The Costs of Repression: A Meta-Analysis on the Relation Between Repressive Coping and Somatic Diseases

Abstract

Objective:
When Freud introduced the term repression, he stated its pathogenic potential. Since then, this notion was adapted and continued to date. Surprisingly, there is no attempt to synthesize research on the effect of repressive coping on somatic diseases quantitatively. The current study closes this gap and examines the association between repressive coping as operationalized by Weinberger, Schwartz, and Davidson (1979) and the incidence of cancer, cardiovascular diseases, asthma, and diabetes by applying meta-analytic techniques.

Method:
An extensive literature search identified 22 studies that met all inclusion criteria and assessed a total of 6,775 participants. The study set was analyzed both as a whole sample (k = 22 studies) and disease-specific (k = 2-10 studies; no study found for diabetes).

Results:
The results of the meta-analysis indicate a higher risk for repressive copers to suffer from one of the investigated diseases, especially cancer (logOdds after diagnosis: 0.41; 95% confidence interval, 0.09-0.73, p = .012) and hypertension (logOdds: 0.59; 95% confidence interval, 0.32-0.86, p < .0001). After application of a continuity correction due to a missing control group the results for coronary heart disease imply an increased risk for nonrepressors to be affected. Results for cardiovascular diseases in general, heart attack, and asthma did not reach significance.

Conclusions:
The current meta-analysis revealed significant associations between repressive coping, cancer, and cardiovascular diseases, especially hypertension. These results add to the notion of repressive coping as a consequence of cancer as well as to its important role for the issue of hypertension.

Full text

Running head: REPRESSIVE COPING AND SOMATIC DISEASES 1
The Costs of Repression: A Meta-Analysis on the Relation between Repressive Coping and
Somatic Diseases
Marcus Mund & Kristin Mitte
Friedrich-Schiller-Universität Jena
Health Psychology, 2012, Volume 31, No. 5, 640–649
COPYRIGHT c
American Psychological Association
http://www.apa.org/pubs/journals/hea/index.aspx
This article may not exactly replicate the final version published in the APA journal.
It is not the copy of record.
Author Note
This article was published Online First November 14, 2011. DOI: 10.1037/a0026257
We are indebted to André Kretzschmar and Dörte Mund for their time and effort at coding,
as well as for their helpful comments on earlier versions of this paper. We also thank Leila Al-Kuwaiti,
Christiane Gentzel and Sarah Neubauer for their valuable comments on the revision of this manuscript.
Kristin Mitte is now at the Department of Psychology, University of Kassel, Kassel, Germany.
Correspondence concerning this article should be addressed to Marcus Mund, Friedrich-Schiller-Universität,
Institut für Psychologie, Humboldstraße 11, 07743 Jena, Germany. E-mail: marcus.mund@gmail.com

REPRESSIVE COPING AND SOMATIC DISEASES 2
Abstract
Objective: When Freud introduced the term repression, he stated its pathogenic potential.
Since then, this notion was adapted and continued to date. Surprisingly, there is no attempt
to synthesize research on the effect of repressive coping on somatic diseases quantitatively.
The current study closes this gap and examines the association between repressive coping
as operationalized by Weinberger et al. (1979) and the incidence of cancer, cardiovascular
diseases, asthma, and diabetes by applying meta-analytic techniques. Methods: An extensive
literature search identified 22 studies that met all inclusion criteria and assessed a total of
6 775 participants. The study set was analyzed both as a whole sample (k=22 studies) and
disease-specific (k=2−10 studies; no study found for diabetes). Results: The results of the
meta-analysis indicate a higher risk for repressive copers to suffer from one of the investigated
diseases, especially cancer (logOdd after diagnosis: 0.41; 95% confidence interval, 0.09–0.73,
p=.012) and hypertension (logOdd: 0.59; 95% confidence interval, 0.32–0.86, p< .0001).
After application of a continuity correction due to a missing control group the results for
coronary heart disease imply an increased risk for non-repressors to be affected. Results
for cardiovascular diseases in general, heart attack and asthma did not reach significance.
Conclusions: The current meta-analysis revealed significant associations between repressive
coping, cancer, and cardiovascular diseases, especially hypertension. These results add to the
notion of repressive coping as a consequence of cancer as well as to its important role for the
issue of hypertension.
Keywords: repressive coping; meta-analysis; cancer; hypertension; cardiovascular
disease.

REPRESSIVE COPING AND SOMATIC DISEASES 3
The Costs of Repression: A Meta-Analysis on the Relation
between Repressive Coping and Somatic Diseases
Introduction
As early as 1894, Freud stated that inhibited negative feelings can lead to severe diseases
like phobias, neuroses, or conversion hysteria (Freud, 1925/1999). While the former two affect
the psyche of a person, conversion hysteria is conceptualized as the transformation of negative
affect into bodily symptoms (Alexander, 1943). The mechanism behind this pathological
process was introduced as ‘repression’ and conceptualized as a way to keep unwanted feelings
unconscious (Freud, 1915/1999). Over the years, Freud’s and Alexander’s original concepts
were rejected (Erdelyi, 2006; Holmes, 1990) and therefore underwent numerous revisions and
reformulations. This led to a considerable amount of constructs labeled ‘repression’ meaning
something different, and of constructs labeled differently but meaning essentially the same
(Garssen, 2007).
Nevertheless, the concept of repression remained popular and scientists became more
interested in it as a dimension of personality. One of the first attempts to identify people high
in this dimension was the Repression-Sensitization scale (R-S scale; Byrne, 1961), which was
a derivative of the MMPI. Repressors classified in this way are said to have a higher threshold
for the perception of negative stimuli whereas sensitizers would approach to those stimuli
cognitively. However, there are some serious problems with the R-S scale concerning its validity
(Slough, Kleinknecht, & Thorndike, 1984) and rationale (Weinberger, 1990). Weinberger,
Schwartz, and Davidson (1979) tried to address these shortcomings by using a measure of
defensiveness (i. e. the tendency trying to be rational, self-protecting and to dismiss feelings)
and a standard trait anxiety measure simultaneously, which enabled them to distinguish between
people who are really not anxious and those who only pretend not to be anxious. Additionally,
they used several physiological measures to test whether those who scored low on anxiety, but
high on defensiveness differ from the other three groups derived by splitting both instruments.

REPRESSIVE COPING AND SOMATIC DISEASES 4
The first group is referred to as repressors or repressive copers1, which are defined as people
who try to avoid the conscious perception of negative feelings (Asendorpf & Scherer, 1983;
Weinberger, 1990). The three other groups—throughout the remainder of this paper referred to
as non-repressors—are named true low anxious, high anxious and defensive high anxious.
Despite stating not to be stressed and to be even less anxious, the repressive copers’ heart
rate and skin conductance increased even above the respective levels of the other three groups
when confronted with a stressful task. This finding was reproduced by other authors that used
different tasks and different physiological measures (Asendorpf & Scherer, 1983).
The dissociation between reported and objectively existing physiological and
psychological arousal might be the effect of the repressors’ tendency to deceive others (Furnham,
Petrides, Sisterson, & Baluch, 2003). However, it seems more plausible to assume a tendency
for self-deception, due to converging evidence concerning repressors’ self-image, coping
styles and memory (for review see Myers, 2010). Additionally, when distinguishing self- from
other-deceivers, most of the repressors are found to belong to the former group (e.g. Derakshan
& Eysenck, 1999).
Despite its lasting popularity and its function as a starting for very fruitful research
(Myers, 2010), the Weinberger et al. approach was criticized for its descriptive character that
does not explain the basal mechanisms of repressive coping. Current theories (Derakshan,
Eysenck, & Myers, 2007; Hock & Krohne, 2004; Mendolia & Baker, 2008) addressed these
shortcomings and shed light on the underlying processes.
Of course, research on repression as a defense mechanism also advanced. The notion
that defense mechanisms work in hierarchically organized clusters is now well established
(Cramer, 1998; Soldz & Vaillant, 1998; Vaillant, Bond, & Vaillant, 1986), with repression located
at an intermediate level of maturity as part of the cluster of neurotic defenses (Vaillant et al.,
1986). Along with the theoretical improvements, new instruments for the assessment of defense
1The terms repressor and repressive coper are used synonymously throughout this paper, albeit there is some
debate whether the repression of negative feelings is a conscious or an unconscious process (Cramer, 1998; Erdelyi,
2006). We use both terms to refer to people applying repressive coping.

REPRESSIVE COPING AND SOMATIC DISEASES 5
mechanisms as conceptualized from a psychodynamic point of view have been introduced, with
the Life Style Index (Conte & Apter, 1995) and the Defense Style Questionnaire (Bond, Gardner,
Christian, & Sigal, 1983) as the most popular. Recently, a Defensive Functioning Scale was also
included in the DSM-IV, which locates repression at the level of mental inhibition (Cramer,
2000). However, beside the fact that repressive coping and the defense of repression both serve
as adaptational processes in response to adversities, their further particulars are very different
(Cramer, 1998).
Repressive Coping and Health
One of the most consistent findings of the research on repressive coping as conceptualized
by Weinberger et al. is the already mentioned discrepancy between reported and objectively
existing response to emotional stimuli or stressful tasks, as long as they are sufficiently
self-relevant (Houtveen, Rietveld, Schoutrop, Spiering, & Brosschot, 2001; Weinberger et
al., 1979). Confronted with such stimuli, repressors typically react with increased heart rate
and decreased skin conductance resistance. People who are told to repress their emotions in
experimental studies show exactly the same features and differ from those who are allowed
to express their feelings as they like (Nyklíˇ
cek, Vingerhoets, & Denollet, 2002). If repressive
copers repressed unwanted feelings permanently (Weinberger, 1990), and if repression of feelings
was associated with the mentioned physiological features, it is close at hand to infer that a high
proportion of repressors should be affected by pathologically high blood pressure or associated
diseases like coronary heart disease (CHD). Indeed, there are numerous studies linking both
(Gleiberman, 2007; Myers et al., 2008) and showing serologically an increased risk for severe
cardiovascular diseases (CVD; Barger, Marsland, Bachen, & Manuck, 2000; Niaura, Herbert,
McMahon, & Sommerville, 1992). It was also shown that their heart rate variability, which is
the time between two subsequent heartbeats and depends on sympathetic and parasympathetic
activity (Acharya, Joseph, Kannathal, Lim, & Suri, 2006), differs from those of non-repressors in
a pathological direction (Fuller, 1992) and that this is predictive for cardiac problems (Acharya et
al., 2006).

REPRESSIVE COPING AND SOMATIC DISEASES 6
Additionally, repressive coping is assumed to be associated with the development of
cancer (Myers et al., 2008). It was always suspected to be part of the ‘cancer personality’, a
configuration of personality traits that differentiates premorbidly between people who will suffer
from cancer once in their life and those who will not. The empirical evidence for this construct is
mixed, with both supporting (Dattore, Shontz, & Coyne, 1980; Kune, Kune, Watson, & Bahnson,
1991; McKenna, Zevon, Corn, & Rounds, 1999) and rejecting (Bleiker & van der Ploeg, 1999;
C. G. Watson & Schuld, 1977) findings. However, most studies are based on the use of the
MMPI or its derivatives like the R-S scale (Dattore et al., 1980; C. G. Watson & Schuld, 1977),
very broad conceptualizations of the repressive coping style (Bleiker & van der Ploeg, 1999;
McKenna et al., 1999) or self-made instruments with unknown psychometric quality (Kune et al.,
1991). Furthermore, most of the self-report measures were administered only once following the
diagnosis. Hence, it is possible that the ‘cancer personality’ is rather a difference in coping with
the diagnosis than a predisposing factor. Nevertheless, concerning the immunological markers
it was shown that repressors are more cancer-prone than non-repressors (Barger et al., 2000;
Jamner, Schwartz, & Leigh, 1988).
The same is true for asthma and diabetes, which both can be linked to several immune
features. The number of eosinophile, for instance, is increased in repressive copers (Jamner et al.,
1988) and in case of an existing asthma can lead to an aggravation of the symptoms. Research on
diabetes shows an increased level of both insulin and glucose in repressors (Barger et al., 2000;
Jamner et al., 1988). Both markers can signal a beginning insulin resistance which is the most
popular feature of type 2 diabetes.
In summary, repressive copers seem to be more predisposed to certain diseases than
non-repressors. But surprisingly, they seem to feel quite comfortable and report only few
symptoms (Boden, Hyland, & Dale, 2005). Until now, it is unclear whether this a) originates
from their high level of defensiveness and the associated tendency to protect themselves from
feeling weak, vulnerable and out of control (Boden et al., 2005), b) it is just an artifact of the
instruments used (Furnham et al., 2003), c) they do not notice their symptoms, or d) they are

REPRESSIVE COPING AND SOMATIC DISEASES 7
just not sick.
Current Study
Given the proposed increased vulnerability, it is all the more surprising that to our best
knowledge there is no meta-analytic attempt to assess whether repressors as conceptualized
by Weinberger et al. (1979) really do suffer more often from several physical diseases than
non-repressors and, if so, to quantify the difference. The current study tries to close this gap,
thereby allowing generalizable evidence.
Method
Inclusion Criteria
Studies had to meet the following criteria to be included in the meta-analysis: a) The
disease under investigation must have been cancer, CVD (i. e. hypertension, CHD or heart
attack), asthma or diabetes. These had been selected because they all have theoretical as well as
physiological associations with repressive coping. b) Because the theoretical framework used is
the repression-approach established by Weinberger et al. (1979), studies must either have utilized
the Weinberger Adjustment Inventory (WAI; Weinberger & Schwartz, 1990) or a combination
of trait anxiety and defensiveness. The WAI was primarily designed to measure an individual’s
socio-emotional adjustment using the two scales distress (i. e. the affective component of
adjustment like depression, low self-esteem, low subjective well-being, proneness to anxiety) and
self-restraint (i. e. ability to suppress egoistic interests in favor of long-term goals and the relation
to others like impulse control, consideration of others, suppression of aggression, responsibility).
For inclusion, at least the proneness to anxiety subscale in combination with the self-restraint
scale must have been administered. As trait anxiety measure, the TMAS (Taylor, 1953), its short
form (Bendig, 1956), the STAI (Spielberger, Gorsuch, & Lushene, 1970) and the R-S scale as
well as their modified forms were accepted. Due to its high correlation with anxiety (e. g. Slough
et al., 1984), the R-S scale was only accepted, if it was combined with a defensiveness scale.
Studies were also included, if they aggregated the anxiety and the defensiveness measure to one

REPRESSIVE COPING AND SOMATIC DISEASES 8
single continuous score.
Concerning defensiveness, the Marlowe-Crowne Social Desirability Scale (MC-SDS;
Crowne & Marlowe, 1960) and the Social-Desirability-Scale-17 (SDS-17; Stöber, 2001) as well
as their modified forms were accepted. The MC-SDS was also used by Weinberger et al. and
in most other studies on repressive coping which adopted their approach. The SDS-17 tries to
update the MC-SDS by including contemporaryly formulated items. The correlation between
both instruments is appropriate (Stöber, 2001).
Recent research attempted to identify repressors using different instruments or different
emotions from the spectrum of negative affectivity (D. Watson & Clark, 1984), which inflamed
a debate about the equivalence of different operationalizations (Mendolia, 1999; Egloff &
Hock, 1999; Furnham & Traynar, 1999). Because the rates of concordance with the original
operationalization were too low in all cases, studies adopting such alternative strategies have
been excluded. Applying such strict criteria ensures to analyze the effects of repressive coping
according to Weinberger et al., and excludes other ‘repression-like’ constructs (Garssen, 2007).
As an additional criterion, c) only studies whose participants were 18 years of age and
older were considered.
Search Strategies
To find as many relevant studies as possible, an extensive literature search was conducted
during May and June 2010. In a first step, the full-text of the databases Psyndex, PubMed,
ScienceDirect, Web of Knowledge, Psycinfo, PsycArticles and the Psychology and Behavioral
Sciences Collection was scrutinized. Additionally, ProQuest and OpenSigle were browsed to
include relevant gray literature. As search terms we applied (repression OR repressive coping
OR defensive coping OR repressor OR represser) AND (cancer OR asthma OR coronary heart
disease OR chd OR blood pressure OR hypertens* OR cardiovascular disease OR diabet*) NOT
(child* OR adolesc* OR juvenil*). In a second step, all possibly relevant studies were searched
thoroughly for more relevant articles that could meet the inclusion criteria as evaluated by title
and context of citation. As third step, a forward search was carried out starting from the articles

REPRESSIVE COPING AND SOMATIC DISEASES 9
of which at least one had to be cited by relevant studies (i. e. Crowne & Marlowe, 1960; Stöber,
2001; Weinberger et al., 1979). Forward search was conducted using PubMed and the Web of
Knowledge. Besides the full-text databases, the International Journal of Behavioral Medicine,
Psychological Inquiry, British Medical Journal, Health Psychology, British Journal of Health
Psychology, Psychology and Health, Psycho-Oncology, Psychosomatic Medicine and Cancer
were separately searched, also manually, if possible. As a last step, authors known to have done
research on the repressive coping style and several diseases were asked for unpublished data.
A flow chart of the conducted literature search is shown in figure 1.
Coding
Data were extracted using a coding form implemented in a database due to error
reduction. Two independent coders repeated the procedure with nine randomly selected studies.
κamounted 0.84 for the variables relevant to the analysis, thus the agreement is considered very
good.
Effect Size and Data Analysis
Because we were particularly interested in whether repressors suffer more often from
certain diseases than non-repressors, odds ratios (OR) were used as effect size, hence they are
the appropriate coefficient for fourfold tables (Haddock, Rindskopf, & Shadish, 1998). Because
of better mathematical properties and similarity to more popular coefficients, we log-transformed
every study-specific OR (logOdds; Haddock et al., 1998). Fourfold tables were created so that
logOdds greater than 0 indicate a higher risk for repressors to suffer from the analyzed disease.
When the selected studies did not include a healthy control group2, a continuity correction was
applied that used the pooled data of the studies with control group. In case of studies using
dimensional measures of repressive coping, the effect sizes were transformed into logOdds.
When multiple effects could be computed (e. g., because multiple variations of splitting the
instruments were used, in order to identify repressors and non-repressors), parameters were
averaged to avoid an overrepresentation of the respective study.
2In the sense of absence of the disease under investigation.

REPRESSIVE COPING AND SOMATIC DISEASES 10
For combining individual effect sizes, random-effects models (RE) should have been
calculated. The RE differs from the fixed-effect model (FE) primarily with respect to the
generalizability of results. While the FE allows inferences only about the included sample of
studies, RE allows drawing unconditional inferences about a complete universe of studies.
Furthermore, the FE assumes a common population effect size and explains variance in
individual effect sizes by sampling error (within-study variance) alone. On the contrary, the RE
takes into account the sampling error and additionally an error resulting from randomly selecting
studies out of a universe of studies (between-study variance). The effect in RE is thought of as
the mean of a distribution of effect sizes (Hedges & Vevea, 1998; Viechtbauer, 2010). Since the
results of RE are only reliable with k>5 studies (Hedges & Vevea, 1998), it was not always
possible to use RE in the disease-specific analyses. All RE-analyses were conducted using the
Restricted Maximum-Likelihood estimator, for it is superior to the DerSimonian-Laird estimator
in most cases and addresses some of the latter’s shortcomings concerning moderate number of
studies (Viechtbauer, 2005).
As measure of heterogeneity, the Q-statistic was used, thereby taking into account its low
power by applying a probability level of .10.
Publication bias was addressed by means of trim-and-fill corrected funnel plots (Duval &
Tweedie, 2000) and normal-quantil plots (Wang & Bushman, 1998). To evaluate the robustness
of the results, fail-safe numbers were calculated using the formulas proposed by Rosenberg
(2005).
All analyses were conducted using the metafor-package for R (R Development Core
Team, 2010; Viechtbauer, 2010).
Results
Study Sample
Table 1 presents the characteristics of all studies that matched the inclusion criteria.
Data of 6 775 participants were integrated in the analysis, 3 883 of them female and
2 843 male; 49 did not indicate gender. Most studies investigated cancer (45%), followed by

REPRESSIVE COPING AND SOMATIC DISEASES 11
hypertension and other CVD. Only two studies dealing with asthma could be identified and
none studying diabetes. Furthermore, it is evident that all studies using a categorical approach
of repressive coping utilized the MC-SDS. This trend towards established instruments is also
evident regarding measures of trait anxiety.
Repressive Coping and Somatic Diseases
The first analysis addressed the question whether the risk of suffering from one of the
investigated diseases is increased for repressors. Across all k=22 studies, an effect size of
0.27 was found (OR =1.31; see table 2 for more details). This value is significant and implies
that the repressors’ risk for suffering from one of the diseases is increased by 31%. This result
was not contradicted by the additionally conducted sensitivity analyses, which used different
estimators or left single studies out. According to the fail-safe-N-statistic, 19 studies would be
necessary to nullify the effect. Applying the criterion proposed by Rosenthal (1979), the mean
effect size can not be considered robust. However, taking into account that only 22 studies met all
inclusion criteria at all, it seems not feasible to assume that another 19 studies are unpublished or
not retrievable.
Because the applied continuity correction might have biased the results, an additional
analysis was conducted that contained only the studies with a healthy control group (Ginzburg,
Solomon, & Bleich, 2002; Jensen, 1987; Kraft, 1998; Kreitler, Chaitchik, & Kreitler, 1993; Perini
et al., 1994; Servaes, Vingerhoets, Vreugdenhil, Keuning, & Broekhuijsen, 1999; Zachariae et al.,
2004) or operationalized repressive coping as a dimensional construct (Nyklíˇ
cek, Vingerhoets,
van Heck, & van Limpt, 1998; Nyklíˇ
cek, Vingerhoets, & van Heck, 1999). This analysis yielded
an effect size of 0.53 (OR =1.70). Again, this result proved to be robust in the sensitivity
analyses and with regard to the fail-safe number (see table 2).
With regard to potential moderators of the effect size, regression analyses were
conducted. Note that it was only possible to calculate regressions with not more than two
predictors due to the small number of studies. Among the studies with a control group, those
conducted in the USA produced greater effect sizes; no other significant moderators were

REPRESSIVE COPING AND SOMATIC DISEASES 12
identified3.
Repressive Coping and Cancer
10 studies focused on cancer (2 015 participants, 95.83% females).
Across all 10 studies, a significant effect size of 0.41 (OR =1.51) resulted. Thus,
for repressors the risk of a cancer diagnosis is increased by 51%. Again, this result proved
to be robust in the conducted sensitivity analyses. With regard to the fail-safe number, 44
studies would be necessary to nullify the effect, which is not a robust number according to the
Rosenthal-criterion. Again, given the fact that only 10 studies met all inclusion criteria at all, it
seems not plausible to assume that another 44 studies are unpublished or not retrievable. Only
two studies (Kreitler et al., 1993; Zachariae et al., 2004) assessed the repressive coping style
before diagnosis. These two studies were analyzed using FE, yielding an overall effect size of
0.20 (OR =1.22), which was not significant (p=.4; see table 2 for more details).
Repressive Coping and CVD
10 studies with 4 664 (40.46% females) participants in total were included that focused
either on hypertension, CHD, or heart attack. The mean effect size of 0.21 (OR =1.23; see table
2) was not significant. This result was confirmed through all sensitivity analyses.
Considering every single disease (see table 2), only the FE-analysis for hypertension
yielded a significant and robust result. The analysis suggests an increased risk of suffering from
higher blood pressure in the repressor group. Nevertheless, due to the small study sample and
the significant heterogeneity of the studies this result can not be generalized (Hedges & Vevea,
1998; Viechtbauer, 2010).
Although the result for the FE-analysis concerning CHD was also significant, thereby
implying an increased risk for non-repressors, it should be interpreted very cautiously for two
reasons. First, all four studies had to be corrected due to a missing control group. Second, the
effect depends on one single study (Denollet, Martens, Nyklíˇ
cek, Conraads, & de Gelder, 2008);
3Other moderators examined were the measure of anxiety, year of publication, sample size, sample age, sample
sex (only female, only male, mixed), disease, number of groups assessed, way of categorization (splitting at median,
mean, a specific reference value or a certain quantile) and categorical vs. dimensional operationalization.

REPRESSIVE COPING AND SOMATIC DISEASES 13
leaving it out led to non-significance.
The analysis for repressive coping and heart attack yielded no significant result. Note that
the continuity correction had to be applied for both studies, too.
Repressive Coping and Asthma
As for heart attack, the two asthma studies had to be corrected due to a missing control
group. Again, the result was not significant (logOdds =−0.15,OR =0.86; see table 2).
Discussion
Although current reviews conclude that there is much evidence for the association
between repressive coping and somatic diseases (Myers, 2010; Myers et al., 2008), there was
not yet a summarizing meta-analysis to quantify this effect. The current paper tried to close this
gap. The main question investigated was whether repressors, according to the conceptualization
established by Weinberger et al. (1979), do have an increased risk for suffering from cancer,
CVD, or asthma.
The first main finding of this study is that the Weinberger et al. approach seems to be
very popular in research, but is rarely adopted for use with patients. Therefore, only 22 studies
could have been identified. In sum, the results imply a significantly increased risk for repressors
to suffer from one of the investigated diseases, especially cancer and elevated blood pressure
or even hypertension. Analyses for the continuity corrected CHD and asthma studies yielded a
significantly increased risk for non-repressors or null effects, respectively.
Stating that repressors have an increased risk of suffering from several diseases provokes
subsequent questions. First, does repressive coping cause the diseases and, if so, what are the
underlying processes? Second, what consequences does a repressive coping style have with
regard to the progression of the disease?
For cancer, the present results imply that repressive coping does not precede the
diagnosis, but is rather a consequence of it. However, only two studies assessed repressive
coping before the diagnosis, hence this interpretation is fragile. Nevertheless, neither Kreitler

REPRESSIVE COPING AND SOMATIC DISEASES 14
et al. (1993) nor Zachariae et al. (2004) were able to predict the later patients by repressive
coping. In contrast, after the diagnosis the risk of being repressive increased. Both research
groups interpreted their results in terms of repressive coping as a consequence of cancer.
Similarly, Drageset and Lindstrøm (2003) could show that among women awaiting the results
of a biopsy those who applied cognitive defenses were less anxious, and they hypothesized
that this mechanism is especially useful in situations without the possibility for active coping.
Indeed, Zachariae et al. (2004) found a decrease of the proportion of repressors 12 weeks after
the diagnosis, which supports the notion of repressive coping as a first self-protective mechanism.
This line of interpretation is also concordant with other current reviews (e. g. Bleiker & van
der Ploeg, 1999). Extending the focus beyond repressive coping leads to similar results with
no single personality trait being found to be carcinogenic (e. g. Hansen, Floderus, Frederiksen,
& Johansen, 2005). Nevertheless, some explication is needed for the point estimate of 0.22
on the logarithmic odds ratio scale resulting from the analysis of cancer and repressive coping
before the diagnosis. If personality was unimportant, a logOdds of 0 would have been expected.
The deviation found (although non-significant) could be ascribed to the high selectivity of the
primary samples. Both studies assessed women undergoing biopsy. Hence, the setting itself—a
hospital, incalculable result—could have fostered repressive tendencies. Furthermore it can
be assumed that the patients belonged to at-risk groups (e. g. genetic predisposition, smokers,
older age) and therefore had a premonition. The repressors in the Zachariae et al. (2004) study
were slightly more convinced of having cancer which again could have activated the repressive
defense. In sum, the slightly increased risk of repressors to be diagnosed with cancer found in the
meta-analysis could be explained by processes that had already started before the primary studies
were even conducted. Future research could benefit from the distinction between state- and
trait-repressive coping (Mendolia, Moore, & Tesser, 1996) to isolate the effects of dispositions
on the development and the course of cancer (and other relevant diseases) from the effects
situations. However, this would necessitate much longer-lasting studies to identify the point in
time when situational effects begin to mix up with dispositional effects.

REPRESSIVE COPING AND SOMATIC DISEASES 15
If repressive coping was a consequence of the diagnosis—what does it mean for the
course of the disease? The preliminary answers are somewhat ambiguous. On the one hand,
Giese-Davis and Spiegel (2002) conclude that repressors do not regularly participate in and
terminate therapy earlier. Jensen (1987) found shorter duration of remission and more recurrence
among the repressive patients. Other studies report higher mortality for repressive or similarly
configured patients (Weihs, Enright, Simmens, & Reiss, 2000). On the other hand, there is
evidence that repressors suffer less under the chemotherapy and ask for support more often
(Ward, Leventhal, & Love, 1988), exhibit less symptoms of acute stress (Fischer Pedersen
& Zachariae, 2010), and experience more time without relapse than non-repressors (Dean &
Surtees, 1989). Finally, other authors could not find any associations between coping styles and
disease course at all (Petticrew, Bell, & Hunter, 2002). Because of these contradicting results it
is nearly impossible to draw any conclusions about repressive coping and disease progression, all
the more because a wide range of conceptualizations and instruments has been used in the cited
studies, which might also have caused the ambiguity of the findings.
Concerning CVD, the meta-analysis showed that repressors’ risk of suffering from at
least elevated blood pressure is increased by 80% compared to non-repressors. Three out of the
four analyzed studies assessed participants without abnormal hypertension, but it is known that
blood pressure increases with age which in turn increases the probability of an early diagnosed
pathological hypertension among the repressors, and consequently the risk for consecutive
illnesses like CHD or heart attack. Furthermore, the time during which anti-hypertensive drugs
have to be taken is prolonged. The processes behind this association are not yet clear. According
to the discontinuity hypothesis (Hock & Krohne, 2004), repressors experience more arousal when
confronted with ambiguous situations which could increase blood pressure via the release of
cortisol (Brown et al., 1996). Another explanation stems from the opioid-hypothesis of repressive
coping (e. g. Jamner et al., 1988). According to this approach, the hypothalamic-pituitary-adrenal
axis (HPA) is supposed to be hyperactive in repressors. Consequently, more opioid is released
which leads to a permanently euphoric state. This would explain why repressors’ in fact higher

REPRESSIVE COPING AND SOMATIC DISEASES 16
trait anxiety does not manifest in questionnaires. Additionally, the HPA-hyperactivity leads to an
increased release of cortisol and in turn to an increased blood pressure. Hence, according to the
opioid-hypothesis, repressive coping and hypertension are two facets of one single process. It
should be noted, however, that although the present results for the other CVD imply an increased
risk for non-repressors to suffer from CHD, they are determined by one single study and were
all continuity corrected. Thus, it is possible that future studies including a healthy control
group could yield even conversed results. However, either outcome is possible and plausible.
First, due to the potentially earlier diagnosis of hypertension in repressors it is conceivable that
preventive actions are taken on time to avoid CHD. This would lead to an increased proportion
of non-repressors among the CHD patients, since they have not had so much time to adapt
their lifestyle to their cardiovascular conditions. Second, an increased risk for repressors could
be a consequence of the prolonged duration of hypertension which leads to an accelerated
development of CHD. However, both assumptions are speculative yet and should be subject to
further research.
With regard to asthma, the main finding is an urgent need for more research. The analysis
showed a slightly increased risk for non-repressors to be affected. Note that the continuity
correction had to be applied for both included studies. Nevertheless, with regard to disease
progression, repressors could be expected to have a better prognosis because many tasks of
disease management support their wish for control (e. g. Weinberger, 1990; Weinberger et
al., 1979), for instance the monitoring of lung volume and the avoidance of asthma triggers.
Indeed, González-Freire, Vázquez-Rodríguez, Marcos-Velázquez, and González de la Cuesta
(2010) noticed that the repressors report less severe asthmatic symptoms than the affected
non-repressors, and Cooke, Myers, and Derakshan (2003) observed a better therapy adherence.
However, both studies admonish caution, because lung volume as an objective marker of disease
severity was more decreased in the repressor groups. González-Freire et al. (2010) also note that
repressors could have applied an optimistic bias concerning their symptoms, which is in line with
earlier research (for review see Myers, 2010).

REPRESSIVE COPING AND SOMATIC DISEASES 17
Limitations
The current paper was strictly guided by the repressive coping concept established by
Weinberger et al. (1979). This line of action seemed appropriate considering the heterogeneity
of related constructs (Garssen, 2007). Nevertheless, future research should try to illuminate
the convergence between several ways of identifying repressors to provide a variety of tools
to bypass repressors’ response styles. This would allow to extend the current analysis as well
as previous work done, which used other constructs, in terms of interpretation. Furthermore, it
could help clearing the contradictions in current research if it could be shown that some of the
operationalizations measure in fact different constructs.
To our surprise, no study focusing repressive coping and diabetes could be found despite
the strong psychophysiological component linking both. Indeed, some authors report about such
studies (Myers et al., 2008), but data, however, were not available.
Moreover, it is still possible that age confounds the results. Erskine, Kvavilashvili,
Conway, and Myers (2007) report a higher prevalence of repressors among elderly. Due to the
lack of group-specific means it was impossible to control for that variable in the current study.
If the results were strongly confounded with age effects, the interpretation would have to be
changed and would then suggest that there is no increased risk for repressors to suffer from
cancer or hypertension, but for older people. Educational status could be another potentially
confounding factor. However, data about coping style and education are reported in two studies
only. While there is evidence for an association in the work of Hunt (2001), there is none in that
of Zachariae et al. (2004). Future research should pay more attention to age and education in
order to partial out their effects and to facilitate conclusions about the main effects of repressive
coping.
From a psychodynamic point of view, the study could be criticized for taking only one
single style of ego defense into account. Current theories postulate that defense mechanisms are
employed in clusters and do not work independently (Cramer, 1998; Vaillant et al., 1986). Hence,
Paika et al. (2010) demonstrated that, for cancer, mechanisms from the same cluster can have

REPRESSIVE COPING AND SOMATIC DISEASES 18
very different influences on the same outcome variable. However, this again leads to the necessity
of establishing a more sophisticated nomological network around the construct of repressive
coping to enable the consideration of several processes simultaneously.
Even though the meta-analysis provides some ideas about the associations between
repressive coping and health it is necessary to conduct more longitudinal and prospective studies
to draw reliable conclusions about antecedent and consecutive effects.
Conclusion
The current meta-analysis closes the gap between the often stated notion of an association
between repressive coping and somatic diseases and the lack of integrated empirical data.
It was shown that repressors according to the definition of Weinberger et al. (1979) have an
increased risk for suffering from a somatic disease, with the most increased risk for hypertension.
Concerning cancer, an association was evident only following the diagnosis.

REPRESSIVE COPING AND SOMATIC DISEASES 19
References
References marked with an asterisk indicate studies included in the meta-analysis.
Acharya, U. R., Joseph, K. P., Kannathal, N., Lim, C. M., & Suri, J. S. (2006). Heart
rate variability: A review. Medical and Biological Engineering and Computing,44,
1031–1051.
Alexander, F. (1943). Fundamental concepts of psychosomatic research: Psychogenesis,
conversion, specificity. Psychosomatic Medicine,5, 205–210.
Asendorpf, J. B., & Scherer, K. R. (1983). The discrepant repressor: Differentiation between
low anxiety, high anxiety, and repression of anxiety by autonomic-facial-verbal patterns of
behavior. Journal of Personality and Social Psychology,45, 1334–1346.
Barger, S. D., Marsland, A. L., Bachen, E. A., & Manuck, S. B. (2000). Repressive coping and
blood measures of disease risk: Lipids and endocrine and immunological responses to a
laboratory stressor. Journal of Applied Social Psychology,30, 1619–1638.
Bendig, A. W. (1956). The development of a short form of the Manifest Anxiety Scale. Journal
of Consulting Psychology,20, 384.
Bleiker, E. M. A., & van der Ploeg, H. M. (1999). Psychosocial factors in the etiology of breast
cancer: Review of a popular link. Patient Education and Counseling,37, 201–214.
Boden, J. M., Hyland, M. E., & Dale, K. L. (2005). Defensiveness and symptom reporting.
Current Psychology,24, 153–170.
Bond, M., Gardner, S. T., Christian, J., & Sigal, J. J. (1983). Empirical study of self-rated defense
styles. Archives of General Psychiatry,40, 333–338.
Brown, L. L., Tomarken, A. J., North, D. N., Loosen, P. T., Kalin, N. H., & Davidson, R. J.
(1996). Individual differences in repressive-defensiveness predict basal salivary cortisol
levels. Journal of Personality and Social Psychology,70, 362–371.
Byrne, D. (1961). The repression-sensitization scale: Rationale, reliability, and validity. Journal
of Personality,29, 334–349.

REPRESSIVE COPING AND SOMATIC DISEASES 20
Conte, H. R., & Apter, A. (1995). The Life Style Index: A self-report measure of ego
defenses. In H. R. Conte & R. Plutchik (Eds.), Ego defenses: Theory and measurement
(pp. 179–201). New York: Wiley.
∗Cooke, L., Myers, L. B., & Derakshan, N. (2003). Lung function, adherence and denial in
asthma patients who exhibit a repressive coping style. Psychology, Health & Medicine,8,
35–44.
Cramer, P. (1998). Coping and defense mechanisms: What’s the difference? Journal of
Personality,66, 919–946.
Cramer, P. (2000). Defense mechanisms in psychology today: Further processes for adaptation.
American Psychologist,55, 637–646.
Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability independent of
psychopathology. Journal of Consulting Psychology,24, 349–354.
Dattore, P. J., Shontz, F. C., & Coyne, L. (1980). Premorbid personality differentiation of cancer
and noncancer groups: A test of the hypothesis of cancer proneness. Journal of Consulting
and Clinical Psychology,48, 388–394.
Dean, C., & Surtees, P. G. (1989). Do psychological factors predict survival in breast cancer?
Journal of Psychosomatic Research,33, 561–569.
∗Denollet, J. (1991). Negative affectivity and repressive coping: Pervasive influence on
self-reported mood, health, and coronary-prone behavior. Psychosomatic Medicine,53,
538–556.
∗Denollet, J. (1999). Repressive coping predicts mortality in patients with coronary heart disease.
Psychosomatic Medicine,61, 125.
∗Denollet, J., Martens, E. J., Nyklíˇ
cek, I., Conraads, V. M., & de Gelder, B. (2008). Clinical
events in coronary patients who report low distress: Adverse effect of repressive coping.
Health Psychology,27, 302–308.
Derakshan, N., & Eysenck, M. W. (1999). Are repressors self-deceivers or other-deceivers?
Cognition and Emotion,13, 1–17.

REPRESSIVE COPING AND SOMATIC DISEASES 21
Derakshan, N., Eysenck, M. W., & Myers, L. B. (2007). Emotional information processing in
repressors: The vigilance-avoidance theory. Cognition and Emotion,21, 1585–1614.
Drageset, S., & Lindstrøm, T. C. (2003). The mental health of women with suspected
breast cancer: The relationship between social support, anxiety, coping and defence in
maintaining mental health. Journal of Psychiatric and Mental Health Nursing,10,
401–409.
Duval, S., & Tweedie, R. (2000). Trim and Fill: A simple Funnel Plot based method of testing
and adjusting for publication bias in meta-analysis. Biometrics,56, 455–463.
Egloff, B., & Hock, M. (1999). When is an agreement between two methods for assessing
coping styles ‘good’? A reply to Mendolia. Personality and Individual Differences,27,
1021–1023.
Erdelyi, M. H. (2006). The unified theory of repression. Behavioral and Brain Sciences,29,
499–551.
Erskine, J. A. K., Kvavilashvili, L., Conway, M. A., & Myers, L. B. (2007). The effects of age on
psychopathology, well-being and repressive coping. Aging & Mental Health,11, 394–404.
∗Fischer Pedersen, A., & Zachariae, R. (2010). Cancer, acute stress disorder, and repressive
coping. Scandinavian Journal of Psychology,51, 84–91.
∗Frasure-Smith, N., Lespérance, F., Gravel, G., Masson, A., Juneau, M., & Bourassa, M. G.
(2002). Long-term survival differences among low-anxious, high-anxious and repressive
copers enrolled in the Montreal Heart Attack Readjustment Trial. Psychosomatic Medicine,
64, 571–579.
Frasure-Smith, N., Lespérance, F., Prince, R. H., Verrier, P., Garber, R. A., Juneau, M., . . .
Bourassa, M. G. (1997). Randomised trial of home-based psychosocial nursing
intervention for patients recovering from myocardial infarction. Lancet,350, 473–479.
Freud, S. (1915/1999). Die Verdrängung [Repression.]. In A. Freud (Ed.), Sigm. Freud:
Gesammelte Werke, chronologisch geordnet: Zehnter Band, Werke aus den Jahren
1913-1917 (pp. 247–261). Frankfurt am Main: Fischer.

REPRESSIVE COPING AND SOMATIC DISEASES 22
Freud, S. (1925/1999). Studien über Hysterie [Studies in hysteria.]. In A. Freud (Ed.), Sigm.
Freud: Gesammelte Werke, chronologisch geordnet: Erster Band, Werke aus den Jahren
1892-1899 (pp. 75–312). Frankfurt am Main: Fischer.
Fuller, B. F. (1992). The effects of stress-anxiety and coping styles on heart rate variability.
International Journal of Psychophysiology,12, 91–96.
Furnham, A., Petrides, K. V., Sisterson, G., & Baluch, B. (2003). Repressive coping style and
positive self-presentation. British Journal of Health Psychology,8, 223–249.
Furnham, A., & Traynar, J. (1999). Repression and effective coping styles. European Journal of
Personality,13, 465–492.
Garssen, B. (2007). Repression: Finding our way in the maze of concepts. Journal of Behavioral
Medicine,30, 471–481.
Giese-Davis, J., & Spiegel, D. (2002). Emotion regulation and metastatic breast cancer.
International Congress Series,1241, 31–35.
∗Ginzburg, K., Solomon, Z., & Bleich, A. (2002). Repressive coping style, acute stress disorder,
and posttraumatic stress disorder after myocardial infarction. Psychosomatic Medicine,64,
748–757.
Gleiberman, L. (2007). Repressive/defensive coping, blood pressure, and cardiovascular
rehabilitation. Current Hypertension Reports,9, 7–12.
∗González-Freire, B., Vázquez-Rodríguez, I., Marcos-Velázquez, P., & González de la Cuesta, C.
(2010). Repression and coping styles in asthmatic patients. Journal of Clinical Psychology
in Medical Settings,17, 220–229.
Haddock, C. K., Rindskopf, D., & Shadish, W. R. (1998). Using odds ratios as effect sizes
for meta-analysis of dichotomous data: A primer on methods and issues. Psychological
Methods,3, 339–353.
Hansen, P. E., Floderus, B., Frederiksen, K., & Johansen, C. (2005). Personality traits, health
behavior, and risk for cancer: A prospective study of a Swedish twin cohort. Cancer,103,
1082–1091.

REPRESSIVE COPING AND SOMATIC DISEASES 23
Hedges, L. V., & Vevea, J. L. (1998). Fixed- and random-effects models in meta-analysis.
Psychological Methods,3, 486–504.
Hock, M., & Krohne, H. W. (2004). Coping with threat and memory for ambiguous information:
Testing the repressive discontinuity hypothesis. Emotion,4, 65–86.
Holmes, D. S. (1990). The evidence for repression: An examination of sixty years of research.
In J. L. Singer (Ed.), Repression and dissociation: Implications for personality theory,
psychopathology, and health (pp. 85–102). Chicago: University of Chicago Press.
Houtveen, J. H., Rietveld, S., Schoutrop, M., Spiering, M., & Brosschot, J. F. (2001). A
repressive coping style and affective, facial and physiological responses to looking at
emotional pictures. International Journal of Psychophysiology,42, 265–277.
∗Hunt, M. E. (2001). Exploring the repressive coping styles of Caucasian and African-American
women with newly diagnosed stage II and recurrent breast cancer: The impact of
religiosity and under-reporting of stressful life events (Unpublished doctoral dissertation).
George Washington University, Washington, D. C..
Jamner, L. D., Schwartz, G. E., & Leigh, H. (1988). The relationship between repressive and
defensive coping styles and monocyte, eosinophile, and serum glucose levels: Support for
the opioid peptide hypothesis of repression. Psychosomatic Medicine,50, 567–575.
∗Jensen, M. R. (1987). Psychobiological factors predicting the course of breast cancer. Journal
of Personality,55, 317–342.
∗Jensen-Johansen, M. B., Christensen, S., Valdimarsdottir, H., Zakowski, S., Bovbjerg, D. H.,
Jensen, A. B., & Zachariae, R. (in preparation). The effect of an expressive writing
intervention on self-reported health among breast cancer patients: Results from a
nationwide randomized clinical trial.
∗King, A. C., Taylor, C. B., Albright, C. A., & Haskell, W. L. (1990). The relationship
between repressive and defensive coping styles and blood pressure responses in healthy,
middle-aged men and women. Journal of Psychosomatic Research,34, 461–471.
∗Kraft, M. E. (1998). The relationship between the repressive coping style and breast cancer

REPRESSIVE COPING AND SOMATIC DISEASES 24
incidence (Unpublished doctoral dissertation). University of California, Irvine.
∗Kreitler, S., Chaitchik, S., & Kreitler, H. (1993). Repressiveness: Cause or result of cancer?
Psycho-Oncology,2, 43–54.
Kune, G. A., Kune, S., Watson, L. F., & Bahnson, C. B. (1991). Personality as a risk factor in
large bowel cancer: Data from the Melbourne Colorectal Cancer Study. Psychological
Medicine,21, 29–41.
McKenna, M. C., Zevon, M. A., Corn, B., & Rounds, J. (1999). Psychosocial factors and the
development of breast cancer: A meta-analysis. Health Psychology,18, 520–531.
Mendolia, M. (1999). Agreement between two methods for identifying dispositional repressors:
A comment on Egloff and Hock (1997). Personality and Individual Differences,27,
1015–1020.
Mendolia, M., & Baker, G. A. (2008). Attentional mechanisms associated with repressive
distancing. Journal of Research in Personality,42, 546–543.
Mendolia, M., Moore, J., & Tesser, A. (1996). Dispositional and situational determinants of
repression. Journal of Personality and Social Psychology,70, 856–867.
∗Mizrahi, J. B. (2003). The impact of family process and repressive coping style on change in
distress after treatment for stage II or III breast cancer (Unpublished doctoral dissertation).
George Mason University, Fairfax.
Myers, L. B. (2010). The importance of the repressive coping style: Findings from 30 years of
research. Anxiety, Stress, & Coping,23, 3–17.
Myers, L. B., Burns, J. W., Derakshan, N., Elfant, E., Eysenck, M. W., & Phipps, S. (2008).
Current issues in repressive coping and health. In A. J. Vingerhoets, I. Nyklíˇ
cek, &
J. Denollet (Eds.), Emotion regulation: Conceptual and clinical issues (pp. 69–86). New
York: Springer.
Niaura, R., Herbert, P. N., McMahon, N., & Sommerville, L. (1992). Repressive coping and
blood lipids in men and women. Psychosomatic Medicine,54, 698–706.
Nyklíˇ
cek, I., Vingerhoets, A., & Denollet, J. (2002). Emotional (non-)expression and health:

REPRESSIVE COPING AND SOMATIC DISEASES 25
Data, questions and challenges. Psychology and Health,17, 517–528.
∗Nyklíˇ
cek, I., Vingerhoets, A. J. J. M., & van Heck, G. L. (1999). Elevated blood pressure and
self-reported symptom complaints, daily hassles, and defensiveness. International Journal
of Behavioral Medicine,6, 177–189.
∗Nyklíˇ
cek, I., Vingerhoets, A. J. J. M., van Heck, G. L., & van Limpt, M. C. A. M. (1998).
Defensive coping in relation to casual blood pressure and self-reported daily hassles and
life events. Journal of Behavioral Medicine,21, 145–161.
Paika, V., Almyroudi, A., Tomenson, B., Creed, F., Kampletsas, E. O., Siafaka, V., . . . Hyphantis,
T. (2010). Personality variables are associated with colorectal cancer patients’ quality
of life independent of psychological distress and disease severity. Psycho-Oncology,19,
273–282.
∗Perini, C., Smith, D. H. G., Neutel, J. M., Smith, M. A., Henry, J. P., Bühler, F. R., . . . Weber,
M. A. (1994). A repressive coping style protecting from emotional distress in low-renin
essential hypertensives. Journal of Hypertension,12, 601–607.
Petticrew, M., Bell, R., & Hunter, D. (2002). Influence of psychological coping on survival
and recurrence in people with cancer: Systematic review. British Medical Journal,325,
1066–1075.
R Development Core Team. (2010). R: A language and environment for statistical computing.
Vienna: R Foundation for Statistical Computing.
Rosenberg, M. S. (2005). The file-drawer problem revisited: A general weighted method for
calculating fail-safe numbers in meta-analysis. Evolution,59, 464–468.
Rosenthal, R. (1979). The ‘file drawer problem’ and tolerance for null results. Psychological
Bulletin,86, 638–641.
∗Servaes, P., Vingerhoets, A. J. J. M., Vreugdenhil, G., Keuning, J. J., & Broekhuijsen, A. M.
(1999). Inhibition of emotional expression in breast cancer patients. Behavioral Medicine,
25, 23–27.
∗Shaw, R. E., Cohen, F., Fishman-Rosen, J., Murphy, M. C., Stertzer, S. H., Clark, D. A., &

REPRESSIVE COPING AND SOMATIC DISEASES 26
Myler, R. K. (1986). Psychologic predictors of psychosocial and medical outcomes in
patients undergoing coronary angioplasty. Psychosomatic Medicine,48, 582–597.
Slough, N., Kleinknecht, R. A., & Thorndike, R. M. (1984). Relationship of the
Repression-Sensitization Scales to anxiety. Journal of Personality Assessment,48,
378–379.
Soldz, S., & Vaillant, G. E. (1998). A 50-year longitudinal study of defense use among inner city
men: A validation of the DSM-IV defense axis. The Journal of Nervous & Mental Disease,
186, 104–111.
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the State-Trait Anxiety
Inventory. Palo Alto: Consulting Psychologists Press.
Stöber, J. (2001). The Social-Desirability-Scale-17 (SDS-17): Convergent validity, discriminant
validity, and relationship with age. European Journal of Psychological Assessment,17,
222–232.
Taylor, J. A. (1953). A personality scale for manifest anxiety. The Journal of Abnormal and
Social Psychology,48, 285–290.
Vaillant, G. E., Bond, M., & Vaillant, C. O. (1986). An empirically validated hierarchy of defense
mechanisms. Archives of General Psychiatry,43, 786–794.
Viechtbauer, W. (2005). Bias and efficiency of meta-analytic variance estimators in the
random-effects model. Journal of Educational and Behavioral Statistics,30, 261–293.
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal of
Statistical Software,36, 1–48.
Wang, M. C., & Bushman, B. J. (1998). Using the Normal Quantile Plot to explore meta-analytic
data sets. Psychological Methods,3, 46–54.
Ward, S. E., Leventhal, H., & Love, R. (1988). Repression revisited: Tactics used in coping with
a severe health threat. Personality and Social Psychology Bulletin,14, 735–746.
Watson, C. G., & Schuld, D. (1977). Psychosomatic factors in the etiology of neoplasms. Journal
of Consulting and Clinical Psychology,45, 455–461.

REPRESSIVE COPING AND SOMATIC DISEASES 27
Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience aversive
emotional states. Psychological Bulletin,96, 465–490.
∗Weihs, K. L., Enright, T. M., Simmens, S. J., & Reiss, D. (2000). Negative affectivity,
restriction of emotions, and site of metastases predict mortality in recurrent breast cancer.
Journal of Psychosomatic Research,49, 59–68.
Weinberger, D. A. (1990). The construct validity of the repressive coping style. In J. L. Singer
(Ed.), Repression and dissociation: Implications for personality theory, psychopathology,
and health (pp. 337–386). Chicago: University of Chicago Press.
Weinberger, D. A., & Schwartz, G. E. (1990). Distress and restraint as superordinate dimensions
of self-reported adjustment: A typological perspective. Journal of Personality,58,
381–417.
Weinberger, D. A., Schwartz, G. E., & Davidson, R. J. (1979). Low-anxious, high-anxious, and
repressive coping styles: Psychometric patterns and behavioral and physiological responses
to stress. Journal of Abnormal Psychology,88, 369–380.
∗Zachariae, R., Jensen, A. B., Pedersen, C., Jørgensen, M. M., Christensen, S., Lassesen, B.,
& Lehbrink, M. (2004). Repressive coping before and after diagnosis of breast cancer.
Psycho-Oncology,13, 547–561.

REPRESSIVE COPING AND SOMATIC DISEASES 28
Table 1
Descriptive data of the selected studies (k = 22 studies).
Neffect sizef
Author(s), year countryasumbfemale male REPcdefensivenessdanxietyelogOdd0logOdd1
Cancer
Fischer Pedersen &
Zachariae, 2010
DK 112 60 52 31 MC-SDS SF-MAS −0.07
Hunt, 2001 USA 127 127 0 39 MC-SDS SF-MAS 0.25
Jensen, 1987 — 86 86 0 42 MC-SDS SF-MAS 0.51
Jensen-Johansen et al., in
preparation
DK 507 507 0 160 MC-SDS TMAS 0.17
Kraft, 1998 USA 220 220 0 59 MC-SDS SF-MAS 1.42
Kreitler et al., 1993 ISR 98 98 0 26 MC-SDS TMAS 0.13 1.03
Mizrahi, 2003 USA 91 91 0 MC-SDS SF-MAS 0.79
Servaes et al., 1999 — 96 96 0 — — WAI −0.07

REPRESSIVE COPING AND SOMATIC DISEASES 29
Table 1 — continued
Author(s), year countryasumbfemale male REPcdefensivenessdanxietyelogOdd0logOdd1
Weihs et al., 2000 USA 32 5 MC-SDS SF-MAS −0.81
Zachariae et al., 2004 DK 646 646 0 164 MC-SDS SF-MAS 0.21 0.69
Hypertension
King, Taylor, Albright, &
Haskell, 1990
USA 120 60 60 29 MC-SDS SF-MAS 1.70
Nyklíˇ
cek et al., 1998 NL 417 90 310 — MC-SDS-mod WAI-mod 0.36
Nyklíˇ
cek et al., 1999 NL 1061 1061 g— MC-SDS WAI-mod 0.54
Perini et al., 1994 USA 87 23 64 13 MC-SDS SF-MAS 1.69
Coronary heart disease
Denollet, 1991 BEL 178 0 178 62 MC-SDS-mod STAI 0.29
Denollet, 1999 — 409 49 360 97 MC-SDS STAI −0.25

REPRESSIVE COPING AND SOMATIC DISEASES 30
Table 1 — continued
Author(s), year countryasumbfemale male REPcdefensivenessdanxietyelogOdd0logOdd1
Denollet et al., 2008 BEL 731 75 656 159 MC-SDS STAI −0.36
Shaw et al., 1986 USA 97 14 83 25 MC-SDS-mod SF-MAS −0.078
Heart attack
Frasure-Smith et al., 2002hCND 1376 473 903 408 MC-SDS STAI 0.07
Ginzburg et al., 2002 ISR 188 43 145 49 MC-SDS TMAS −0.22
Asthma
Cooke et al., 2003 GB 21 15 6 7 MC-SDS SF-MAS 0.13
González-Freire et al., 2010 ESP 75 49 26 18 MC-SDS STAI −0.25
Note. a: DK: Denmark; ISR: Israel; NL: Netherlands; BEL: Belgium; CND: Canada; GB: Great Britain; ESP: Spain; —: unknown;
b: not all participants indicated gender, thus data can differ; c: REP: repressors; d: MC-SDS: Marlowe-Crowne Social Desirability
Scale, MC-SDS-mod: modified version of MC-SDS; e: SF-MAS: Short Form Manifest Anxiety Scale, TMAS: Taylor Manifest
Anxiety Scale, WAI: Weinberger Adjustment Inventory, WAI-mod: modified version of WAI; f: 0 indicates effect size before
diagnosis, 1 indicates effect size after diagnosis of disease; g: data of 906 male participants had been requested, but were unavailable
until completion of the meta-analysis; h: descriptive data taken from Frasure-Smith et al. (1997).

REPRESSIVE COPING AND SOMATIC DISEASES 31
Table 2
Statistical data of the analyses.
CI95% Heterogeneity
Model k M(E S)min max p Q pQFSN
RE complete 22 0.27 0.04 0.49 .0197 75.06 <.0001 19
RE complete CG 9 0.53 0.19 0.86 .002 16.17 .04 61
FE cancer t02 0.20 −0.26 0.66 .4 0.02 .89 —
RE cancer t110 0.41 0.09 0.73 .012 22.29 .01 44
RE CVD 10 0.21 −0.14 0.55 .24 42.81 <.0001 —
FE Hypertension 4 0.59 0.32 0.86 <.0001 10.25 .02 12
FE CHD 4 −0.22 −0.39 −0.05 .0095 6.80 .08 10
FE Heart attack 2 0.05 −0.11 0.21 .527 0.69 .40 —
FE Asthma 2 −0.15 −0.78 0.47 .63 0.26 .61 —
Note. k: Number of studies; RE: Random-Effects Model; FE: Fixed-Effect Model; M(ES): mean
effect size (logOdd s); FSN: fail-safe number according to Rosenberg (2005).

REPRESSIVE COPING AND SOMATIC DISEASES 32
Figure 1: Flow chart of the conducted literature search.