & Gilbert, 2014). This is
an important consideration and highlights the need to ensure that efforts
to measure levels of compassion among individuals do not serve to over-
state individual deficits while deflecting attention from the broader im-
pact of resourcing constraints and wider organizational changes.
4.2. Future research
This review has argued that currently no psychometrically robust
self- or observer-rated measure of compassion exists, despite wide-
spread interest in measuring and enhancing compassion towards
self and others. Future research should t herefore focus on developing
a psychometrically robust questionnaire-based measure of compas-
sion, while keeping in mind the complexities around measuring
this construct. It will subsequently be of value for future research
to identify interventions (at both an individual level and organiza-
tional level) that have the potential to enhance compassion and ex-
amine whether changes in compassion mediate the outcomes of
these interventions.
Although our reviewprovides a foundation for progressing research
into compassion, it represents a starting point. Futurework should artic-
ulate theory driven hypotheses that test the relationships between key
constructs and the validity of our five-element definition of compassion.
This will generate important new knowledge about how these different
elements interact to give rise to compassion. It may be that some
elements are facilitators of compassion or emergent factors rather
than defining features.
Using a range of designs (including prospective and experimental
designs), and triangulation of measurement to include behavioral
(e.g., observable compassionate responses), bio-behavioral measures
(e.g., as derived from Gilbert's theory) and self-report measures, will
further aid the development of theory and understanding. It is likely
that this will have real practical implications for how best to cultivate
compassion in ways that support resilience and well being at both
personal and societal levels.
4.3. Conclusion
In recent years, compassion has received increased scientific
interest. Compassion has been defined here, in line with the literature,
as involving five elements: recognizing suffering in others; understand-
ing the common humanity of this suffering; feeling emotionally con-
nected with the person who is suffering; tolerating difficult feelings
that may arise; and acting or being motivated to act to help the person.
A systematic search of measures of compassion was undertaken but all
of the identified measures were found to have notable psychometric
weaknesses. This is a serious limitation in the field. For example, with-
out adequate measures, we cannot determine with any confidence
levels of compassion or whether interventions designed to enhance
compassion are effective. Therefore, we now call for empirical testing
of our five element definition and the development of a measure of
compassion, following good practice guidelines to identify items and
to test its psychometric properties.
26 C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
References
AmericanMedical Association (2001). Principles of medical ethics. Retrieved from http://
www.ama-assn.org/ama/pub/p hysician-resources/medical-ethics/code-medical-
ethics/principles-medical-ethics.page
Ardelt, M. (2003). Empirical Assessment of a Three-Dimensional Wisdom Scale. Research
on Aging,Vol. 25,275–324.
Barker, C., Pistrang,N., & Elliott, R. (2002).Research methodsin clinical psychology. London:
Wiley.
Burnell, L., & Agan, D. L. (2013). Compassionate care: Can it be defined and measured?
The development of the Compassionate Care Assessment Tool. International Journal
of Caring Sciences,6(2), 180–187.
Cameron,R. A., Mazer, B. L., DeLuca, J. M., Mohile, S. G.,& Epstein, R. M. (2015).In search of
compassion: a new taxonomy of compassionate physician behaviours. Health
Expectations,18,1672–1685. http://dx.doi.org/10.1111/hex.12160.
Cassell, E. J. (2002). Compassion. In C. R. Synder, & S. J. Lopez (Eds.), Handbook of positive
psychology (pp. 434–445). Oxford: Oxford University Press.
Chadwick, P., Hemb er, M., Symes, J., Peters, E., Kuipers, E., & Dagnan, D. (2008).
Responding mindfully to unpleasant thoughts and images: Reliability and validity
of the Southampt on mindfulness qu estionnaire (SM Q). British Journal of Clinical
Psychology,47(4), 451–455.
CoED Foundation (2014). Bringing compassion into education and learning: The second
strategic plan. Retrieved fro m http://www.c oedfoundation.org.uk/pdfs/CoEd_
Foundation_strategic_plan.pdf
Cosley, B. J., McCoy, S. K., Saslow, L. R., & Epel, E. S. (2010). Is compassion for others stress
buffering? Consequences of compassion and social support for physiological reactiv-
ity to stress. Journal of Experimental Social Psychology,46,816–823.
Costa,J., Marôco,J., Pinto-Gouveia, J.,Ferreira, C.,& Castilho, P.(2015). Validation of thepsy-
chometric properties of the self-compassion scale. Testing the factorial validity and
factorial invariance of themeasure among borderlinepersonalitydisorder, anxiety dis-
order, eating disorder and general populations. Clinical Psychology & Psychotherapy.
Crawford, P., Brown, B., Kvangarsnes, M., & Gilbert, P. (2014). The design of compassion-
ate care. JournalofClinicalNursing,23(23–24), 3589–3599.
Darwin,C. (1871). The descentof man, and selectionin relation to sex. London: John Murray.
Davis, M. A. (1980). A multidimensional approach to individual differences in empathy.
JSAS Catalogue of Selected Documents in Psychology,10,85.
Davis, M. H. (1983). Mea suring individu al differences in empathy: Evidence for a
multidimensional approach. Journal of Personality and Social Psychology,44,1
13–126.
de Waal, F. (2009). The age of empathy. New York: Harmony.
Department of Health (2013). The handbook to the NHS constitution. Retrieved from
https://www.gov.uk/government/publications/the-nhs-constitution-for-england
Dewar, B., Pullin, S., & Tocheris, R. (2011). Valuing compassion through definition and
measurement. Nursing Management,17(9), 32–37.
Dreher, D. E., Holloway, K. A., & Schoenfelder, E. (2007). The vocation identity question-
naire: Measuring the sense of calling. Research in the Social ScientificStudyof
Religion,18,99.
Eisenberg, N., Guthrie, I. K., Cumberland, A., Murphy, B. C., Shepard,S. A., Zhou, Q., & Carlo,
G. (2002). Prosocial development in early adulthood: A longitudinal study. Journal of
Personality and Social Psychology,82(6), 993–1006.
Eisenberg, N., VanSchyndel, S. K., & Hofer, C. (2015). The association of maternal sociali-
zation in childhood and adolescence with adult offsprings' sympath y/caring.
Developmental Psychology,51(1), 7–16.
Epstein, R. M., Franks, P., Shields, C. G., Meldrum, S. C., Miller, K. N., Campbell, T. L., &
Fiscella, K. (2005). Patient-centered communication and diagnost ic testing. The
Annals of Family Medicine,3(5), 415–421.
Feldman, C., & Kuyken, W. (2011). Compassion in the landscape of suffering.
Contempo rary Buddhism,12(1), 143–155.
Galek, K., Flannelly, K. J., Vane, A., & Galek, R. M. (2005). Assessing a patient's spiritual
needs: A comprehensive instrument. Holistic Nursing Practice,19(2), 62–69.
Gilbert, P. (2005). Compassion and cruelty: A biopsychosocial approach. In P. Gilbert
(Ed.), Compassion: Conceptualisations, research and use in psychotherapy (pp. 9–74).
London: Routledge.
Gilbert, P. (2009). The compassionate mind: A new approach to life's challenges. London:
Constable and Robinson.
Gilbert, P. (2010). The compassionate mind. London: Constable & Robinson Ltd.
Gilbert,P.,Clarke,M.,Hempel,S.,Miles,J.N.V.,&Irons,C.(2004).Criticizing and
reassuring oneself: An exploration of forms, styles and reasons in female students.
British Journal of Clinical Psychology,43(1), 31–50.
Gillath, O., Shaver, P. R., & Mikulincer, M. (2005). An attachment-theoretical approach to
compassion and altruism. In P. Gilbert (Ed.), Compassion: Conceptualisations, research
and use in psychotherapy (pp. 122–147). London: Routledge.
Goetz, J. L., Keltner, D., & Simon-Thomas, E. (2010). Compassion: An evolutionary analysis
and empirical review. Psychological Bulletin,136(3), 351–374.
Griffin, D. W., & Bartholomew, K. (1994). Relationship Scales Questionnaire. London, UK:
Jessica Kingsley Publishers.
Hacker, T. (2008). The relational compassion scale: Development and validation of a new
self-rated scale for the assessment of self-other compassion. (Doctoral dissertation)
University of Glasgow.
Hogan, R. (1969). Development of an empathy scale. Journal of Consulting and Clinical
Psychology,33(3), 307–316.
Hwang, J. Y., Plante, T., & Lackey, K. (2008). The development of the Santa Clara brief
compassion scale: An abbreviation of Sprecher and Fehr's compassionate love scale.
Pastoral Psychology,56(4), 421–428.
Kanov, J. M., Maitlis, S., Worline, M. C., Dutton, J. E., Frost, P. J., & Lilius, J. M. (2004). Com-
passion in Organizational Life. American Behavioral Scientist47(6), 808–827. http://
dx.doi.org/10.1177/0002764203260211.
Keltner, D. (2009). Born to be good: The science of a meaningful life. New York: Norton.
Klimecki,O. M., Leiberg, S., Ricard, M., & Singer,T. (2014). Differential pattern of function-
al brain plasticity after comp assion and empathy training. Social Cognitive and
Affective Neuroscience,9(6), 873–879.
Dalai Lama (1995). The power of compassion. New Delhi: HarperCollins.
Dalai Lama (2002). How to practice: The way to a meaningful life. New York: Pocket Books.
Dalai Lama (2005). Essence of the heart sutra: The Dalai Lama's heart of wisdom teachings.
USA: Wisdom Publications.
Lazarus, R. S. (1991). Emotion and adaptation. Oxford: Oxford University Press.
Lee, R. M., & Robbins, S. B. (1995). Measuring belongingness: The Social Connectedness
and the Social Assurance scales. Journal of Counseling Psychology,42(2), 232.
López, A., Sanderman,R., Smink, A., Zhang, Y.,van Sonderen, E., Ranchor, A., & Schroevers,
M. J. (2015). A reconsideration of the Se lf-Compassion Scale's total score: Self-
compassion versus self-criticism. PloS One,10(7), e0132940.
Lown, B. A., Muncer, S. J., & Chadwick, R. (2015). Can compassionate healthcare be
measured? The Schwartz Center Compassionate Care Scale™.Patient Education and
Counseling,98,1005–1010.
MacBeth, A., & Gumley, A. (2012). Exploring compass ion: A meta-analysis of the
association bet ween self-compassion and psychopathology. Clinical Psychology
Review,32,545–552.
Martins, D., Nicholas, N. A., Shaheen, M., Jones, L., & Norris, K. (2013). The development
and evaluation of a com passion scale. Journal of Health Care for the Poor and
Underserved,24(3), 1235–1246.
Mehrabian, A., & Epstein, N. (1972). A measure of emotional emp athy. Journal of
Personality,40(4), 525–543.
Mikulincer, M., Shaver, P. R., Gillath, O., & Nitzberg, R. A. (2005). Attachment, Caregiving,
and Altruism: Boosting Attachment Security Increases Compassion and Helping.
Journal of Personality and Social Psychology,89(5), 817–839.
Najjar, N., Davis, L. W., Beck-Coon, K., & Carney Doebbeling, C. (2009). Compassion
fatigue: A review of the research to date and relevanc e to cancer-care providers.
Journal of Health Psychology,14,267–277.
Neff,K.D.(2003a).Self-compassion: An alternati ve conceptualization of a health y
attitude toward oneself. Self and Identity,2(2), 85–101.
Neff, K. D. (2003b). The development and validation of a sca le to measure self-
compassion. Self and Identity,2(3), 223–250.
Neff, K. D. (2015).The self-compassion scale is a valid and theoretically coherent measure
of self-compassion. Mindfulness,1–11.
Neff, K. D., & Pommier, E. (2013). The relationship between self-compassion and other-
focused concern among college undergraduates, community adults, and practicing
meditators. Self and Identity,12(2), 160–176.
Norko, M. A. (2005). Commentary: Compassion at the core of forensic ethics. Journal of the
American Academy of Psychiatry and the Law,33,386–389.
Pearson, A. (2006). Powerful caring. Nursing Standard (Royal College of Nursing (Great
Britain): 1987),20(48), 20–22.
Penner, L. A., Fritzsche, B. A., Craiger, J. P., & Freifeld, T. R. (1995). Measuring the prosocial
personality. In J.Butcher, & C. D.Spielberger (Eds.), Advances in personality assessment,
Vol. 10 Hillsdale, NJ: LEA.
Plante, T. G., & Boccaccini, M. T. (1997). The Santa Clara strength of religious faith ques-
tionnaire. Pastoral Psychology,45(5), 375–387.
Pommier, E. A. (201 0). The compass ion scale (Order N o. 3445994). (Available from
ProQuest Dissertations & Theses Global. (855633530)).
Raes, F., Pommier, E., Neff, K. D., & Van Gucht, D. (2011). Con struction and fa ctorial
validation of a short form of the self- compassion scal e. Clinical Psych ology &
Psychotherapy,18(3), 250–255.
Redelmeier, D. A., Molin, J. P.,& Tibshirani, R. J. (1995). A randomised trial of compassion-
ate care for the homeless in an emergency department. The Lancet,345(8958),
1131–1134.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton UP.
Salovey, P., Mayer, J. D., Goldman, S. L., Turvey, C., & Palfai, T. P. (1995). Emotional atten-
tion, clarity, and repair: Exploring emotional intelligence using the Trait Meta-Mood
Scale. Emotion, disclosure, and health,125,154.
Schieman, S., & Van Gundy,K. (2000). The personaland social links between ageand self-
reported empathy. Social Psychology Quarterly,63,152–174.
Sprecher,S., & Fehr, B. (2005). Compassionate love for close others and humanity. Journal
of Social and Personal Relationships,22(5), 629–651.
Stanton, A. L., Kirk, S. B., Cameron, C. L., & Danoff-Burg, S. (2000). Coping through emo-
tional approach: Scale construction and validation. Journal of Personality and Social
Psychology,78,1150–1169.
Terwee, C. B., Bot, S. D. M., deBoer, M. R., van der Windt, D.A. W. M., Knol, D. L., Dekker, J.,
...deVet,H.C.W.(2007).Quality criteria were proposed for measurement properties
of health status questionnaires. Journal of Clinical Epidemiology,60(1), 34–42.
Underwood, L. G., & Teresi, J. A. (2002). The dailyspiritual experience scale: development,
theoretical description, reliability, exploratory factor analysis, and preliminary con-
struct validity using health-related data. Annals of Behavioral Medicine,24(1), 22–33.
Wang, S. (2005). A conceptual framework for integrating research related to the physiology
of compassion and the wisdom of Buddhist teachings. In P. Gilbert (Ed.), Compassion:
Conceptualisations, research and use in psychotherapy (pp. 75–120). London: Routledge.
Williams,M. J., Dalgleish,T., Karl, A., & Kuyken,W. (2014). Examining the factor structures
of the five facet mindfulness questionnaire and the self- compassion scale.
Psychological Assessment,26(2), 407–418.
Wispe, L. (1991). The psychology of sympathy. New York: Plenum.
Wu, Y., Larrabee, J. H., & Putman, H. P. (2006). Caring behaviors inventory: A reduction of
the 42-item instrument. Nursing Research,55(1), 18–25.
27C. Strauss et al. / Clinical Psy chology Revi ew 47 (2016) 15–27
