Content uploaded by Myriam Mongrain
Author content
All content in this area was uploaded by Myriam Mongrain on Feb 22, 2014
Content may be subject to copyright.
PLEASE SCROLL DOWN FOR ARTICLE
!"#$%&'(#)*+%,&$%-.,/*.&-+-%012%
34"&5#'&6%7+&"%89:
;/2%
<=%;)(.0+'%<>=>
?))+$$%-+(&#*$2%
?))+$$%@+(&#*$2%3$A0$)'#5(#./%/AB0+'%C<DE>FGHF:
IA0*#$"+'%
J.A(*+-K+
L/M.'B&%7(-%J+K#$(+'+-%#/%N/K*&/-%&/-%O&*+$%J+K#$(+'+-%PAB0+'2%=>Q<CFE%J+K#$(+'+-%.MM#)+2%R.'(#B+'%S.A$+6%HQT
E=%R.'(#B+'%4('++(6%7./-./%O=!%HUS6%VW
!"+%U.A'/&*%.M%I.$#(#X+%I$1)".*.K1
IA0*#)&(#./%-+(&#*$6%#/)*A-#/K%#/$('A)(#./$%M.'%&A(".'$%&/-%$A0$)'#5(#./%#/M.'B&(#./2
"((52YY,,,9#/M.'B&,.'*-9).BY$B55Y(#(*+Z)./(+/([(Q<EC<=<GH
!"+%0+/+M#($%.M%$+*MT).B5&$$#./%&/-%.5(#B#$B%+\+')#$+$%M.'%#/-#X#-A&*$
XA*/+'&0*+%(.%-+5'+$$#./
7+&"%89%4"&5#'&&]%R1'#&B%R./K'&#/&
&%@+5&'(B+/(%.M%I$1)".*.K16%^.'_%V/#X+'$#(16%8+"&X#.A'&*%4)#+/)+$%8A#*-#/K6%!.'./(.6%;P%RHU%=IH6
`&/&-&
;/*#/+%5A0*#)&(#./%-&(+2%<>%;)(.0+'%<>=>
!.%)#(+%("#$%?'(#)*+%4"&5#'&6%7+&"%89%&/-%R./K'&#/6%R1'#&Ba<>=>b%c!"+%0+/+M#($%.M%$+*MT).B5&$$#./%&/-%.5(#B#$B%+\+')#$+$
M.'%#/-#X#-A&*$%XA*/+'&0*+%(.%-+5'+$$#./c6%!"+%U.A'/&*%.M%I.$#(#X+%I$1)".*.K16%F2%F6%HQQ%d%HDC
!.%*#/_%(.%("#$%?'(#)*+2%@;L2%=>9=>D>Y=QEHCQG>9<>=>9F=GQGH
VJ72%"((52YY-\9-.#9.'KY=>9=>D>Y=QEHCQG>9<>=>9F=GQGH
Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
The Journal of Positive Psychology
Vol. 5, No. 5, September 2010, 377–389
RESEARCH ARTICLE
The benefits of self-compassion and optimism exercises for individuals vulnerable to depression
Leah B. Shapira*and Myriam Mongrain
Department of Psychology, York University, Behavioural Sciences Building, Room 297, 4700 Keele Street,
Toronto, ON M3J 1P3, Canada
(Received 5 September 2009; final version received 27 July 2010)
The effectiveness of two online exercises intended to help individuals experience (1) self-compassion (n¼63) and
(2) optimism (n¼55) were compared to a control intervention where participants wrote about an early memory
(n¼70). A battery of tests was completed at 1 week following the exercise period, and at 1-, 3-, and 6-month
follow-ups. Both active interventions resulted in significant increases in happiness observable at 6 months and
significant decreases in depression sustained up to 3 months. The interventions were examined in relationship to
dependency and self-criticism, both related to vulnerability to depression. Individuals high in self-criticism
became happier at 1 week and at 1 month in the optimism condition in the repeated measures analysis.
A sensitivity test using multi-level modeling failed to replicate this effect. More mature levels of dependence
(connectedness) were related to improvements in mood up to 6 months in the self-compassion condition.
This study suggests that different personality orientations may show greater gains from particular types of
positive psychology interventions.
Keywords: positive-psychology exercises; self-compassion; optimism; mature and immature dependence;
self-criticism
Introduction
A call has been made by proponents of positive
psychology to empirically study human strengths and
positive emotional well-being, rather than focusing
exclusively on experiences of suffering (Rashid, 2009;
Seligman & Csikszentmihalyi, 2000; Seligman, Steen,
Park, & Peterson, 2005). Previous research has high-
lighted the beneficial effects of positive psychology
exercises, including a recent meta-analysis indicating
that these interventions, overall, increase well-being
and reduce depressive symptoms (Sin & Lyubomirsky,
2009). Mitchell, Stanimirovic, Klein, and Vella-
Brodrick (2009) investigated three positive psychology
interventions disseminated via the internet, and despite
mixed results, reported increased well-being as a result
of a strength-based online exercise. This study exam-
ined the efficacy of two online self-help interventions
that have not been empirically studied to date,
intended to teach individuals to utilize their own
resources in order to enhance emotional well-being,
which we defined as an increase in happiness and
decrease in depression. We were also interested in
determining whether those at risk for depression
may profit from these exercises and whether there
may be differential responses based on the depressive
personality variables tested.
Numerous studies using both clinical and
non-clinical samples have established a solid founda-
tion for the vulnerability to depression entailed by
dependent and self-critical personality styles
(Mongrain & Leather, 2006; Nietzel & Harris, 1990;
Zuroff, Mongrain, & Santor, 2004). Specifically,
dependency and self-criticism have been associated
with both dysphoria and severity of depression in
cross-sectional and longitudinal studies (Luyten et al.,
2007; Mongrain & Leather, 2006; Mongrain, Lubbers,
& Struthers, 2004). Santor and Patterson (2004) found
that these personality orientations were prospectively
related to both the number and duration of mood
disturbances, and Mongrain and Leather (2006) found
neediness and self-criticism to prospectively predict
future episodes of depression.
Self-critical individuals have a self-evaluative and
self-abating nature and often feel extreme guilt and
shame for not living up to the demanding standards
they set for themselves (Blatt, 1974). When self-critics
perceive that they have failed, they criticize them-
selves in a hostile manner, generating feelings of
worthlessness and perpetuating negative affect.
Dependent individuals are characterized by signif-
icant concerns about abandonment and the dissolution
of interpersonal relationships (Blatt, 1974). Their main
*Corresponding author. Email: lshapira@yorku.ca
ISSN 1743–9760 print/ISSN 1743–9779 online
!2010 Taylor & Francis
DOI: 10.1080/17439760.2010.516763
http://www.informaworld.com
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
goal is to be nurtured, loved, and protected (Bornstein,
1992). Two facets of the dependency factor on the
Depressive Experiences Questionnaire (DEQ; Blatt,
d’Afflitti, & Quinlan, 1976) have been further identi-
fied. One facet of dependency is an immature form
involving intense fears of abandonment and rejection,
a sense of helplessness, and extreme anxiety surround-
ing the loss of interpersonal relationships (Blatt,
Zohar, Quinlan, Zuroff, & Mongrain, 1995; Rude &
Burnham, 1995). This form of dependency, coined as
‘neediness’ (Rude & Burnham, 1995), has been asso-
ciated with more severe pathology and a greater risk
for depressive recurrences (Schulte, Mongrain, &
Flora, 2008). Mature dependency, referred to as
‘connectedness,’ (Rude & Burnham, 1995) involves
more adaptive correlates, but is associated with an
inordinate need to please others (Schulte et al., 2008).
Connected individuals can attain positive and trusting
interactions with others, although they still experience
anxiety around specific relationships (Blatt et al., 1995;
Whiffen, Aube, Thompson, & Campbell, 2000).
Finding therapeutic interventions that enhance
well-being for those at risk to depression is integral,
in efforts to prevent distress from escalating to clinical
levels. Recently, self-compassion has been proposed as
an important tool to improve mood and well-being,
and has been incorporated within the therapy setting
(Gilbert & Irons, 2005; Neff, 2003). Self-compassion
has a long historical tradition in Eastern healing
practices and has been described by Neff (2003) as
a perspective in which one espouses a positive view
of himself and his emotional experiences.
Self-compassionate persons, accepting both suffering
and loss, allow themselves to non-judgmentally engage
in self-appraisals rooted in warmth and kindness.
Being compassionate toward the self has been
correlated with positive mental health outcomes,
including lower levels of depression, less self-evaluative
anxiety, less rumination, lowered negative affect, and
higher life satisfaction (Leary, Tate, Adams, Allen, &
Hancock, 2007; Neff, 2003). In a correlational study,
Neff, Rude, and Kirkpatrick (2007) also reported that
self-compassion was linked to a variety of indicators of
emotional well-being, ranging from positive emotions
and optimism to greater ambition and inquisitiveness.
Given these promising findings, further empirical
research into the benefits of self-compassion as a tool
to increase emotional well-being is necessary.
Optimism has also been lauded as important to
enhance emotional well-being. Essentially, optimists
are individuals who generally have positive expecta-
tions about what will happen to them in their lives
(Scheier & Carver, 1985). In addition to numerous
health benefits that accompany being optimistic, a
plethora of studies indicate that optimists are better
able to cope with adversity across a variety of life
domains, using more active and problem-focused
coping techniques (Aspinwall & Taylor, 1992; Nes &
Segerstrom, 2006). Importantly, optimism has also
been associated with lowered psychological distress,
including lower levels of anxiety and depressive
symptoms, and greater subjective well-being (Scheier,
Carver, & Bridges, 2001). Therefore, an intervention
that teaches individuals to visualize a better future
should thus be particularly helpful.
Recently, exercises involving the cultivation of
optimism through the visualization of a positive
potential future were empirically investigated. King
(2001), for example, found that compared to other
control writing interventions, undergraduates who
wrote about their future from a positive perspective
for 4 days, increased in subjective well-being after
3 weeks and evidenced fewer incidences of illnesses at
5 months post-test. Given these encouraging findings,
further research is needed into the effectiveness of such
future-oriented exercises, albeit with a more diversified
sample.
O’Hanlon (2006, 2007) described a letter writing
exercise to be utilized in therapy that also focuses on
visualizing a better future. In this intervention, indi-
viduals imagine their future in which issues that are
bothering them at the moment have been resolved,
opening up the possibility that change in their current
situation is attainable, and helping individuals coach
themselves in attaining their goals. Individuals then
write a letter from their future self where these issues
have been settled, describing in detail how they
achieved this positive outcome, and provide themselves
with compassionate advice. To our knowledge, no
study has yet directly examined the effectiveness of this
exercise involving letter writing from a future self.
The current study
This research examined the effect of two interventions
delivered over the internet and designed to develop
self-compassion and optimistic thinking. The interven-
tion consisted of daily exercises delivered over a 7-day
period. Participants in the ‘self-compassion’ condition
wrote a letter about a distressing event that happened
during the day, providing compassion to themselves.
Those in the ‘optimism’ condition visualized a future
where current issues were resolved and gave themselves
advice on how to get there, also in the form of a letter.
These two experimental groups were compared to a
control condition, where participants wrote freely
about an early memory (based on Seligman et al.,
2005). This condition was utilized to control for
positive expectancies for change or demand character-
istics of the study, and to rule out that any improve-
ments in well-being were merely due to the act of
writing or the passage of time. Participants completed
baseline measures of mood and personality and were
378 L.B. Shapira and M. Mongrain
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
followed for a period of 6 months to determine
sustained improvements in emotional well-being.
It was expected that at the conclusion of the 1 week
study period, participants in the self-compassion con-
dition and optimism condition would show greater
increases in emotional well-being, that is, increases in
happiness and decreases in depressive symptoms,
compared to the control condition. These gains were
expected to remain throughout follow-ups at 1-, 3-,
and 6-month post-test. These hypotheses are supported
by previous research examining similar interventions
(Leary et al., 2007; Seligman et al., 2005; Sheldon &
Lyubomirsky, 2006).
This study also explored the possibility that these
interventions would address problematic inner dynam-
ics experienced by individuals presenting with a
vulnerability to depression. We hypothesized that
therapeutic techniques intended to engender self-
compassion would advantageously target documented
deficiencies associated with self-criticism. Self-compas-
sion mitigates negative reactions to distressing events,
particularly when coping with failure and rejection
(Leary et al., 2007), two salient issues for self-critical
and dependent individuals, respectively. More specifi-
cally, self-critics harbor perfectionistic dysfunctional
beliefs (Mongrain & Zuroff, 1989) and have difficulties
engendering feelings of self-compassion (Gilbert,
Baldwin, Irons, Baccus, & Palmer, 2006), both of
which were specifically targeted in the self-compassion
exercise. Thus, a non-judgmental and self-compassion-
ate stance toward oneself should be particularly
valuable for self-critics.
The predictions for dependency incorporated the
distinction made between mature and immature levels
of this trait. Connected individuals are high in com-
munion (Zuroff, Moskowitz, & Cote, 1999) and are
related to both exploitable and loving traits (Pincus &
Wilson, 2001). Consequently, one would expect that
connected individuals should have the ability to
engender self-compassion and profit from its associ-
ated benefits. Developing the ability to soothe oneself
should be helpful for needy individuals who tend to
rely excessively on obtaining reassurance from others
(Mongrain, 1998). The ability to self-soothe could also
improve the relationships of needy individuals (Coyne,
1976; Joiner, Metalsky, F. Gencoz & T. Gencoz, 2001).
However, the self-compassion exercise could be very
difficult for those high on neediness as they may not
have the inner resources necessary to abandon their
reassurance-seeking strategies. Furthermore, since
needy individuals require the presence of others to
maintain emotional well-being (Bornstein, 1992), an
exercise that does not address this need may be
minimally therapeutically relevant to needy individ-
uals. Therefore, we expected a self-compassion-focused
exercise to be particularly relevant for those exhibiting
connected or self-critical traits, while the potential
benefits for needy individuals were more exploratory.
The second intervention aimed at imagining a more
positive future and giving oneself advice and support
should also be particularly helpful for individuals
vulnerable to depression given the negative view of
one’s self, world and future associated with this
condition (Beck, Rush, Shaw, & Emery, 1979). Such
optimistic thinking would be expected to be particu-
larly beneficial for self-critics who emphasize and focus
on their flaws, teaching these individuals more con-
structive ways of construing current issues in their
lives, while still preserving their autonomy. Self-critical
people also typically engage in avoidant coping
(Dunkley, Zuroff, & Blankstein, 2003), and the opti-
mism exercise could teach these individuals more active
coping strategies for currently distressing events.
For connected individuals, we theorized that they
would have the inner resources necessary to engage in
and benefit from the optimism exercise, as this
personality style has been associated with less
psychopathology than neediness (McBride, Zuroff,
Bacchiochi, & Bagby, 2006). Theoretically, this type
of exercise could also be useful for needy individuals,
who tend to feel helpless, by teaching them active
coping skills to deal with difficult life situations
(Mongrain, 1998). However, given their heightened
levels of psychopathology and anxious attachment
style (Whiffen et al., 2000), needy individuals may find
it too difficult or anxiety provoking to move away
from the typical reassurance-seeking strategies that
they utilize.
Method
Participants
The internet sample was recruited primarily through
advertisements placed on Facebook and was open to
all Canadians. Participants (n¼1002) were primarily
female (n¼817), male (n¼164), Caucasian (79.4%),
and earned an average yearly income between $30,000
and $40,000. The age of participants ranged from 18 to
72 (M¼34 years). There was no exclusionary criterion
for the study except for being at least 18 years of age
and Canadian. Baseline mean scores for depressive
symptoms and happiness were 21.5 and 2.62, respec-
tively, with the former suggesting a moderately
distressed sample.
Participant remuneration
Upon completion of the 1 week exercise and post-test
measures, participants registering in the first phase of
the study received 30 dollars as remuneration (n¼324).
In order to increase sample size, data continued to be
collected but subsequent participants were instead
The Journal of Positive Psychology 379
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
entered into a $1000 draw (n¼678). All participants
regardless of date of entry into the study were asked to
complete follow-up assessments at 1-week post-test,
and 1, 3, and 6 months later. After each completed
follow-up, participants were entered into a $1000
raffle.
Measures
Depressive Experiences Questionnaire
(Blatt et al., 1976)
The DEQ is a 66-item questionnaire that is a well
known, reliable, and valid measure of self-criticism and
dependency (Zuroff et al., 2004). The dependency
subscale has been further distinguished at two levels, a
mature subtype (‘connectedness’) and a more imma-
ture form (‘neediness’; Blatt et al., 1995; Rude &
Burnham, 1995). Previous research has found strong
test–retest reliability (Bagby, Parker, Joffe, & Buis,
1994), and considerable internal consistency, with
Cronbach’s alphas ranging between 0.80 to 0.81 for
dependency and 0.75 to 0.77 for self-criticism (Zuroff,
Quinlan, & Blatt, 1990).
Center for Epidemiological Studies Depression Scale
(Radloff, 1977)
This scale (shortly CES-D) is a widely used measure
consisting of 20 items assessing the prevalence of
depressive symptoms over the previous week. A
4-point scale was utilized, ranging from zero (less
than 1 day) to four (most or all of the time). The
CES-D has reportedly strong internal consistency
(Schulte et al., 2008), as well as strong internal
reliability (Radloff, 1977). The internal consistency
for the CES-D in this study was 0.80. The baseline
mean CES-D scores in our sample were in the
moderate range (M¼21.5).
Steen Happiness Index (Seligman et al., 2005)
This index (shortly SHI) is a new measure intending to
measure changes in happiness, in particular, increases
in happiness. The SHI is comprised of 20 items based
on the structure of the Beck Depression Inventory
(BDI), and taps into Seligman’s definitions of happi-
ness including an assessment of pleasure, engagement,
and meaning in life (BDI; Beck, Ward, Mendelson,
Mock, & Erbaugh, 1961; Seligman et al., 2005).
Participants were required to select one of five state-
ments that best depicted their current states, ranging
from a negative to a positive response (e.g., one ¼I am
joyless; five ¼Almost everything about my life fills me
with joy). Seligman et al. (2005) reported convergent
validity with other happiness measures, including
Lyubomirsky and Lepper’s General Happiness
Scale (1999) and Fordyce’s (1977) Happiness Scale.
Peterson and Park (2009) reported that the SHI
evidenced strong internal consistency and test–retest
reliability over time (from the Seligman et al., 2005
data). In this research, internal consistency for the SHI
was very high at 0.95.
Procedure
This study was conducted online through a website
designed by the research team (www.projecthopecanada.
com). The exercises were part of a larger study of online
self-help interventions called ‘Project HOPE’
(Harnessing One’s Personal Excellence). A similar
American research project (The Trustees of the
University of Pennsylvania, 2006) was consulted for
web-based design strategies.
After registering online and indicating informed
consent, participants (n¼1002) were asked to complete
a battery of questionnaires before being randomly
assigned via a computer program to one of the
three conditions (self-compassion, n¼327; optimism
n¼322; or control condition, n¼353). Participants
were briefly informed of their exercise and were told
that (1) starting the next night, they would have to log
on to the website every night for 7 days to complete
their exercises, and (2) on the seventh day they would
complete another battery of questionnaires.
Participants assigned to the self-compassion condi-
tion were provided with a rationale and instructions
for engaging in a daily psychological exercise promot-
ing a supportive, caring, and compassionate stance
toward the self. Participants were asked to think about
an event that occurred that day which was distressing
and left them feeling upset. Next, they were asked to
write a one paragraph letter to themselves in the first
person about the situation. The following instructions
were provided to participants:
To start writing your own letter, try to feel that part of
you that can be kind and understanding of others.
Think about what you would say to a friend in your
position, or what a friend would say to you in this
situation. Try to have understanding for your distress
(e.g., I am sad you feel distressed . . . ) and realize your
distress makes sense. Try and be good to yourself. We
would like you to write whatever comes to you, but
make sure this letter provides you with what you think
you need to hear in order to feel nurtured and soothed
about your stressful situation or event. This letter may
take about 5–15 min to write, and there is no ‘right’ or
‘wrong’ way of doing it.
Participants in the optimism condition were asked
to imagine a positive future in a variety of domains, for
example, one’s life in general, family relationships, or
work and school life. In writing, participants elabo-
rated on this positive future, where current issues were
resolved, and were asked to give themselves sage advice
from their future self. Participants were provided with
380 L.B. Shapira and M. Mongrain
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
a rationale and the following instructions for engaging
in this daily exercise:
Imagine yourself in the future (6 months/1 year/
2 years/5 years/10 years from now – Pick a time
frame that makes sense to you). Imagine you are in a
better place where you have resolved some of the issues
that are concerning you now.
(1) Describe where you are, what you are doing, and
what is happening in your life. Enrich with as
much detail as possible.
(2) Tell yourself the crucial things you realized or the
critical steps you took to get there. Give yourself
some sage and compassionate advice from a better
future.
Participants in the control early memory condition
were asked to think about an early memory and write
about it in as much detail as possible. It was explained
to participants that early life experiences can be
influential later in life, and that reflecting upon and
writing about these experiences may help them gain
insight and understanding into whom they are and
contribute to overall well-being. The instructions for
this exercise were explained:
Describe an early memory in as much detail as
possible. What were you doing? What were you were
feeling? Who else was with you? (If you cannot
remember some of the details, that is OK. Just type
down what you can remember).
Once participants completed the baseline and
1-week assessments, they were sent e-mails at 1-, 3-,
and 6-month post-test encouraging them to return to the
website for a follow-up assessment. Follow-up ques-
tionnaires included the CES-D and SHI. E-mails were
sent at 2 and 4 months post-test with a copy of their
exercise instructions, encouraging participants to con-
tinue practicing the exercise if they found it beneficial.
Data-analytic strategy
The first approach to the data analyses involved
repeated measures analysis of variance (ANOVAs).
This is in keeping with studies with a very similar
design (e.g., Seligman et al., 2005) and appeared to be a
legitimate first step in exploring the data. A drawback
of this approach is that participants with missing data
points are excluded. Those who drop out may not be a
random subset of the full sample, compromising the
generalizability of the results particularly in studies
with high attrition rates. To address these limitations, a
second statistical approach was employed and the
results were contrasted to those obtained with the
repeated measures ANOVA to determine the robust-
ness of our findings.
Imputation methods (intent-to-treat analyses) have
been widely used in controlled clinical trials to address
issues with high attrition rates and missing data
(Lachin, 2000). For example, Last Observation
Carried Forward (LOCF) is often used in clinical
research where participants’ missing values at follow-
up are replaced with their last measurement scores. The
main advantage of this approach is that no participant
is eliminated. It is also one of the few options available
for longitudinal studies with high attrition rates and
small sample sizes (e.g., Mitchell et al., 2009). A
drawback to this approach is the assumption that
participants with only one observation at baseline
would have been ‘treatment failures’ (i.e., their baseline
values are carried forward). This is a conservative
assumption given that those participants may have
improved had they completed the study leading to an
inflated likelihood of making a Type II error in the
evaluation of a given intervention. To avoid ad hoc
imputations and the restrictive assumptions of
intent-to-treat analyses, we considered an alternative
statistical approach.
A recent study compared intent-to-treat analysis to
a mixed model approach in longitudinal clinical trials
with missing data (Chakraborty & Gu, 2009). Based on
a detailed investigation using simulation studies with
different missing value scenarios, linear mixed models
were found to be more powerful than ad hoc imputa-
tion methods. In the case of studies with a high
percentage of missing values, mixed models were
deemed superior. This technique uses all points in a
data set and makes weighted projections for partici-
pants with missing values based on the sample’s overall
trajectory. Time is modeled more accurately, which is
important when follow-up assessments are unevenly
spaced over the study period. Mixed models also
account for the variability among participants in their
trajectories over time, providing more reliable esti-
mates of change. For these reasons, a mixed model
approach was adopted and the results were compared
to those obtained using repeated measures ANOVAs.
Results
Completers vs. non-completers across time
A total of 1002 participants were randomized to one of
the three conditions and they completed the baseline
assessment. There were no baseline differences in
depressive symptoms, happiness, self-criticism, con-
nectedness, neediness, age, or income across the
experimental groups. Of these, 653 (65.2%) individuals
completed the 1-week assessment, rendering them
eligible to complete the three follow-up assessments.
Complete data at all five time points (baseline, 1 week,
1 month, 3 months, 6 months) was available for 203
participants on the CES-D, and for 197 participants on
the SHI. These individuals were categorized as
‘completers,’ regardless of the number of days they
engaged in the exercise over the active intervention
The Journal of Positive Psychology 381
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
period. Thus, the drop-out rate for the study was high
(79.7%; 799 out of 1002).
Independent sample t-tests were conducted at
6 months to discern whether there were differences
between those who completed the study and those who
did not, as our repeated measures analyses only
included participants who completed questionnaires
at all time points. Those who completed the 6-month
follow-up were older, t(986) ¼"4.87, p50.001
(two-tailed), lower on neediness, t(345.24) ¼2.73,
p¼0.01 (two-tailed), and scored lower on the
CES-D, t(361.01) ¼1.99, p¼0.05 (two-tailed) at base-
line. Thus, participants who adhered to the entire
project and remained until the 6-month assessment
were less needy, less depressed, and older at baseline.
These findings suggest that there were systematic
factors involved in the drop-out rate that require
consideration in the interpretation of the study find-
ings and generalizability of the results.
1
Correlations
As can be seen in Table 1, income was positively
correlated with happiness but negatively correlated
with depressive symptoms, neediness, connectedness,
and self-criticism. Participant’s age was negatively
correlated with neediness and self-criticism. Whether
or not participants were paid (32% of the sample
received remuneration) was negatively associated with
depressive symptoms at baseline, as well as connect-
edness. The extent to which participants adhered to
their exercise, as measured by logging on to the
website and electronically submitting an exercise (i.e.,
‘0’ ¼participant did not complete exercise on any night
and ‘1’ ¼participant completed the exercise on one or
more nights) was also positively correlated with being
paid. Thus, given the association between mood,
personality, age, income, adherence, and payment
status, the demographic variables were controlled for
in subsequent analyses to remove these potential design
and demographic confounds.
Repeated measures ANOVAs
Repeated measures ANOVA’s were conducted to
investigate if well-being over time (baseline, 1-week,
1-month, 3-months, 6-months) was predicted by inter-
vention (self-compassion, optimism, and early memo-
ries) and/or personality (neediness, connectedness, and
self-criticism
2
). Separate repeated measures ANOVAs
were performed for the CES-D and SHI at all time
points. Self-criticism, neediness, and connectedness
were split into high and low scores based on their
median. All data were checked for normality and
outliers were detected and removed on the SHI.
To account for the numerous simple effect contrasts,
alpha was set to 0.025. We also checked for homoge-
neity of slopes and no interactions were significant.
Depressive symptoms
The repeated measures analysis with the CES-D as the
outcome measure at all five time points indicated a
significant Time by condition interaction,
F(8, 370) ¼2.15, p¼0.03, !2
p¼0.04, n¼203. This sug-
gests that changes in depressive symptoms occurred
over time depending on the condition assigned to
Table 1. Correlations among study variables at baseline for completers (n¼1002).
1 2 3 4 5 6 7 8 9 10
1. Condition 1.0
2. Paid "0.03 1.0
3. Adherence
a
"0.06 0.31** 1.0
4. Income (n¼974) "0.00 "0.04 "0.02 1.0
5. Age (n¼988) "0.01 "0.05 0.06 0.36** 1.0
6. Neediness
b
(n¼999) 0.03 "0.06 "0.06* "0.21** "0.18** 1.0
7. Connectedness
c
(n¼999) 0.06 "0.07* "0.05 "0.07* "0.04 0.40** 1.0
8. Self-criticism
d
(n¼999) 0.03 "0.06 "0.03 "0.14** "0.09** 0.50** 0.34** 1.0
9. Depressive symptoms
e
(Baseline)
"0.00 "0.12** "0.06 "0.15** 0.00 0.44** 0.27** 0.58** 1.0
10. Happiness
f
(Baseline) "0.02 0.08* 0.06 0.19** 0.01 "0.51** "0.23** "0.65* "0.66** 1.0
Means and frequencies 5.52 32%
paid
20.3%
adhered
$30,00–40,000
(3.84)
33.55 0.21 "0.33 0.33 21.5 2.62
SD 3.71 0.43 11.59 0.83 0.92 1.02 14.57 0.76
Notes: SD, Standard deviations.
a
Coded ‘0’ ¼did not complete any exercises or ‘1’ ¼completed at least one daily exercise.
b,c,d
Neediness, connectedness, and self-criticism scale from the DEQ (Blatt et al., 1976).
e
From the CES-D (Radloff, 1977).
f
From the SHI (Seligman et al., 2005).
*p50.05; **p50.01.
382 L.B. Shapira and M. Mongrain
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
participants (Figure 1). Simple effect contrasts
indicated that at 1 month, t(186) ¼"3.10, p50.001,
and 3 months, t(186) ¼"2.93, p50.001, individuals in
the optimism condition were less depressed than those
in the early memories condition (see Table 2 for means
and standard errors (SEs) of each simple effect contrast
for depressive symptoms).
3
At 3 months, those in the
self-compassion condition were less depressed than
individuals in the early memories condition,
t(186) ¼"2.60, p¼0.01. This provides preliminary
evidence for the effectiveness of the active
interventions.
The repeated measures analysis including the
CES-D as the outcome measure also produced a
significant interaction between Time, self-criticism, and
connectedness, F(4, 184) ¼2.43, p¼0.05, !2
p¼0.05.
This suggests that self-criticism and connectedness
predicted varying levels of change over time, regardless
of condition. Simple effect contrasts revealed that from
baseline to 3 months, t(186) ¼3.64, p50.001,
*^
^
0
5
10
15
20
25
30
35
Baseline 1 Week 1 Month 3 Months 6 Months
Time
Depressive symptoms
Early memories (n=70)
Self-compassion (n=63)
Optimism (n=55)
Figure 1. Bars represent the mean score on the CES-D (Radloff, 1977) for each condition at each time point. A significant time
by condition effect was found, F(8, 370) ¼2.15, p¼0.04. Post hoc simple effect contrasts showed that compared to the early
memories condition, individuals in the optimism condition showed significantly greater reductions in depressive symptoms at
1 month and 3 months (^). Those in the self-compassion condition showed significantly greater reductions in depressive
symptoms at 3 months compared to the control condition (*).
Table 2. Means, SEs, and p-values of simple effect contrasts on the CES-D.
Mean 1 (SE) Mean 2 (SE) p
Time #Condition
(1) Optimism vs. (2) Early memories – 1 month 14.85 (1.36) 20.35 (1.12) 50.001
(1) Optimism vs. (2) Early memories – 3 months 15.34 (1.44) 20.85 (1.19) 50.001
(1) Self-compassion vs. (2) Early memories – 3 months 16.26 (1.30) 20.85 (1.19) 0.01
Time #Self-criticism #Connectedness
(1) High CON/High SC vs. (2) High CON – 3 months 22.23 (1.56) 13.94 (1.55) 50.001
(1) High CON/High SC vs. (2) High SC – 3 months 22.23 (1.56) 15.99 (1.71) 0.01
Time #Condition #Connectedness
High CON: (1) Self-compassion vs. (2) Early memories – 3 months 14.41 (1.87) 20.25 (1.70) 0.02
High CON: (1) Self-compassion vs. (2) Optimism – 6 months 14.44 (2.00) 24.89 (2.25) 50.001
Low CON: (1) Optimism vs. (2) Early memories – 1 month 12.88 (1.84) 20.39 (1.62) 50.001
Low CON: (1) Optimism vs. (2) Early memories – 3 months 12.34 (1.98) 21.06 (1.74) 50.001
Low CON: (1) Optimism vs. (2) Early memories – 6 months 14.19 (2.21) 21.58 (1.87) 0.01
Low CON: (1) Optimism vs. (2) Self-compassion – 1 month 12.88 (1.84) 20.08 (1.79) 0.01
Notes: CES-D values are taken from (Radloff, 1977); CON, connectedness; SC, self-criticism.
The Journal of Positive Psychology 383
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
individuals high in both connectedness and self-
criticism showed less change in terms of depressive
symptoms than highly connected individuals. At 3
months, those high on both personality variables also
showed less improvement than those who were self-
critical, t(186) ¼2.77, p¼0.01. In sum, across condi-
tions, individuals high on both vulnerability measures
had greater difficulty benefiting from the exercise
interventions, and showed less positive change in
terms of depressive symptoms than those low on one
or both of these vulnerability markers.
Furthermore, a Time by condition by connected-
ness interaction was found for depressive symptoms,
F(8, 370) ¼2.12, p¼0.03, !2
p¼0.04. This indicated that
the impact of the exercises on depression depended on
whether an individual was high or low in connected-
ness. Simple effect contrasts revealed that compared to
the early memories condition, participants high on
connectedness profited more, or showed greater
decreases in depressive symptoms in the self-compas-
sion condition at 3 months, t(186) ¼"2.33, p¼0.02.
Furthermore, those high in connectedness profited
more from the self-compassion exercise compared to
the optimism condition at 6 months, t(186) ¼"3.45,
p50.001. In contrast, individuals who were low in
connectedness profited more in the optimism condi-
tion compared to those in the early memories
condition at 1 month, t(186) ¼"3.06, p50.001,
3 months, t(186) ¼"3.30, p50.001, and 6 months,
t(186) ¼"2.61, p¼0.01. Moreover, those low on
connectedness also showed greater decreases in depres-
sion in the optimism condition compared to the self-
compassion condition at 1 month, t(186) ¼"2.81,
p¼0.01. In summary, those high in connectedness
profited most over time from the self-compassion
exercise when compared to the optimism and control
conditions. Conversely, those low on connectedness
profited more from the optimism exercise compared to
the early memories and self-compassion conditions.
No significant main or interaction effects for neediness
were found.
Happiness
A Time by condition interaction was present,
F(8, 358) ¼2.56, p¼0.01, !2
p¼0.05, n¼197, indicating
that changes in happiness over Time was predicted by
exercise condition (Figure 2). Simple effect contrasts
revealed that at 1 week, t(180) ¼2.65, p¼0.01,
3 months, t(180) ¼3.18, p50.001, and 6 months,
t(180) ¼2.45, p¼0.02, individuals in the optimism
condition were happier than those in the early mem-
ories condition (see Table 3 for means and SEs of each
simple effect contrast for happiness). At 3 months,
t(180) ¼2.45, p¼0.02, and 6 months, t(180) ¼3.20,
p50.001, those in the self-compassion condition were
happier than those in the early memory condition,
^
^
^
2.6
2.8
3
3.2
Baseline 1 Week 1 Month 3 Months 6 Months
Time
Happiness
Early memories (n=70)
Self-compassion (n=63)
Optimism (n=55)
**
Figure 2. Bars represent the mean score on the SHI (Seligman et al., 2005) for each condition at each time point. A significant
time by condition effect was evidenced, F(8, 358) ¼2.56, p¼0.01. Post hoc simple effect contrasts showed that compared to the
early memories condition, individuals in the optimism condition (^) showed significantly greater increases in happiness at 1 week,
3 months, and 6 months. Those in the self-compassion condition (*) showed significantly greater increases in happiness at
3 months, and 6 months, compared to the control condition.
384 L.B. Shapira and M. Mongrain
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
again providing evidence for the effectiveness of the
interventions.
A Time by condition by self-criticism interaction
was also present, F(8, 358) ¼2.00, p¼0.05, !2
p¼0.04,
revealing that over time, changes in happiness
depended on both the condition assigned in interaction
with levels of self-criticism. Simple effect contrasts
indicated that at 1 week, t(180) ¼3.84, p50.001, and
1 month, t(180) ¼2.73, p¼0.01, individuals high in
self-criticism became happier in the optimism condi-
tion compared to the early memories condition. At
1 week, individuals high in self-criticism were happier
in the optimism condition as opposed to the
self-compassion condition, t(180) ¼2.93, p50.001.
A Time by condition by connectedness interaction
was also found for happiness, F(8, 358) ¼2.13,
p¼0.03, !2
p¼0.05. This suggests that the changes
over time depended on condition for high or low levels
of connectedness. Simple effect contrasts showed that
those who were high in connectedness were happier in
the self-compassion condition than in the early mem-
ories condition at 3 months, t(180) ¼2.33, p¼0.01,
and 6 months, t(180) ¼3.59, p50.001. Additionally,
those who were high in connectedness were happier in
the self-compassion condition compared to the opti-
mism condition at 6 months, t(180) ¼3.20, p50.001.
Conversely, at 1 month, t(180) ¼3.21, p50.001,
3 months, t(180) ¼3.58, p50.001, and 6 months,
t(180) ¼3.49, p50.001, individuals who were low in
connectedness were happier in the optimism condition
compared to the early memories condition. Individuals
who were low in connectedness were also happier in the
optimism condition compared to the self-compassion
condition at 3 months, t(180) ¼2.55, p¼0.01, and
6 months, t(180) ¼2.37, p¼0.02. As with depressive
symptoms, individuals high in connectedness showed a
greater increase in happiness in the self-compassion
intervention. Conversely, those low on this personality
variable profited more from the optimism condition.
As we found with depressive symptoms, immature
dependence did not affect responses to the
interventions.
Sensitivity analysis using multi-level modeling
The previous results included completers only, or
participants who had provided data at all time
points. Given the high attrition rate in our sample,
the generalizability of our findings is compromised.
A sensitivity analysis involving multi-level modeling
was conducted to determine the robustness and
reliability of our previous findings. The multi-level
models were run on participants who provided data at
baseline and 1 week (n¼639), in order to preserve a
more balanced data set and produce reliable estimates.
Maximum likelihood estimation was used to estimate
parameters in the models.
Depressive symptoms
The model tested included all of the predictors
previously reported for the repeated measures analyses,
and included as fixed effects: income, adherence,
payment status, age, connectedness, self-criticism,
Table 3. Means, SEs, and pvalues of simple effect contrasts on the SHI.
a
Mean 1 (SE) Mean 2 (SE) p
Time #Condition
(1) Optimism vs. (2) Early memories – 1 week 2.96 (0.05) 2.78 (0.04) 0.01
(1) Optimism vs. (2) Early memories – 3 months 3.01 (0.07) 2.74 (0.06) 50.001
(1) Optimism vs. (2) Early memories – 6 months 2.87 (0.07) 2.64 (0.06) 0.02
(1) Self-compassion vs. (2) Early memories – 3 months 2.94 (0.06) 2.74 (0.06) 0.02
(1) Self-compassion vs. (2) Early memories – 6 months 2.93 (0.07) 2.64 (0.06) 50.001
Time #Condition #Self-criticism
High SC
b
: (1) Optimism vs. (2) Early memories – 1 week 3.06 (0.08) 2.65 (0.07) 50.001
High SC: (1) Optimism vs. (2) Early memories – 1 month 3.06 (0.10) 2.73 (0.08) 0.01
High SC: (1) Optimism vs. (2) Self-compassion – 1 week 3.06 (0.08) 2.75 (0.07) 50.001
Time #Condition #Connectedness
High CON
c
: (1) Self-compassion vs. (2) Early memories – 3 months 3.03 (0.08) 2.75 (0.08) 0.01
High CON: (1) Self-compassion vs. (2) Early memories – 6 months 3.11 (0.09) 2.67 (0.08) 50.001
High CON: (1) Self-compassion vs. (2) Optimism – 6 months 3.11 (0.09) 2.66 (0.10) 50.001
Low CON: (1) Optimism vs. (2) Early memories – 1 month 3.05 (0.09) 2.68 (0.07) 50.001
Low CON: (1) Optimism vs. (2) Early memories – 3 months 3.17 (0.09) 2.73 (0.08) 50.001
Low CON: (1) Optimism vs. (2) Early memories – 6 months 3.09 (0.10) 2.61 (0.09) 50.001
Low CON: (1) Optimism vs. (2) Self-compassion – 3 months 3.17 (0.09) 2.84 (0.09) 0.01
Low CON: (1) Optimism vs. (2) Self-compassion – 6 months 3.09 (0.10) 2.75 (0.10) 0.02
Notes:
a
From the SHI (Seligman et al., 2005).
b
SC, self-criticism.
c
CON, connectedness.
The Journal of Positive Psychology 385
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
exercise condition, and higher order interaction effects
between time, condition, and personality. Inspection of
the random effects for this model revealed that the rate
of change among participants, or the slope of the
trajectories varied significantly (Estimate ¼0.92,
SE ¼0.41, z¼2.23, p50.05), controlling for all vari-
ables in the model. As well, participants’ initial status
in terms of depression showed significant variability
(Estimate ¼62.38, SE ¼5.21, z¼11.98, p50.0001).
However, the rate of change for participants did not
covary with their baseline levels of depression
(Estimate ¼"0.59, SE ¼1.25, z¼"0.47, p¼0.64).
This suggests that participants’ depressive status at
the outset of the study did not systematically impact
their rate of change over time.
In terms of the fixed effects, a Time by connected-
ness interaction was obtained (Estimate ¼"0.51,
SE ¼0.25, t¼"1.99, p50.05), and inspection of the
estimates indicated that highly connected individuals
showed greater decreases in depression than those low
on connectedness. Those higher on mature dependency
(connectedness) appear to profit more from the inter-
ventions overall. There was also a Time by connect-
edness by condition interaction effect that was
marginal (Estimate ¼"0.66, SE ¼0.37, t¼1.80,
p¼0.07). Inspection of the estimates indicated that
connected individuals showed greater decreases in
depression in the self-compassion exercise compared
to the optimism intervention. (The optimism condition
was the reference group for interpreting the fixed
effects.) The superiority of the self-compassion condi-
tion for connected individuals was also demonstrated
with the repeated measures ANOVA.
Happiness
The model for happiness included as fixed effects:
income, adherence, payment status, age, connected-
ness, self-criticism, exercise condition, and higher order
interaction effects between time, condition, and per-
sonality. Inspection of the random effects for this
model revealed that the rate of change
(Estimate ¼0.003, SE ¼0.0008, z¼3.67, p50.0001)
and baseline values for happiness (Estimate ¼0.28,
SE ¼0.02, z¼15.22, p50.0001) showed significant
variability across participants. However, the rate of
change for participants did not covary with their
baseline levels of happiness (Estimate ¼0.003,
SE ¼0.003, z¼0.82, p¼0.41). This indicates that
participants’ happiness scores at the outset of the
study did not systematically impact their trajectories or
rate of change over time.
A Time by condition interaction effect was
obtained (Estimate ¼0.05, SE ¼0.01, t¼3.21,
p50.01), and inspection of the estimates indicated
that the self-compassion exercise produced greater
increases in happiness than the control condition
(reference group). There was also a trend for the
optimism condition to show greater improvements in
happiness compared to the controls (Estimate ¼0.03,
SE ¼0.02, t¼1.66, p¼0.09). A Time by connectedness
effect was also obtained (Estimate ¼0.03, SE ¼0.01,
t¼2.38, p50.05) and inspection of the estimates
indicated that highly connected individuals showed
greater increases in happiness overall. Those higher on
mature dependency therefore appear to profit more
from the interventions, overall both in terms of
reduction of depressive symptoms, and improvements
in happiness.
A significant Time by condition by connectedness
interaction was also obtained (Estimate ¼0.03,
SE ¼0.02, t¼1.93, p¼0.05). Using the control group
as the reference point, connected individuals became
happier in the self-compassion condition than in the
control condition. Furthermore, connected individuals
showed greater increases in happiness in the
self-compassion condition compared to the optimism
condition (Estimate ¼0.05, SE ¼0.02, t¼2.85,
p50.01; reference group is the optimism condition).
Echoing the findings previously reported using the
repeated measures ANOVA, individuals with mature
levels of dependency (high on connectedness) made
greater gains in happiness when assigned to a self-
compassion exercise, compared to a control or opti-
mism exercise. No effects for self-criticism were
obtained using the multi-level modeling.
Discussion
This study demonstrated that two brief exercises prac-
ticed on a daily basis for 1 week can lead to increases in
emotional well-being that are sustained over time. That
is, in comparison to the early memories control inter-
vention, repeated measures analyses revealed that
individuals who practiced self-compassion were less
depressed up to 3 months following the exercise period,
and happier up to 6 months. Multi-level modeling
confirmed the improvements in happiness, regardless of
initial levels of happiness. Similarly, individuals who
visualized a positive future showed reductions in
depressive symptoms for up to 3 months and were
happier up to 6 months later. Using multi-level
modeling, the optimism condition also tended to lead
to greater increases in happiness compared to the
control condition. Overall, these results echo previous
findings that optimistic thinking and being
self-compassionate can have advantageous psycholog-
ical benefits (King, 2001; Leary et al., 2007).
Common elements in both active interventions may
have contributed to their efficacy. First, both the self-
compassion and optimism exercises endeavored to
engender hope by teaching individuals to think more
positively about their current distress and their future,
386 L.B. Shapira and M. Mongrain
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
respectively. In line with the ‘broaden and build
theory,’ through generating positive mind states, the
exercises likely increased one’s capacity to counteract
negative emotions and cope with problems when they
arose, in turn further promoting emotional well-being
(Fredrickson, 1998).
Specific therapeutic ingredients unique to each
active intervention may also have contributed to their
positive influence on well-being. In the self-compassion
condition, individuals were asked to engage in expres-
sive writing about a currently distressing topic. When
compared to writing about neutral topics or events in
the past, writing about a distressing issue has benefits
on both physiological and psychological health
(Smyth, 1998). Alternatively, in the optimism exercise,
individuals were asked to imagine being in a place
where they had resolved concerning issues, and to
comment on the critical steps they took to achieve this
resolution. This exercise may have helped participants
engage in active coping by delineating the next steps
they needed to take to achieve their ideal goal state.
This is important as goals that are meaningful and
have clarity are associated with psychological health
(Emmons, 1986; King, 2001). Thus, both common
elements in the interventions, as well as specific aspects
inherent in each exercise, likely contributed to
increases in emotional well-being.
We hypothesized that personality style would influ-
ence outcome depending on what exercise participants
were assigned to. Specifically, we expected self-critics
would benefit psychologically from the self-compassion
and optimism condition. Our findings only partially
support our predictions. That is, in our repeated
measures analyses using completers only who were
older and less depressed, we found some evidence
for the superiority of the optimism condition for
self-critical participants. The positive outcome for this
intervention for completers could lie in its curtailing
effect on rumination, that is, by teaching self-critics to
move away from their tendency to ruminate on past
failures or perceived inadequacies. Optimism is also
associated with approach-oriented coping strategies
(Nes & Segerstrom, 2006), and the optimism exercises
may have been particularly helpful for self-critics by
requiring them to engage in more adaptive coping
strategies. However, our multi-level modeling findings
failed to replicate the effects for self-criticism.
Therefore, we remain tentative in our conclusions
regarding the usefulness of an optimism exercise for
those who are self-critical.
Both our data analytic strategies indicated that
individuals who were high in connectedness profited
most from the self-compassion exercise. Connected
individuals enjoyed greater happiness and reductions
in depression when they were assigned to the self-
compassion condition compared to the early memories
and optimism condition. Conversely, neediness was
not related to any incremental improvements in mood
across both analytic approaches, indicating that the
self-compassion intervention may not be particularly
helpful for those with immature dependence. It is thus
worth noting that a distinction between mature
(connected) and immature (needy) dependence may
lie in the ability to generate compassionate feelings
toward the self. Connected individuals are able to
nurture others, and have the capacity to establish
reciprocal bonds. Hence, this ability could likely be
extended to themselves, helping to explain the thera-
peutic effects of this intervention for both completers
and non-completers, particularly in terms of happiness.
In summary, the current findings highlight the
impact of individual differences on the success of
various positive psychology interventions, and suggest
that exercises could be matched to individuals according
to certain parameters (e.g., achievement and goal-
oriented exercises vs. relational exercises), in efforts to
enhance emotional well-being and potentially bypass
the downward spiral of depression. The results have
practical implications for use in counseling, highlighting
that it is possible to augment emotional well-being in
moderately distressed individuals if they are willing to
engage in positive psychology interventions.
Limitations
Participant attrition is a common problem in web-
based intervention studies (e.g., Mitchell et al., 2009,
who report an 83% attrition rate at 3 months), and was
an important limitation in our study. We endeavored
to account for attrition by utilizing multi-level model-
ing in our analyses. Additionally, our sample was
relatively distressed, with CES-D scores at baseline
that were in the moderate range. Therefore, we may
have lost participants because they suffered from the
motivational deficits that often accompany depression.
It is also important to note that with online self-help
exercises, an impersonal internet interface may con-
tribute to attrition. Looking forward, further research
specifying what works best for whom and introducing
greater variety in exercises assigned could help to
increase compliance and motivation (Lyubomirsky,
Sheldon, & Schkade, 2005; Mongrain, 2009).
Conclusions
This study provides some evidence for the effectiveness
of brief self-help online exercises based on the main
tenets of positive psychology. The fact that these self-
help techniques were online means they have the
potential to easily access many people in need in a
standardized and anonymous format for a greatly
reduced cost, which is critical given the increasing
burden on managed care. The extent to which a
The Journal of Positive Psychology 387
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
positive self-help exercise matches the propensities of
the person who engages in it is an important factor that
influences well-being outcome (Dickerhoof, 2007;
Mongrain, 2009). Future research should strive to
identify which specific individual difference variables
(such as an interpersonal vs. an achievement orienta-
tion) influence what types of self-help exercises, in
order to further improve the ‘fit’ between the needs of
an individual and a specific exercise.
Notes
1. Seligman et al. (2005) reported a considerably lower
drop-out rate (29%). It is important to note sample
differences between the two studies. Participants in this
project were moderately depressed at baseline while
those in the Seligman et al. (2005) study were mildly
depressed at baseline, as measured by the CES-D.
Additionally, the current sample was poorer, younger,
and less educated (Seligman et al., 2005). The average
age of our sample was 34 years while in the Seligman
et al. (2005) study 64% of participants were between 35
and 54 years of age. Given that age (being younger) and
levels of depressed symptoms (more depressed) were
systematically associated with a higher drop-out rate,
these factors were against us in this project.
Furthermore, since the sample for Seligman’s project
consisted of individuals who frequented the website
designed for Seligman’s book, we have reason to believe
they may have been more motivated to comply and
complete his study.
2. Neediness and connectedness were examined in separate
repeated measures ANOVA models due to their consid-
erable overlap (Table 1; r¼0.40).
3. For each simple effect contrast, scores at baseline were
entered as a covariate in order to detect changes in well-
being from baseline to each point in time.
References
Aspinwall, L.G., & Taylor, S.E. (1992). Modeling cognitive
adaptation: A longitudinal investigation of the impact of
individual differences and coping on college adjustment
and performance. Journal of Personality and Social
Psychology, 63, 989–1003.
Bagby, M.R., Parker, J.D.A., Joffe, R.T., & Buis, T. (1994).
Reconstruction and validation of the depressive experi-
ences questionnaire. Assessment, 1, 59–68.
Beck, A.T., Rush, A.J., Shaw, B.E., & Emery, G. (1979).
Cognitive therapy of depression. New York: Guilford Press.
Beck, A.T., Ward, C.H., Mendelson, M.N., Mock, J., &
Erbaugh, J. (1961). An inventory for measuring depres-
sion. Archives of General Psychiatry, 4, 561–571.
Blatt, S.J. (1974). Levels of object representation in anaclitic
and introjective depression. Psychoanalytic Study of the
Child, 24, 107–157.
Blatt, S.J., d’Afflitti, J.P., & Quinlan, D.M. (1976).
Experiences of depression in normal young adults.
Journal of Abnormal Psychology, 85, 383–389.
Blatt, S.J., Zohar, A.H., Quinlan, D.M., Zuroff, D.C., &
Mongrain, M. (1995). Subscales within the dependency
factor of the Depressive Experiences Questionnaire.
Journal of Personality Assessment, 67, 52–71.
Bornstein, R.F. (1992). The dependent personality:
Developmental, social, and clinical perspectives.
Psychological Bulletin, 112, 3–23.
Chakraborty, H., & Gu, H. (2009, March). A mixed model
approach for intent-to-treat analysis in longitudinal clinical
trials with missing values. Research Triangle Park, NC:
RTI Press, RTI Press Publication no. MR-0009-0903.
Coyne, J.C. (1976). Toward an interactional description of
depression. Psychiatry: Journal for the Study of
Interpersonal Processes, 39, 28–40.
Dickerhoof, R. (2007). Expressing optimism and gratitude:
A longitudinal investigation of cognitive strategies to
increase well-being. Dissertation Abstracts International
Section B: The Sciences and Engineering, 68(6-B), 4174,
(UMI no. 3270426).
Dunkley, D., Zuroff, D.C., & Blankstein, K.R. (2003). Self-
critical perfectionism and daily affect: Dispositional and
situational influences on stress and coping. Journal of
Personality and Social Psychology, 84, 234–252.
Emmons, R.A. (1986). Personal strivings: An approach to
personality and subjective well-being. Journal of
Personality and Social Psychology, 51, 1058–1068.
Fordyce, M.W. (1977). Development of a program to
increase personal happiness. Journal of Counseling
Psychology, 24, 511–520.
Fredrickson, B.L. (1998). What good are positive emotions?
Review of General Psychology, 2, 300–319.
Gilbert, P., Baldwin, M., Irons, C., Baccus, J.R., & Palmer,
M. (2006). Self-criticism and self-warmth: An imagery
study exploring their relation to depression. Journal of
Cognitive Psychotherapy: An International Quarterly, 20,
183–200.
Gilbert, P., & Irons, C. (2005). Focused therapies and
compassionate mind training for shame and self-attacking.
In P. Gilbert (Ed.), Compassion: Conceptualisations,
research and use in psychotherapy (pp. 263–325). London:
Routledge.
Joiner, T.E., Metalsky, G.I., Gencoz, F., & Gencoz, T.
(2001). The relative specificity of excessive reassurance-
seeking to depressive symptoms and diagnoses among
clinical samples of adults and youth. Journal of
Psychopathology and Behavioral Assessment, 23, 35–41.
King, L.A. (2001). The health benefits of writing about life
goals. Personality and Social Psychology Bulletin, 27,
798–807.
Lachin, J.M. (2000). Statistical considerations in the
intent-to-treat principle. Controlled Clinical Trials, 21,
167–189.
Leary, M.R., Tate, E.B., Adams, C.E., Allen, A.B., &
Hancock, J. (2007). Self-compassion and reactions to
unpleasant self-relevant events: The implications of treat-
ing oneself kindly. Journal of Personality and Social
Psychology, 92, 887–904.
Luyten, P., Sabbe, B., Blatt, S.J., Meganck, S., Jansen, B.,
De Grave, C., ..., Corveleyn, J. (2007). Dependency and
self-criticism: Relationship with major depressive disorder,
severity of depression, and clinical presentation.
Depression and Anxiety, 24, 586–596.
388 L.B. Shapira and M. Mongrain
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
Lyubomirsky, S., & Lepper, H.S. (1999). A measure
of subjective happiness: Preliminary reliability and
construct validation. Social Indicators Research, 46,
137–155.
Lyubomirsky, S., Sheldon, K., & Schkade, D. (2005).
Pursuing happiness: The architecture of sustainable
change. Review of General Psychology, 9, 111–131.
McBride, C., Zuroff, D., Bacchiochi, J., & Bagby, M. (2006).
Depressive Experiences Questionnaire: Does it measure
maladaptive and adaptive forms of dependency? Social
Behaviour and Personality, 34, 1–16.
Mitchell, J., Stanimirovic, R., Klein, B., & Vella-Brodrick,
D. (2009). A randomised controlled trial of a self-guided
internet intervention promoting well-being. Computers in
Human Behavior, 25, 749–760.
Mongrain, M. (1998). Parental representations and support-
seeking behaviours related to dependency and self-
criticism. Journal of Personality, 66, 151–173.
Mongrain, M. (2009, August). Positive psychology exercises:
Finding what works best for whom. Symposium presented
at the annual meeting of the American Psychological
Association, Toronto, Canada.
Mongrain, M., & Leather, F. (2006). Immature dependence
and self-criticism predict the recurrence of major
depression. Journal of Clinical Psychology, 62, 705–713.
Mongrain, M., Lubbers, R., & Struthers, W. (2004). The
power of love: Mediation of rejection in roommate
relationships of dependents and self-critics. Personality
and Social Psychology Bulletin, 30, 94–105.
Mongrain, M., & Zuroff, D.C. (1989). Cognitive vulner-
ability to depressed affect in dependent and self-critical
college women. Journal of Personality Disorders, 3,
240–251.
Neff, K.D. (2003). The development and validation of a scale
to measure self-compassion. Self and Identity, 2, 223–250.
Neff, K.D., Rude, S.S., & Kirkpatrick, K.L. (2007). An
examination of self-compassion in relation to positive
psychological functioning and personality traits. Journal of
Research in Personality, 41, 908–916.
Nes, L.S., & Segerstrom, S.C. (2006). Dispositional optimism
and coping: A meta-analytic review. Personality and Social
Psychology, 10, 235–251.
Nietzel, M.T., & Harris, M.J. (1990). Relationship of
dependency and achievement/autonomy to depression.
Clinical Psychology Review, 10, 279–297.
O’Hanlon, B. (2006). Change 101. New York: Norton.
O’Hanlon, B. (2007, May). Applying the science of happiness
in the consulting room. Workshop ‘Positive psychology into
practical therapy interventions: Putting strength-based
research to work’ presented by Leading Edge Seminars
at the YMCA, Toronto, Ontario.
Peterson, C., & Park, N. (2009). The authentic happiness
inventory. Unpublished manuscript, University of
Michigan.
Pincus, A.L., & Wilson, K.R. (2001). Interpersonal varia-
bility in dependent personality. Journal of Personality, 69,
223–251.
Radloff, L.S. (1977). The CES-D scale: A self-report
depression scale for research in the general population.
Applied Psychological Measurement, 1, 385–401.
Rashid, T. (2009). Positive interventions in clinical practice.
Journal of Clinical Psychology, 65, 461–466.
Rude, S.S., & Burnham, B.L. (1995). Connectedness and
neediness: Factors of the DEQ and SAS dependency
scales. Cognitive Therapy and Research, 19, 323–340.
Santor, D., & Patterson, R. (2004). Frequency and duration
of mood fluctuations: Effects of dependency, self-criticism,
and negative events. Personality and Individual Differences,
37, 1667–1680.
Scheier, M.F., & Carver, C.S. (1985). Optimism, coping, and
health: Assessment and implications of generalized out-
come expectancies. Health Psychology, 4, 219–247.
Scheier, M.F., Carver, C.S., & Bridges, M.W. (2001).
Optimism, pessimism, and psychological well-being.
In E. Chang (Ed.), Optimism and pessimism: Implications
for theory, research, and practice (pp. 189–216).
Washington, DC: American Psychological Association.
Schulte, F.S., Mongrain, M., & Flora, D.B. (2008). Healthy
and unhealthy dependence: Implications for major depres-
sion. British Journal of Clinical Psychology, 47, 341–353.
Seligman, M.E.P., & Csikszentmihalyi, M. (2000). Positive
psychology. American Psychologist, 55, 5–14.
Seligman, M., Steen, T., Park, N., & Peterson, C. (2005).
Positive psychology progress: Empirical validation of
interventions. American Psychologist, 60, 410–421.
Sheldon, K.M., & Lyubomirsky, S. (2006). How to increase
and sustain positive emotion: The effects of expressing
gratitude and visualizing best possible selves. The Journal
of Positive Psychology, 1, 73–82.
Sin, N.L., & Lyubomirsky, S. (2009). Enhancing well-being
and alleviating depressive symptoms with positive psychol-
ogy interventions: A practice-friendly meta-analysis.
Journal of Clinical Psychology, 65, 467–487.
Smyth, J. (1998). Written emotional expression: Effect sizes,
outcome types and moderating variables. Journal of
Consulting and Clinical Psychology, 66, 174–184.
The Trustees of the University of Pennsylvania (2006).
Authentic happiness. Retrieved from www.
authentichappiness.com
Whiffen, V.E., Aube, J.A., Thompson, J.M., & Campbell,
T.L. (2000). Attachment beliefs and interpersonal contexts
associated with dependency and self-criticism. Journal of
Social and Clinical Psychology, 19, 184–205.
Zuroff, D.C., Mongrain, M., & Santor, D.A. (2004).
Conceptualizing and measuring personality vulnerability
to depression: Comment on Coyne and Whiffen (1995).
Psychological Bulletin, 130, 489–511.
Zuroff, D.C., Moskowitz, D.S., & Cote, S. (1999).
Dependency, self-criticism, interpersonal behaviour, and
effect: Evolutionary perspectives. British Journal of
Clinical Psychology, 28, 231–250.
Zuroff, D.C., Quinlan, D.M., & Blatt, S.J. (1990).
Psychometric properties of the Depressive Experiences
Questionnaire in a college population. Journal of
Personality Assessment, 55, 65–72.
The Journal of Positive Psychology 389
Downloaded By: [Shapira, Leah B.] At: 18:13 21 October 2010
