Clinical Gerontologist

ISSN: 0731-7115 (Print) 1545-2301 (Online) Journal homepage: http://www.tandfonline.com/loi/wcli20

Self-Compassion, Coping Strategies, and Caregiver

Burden in Caregivers of People with Dementia

Joanna Lloyd, Jane Muers, Tom G Patterson & Magdalena Marczak

To cite this article: Joanna Lloyd, Jane Muers, Tom G Patterson & Magdalena Marczak (2018):
Self-Compassion, Coping Strategies, and Caregiver Burden in Caregivers of People with Dementia,
Clinical Gerontologist, DOI: 10.1080/07317115.2018.1461162

To link to this article: https://doi.org/10.1080/07317115.2018.1461162

Accepted author version posted online: 30

Apr 2018.
Published online: 03 May 2018.

Submit your article to this journal

Article views: 6

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=wcli20
CLINICAL GERONTOLOGIST
https://doi.org/10.1080/07317115.2018.1461162

Self-Compassion, Coping Strategies, and Caregiver Burden in Caregivers of

People with Dementia
Joanna Lloyda,b, Jane Muersa,b, Tom G Pattersona,b, and Magdalena Marczaka,b
a
School of Psychological, Social and Behavioural Sciences, Coventry University, Coventry, UK; bFaculty of Health and Life Sciences, Coventry
University, Coventry, UK

ABSTRACT KEYWORDS
Objective: Caring for someone with dementia can have negative consequences for caregivers, Caregivers; caregiver
a phenomenon known as caregiver burden. Coping strategies influence the impact of caregiv- burden; coping strategies;
ing-related stress. Specifically, using emotion-focused strategies has been associated with dementia; self-compassion
lower levels of burden, whereas dysfunctional strategies have been related to increased
burden. The concept of self-compassion has been linked to both positive outcomes and the
coping strategies that are most advantageous to caregivers. However, as yet, no research has
studied self-compassion in caregivers. Therefore, the aim of this study was to explore the
relationship between self-compassion, coping strategies and caregiver burden in dementia
caregivers.
Method: Cross-sectional survey data was collected from 73 informal caregivers of people with
dementia recruited from post-diagnostic support services and caregiver support groups.
Results: Self-compassion was found to be negatively related to caregiver burden and dysfunc-
tional coping strategies and positively related to emotion-focused coping strategies.
Dysfunctional strategies mediated the relationship between self-compassion and caregiver
burden, whereas emotion-focused strategies did not.
Conclusion: Caregivers with higher levels of self-compassion report lower levels of burden and
this is at least partly due to the use of less dysfunctional coping strategies.
Clinical Implications: Interventions that develop self-compassion could represent a useful
intervention for struggling caregivers.

Introduction Dementia Context

Caring for a person with dementia can be Dementia is a disorder that involves a global decline
stressful and can result in negative physical in intellectual functioning affecting memory, plan-
and psychological consequences for carers, a ning, judgement, and self-care skills and also affects
phenomenon known as carer burden. Coping personality and behavior (American Psychological
strategies have a mediating role in the impact of Association (APA), 2000). There are a number of
caregiving-related stress. Specifically, using emo- causes with the most common including Alzheimer’s
tion-focused strategies has been associated with disease, vascular dementia, and Lewy body disease.
lower levels of burden, whereas dysfunctional There are currently over 46 million people living
strategies have been related to increased burden. with dementia worldwide and this figure is predicted
The concept of self-compassion (being kind to to rise reaching 131.5 million by 2050 (Alzheimer’s
oneself when things go wrong) has been linked Disease International, 2015). The majority of these
to positive outcomes along with the coping stra- people are cared for by informal caregivers such as
tegies that appear to be most beneficial to carers family members, friends, or neighbors (Knapp &
(Neff, Hsieh, & Dejitterat, 2005). This study Prince, 2007).
aimed to explore the relationship between self- Caregivers of a person with dementia have to
compassion, coping strategies, and carer burden cope with the cognitive decline and behavioral
in dementia caregivers. changes that accompany the condition, whilst

CONTACT Joanna Lloyd [email protected] Faculty of Health and Life Sciences, Coventry University, Priory Street, Coventry, CV1 5FB
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/wcli.
© 2018 Taylor & Francis Group, LLC
2 J. LLOYD ET AL.

also having to manage the loss of the relationship model (Lazarus & Folkman, 1984). This suggests
with the person as they used to be. It is therefore that stressful events alone do not determine the
not surprising that research indicates that caring intensity of the negative outcome. Instead, the
for a friend or relative with dementia can be impact of stress is mediated by the person’s
stressful and detrimental to both the caregivers’ appraisal of the stressor and the coping resources
physical and psychological wellbeing (Bell, Araki, they employ to manage it. This model has been
& Neumann, 2001; Etters, Goodall & Harrison, extended specifically to understand the process of
2008; Gallagher-Thompson & Powers, 1997). caregiver stress (Pearlin, Mullan, Semple, & Skaff,
The concept of caregiver burden has received 1990). The authors suggest that a number of
considerable attention in the literature. It is a domains make up this process and these interact
complex construct resulting from the interaction on multiple levels forming a complex process that
between patient risk factors (e.g., level of impair- varies widely among caregivers. The authors sug-
ment, social support) and caregiver risk factors gest that four domains make up this process: back-
(e.g., poor health, length of time caregiving) ground and contextual factors, the stressors,
(Adelman, Tmanova, Delgado, Dion, & Lachsm, mediators of stress, and the outcomes. Contextual
2014; Gaugler et al., 2011). The relational context factors are concerned with key characteristics of
of the caregiving situation is also key with lower the caregiver such as age, gender, and social eco-
levels of burden being associated with higher satis- nomic status as well as caregiver related informa-
faction with the relationship between caregiver tion such as length of time caring. Stressors are
and the person with dementia (Springate & considered to fall into two categories: primary
Tremont, 2014). Furthermore, authors have distin- stressors which are defined as the physical
guished between objective burden, the physical demands of the caregiving role and secondary
aspects of caregiving and subjective burden, the stressors, which are the psychological strains asso-
psychological consequences of being a caregiver ciated with caregiving. Mediating factors include
(Van Der Lee, Bakker, Duivenvoorden, & Dröes, coping strategies and social support and outcomes
2014; Zarit & Zarit, 1982). These appear to have involve the mental and physical health of the care-
differing if overlapping predictors (Pinquart & givers as well as their ability to continue in their
Sorensen, 2011). For example, the degree of sever- role. It is hypothesized that these domains interact
ity of behavioral issues displayed by the person on multiple levels forming a complex process that
with dementia is a key predictor of both subjective varies widely among caregivers.
and objective burden whereas higher caregiver This model has been widely used in the litera-
education is associated with higher subjective bur- ture to underpin research into caregiver burden.
den, but fewer caregiver hours, a measure of objec-
tive burden (Hughes et al., 2014).
Coping Strategies
Burden is also both a consequence of caregiving
and a predictor of poorer mental health for care- Coping strategies are the means by which people
givers (Razani et al., 2014) and poorer outcomes manage stress. There have been numerous
for people with dementia (Afram et al., 2014). attempts in the literature to define and organize
However, despite this association, not all care- different categories of coping strategies (for a
givers appear to suffer in these ways and many review see Skinner, Edge, Altman, & Sherwood,
cope well with their role (Kramer, 1997a). 2003). The most consistently used were initially
Understanding the processes that underpin these proposed by Lazarus and Folkman (1984) as part
different reactions has therefore been a priority for of the stress-process model. Two broad categories
researchers in this area. of strategies were suggested: emotion-focused and
problem-focused. Emotion-focused strategies refer
to processes that serve to reduce the emotional
Stress-Process Model
distress associated with the stressor, for example
The most widely used paradigm for understanding through acceptance, positive restructuring, and
how people cope with stress is the stress-process humor. Problem-focused strategies look to try to
 CLINICAL GERONTOLOGIST 3

change the situation for the better. These include 2003; Li, Cooper, Bradley, Shulman &
generating alternative solutions, planning and tak- Livingstone, 2012). It seems likely that many pro-
ing action to resolve or circumvent the stressor. blems faced by caregivers of people with dementia
Which strategy is beneficial depends on the nature are intractable, especially as the duration of caring
and context of the stressor. increases, and therefore people need to adapt emo-
Research has sought to identify distinct types of tionally. Longitudinal research suggests that emo-
strategies that serve particular functions, while also tion-focused strategies buffer caregivers from
aiming to distinguish between helpful and unhelp- developing higher anxiety and increased feelings
ful coping (Carver, Scheier, & Weintraub, 1989). of burden over time (Cooper, Katona, Orrell, &
Using evidence from Lazarus and Folkman’s Livingston, 2008; Vitaliano, Russo, Young, & Teri,
(1984) model as well as a model of behavioral 1991).
self-regulation (Carver & Sheier, 1981), Carver
and colleagues (1989), identified 14 distinct strate-
Summary
gies. These have since been grouped by other
researchers into three categories: emotion-focused The type of coping strategies used by caregivers
strategies including acceptance, emotional support, appears to influence the impact of the stress of the
humor, positive reframing, and religion; problem- role. In particular, emotion-focused strategies
focused strategies, including active coping, instru- seem to buffer caregivers from the negative impact
mental support, and planning; and a third cate- of stress whereas dysfunctional strategies leave
gory, dysfunctional coping represented less helpful caregivers more susceptible to it. Therefore, it
strategies. These include behavioral disengage- seems appropriate to investigate factors that pro-
ment, denial, self-distraction, self-blame, substance mote adaptive emotion-focused strategies and dis-
use and emotional venting (Coolidge, Segal, Hook, courage dysfunctional coping in order to develop
& Stewart, 2000). interventions that can promote wellbeing and
reduce feelings of burden amongst caregivers of
people with dementia. The present study aims to
Caregivers and Coping Styles
build on the literature by investigating the role of
The three coping styles have all been studied in self-compassion in the caregiver stress process.
carers of people with dementia with emotion-
focused strategies and dysfunctional strategies
Self-Compassion
showing the strongest association with caregiver
burden. At the heart of the concept of self-compassion is
Dysfunctional coping strategies have been consis- the idea of treating oneself kindly when things go
tently linked to higher levels of depression (Kim, wrong. In the same way that people can show
Knight, & Longmire, 2007; Li, Cooper, Bradley, compassion towards others in times of difficulty,
Schulman, & Livingston, 2012), anxiety (Cooper those who are self-compassionate respond to their
et al., 2010) and caregiver burden (Wright, Lund, own problems with self-directed kindness as
Caserta, & Pratt, 1991) as well as lower satisfaction opposed to being self-critical and judgemental
with life (Sun, Kosberg, Kaufman, & Leeper, 2010). (Neff, 2003).
Furthermore, use of more dysfunctional strategies Self-compassion has been conceptualized as
has been shown to mediate the relationship between having three components each of which has two
stressors and depression, anxiety and burden in care- parts, the presence of one and the negation of the
givers, both in cross-sectional (Mausbach et al., other: (a) being kind to oneself as opposed to
2006) and longitudinal studies (Vedhara, Shanks, being self-critical; (b) accepting ones failings as
Wilcock, & Lightman, 2001). part of the larger human experience rather than
However, using more emotion-focused coping seeing them as isolating; and (c) mindfully holding
strategies, such as acceptance, has been linked to painful thoughts and feelings in awareness as
lower levels of depression and anxiety in caregivers opposed to avoiding or over-identifying with
of people with dementia (Kneebone & Martin, them (Neff, 2003a).
4 J. LLOYD ET AL.

Current research indicates that those high in explored this further found that students who
self-compassion tend to score highly on other were higher in self-compassion responded more
measures of wellbeing. Self-compassion has been adaptively and resiliently in the face of a per-
associated with higher life satisfaction and subjec- ceived academic failure (Neff et al., 2005).
tive wellbeing as well as lower anxiety and depres- Notably, they were more likely to use emotion-
sion (Neely, Schallert, Mohammed, Roberts, & focused strategies, such as acceptance or positive
Chen, 2009; Neff, 2004). These positive associa- reinterpretation and were less likely to use dys-
tions appear to continue across the lifespan with functional avoidant strategies such as denial or
studies demonstrating that self-compassion can be mental disengagement. This suggests that one
a predictor of dimensions of positive aging such as mechanism by which self-compassion may act
ego integrity and meaning in life (Philips & is through influencing and adapting coping
Fergusson, 2013). strategies.
Furthermore, self-compassion has been investi-
gated in relation to caregiving. Studies have shown
that professional caregivers (nurses) who are high Summary
in self-compassion deliver more compassionate
Studies indicate that self-compassion is linked to
care and are more resilient against burnout than
positive outcomes across the lifespan and for care-
those lower in self-compassion (Durkin,
givers. In addition, inducing a compassionate
Beaumont, Martin, & Carson, 2016). This trend
mindset appears to be a promising intervention
also appears to be relevant to informal caregivers.
for those experiencing psychological difficulties
Self-compassion was positively related to life satis-
including caregiver stress. Despite this encoura-
faction and hope, and negatively related to depres-
ging evidence, there has been little research thus
sion and stress in parents of children with an
far looking at self-compassion in caregivers of
Autistic Spectrum Disorder (Neff & Faso, 2014).
people with dementia. However, research linking
Preliminary research findings also suggest that a
self-compassion and coping supports this as a
compassionate mindset can be developed (Adams
viable avenue for investigation.
& Leary, 2007). As such, clinical interventions
aimed at increasing self-compassion have begun
to emerge. Initial results suggest that these may
Rationale for Current Study
be helpful in reducing symptoms of depression,
anxiety and self-criticism as well as improving The evidence above suggests that an exploration
participants’ ability to self-soothe (Gilbert & of the relationship between self-compassion,
Proctor, 2006). One intervention study has inves- coping and caregiver burden would inform
tigated the a yoga and compassion meditation understanding of the caregiver stress process.
programme for caregivers of people with Not only does self-compassion demonstrate sig-
Alzheimer’s Disease. Results suggest that care- nificant positive associations with indices of
givers found this useful with the intervention wellbeing, it also appears to be related to better
group demonstrating significant improvements in caregiver outcomes and coping strategies that
quality of life, attention, vitality and self-compas- are most adaptive for caregivers. As a trait,
sion (Danucalov, Kozasa, Alfonso, Galduroz & low levels of self-compassion could signify vul-
Leite, 2016). nerability to caregiver burden and when
Research has shown that people higher in self- induced, a compassionate mindset could repre-
compassion experience less anxiety when con- sent a potential therapeutic intervention to
fronted with stressful events than people low in improve caregiver quality of life. Furthermore,
self-compassion, even when self-esteem is both self-compassion and coping strategies
accounted for (Neff, Kirkpatrick, & Rude, could be considered to form part of the ‘med-
2007). This suggests that self-compassion buffers iating factors’ section of the stress-process
people from the effects of stress and could be model, with carer burden being an outcome,
involved in the coping process. A study that thus the present study could add clarification
 CLINICAL GERONTOLOGIST 5

to processes that occur within and between 20 years and engaging in at least 5 hours of car-
aspects of this model. egiving activities per week (Ablitt, Jones, & Muers,
2008).
Klein (1998) recommends that when using
Aims and Hypotheses
regression analysis, such as that used in media-
The overall aim of the proposed study is therefore tion analysis, 20 participants per variable inves-
to explore the relationship between self-compas- tigated should be sufficient to assess significance.
sion, dysfunctional coping and emotion-focussed The model under investigation has one predictor
coping and caregiver burden. In line with the variable, two possible mediators and one out-
rationale outlined above, four hypotheses are come variable. Therefore, according to this
proposed: recommendation, a sample size of 80 was
deemed sufficient.
H1: Self-compassion will be negatively related to Support services were attended by the
caregiver burden. researcher. The study was explained to care-
givers and questionnaire packs, containing infor-
H2: Self-compassion will be positively related to mation sheets, consent forms and the measures
emotion-focused coping strategies. were handed out to those interested. Additional
packs were mailed to past participants of a local
H3: Self-compassion will be negatively related to caregiver course with a cover letter explaining
dysfunctional coping strategies. the study. Caregivers completed the measures at
home and returned them to the researcher along
H4: The type of coping strategy used will mediate with a signed consent form in a stamped
the relationship between self-compassion and care- addresses envelope provided.
giver burden. A total of 233 packs were distributed. 75 of
these were returned with complete data, repre-
senting a response rate of 32.6%. Of these, 2 did
Methods not meet inclusion criteria, resulting in a sample
Design of 73 caregivers. There are a number of reasons
why caregivers may have chosen not to partici-
A cross-sectional survey research design was used to pate including being busy with caregiving duties,
explore participant’s perceptions of self-compassion, not prioritizing research, and the fact that many
coping strategies and caregiver burden. of them received the study survey in the mail
and therefore had limited opportunities to ask
questions. Research indicates that characteristics
Participants and Procedure
of responders and non-responders actually vary
Ethical approval for the present study was granted little overall, with the main difference being that
by Coventry University Ethics Committee, a responders tend to have a higher involvement in
National Health Service Research Ethics caregiving (Oldenkamp, Wittek, Hagedoorn,
Committee and an NHS Trust Research and Stolk, & Smidt, 2016). The majority were female
Development service. Recruitment took place (n = 54) which is consistent with previous care-
between August 2013 and January 2014. giver research. Ages ranged from 39 to 87 years
Caregivers were recruited from post-diagnostic (mean = 67.21, SD = 11.47). The majority of
support services run within a UK NHS Mental participants described their ethnicity as White
Health Trust as well as from third Sector support British (94.5%) with the remainder identifying
services including the Alzheimer’s Society and themselves as White non-European (2.4%),
independent caregiver groups. To be eligible for other (1.4%) or did not want to say (1.4%).
inclusion in the study, participants were to be self- Over two-thirds of caregivers were spouses
identified caregivers of a friend of relative with (69.9%) with the next biggest group being chil-
dementia. They were to be over the age of dren (20.5%). The remainder were siblings
6 J. LLOYD ET AL.

Table 1. Means and standard deviations of study variables.

Self-compassion (SCS-SF) Emotion-focused Coping Dysfunctional coping Caregiver Burden (ZBI)
Gender
Male (N = 19) 42.89 (7.72) 24.74 (5.13) 19.16 (4.31) 20.74 (9.64)
Female (N = 52) 37.31 (8.09) 23.56 (4.91) 20.94 (5.30) 24.71 (8.48)
Relationship to Care Recipient
Spouse (N = 51) 39.61 (8.31) 23.78 (5.10) 20.14 (5.15) 24.41 (9.10)
Child (N = 15) 36.80 (8.73) 23.80 (4.68) 20.87 (4.98) 21.27 (8.61)
Sibling (N = 2) 35.00 (1.41) 20.50 (6.36) 25.00 (7.07) 20.50 (.71)
Other (N = 3) 37.67 (10.02) 28.00 (1.73) 21.00 (4.36) 24.67 (11.06)
Age, years
< 65 (N = 26) 38.73 (8.40) 23.96 (5.00) 21.54 (5.12) 22.23 (10.11)
66–73 (N = 23) 36.65 (7.94) 24.43 (4.69) 22.35 (6.98) 27.48 (7.64)
74+ (N = 24) 40.50 (8.31) 23.25 (5.74) 18.67 (4.91) 20.29 (8.48)
Education Level
None (N = 16) 38.06 (7.50) 23.00 (5.05) 21.69 (7.67) 24.63 (9.00)
O Level/GCSE (N = 23) 41.78 (9.10) 24.28 (4.67) 19.52 (5.58) 19.48 (9.12)
Higher School Cert. (N = 2) 32.50 (2.12) 18.50 (4.95) 22.50 (.71) 15.00 (21.21)
A Level/School Cert. (N = 12) 34.92 (6.17) 24.83 (5.02) 20.83 (3.07) 26.17 (8.46)
Degree (N = 12) 39.25 (9.61) 25.00 (6.34) 21.83 (5.36) 27.50 (5.84)
Other (N = 8) 37.13 (6.90) 22.13 (4.58) 21.13 (7.59) 22.63 (9.78)

(4.1%) or ‘other’ (5.5%) including daughters-in- Participants respond on a 4-point Likert-type scale
law and friends (see Table 1). where 1 = “I haven’t been doing this at all” and
4 = “I’ve been doing this a lot.” The present study is
concerned with emotion-focused and dysfunctional
Materials strategies, and therefore only these subscales were
Self-Compassion used. These have been investigated with caregivers of
The 12-item Self-Compassion Scale – Short Form people with dementia and demonstrate good internal
(SCS-SF; Raes, Pommier, Neff, & Van Gucht, 2011) consistencies (Emotion-Focused α = .72 and
was used to assess self-compassion. Example items Dysfunctional α = .75; Cooper, Katona, &
included, ‘I try to be understanding and patient Livingston, 2008). The present study demonstrated
towards those aspects of my personality that I don’t internal reliability similar to previous research
like,’ and ‘when something painful happens, I try to (Emotion-Focused α = .67 and Dysfunctional α = .74).
take a balanced view of the situation.’ Responses were
recorded on a 5-point Likert-type scale where Caregiver burden
1 = almost always and 5 = almost never. Negative The short form of the Zarit Burden Interview (Bédard
items were reversed scored, and all items were et al., 2001) was used to assess caregiver burden. The
summed to create an overall self-compassion score. original Zarit Burden Interview (ZBI; Zarit, Orr, &
The index has demonstrated good internal consis- Zarit, 1985) is the most consistently used measure in
tency (α = .86) and showed a near perfect correlation the dementia caregiver literature. It considers the most
for with the full scale (r = .98). In the present study, the common problem areas reported by caregivers includ-
internal reliability was α = .55. ing health, psychological wellbeing, finances, social life
and relationship with the care recipient. Participants
Coping strategies record how frequently they experience these issues on
Two subscales of the Brief COPE (Coping a 5-point Likert-type scale where 0 = never and
Orientations to Problems Experienced; Carver, 1997) 4 = nearly always. It demonstrated strong internal
were used. This self-report measure is a shortened consistency with an α coefficient regularly in the .90s
version of the COPE index (Carver et al., 1989) of 28 (McConaghy & Caltabiano, 2005). The short form is a
items, which assesses 14 different coping strategies. 12-item version that has shown excellent correlations
These strategies can be averaged into 3 subscales: with the original (r = .92-.97, Bédard et al., 2001). For
emotion-focused coping, problem-focused coping, the present study, this measure demonstrated good
and dysfunctional coping (Coolidge et al., 2000). internal reliability (α = .87).
 CLINICAL GERONTOLOGIST 7

Results dysfunctional coping strategies and positively corre-

Descriptive Statistics
lated with emotion-focused coping strategies.
Given that levels of self-compassion were signifi-
Analysis was conducted using SPSS version 20. cantly different between men and women in the
Descriptive statistics are presented in Table 1. On sample, further analysis was carried out to control
average, the sample exhibited mild levels of caregiver for gender. Results indicate that controlling for gen-
burden (where 21–40 = mild to moderate burden). der using partial correlation did not affect the rela-
There were no significant differences in the scores tionship between self-compassion and either
between participants of different ages, education emotion-focused (r(68) = .29 (p < .001) or dysfunc-
level, or differing relationship with the care recipient. tional strategies (r(68) = -.53 (p < .001).
However, there was a significant difference between Further to the hypothesized findings, it is note-
scores for men and women on the self-compassion worthy that emotion-focused strategies and dys-
scale (t (69) = 2.69, p = .01) with women scoring functional strategies were not related to each other
significantly less than men. This is in keeping with and only dysfunctional coping strategies were sig-
other research (e.g., Hwang, Kim, Yang, & Yang, nificantly correlated with caregiver burden.
2016). There were no other significant differences
between men and women in their scores on other
measures. Mediation Analysis
The fourth hypothesis predicted that both types of
coping strategies would mediate the relationship
Correlation Analysis between self-compassion and caregiver burden.
Mediation analysis was conducted using an SPSS
Pearson’s correlations between the study variables macro called PROCESS developed by Preacher and
are presented in Table 2. Results of the correlation Hayes (2008). This method uses a bootstrapping
analysis support the first hypothesis, that self-com- procedure to obtain estimates and confidence
passion would be negatively related to caregiver intervals around the indirect effects. Prior to con-
burden. In the total sample, higher levels of self- ducting the regression equations required for med-
compassion were associated with lower levels of iation analysis, the data were screened to
caregiver burden as indicated by the significant determine whether they satisfied the assumptions
negative correlation. of multiple regression analysis. Cook’s D indicated
The second and third hypotheses predicted the that there were two multivariate outliers which
relationship between self-compassion and coping, were subsequently removed from the analysis. All
suggesting that self-compassion would be positively other assumptions were met.
related to emotion-focused coping strategies and The method involves a number of steps relat-
negatively related to dysfunctional coping. These ing to Figure 1. The first is to estimate the effect
hypotheses were also upheld with results showing of self-compassion on both emotion-focused
that self-compassion was negatively correlated with coping and dysfunctional coping (a & d). The
second then requires the estimation of the effects
Table 2. Pearson’s correlations of study variables. of both types of coping on caregiver burden
Emotion- while controlling for the effect of self-compas-
Self- focused Dysfunctional Caregiver sion (b & e). Thirdly, the indirect effect of self-
Variable compassion coping coping Burden
Self- 1
compassion on caregiver burden through each
compassion type of coping is calculated. Finally, a confidence
Emotion- .303** 1 interval (CI) is derived from the empirically
focused
Coping derived bootstrapped sampling distribution of
Dysfunctional −.489** −.028 1 “ab” and “de.” In this case, the total effect was
Coping
Caregiver −.541** −.024 .444** 1
defined as the sum of the indirect effect and
Burden direct effect in a given model (ab+c’ and de
**Correlation is significant at the .01 level (2-tailed). +f’). Using the bootstrap sample, the indirect
8 J. LLOYD ET AL.

c’
Self-compassion Caregiver Burden

a b
Dysfunctional
coping

f’
Self-compassion Caregiver Burden

d e
Emotion-
focused
coping

Figure 1. Illustration of the direct and indirect effects of self-compassion on caregiver burden. c & f = total effect, c’ & f’ = direct
effect, ab & de = indirect effect.

effect (ab & de) or the product of the two Post-hoc Analysis
regression coefficients between self-compassion
Given the significant difference between males and
and caregiver burden through coping style was
females scores on the self-compassion scale, med-
calculated. If the 95% bias-corrected confidence
iation analysis was performed grouped by gender.
interval for the parameter estimate did not con-
No significant differences were found in terms of
tain zero, then the indirect effect was statistically
standardized beta coefficient values within the
significant, and indirect effect was demonstrated
models.
(Preacher & Hayes, 2008).
The total effect of self-compassion on caregiver
burden was significant (b = -.33, t = −5.39, p < .001).
Table 3 shows the direct and indirect effects of self- Discussion
compassion on carer burden with the mediator vari- The aim of the present study was to explore the
ables taken in account. After adjusting for the indirect relationship between self-compassion, coping strate-
effects of the mediator variables, the direct effect gies and caregiver burden. Self-compassion was sig-
remained significant only for dysfunctional coping nificantly related to caregiver burden as predicted by
strategies suggesting partial mediation. Results indi- the first hypothesis. Caregivers who reported high
cate that Emotion-focused coping does not appear to levels of self-compassion experienced less caregiver
act as a mediator. burden than those lower in self-compassion. The sec-
Examining the 95% BCa Confidence Interval con- ond and third hypotheses were also upheld with
firms that self-compassion has a significant indirect higher levels of self-compassion being associated
effect on carer burden through the mediating variable with the use of more emotion-focused coping strate-
of dysfunctional coping strategies (b = -.23 (95% CI gies and fewer dysfunctional coping strategies and vice
-.37, -.12)). versa. The final hypothesis, that the relationship

Table 3. The direct and indirect effects of self-compassion on perceived burden.

Direct Effect: Self-Compassion on Perceived Burden Indirect Effect: Self-Compassion on Perceived Burden
B SE 95% CI B Boot SE Boot 95%CI
Dysfunctional coping −.43* .11 −.65 to -.20 −.23** .06 −.37 to -.12
Emotional coping .0043 .0048 −.0031 to .0166 .0360 .0391 −.0246 to .1361
*p < .001.
**Statistical software did not distinguishe p- values for <.05 for indirect effects.
 CLINICAL GERONTOLOGIST 9

between self-compassion and caregiver burden would were not related to caregiver burden or to dysfunc-
be mediated by the type of coping strategy used was tional strategies. This is not consistent with pre-
partly upheld. Dysfunctional strategies were shown to vious research, which has found emotion-focused
be a partial mediator whereas emotion-focused strate- strategies to correlate significantly with both dys-
gies were not. These results demonstrate that care- functional strategies and caregiver burden
givers who have higher levels of self-compassion are (Cooper, Katona & Livingstone, 2008; Cooper
more likely to use adaptive emotion-focused strategies et al., 2008). However, the same research also
and less likely to use dysfunctional ones. However, demonstrated the emotion-focused strategies tend
only the use of dysfunctional coping strategies con- to behave differently to other types of coping
tributed to increased feelings of caregiver burden. strategy in terms of their relationships with care-
Higher levels of self-compassion reduce the likelihood giver burden and other caregiving factors.
of these strategies being used, protecting caregivers Therefore further research exploring emotion-
from increased burden. focused strategies is warranted.
The results of the present study are consistent with
Pearlin and colleagues’ (1990) stress-process model.
Limitations
Specifically, the results appear to clarify a process
within the mediating factors section of the model. The present study used an opportunity sample of self-
This section of the model includes a broad range of identified caregivers and therefore may not be repre-
factors including stable personality factors and more sentative of all caregivers. Furthermore, the over-
flexible coping responses as well as social support. whelming majority of participants were White
Self-compassion and coping strategies could both be British in ethnicity and therefore the findings of this
considered factors in this section with caregiver bur- study may not generalize to caregivers of other ethnic
den being an outcome variable. This further high- backgrounds.
lights complexity of the caregiving stress process and The present study also relied on self-report mea-
suggests that the mediating factors aspect of this sures and therefore could have been affected by
model merits further attention in future research. socially desirable responding. In addition, caregivers
Previous research has shown that high levels of were not asked about how they cope in specific situa-
self-compassion can buffer people from the effects tions. Instead the brief COPE is a general measure of
of stress (Neff et al., 2007). This is the first study to coping and therefore responses given may not accu-
show the significant relationship between self- rately represent caregiver behavior (Skinner et al.,
compassion and caregiver burden. Self-compas- 2003). A further limitation linked to the measures is
sion appears to protect caregivers from the burden that the internal reliability for the self-compassion
associated with caring. scale was lower than has been demonstrated in pre-
The finding that self-compassion predicts the use of vious studies. Future studies may benefit from using
more emotion-focused coping strategies and fewer the longer version of the measure which has more
dysfunctional ones supports previous research. Neff items per subscale.
and colleagues (2005) found that students with higher The current study employed a cross-sectional
levels of self-compassion responded to a perceived design and therefore it is not possible to demonstrate
academic failure with more emotion based strategies causality. The use of the mediational model is consid-
such as acceptances and positive reframing and fewer ered causal modelling and relies on the variables being
dysfunctional avoidant strategies. The current arranged in the right order. This can be theory driven
research extends this finding by establishing that the and also inferred from previous research. The model
relationship between self-compassion and coping stra- presented in the current study is in line with Pearlin
tegies exists in participants of a broader ranges of ages and colleagues’ (1990) stress-process model.
and education levels and when the stressor is ongoing, Furthermore, self-compassion has been shown to be
as it is in the caregiving situation. related to coping strategies (Neff et al., 2005) and both
Only dysfunctional strategies were found to types of coping strategies have been causally related to
mediate the relationship between self-compassion carer burden (Kneebone & Martin, 2003). Therefore,
and caregiver burden. Emotion-focused strategies it seems likely that the order of variables was correctly
10 J. LLOYD ET AL.

laid out in the mediation analysis for the present study, 2001). It would be interesting to explore the role of
however further experimental research to investigate self-compassion in the differing experiences.
these relationships would be useful.
A general criticism that has been raised with regard
to caregiver research is that caregivers are often treated Clinical Implications
as a homogenous group, when there are often signifi- The finding that self-compassion is signifi-
cant within group differences (Gottlieb & Wolfe, cantly related to caregiver burden highlights
2002). The current study attempted to overcome this an opportunity for both assessment and inter-
by only recruiting caregivers of people with dementia vention in clinical work with caregivers of
and collecting caregiver demographic information to people with dementia. Low levels of self-com-
account for possible confounding variables. However, passion could represent a useful indicator of
not all potentially confounding variables could be people currently experiencing high levels of
covered. In particular, it would have been helpful to burden or those at risk of becoming burdened.
have recorded length of time caring as this is likely to Therefore, introducing an assessment of care-
have an impact on level of burden and coping strate- giver self-compassion could provide useful
gies employed. information about current or future support
needs.
In terms of intervention, compassion-based
Research Implications
therapies are becoming increasingly recognized
Further research exploring the relationships as a helpful and accessible means of supporting
between self-compassion and other variables people with a wide variety of clinical presenta-
within the stress-process model could be a useful tions (Gilbert & Proctor, 2006). These aim to
means of extending this research. For example, help people to develop a more compassionate
examining whether self-compassion demonstrates mindset particularly through reducing self-cri-
a similar or different relationship with objective ticism and self-blame. This appears to relate to
burden as opposed to subjective burden. In addi- the first dimension of self-compassion (Neff,
tion, longitudinal studies would be helpful to con- 2003a) as well as one of the dysfunctional
firm the direction of causality of the variables. The coping strategies (self-blame) and therefore
findings of the present study also do not preclude could be a useful intervention for caregivers
the possibility that other factors may mediate the of people with dementia. More recent develop-
relationship between self-compassion and carer ments include an 8 session mindful self-com-
burden and these would be worthy of passion program aimed specifically at
investigation. increasing self-compassion in both the general
Further research is also needed to explore the con- population and clinical populations (Neff &
ceptualisations, definitions and relationships of the Germer, 2013). Trials of the program have so
different coping strategies to ascertain why emotion- far demonstrated favorable outcomes and so in
focused strategies appear to behave differently to other light of the findings of the present study could
strategies within the stress-process model. be worth extending to caregivers of people
Additionally, it will be important to explore these with dementia. Furthermore, the finding that
variables in specific groups of caregivers with differing dysfunctional strategies mediate the relation-
relationships (e.g., spouses, adult child) and whether ship between self-compassion and burden sug-
or not the relationship between self-compassion and gests that treatment packages that aim to
caregiver burden extends to caregivers of people with reduce the use of dysfunctional coping strate-
other physical or mental health difficulties. It would gies could also be an effective way to support
also be useful to expand this study to caregivers of caregivers of people with dementia.
different ethnic backgrounds. Research has shown
that, generally, people from non-white ethnic back-
grounds tend to appraise caregiving as less stressful ● When assessing caregiver need, low levels
than their white counterparts (Janevic & Connell, of self-compassion could be a useful
 CLINICAL GERONTOLOGIST 11

indicator of caregivers currently experien- American Medical Directors Association, 15(2), 108–116.
cing high levels of burden or at risk of doi:10.1016/j.jamda.2013.09.012
Alzheimer’s Disease International. (2015). Retrieved August
becoming burdened.
22, 2016, from https://www.alz.co.uk/research/
● Compassion-based therapies could repre- WorldAlzheimerReport2015.pdf
sent a useful intervention for caregivers of American Psychiatric Association. (2000). Diagnostic and
people with dementia who are struggling statistical manual of mental disorders, Fourth Edition:
within their role. DSM-IV-TR®. Washington DC: American Psychiatric Pub.
● Developing interventions that aim to Bédard, M., Molloy, D. W., Squire, L., Dubois, S., Lever, J. A.,
& O’Donnell, M. (2001). The Zarit Burden Interview: A
reduce the use of dysfunctional coping stra-
new short version and screening version. The
tegies could also be an effective way to Gerontologist, 41, 652–657. doi:10.1093/geront/41.5.652
support caregivers of people with dementia. Bell, C. M., Araki, S. S., & Neumann, P. J. (2001). The
association between caregiver burden and caregiver
health-related quality of life in Alzheimer disease.
Alzheimer Disease and Associated Disorders, 15, 129–136.
Retrieved from http://arldocdel.iii.com/940611.pdf
Conclusion Carver, C. S. (1997). You want to measure coping but your
protocol’s too long: Consider the brief COPE.
The present study aimed to explore the relation- International Journal of Behavioural Medicine, 4, 92–100.
ship between self-compassion, coping strategies, doi:10.1207/s15327558ijbm0401_6
and carer burden. The results provide the first Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989).
indication that self-compassion is related to carer Assessing coping strategies: A theoretically based
burden and that this relationship is mediated, at approach. Journal of Personality and Social Psychology,
56, 267–283. doi:10.1037/0022-3514.56.2.267
least in part, by the influence of dysfunctional Carver, C. S., & Sheier, M. F. (1981). Attention and self-
coping strategies. More research is needed to regulation: A control-theory approach to human behavior.
explore the role of self-compassion in carers of New York, USA: Springer.
people with a range of physical and mental health Coolidge, F. L., Segal, D. L., Hook, J. N., & Stewart, S. (2000).
difficulties as well as in carers of different ethnic Personality disorders and coping among anxious older
backgrounds. However, the present findings pro- adults. Journal of Anxiety Disorders, 14, 157–172.
doi:10.1016/S0887-6185(99)00046-8
vide a useful starting point from which to begin to Cooper, C., Katona, C., & Livingston, G. (2008). Validity and
develop compassion-based assessments and inter- reliability of the brief COPE in caregivers of people with
ventions for carers, which could serve to reduce dementia. The Journal of Nervous and Mental Disease, 196,
feelings of burden and enable carers to continue in 838–843. doi:10.1097/NMD.0b013e31818b504c
their role in a way that is manageable for both Cooper, C., Katona, C., Orrell, M., & Livingston, G. (2008).
themselves and the person they care for. Coping strategies, anxiety and depression in caregivers of
people with Alzheimer’s disease. International Journal of
Geriatric Psychiatry, 23, 929–936. doi:10.1002/gps.2007
Cooper, C., Selwood, A., Blanchard, M., Walker, Z., Blizard,
References
R., & Livingston, G. (2010). The determinants of family
Ablitt, A., Jones, G., & Muers, J. (2008). Awareness of carer caregivers’ abusive behaviour to people with dementia:
distress in people with dementia. International Journal of Results of the CARD study. Journal of Affective Disorders,
Geriatric Psychiatry, 25, 1246–1252. doi:10.1002/gps.2461 121, 136–142. doi:10.1016/j.jad.2009.05.001
Adams, C. E., & Leary, M. R. (2007). Promoting self-com- Danucalov, M., Kozasa, E., Afonso, R., Galduroz, J., & Leite,
passionate attitudes toward eating among restrictive and J. (2016). Yoga and compassion meditation program
guilty eaters. Journal of Social and Clinical Psychology, 26, improve quality of life and self-compassion in family care-
1120–1144. doi:10.1521/jscp.2007.26.10.1120 givers of patients with Alzheimer’s Disease: A randomized
Adelman, R., Tmanova, L., Delgado, D., Dion, S., & Lachsm, control trial. Geriatrics Gerontology International.
M. (2014). Caregiver burden: A clinical review. Journal of doi:10.1111/ggi.12675
the American Medical Association, 311(10), 1052–1059. Durkin, M., Beaumont, E., Martin, C. J. H., & Carson, J.
doi:10.1001/jama.2014.304 (2016). A pilot study exploring the relationship between
Afram, B., Stephan, A., Verbeek, H., Bleijlevens, M., Suhonen, self-compassion, self-judgement, self-kindness, compas-
R., Sutcliffe, C., . . . Hamers, J. (2014). Reasons for institu- sion, professional quality of life and wellbeing among UK
tionalization of people with dementia: Informail caregiver community nurses. Nurse Education Today, 46, 109–114.
reports from 8 European countries. Journal of the doi:10.1016/j.nedt.2016.08.030
12 J. LLOYD ET AL.

Etters, L., Goodall, D., & Harrison, B. (2008). Caregiver Li, R., Cooper, C., Bradley, J., Schulman, A., & Livingston, G.
burden among dementia patient caregivers: A review of (2012). Coping strategies and psychological morbidity in
the literature. Journal of the American association of nurse family carers of people with dementia: A systematic review
practitioners. 20(8), 423–428. doi:10.1111/j.1745- and meta-analysis. Journal of Affective Disorders, 139, 1–
7599.2008.00342.x 11. doi:10.1016/j.jad.2011.05.055
Gallagher-Thompson, D., & Powers, D. V. (1997). Primary Mausbach, B. T., Aschbacher, K., Patterson, T. L., Ancoli-
stressors and depressive symptoms in caregivers of demen- Israel, S., Von Kanel, R., Mills, P. J., . . . Grant, I. (2006).
tia patients. Aging and Mental Health, 1, 248–255. Avoidant coping partially mediates the relationship
doi:10.1080/13607869757146 between patient problem behaviors and depressive symp-
Gaugler, J., Wall, M., Kane, R., Menk, J., Sarsour, K., toms in spousal Alzheimer caregivers. American Journal of
Johnston, J., . . . Newcomer, R. (2011). Does caregiver Geriatric Psychiatry, 14, 299–306. doi:10.1097/01.
burden mediate the effects of behavioural disturbance on JGP.0000192492.88920.08
nursing home admission? The American Journal of McConaghy, R., & Caltabiano, M. L. (2005). Caring for a
Geriatric Psychiatry, 19, 497–506. doi:10.1097/ person with dementia: Exploring relationships between
JGP.0b013e31820d92cc perceived burden, depression, coping and wellbeing.
Gilbert, P., & Procter, S. (2006). Compassionate mind train- Nursing and Health Studies, 7, 81–91. doi:10.1111/j.1442-
ing for people with high shame and self-criticism: 2018.2005.00213.x
Overview and pilot study of a group therapy approach. Neely, M. E., Schallert, D. L., Mohammed, S. S., Roberts, R.
Clinical Psychology and Psychotherapy, 13, 353–379. M., & Chen, Y. J. (2009). Self-kindness when facing stress:
doi:10.1002/cpp.507 The role of self-compassion, goal regulation, and support
Gottlieb, B. H., & Wolfe, J. (2002). Coping with family in college students’ well-being. Motivation and Emotion,
caregiving to people with dementia: A critical review. 33, 88–97. doi:10.1007/s11031-008-9119-8
Aging and Mental Health, 6, 325–342. doi:10.1080/ Neff, K. D. (2003). The development and validation of a scale
1360786021000006947 to measure self-compassion. Self and Identity, 2, 223–250.
Hughes, T., Black, B., Albert, M., Gitlin, L., Johnson, D., doi:10.1080/15298860309027
Lyketsos, C., & Samus, Q. (2014). Correlates of objective Neff, K. D. (2004). Self-compassion and psychological well-
and subjective measures of caregiver burden among demen- being. Constructivism in the Human Sciences. 9, 27–37.
tia caregivers: Influence of unmet patient and caregiver Retrieved from http://self-compassion.org/UTserver/pubs/
dementia-related care needs. International Psychogeriatrics, listofpublications.htm
26, 1875–1883. doi:10.1017/s104160214001240 Neff, K. D., & Faso, D. J. (2014). Self-Compassion and Well-
Hwang, S., Kim, G., Yang, J. W., & Yang, E. (2016). The Being in Parents of Children with Autism. Mindfulness, 1–
moderating effects of age on the relationships of self-com- 10. doi:10.1007/s12671-014-0359-2
passion, self-esteem, and mental health. Japanese Neff, K. D., & Germer, C. K. (2013). A pilot study and
Psychological Research. doi:10.1111/jpr.12109 randomized control trial of the midful self-compassion
Janevic, M., & Connell, C. (2001). Racial, ethnic and cul- program. Journal of Clinical Psychology, 69, 28–44.
tural differences in the dementia caregiving experience. doi:10.1002/jclp.21923
The Gerontologist, 41, 334–347. doi:10.1093/geront/ Neff, K. D., Hsieh, Y. P., & Dejitterat, K. (2005). Self-compas-
41.3.334 sion, Achievement Goals, and Coping with Academic
Kim, J. H., Knight, B. G., & Longmire, C. V. F. (2007). The Failure. Self and Identity, 4, 263–287. doi:10.1080/
role of familism in stress and coping processes among 13576500444000317
African American and white dementia caregivers: Effects Neff, K. D., Kirkpatrick, K., & Rude, S. S. (2007). Self-com-
on mental and physical health. Health Psychology, 26, 564– passion and its link to adaptive psychological functioning.
576. doi:10.1037/0278-6133.26.5.564 Journal of Research in Personality, 41, 139–154.
Klein, R. B. (1998). Principles and practice of structural equa- doi:10.1016/j.jrp.2006.03.004
tion modeling. New York, USA: Guilford Press. Oldenkamp, M., Wittek, R. P. M., Hagedoorn, M., Stolk, R.
Knapp, M., & Prince, M. (2007). Dementia UK: Report to the P., & Smidt, N. (2016). Survey nonresponse among
Alzheimer’s Society. London, United Kingdom: Alzheimer’s informal caregivers: Effects on the presence and magni-
Society. Retrieved from http://www.alzheimers.org.uk/site/ tude of associations with caregiver burden and satisfac-
scripts/download_info.php?fileID=2 tion. BMC Public Health, 16, 480. doi:10.1186/s12889-
Kneebone, I. I., & Martin, P. R. (2003). Coping and caregivers 016-2948-6
of people with dementia. British Journal of Health Pearlin, L. I., Mullan, J. T., Semple, S. J., & Skaff, M. M.
Psychology, 8, 1–17. doi:10.1348/135910703762879174 (1990). Caregiving and the stress process: An overview
Kramer, B. J. (1997a). Gain in the caregiving experience: of concepts and their measures. The Gerontologist, 30,
Where are we? What next? The Gerontologist, 37, 218– 583–594. doi:10.1093/geront/30.5.583
232. doi:10.1093/geront/37.2.218 Philips, W., & Fergusson, S. (2013). Self-compassion: A
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and resource for positive aging. The Journals of Gerontology:
coping. New York, USA: Springer. Series 1, 68, 529–539. doi:10.1093/geronb/gbs091
 CLINICAL GERONTOLOGIST 13

Pinquart, M., & Sorensen, S. (2011). Spouses, adult children, Sun, F., Kosberg, J. I., Kaufman, A. V., & Leeper, D. (2010).
and children-in-law as caregivers of older adults: A meta- Coping strategies and caregiving outcomes among rural
analytic comparison. Psychology and Aging, 26, 1–14. dementia caregivers. Journal of Gerontological Social Work,
doi:10.1037/a0021863 53, 547–567. doi:10.1080/01634372.2010.496823
Preacher, K., & Hayes, A. (2008). Asymptotic and resampling Van Der Lee, J., Bakker, T., Duivenvoorden, H., & Dröes, A.
strategies for assessing and comparing indirect effects in (2014). Multivariate models of subjective caregiver burden
multiple mediator models. Behaviour research methods. 40, in dementia: A systematic review. Ageing Research Reviews,
879–891. https://doi.org/10.3758/BRM.40.3.879 15, 76–93. doi:10.1016/j.arr.2014.03.0031
Raes, F., Pommier, E., Neff, K. D., & Van Gucht, D. (2011). Vedhara, K., Shanks, N., Wilcock, G., & Lightman, S. L.
Construction and factorial validation of a short form of the (2001). Correlates and predictors of self-reported psycho-
Self-Compassion Scale. Clinical Psychology and logical and physical morbidity in chronic caregiver stress.
Psychotherapy., 18, 250–255. doi:10.1002/cpp.702 Journal of Health Psychology, 6, 101–119. doi:10.1177/
Razani, J., Corona, R., Quilici, J., Matevosyan, A., Funes, C., 135910530100600108
Larco, A., . . . Lu, P. (2014). The effects of declining func- Vitaliano, P. P., Russo, J., Young, H. M., & Teri, L. (1991).
tioning abilities in dementia patients and increases in Predictors of burden in spouse caregivers of individuals with
psychological distress on caregiver burden over a one- Alzheimer’s disease. Psychology of Aging, 6, 392–402.
year period. Clinical Gerontologist, 27, 235–252. doi:10.1037/0882-7974.6.3.392
doi:10.1080/07317115.2014.885920 Wright, S. D., Lund, D. A., Caserta, M. S., & Pratt, C. (1991).
Skinner, E., Edge, K., Altman, J., & Sherwood, H. (2003). Coping and caregiving wellbeing: The impact of maladap-
Searching for the structure of coping: A review and cri- tive strategies. Journal of Gerontological Social Work, 17,
tique of category systems for classifying ways of coping. 75–91. doi:10.1300/J083v17n01_07
Psychological Bulletin, 129, 216–269. doi:10.1037/0033- Zarit, S. H., Orr, N. K., & Zarit, J. M. (1985). The hidden victims of
2909.129.2.216 Alzheimer’s disease: Families under stress. New York, USA:
Springate, B., & Tremont, G. (2014). Dimensions of care- New York University Press.
giver burden in dementia: Impact of demographic, Zarit, S. H., & Zarit, J. M. (1982). Families under stress:
mood, and care recipient variables. American Journal of Interventions for caregivers of senile dementia.
Geriatric Psychiatry, 22, 294–300. doi:10.1016/j. Psychotherapy: Theory, Research and Practice, 19, 461–471.
jagp.2012.09.006 doi:10.1037/h0088459