Content uploaded by Martin M Smith
Author content
All content in this area was uploaded by Martin M Smith on Dec 05, 2017
Content may be subject to copyright.
Short Communication
Informant reports add incrementally to the understanding of the
perfectionism–depression connection: Evidence from a prospective longitudinal
study
Simon B. Sherry
a,b,
⇑
, Logan J. Nealis
a
, Matthew A. Macneil
a
, Sherry H. Stewart
a,b
, Dayna L. Sherry
c
,
Martin M. Smith
a
a
Department of Psychology, Dalhousie University, 1355 Oxford Street, Halifax, Nova Scotia, Canada B3H4R2
b
Department of Psychiatry, Dalhousie University, 5909 Veteran’s Memorial Lane, Halifax, Nova Scotia, Canada B3H2E2
c
Queen Elizabeth II Health Sciences Centre, 1276 South Park Street, Halifax, Nova Scotia, Canada B3H2Y9
article info
Article history:
Received 16 July 2012
Received in revised form 17 December 2012
Accepted 9 January 2013
Available online 1 February 2013
Keywords:
Perfectionism
Self-criticism
Depression
Informant
Longitudinal
abstract
Perfectionism is a putative risk factor for depressive symptoms. However, most research in this area uses
cross-sectional designs (which fail to address temporal precedence) and mono-source designs (which are
influenced by various biases). The present study overcomes these limitations by using a novel design
involving both self- and informant reports of self-critical perfectionism (i.e., negative reactions to per-
ceived failures, concern over others’ criticism and expectations, doubts about performance abilities,
and intense self-rebuke). It was hypothesized that self- and informant reports of self-critical perfection-
ism would correlate moderately and that self- and informant reports of self-critical perfectionism would
predict increases in depressive symptoms over time. A sample of 155 target participants and 588 infor-
mants was recruited and studied using a prospective longitudinal design. All study hypotheses were
supported, including evidence that self- and informant reports of self-critical perfectionism each add
incrementally to the understanding of the self-critical perfectionism–depressive symptoms connection.
Informant reports may provide a more complete picture of the self-critical perfectionist and her or his
vulnerability to depressive symptoms.
Ó2013 Elsevier Ltd. All rights reserved.
1. Introduction
Perfectionism is a putative risk factor for depressive symptoms.
Evidence suggests this link persists after controlling for established
predictors of depressive symptoms such as neuroticism (Mackinnon
et al., 2012). Notwithstanding, major gaps still exist in our under-
standing of the perfectionism–depressive symptom connection.
Nearly all studies in this area use cross-sectional, mono-
source designs. Cross-sectional designs fail to address temporal
precedence. Mono-source designs are potentially problematic,
especially when studying undesirable or maladaptive traits
(Klonsky, Oltmanns, & Turkheimer, 2002). Self-report question-
naires assessing people high in perfectionism may be influenced
by self-presentational biases (e.g., appearing perfect) or defensive
responses that promote or conceal certain traits. People high in
perfectionism may become so accustomed to their behavior
(e.g., unrealistic self-expectations) that they fail to accurately
recall or report it. Informant reports can overcome biases in or
limitations of self-perceptions (Vazire, 2006). For example, infor-
mant ratings of perfectionism help to bypass self-enhancing or
self-deprecating biases.
We know of only one perfectionism study involving informants
(Flett, Besser, & Hewitt, 2005). These authors used a cross-
sectional, single informant design where one friend informed on
the other; they found self- and informant reports of perfectionism
were correlated with each other and with depressive symptoms.
Although this study improved on mono-source designs, multiple
informants produce a more reliable, encompassing assessment of
target participants’ traits (Vazire, 2006). Multiple informants par-
ticipate in various roles and situations with target participants;
for example, romantic partners observe different behaviors than
coworkers.
Our study overcomes limitations of prior research. Specifically,
we collected self-reports of perfectionism as well as informant
reports of perfectionism from multiple sources and tested whether
these reports predicted longitudinal increases in depressive
symptoms.
0191-8869/$ - see front matter Ó2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.paid.2013.01.002
⇑
Corresponding author at: Department of Psychology, Dalhousie University,
1355 Oxford Street, Halifax, Nova Scotia, Canada B3H4R2. Tel.: +1 902 494 8070;
fax: +1 902 494 6585.
E-mail address: simon.sherry@dal.ca (S.B. Sherry).
Personality and Individual Differences 54 (2013) 957–960
Contents lists available at SciVerse ScienceDirect
Personality and Individual Differences
journal homepage: www.elsevier.com/locate/paid
1.1. Self-critical perfectionism and depressive symptoms
Three prominent theoretical models in perfectionism research
include psychodynamic models (Blatt, 1995), cognitive-behavioral
models (Frost, Marten, Lahart, & Rosenblate, 1990), and interper-
sonal models (Hewitt & Flett, 1991). Self-critical perfectionism
(negative reactions to perceived failures, concern over others’ crit-
icism and expectations, doubts about performance abilities, and
intense self-rebuke) represents a theoretically based, empirical
synthesis of these models (Dunkley, Zuroff, & Blankstein, 2003).
Self-critical perfectionism is conceptualized as a risk factor that
comes before and contributes to increases in depressive symptoms.
Consistent with this conceptualization, clinical observations (Blatt,
1995) and empirical findings (Dunkley et al., 2003; McGrath et al.,
2012) suggest self-critical perfectionism encapsulates core attri-
butes (e.g., harsh self-criticism) and central preoccupations (e.g.,
evaluative fears) of people at risk for depressive symptoms.
Building upon Flett et al. (2005), we hypothesized Wave 1 self-
and informant reports of self-critical perfectionism would correlate
moderately. We also hypothesized Wave 1 informant reports of
self-critical perfectionism would predict Wave 2 depressive symp-
toms after controlling for Wave 1 self-reports of self-critical perfec-
tionism and Wave 1 depressive symptoms. Based on McGrath et al.
(2012), we hypothesized Wave 1 self-reports of self-critical perfec-
tionism would predict Wave 2 depressive symptoms after control-
ling for Wave 1 informant reports of self-critical perfectionism and
Wave 1 depressive symptoms. These hypotheses draw on research
suggesting each source (self- and informant reports) has unique
information to convey that is important in predicting depressive
symptoms (Vazire, 2006).
2. Method
2.1. Participants
A sample of 155 undergraduates (119 women) was recruited
from Dalhousie University. Participants averaged 20.65 years of
age (SD = 3.03) and were mostly (70.3%) of European descent. We
call these 155 participants targets. A sample of 588 participants
(364 women) was recruited from friends, family, etc. of targets.
We call these 588 participants informants; 39.3% of informants
were friends, 16.3% were mothers, 9.7% were fathers, and 34.7%
were involved in other types of relationships with targets (e.g., girl-
friend). Informants averaged 30.20 years of age (SD = 14.82), and
were mostly (73.3%) of European descent. Informants indicated
they had face-to-face contact with targets an average of 3.70 days
per week (SD = 2.57) and had some other form of contact with tar-
gets (e.g., phone) an average of 5.13 days per week (SD = 2.05).
Most informants (50.9%) reported they knew targets ‘‘extremely
well’’ (M= 3.39, SD = 0.72) on a 5-point scale from 1 (not at all)to
5(extremely well). Informants indicated the average length of their
relationship with targets was 10.46 years (SD = 8.97).
2.2. Measures
2.2.1. Self-critical perfectionism
Self-critical perfectionism was measured using the 9-item self-
criticism subscale (SC; e.g., ‘‘I often find I don’t live up to my
ideals’’) of the Depressive Experience Questionnaire (DEQ; Bagby,
Parker, Joffe, & Buis, 1994), the 5-item short form of the concern
over mistakes subscale (COM; e.g., ‘‘If I fail at school, I am a failure
as a person’’) of Frost’s Multidimensional Perfectionism Scale (FMPS;
Frost et al., 1990), the 4-item doubts about actions subscale (DAA;
e.g., ‘‘I have doubts about the things I do’’) of the FMPS, and the
5-item short form of the socially prescribed perfectionism subscale
(SPP; e.g., ‘‘My family expects me to be perfect’’) of Hewitt and
Flett’s Multidimensional Perfectionism Scale (HFMPS; Hewitt & Flett,
1991). Studies support the reliability and validity of these sub-
scales (McGrath et al., 2012).
Informant reports of self-critical perfectionism were assessed
with modified versions of the DEQ-SC, FMPS-COM, FMPS-DAA,
and HFMPS-SPP. Self-report items (e.g., ‘‘People expect more from
me than I am capable of giving’’) were modified into informant re-
port items (e.g., ‘‘They believe that people expect more from them
than they are capable of giving’’). The number of items (e.g., five
items for the socially prescribed perfectionism subscale) and the
item response options (e.g., 1 = strongly disagree;7=strongly agree)
were unmodified for informant measures. These four informant
measures were created for the present study. Research on their
reliability and validity is therefore unavailable.
2.2.2. Depressive symptoms
Depressive symptoms were measured using the 7-item depres-
sion subscale (e.g., ‘‘I felt down-hearted’’) of the Depression Anxiety
Stress Scale (DASS-D; Lovibond & Lovibond, 1995). Evidence sup-
ports the reliability and validity of this subscale, including research
suggesting the DASS-D specifically measures depressive symptoms
apart from anxiety symptoms and stress (Lovibond & Lovibond,
1995).
2.3. Procedure
Dalhousie University’s Ethics Board approved this study. Our
study involved two waves separated by 28 days. Targets completed
measures of self-critical perfectionism at Wave 1 and measures of
depressive symptoms at Wave 1 and 2. Informants completed
measures of self-critical perfectionism at Wave 1. After Wave 2,
targets received $30 and a 3.0% bonus credit toward a psychology
course or $45; informants were entered in 1 of 20 draws for $50.
Targets provided a list of five informants before starting our study.
Informants met three inclusion criteria: knowing the target for at
least three months, interacting with the target at least twice a
week, and knowing the target reasonably well. Informants were
contacted via email and invited to complete an internet-based
questionnaire. Informants were emailed three times to maximize
participation. Not all informants who were invited to participate
completed our study, meaning the number of informants per target
varied somewhat; 75.9% (588 of 775) of informants invited to par-
ticipate completed our study. For each target, there was an average
of 3.89 informants (SD = 1.36). Overall, 155 (100%) targets finished
Wave 1 and 152 targets (98.1%) finished Wave 2. Not all targets
completed Wave 2 exactly when requested (i.e., 28 days after
Wave 1). On average, Wave 2 occurred 30.11 (SD = 1.88) days after
Wave 1.
2.4. Data analysis
Missing data were minimal for targets and informants (<2.5%).
We imputed missing data with an expectation maximization algo-
rithm. Hierarchical multiple regression analyses were also used to
test hypotheses.
3. Results
3.1. Descriptive statistics
Means for self-report measures were within one standard devi-
ation of means from past studies of undergraduates (McGrath
et al., 2012), indicating consistency with past studies using similar
samples. Alpha reliabilities for self- and informant report measures
958 S.B. Sherry et al. / Personality and Individual Differences 54 (2013) 957–960
were acceptable (P.86; see Table 1); alpha reliabilities for self-
report measures resembled prior work (McGrath et al., 2012).
Comparison means and alphas for informant report measures are
not available as our study represents the first time these informant
report measures have been used.
3.2. Correlations
Indicators of self-critical perfectionism were standardized and
summed before analyses. We aggregated across informant reports
since aggregation of reports from multiple informants leads to
more reliable assessments of personality (Vazire, 2006). For exam-
ple, if informant reports from a roommate, friend, mother, and girl-
friend were available for a target participant, then we combined all
four reports into one aggregated informant report. This resulted in
151 aggregated informant reports, meaning 97.4% of our 155 tar-
gets provided us with at least one informant.
As hypothesized, Wave 1 self- and informant reports of
self-critical perfectionism were correlated (see Table 1). Wave 1
self- and informant reports of self-critical perfectionism were also
correlated with Wave 1 and 2 depressive symptoms.
3.3. Hierarchical multiple regression analyses
As Panel 1 of Table 2 shows, a hierarchical multiple regression
analysis was conducted with Wave 2 depressive symptoms as
the outcome. Wave 1 depressive symptoms were entered as a pre-
dictor in Step 1, followed by Wave 1 self-critical perfectionism
(self-report) in Step 2, and Wave 1 self-critical perfectionism
(informant report) in Step 3. As hypothesized, Wave 1 informant
reports of self-critical perfectionism significantly predicted incre-
mental variance in Wave 2 depressive symptoms after controlling
for Wave 1 self-reports of self-critical perfectionism and Wave 1
depressive symptoms.
As Panel 2 of Table 2 illustrates, another hierarchical multiple
regression analysis was conducted with Wave 2 depressive symp-
toms as the outcome. Wave 1 depressive symptoms were entered
as a predictor in Step 1, followed by Wave 1 self-critical perfection-
ism (informant report) in Step 2, and Wave 1 self-critical perfec-
tionism (self-report) in Step 3. Consistent with hypotheses, Wave
1 self-reports of self-critical perfectionism significantly predicted
incremental variance in Wave 2 depressive symptoms even after
controlling for Wave 1 informant reports of self-critical perfection-
ism and Wave 1 depressive symptoms.
4. Discussion
Though self-perceptions can be accurate, a person’s position as
expert on his or her traits can be compromised by biases (e.g., self-
serving biases). Using self-report questionnaires, Dunkley et al.
(2003) and McGrath et al. (2012) found that people high in self-
critical perfectionism are at risk for depressive symptoms. Our
study complements and extends this research by showing that
(a) self- and informant reports of self-critical perfectionism were
moderately correlated and (b) self- and informant reports of self-
critical perfectionism each add incrementally to the understanding
of the perfectionism-depressive symptoms connection.
Our study supported the convergent validity of the self-critical
perfectionism construct by demonstrating self- and informant
reports correlated moderately (r= .35). This finding is congruent
with Flett et al. (2005), who found self- and informant reports
of perfectionism correlated moderately in a single informant
study where one friend informed on the other. This finding also
intersects with a wider literature suggesting self- and informant
Table 1
Means, standard deviations, alpha reliabilities, and correlations.
Variable MSD
a
12 3 4
1. Self-critical perfectionism
(Wave 1; self-report) – – .90 – .35
⁄⁄⁄
.53
⁄⁄⁄
.42
⁄⁄⁄
2. Self-critical perfectionism
(Wave 1; informant report) – – .95 – .20
⁄
.32
⁄⁄⁄
3. Depressive symptoms
(Wave 1; self-report) 4.01 3.98 .86 – .45
⁄⁄⁄
4. Depressive symptoms
(Wave 2; self-report) 3.87 4.07 .88 –
Note: Self- and informant report subscales comprising self-critical perfectionism were standardized and summed. Self-critical perfectionism has a mean of 0 and standard
deviation of 1.
⁄
p< .05,
⁄⁄
p< .01,
⁄⁄⁄
p< .001.
Table 2
Hierarchical multiple regression analyses predicting wave 2 depressive symptoms.
Panel 1 Panel 2
D
R
2
D
Fb
D
R
2
D
Fb
Step 1 .21 38.83
⁄⁄⁄
Step 1 .21 38.83
⁄⁄⁄
Depressive symptoms .46
⁄⁄⁄
Depressive symptoms .46
⁄⁄⁄
(Wave 1; self-report) (Wave 1; self-report)
Step 2 .05 9.34
⁄⁄
Step 2 .06 11.12
⁄⁄⁄
Self-critical perfectionism .26
⁄⁄
Self-critical perfectionism .24
⁄⁄⁄
(Wave 1; self-report) (Wave 1; informant report)
Step 3 .03 6.66
⁄
Step 3 .02 4.95
⁄
Self-critical perfectionism .19
⁄
Self-critical perfectionism .19
⁄
(Wave 1; informant report) (Wave 1; self-report)
Note: N = 151.
⁄
p< .05,
⁄⁄
p< .01,
⁄⁄⁄
p< .001.
S.B. Sherry et al. / Personality and Individual Differences 54 (2013) 957–960 959
reports of disordered personality traits correlate moderately
(Klonsky et al., 2002).
Self-critical perfectionism represents a constellation of traits
with salient self-evaluative features (e.g., intense self-reproach)
and social-cognitive features (e.g., perceiving that others demand
perfection of oneself). These traits may be conceptualized as inter-
nal mental representations of the self and other people (Blatt,
1995). So, what is it informants see when they think they see a per-
fectionist? Informants may be directly observing (or indirectly
inferring) targets’ self-critical perfectionism through verbally ex-
pressed self-criticism, repetitive checking behaviors, or overt inter-
personal behaviors with perfectionistic themes (e.g., complaints
regarding pressures from others).
Results also indicated informant reports of self-critical perfec-
tionism provide information that is non-redundant with self-
reports of self-critical perfectionism. Wave 1 informant reports of
self-critical perfectionism predicted Wave 2 depressive symptoms
after controlling for Wave 1 self-reports of self-critical perfection-
ism and Wave 1 depressive symptoms, thereby supporting our
hypothesis. These analyses represent a strict test of the self-critical
perfectionism–depressive symptoms connection and support our
assertion self-critical perfectionism may be a premorbid personal-
ity structure that places people at risk for depressive symptoms
(McGrath et al., 2012). Researchers and clinicians should be aware
informant reports of self-critical perfectionism can contribute
incrementally to the understanding of participants and patients.
Tests of self-critical perfectionism as a risk factor for depressive
symptoms are scarce. Consistent with Dunkley et al. (2003), McG-
rath et al. (2012), and our hypotheses, Wave 1 self-reports of self-
critical perfectionism predicted Wave 2 depressive symptoms after
taking into account Wave 1 informant reports of self-critical per-
fectionism and Wave 1 depressive symptoms. Our research also
converges with Dunkley et al. (2003) and McGrath et al. (2012)
to suggest aggregating multiple, lower-order perfectionism facets
into a single, higher-order perfectionism domain (i.e., self-critical
perfectionism) represents one effective way to operationalize the
perfectionism construct.
Our study has limitations. Our sample was primarily female and
Caucasian. It is unclear whether our results generalize to other
samples. Informants were not selected randomly from targets’ so-
cial networks. Targets may have selected informants who viewed
them positively. We also did not collect informant reports of
depressive symptoms, meaning our outcome variable is more
susceptible to self-report biases. Moreover, informants reported
on targets’ internal traits; some of these traits could not be directly
observed (e.g., self-criticism). Informants may be better judges of
external traits (e.g., behaviors) compared to internal traits (e.g.,
thoughts). Informants are also not without potential biases (e.g.,
parents might have ego-protective biases in reporting on children)
or informants may simply lack information about targets.
Self-reports are subject to biases (e.g., defensive responses)
which may compromise the accuracy of information obtained, espe-
cially for participants with undesirable or maladaptive traits such
as self-critical perfectionism (Klonsky et al., 2002). In our study,
self- and informant reports of self-critical perfectionism converged
moderately, and both approaches to assessing self-critical perfec-
tionism predicted longitudinal increases in depressive symptoms.
Acknowledgments
This research was funded by Grants from the Capital Health Re-
search Fund and Social Sciences and Humanities Research Council
of Canada. Funding sources were not involved in writing this
manuscript.
References
Bagby, R., Parker, J., Joffe, R., & Buis, T. (1994). Reconstruction and validation of the
depressive experiences questionnaire. Assessment, 1, 59–68.
Blatt, S. (1995). The destructiveness of perfectionism. American Psychologist, 50,
1003–1020.
Dunkley, D., Zuroff, D., & Blankstein, K. (2003). Self-critical perfectionism and daily
affect. Journal of Personality and Social Psychology, 84, 234–252.
Flett, G., Besser, A., & Hewitt, P. (2005). Perfectionism, ego defense styles, and
depression: A comparison of self-reports versus informant ratings. Journal of
Personality, 73, 1355–1396.
Frost, R., Marten, P., Lahart, C., & Rosenblate, R. (1990). The dimensions of
perfectionism. Cognitive Therapy and Research, 14, 449–468.
Hewitt, P., & Flett, G. (1991). Perfectionism in the self and social contexts. Journal of
Personality and Social Psychology, 60, 456–470.
Klonsky, E., Oltmanns, T., & Turkheimer, E. (2002). Informant-reports of personality
disorder. Clinical Psychology, 9, 300–311.
Lovibond, P., & Lovibond, S. (1995). The structure of negative emotional states.
Behaviour Research and Therapy, 33, 335–343.
Mackinnon, S., Sherry, S., Antony, M., Stewart, S., Sherry, D., & Hartling, N. (2012).
Caught in a bad romance. Perfectionism, conflict and depression in romantic
relationships. Journal of Family Psychology, 26, 215–225.
McGrath, D., Sherry, S., Stewart, S., Mushquash, A., Allen, S., Nealis, L., et al. (2012).
Reciprocal relations between self-critical perfectionism and depressive
symptoms. Canadian Journal of Behavioural Science, 44, 169–181.
Vazire, S. (2006). Informant reports: A cheap, fast, and easy method for personality
assessment. Journal of Research in Personality, 40, 472–481.
960 S.B. Sherry et al. / Personality and Individual Differences 54 (2013) 957–960