|
Abstract
Research has revealed that shy people
are self-critical, blame themselves for perceived social inadequacy, and experience more shame than controls.
These tendencies are greater in extremely shy
people who are undergoing treatment. A related question that warrants
consideration is whether shy people are critical
of others, and if this tendency is also greater for shy people in
treatment. If so, the interpersonal consequences of mistrust and
resentment are likely to be different from the
consequences of perceived inadequacies in the self, and call for
different kinds of cognitive and behavioral strategies to overcome
these tendencies. Previous studies have shown
that chronically shy adults in treatment report significant
interpersonal mistrust and resentment, and shy
college students report more negative thoughts about others than
non-shy students. In order to determine the
extent to which these thoughts and emotions predicted maladaptive behavior and interpersonal problems, we
utilized the initial evaluation results of a
sample of shyness clinic clients. We hypothesized that clients’ scores
on measures of shame would correlate with
measures related to mistrust, resentment, and interpersonal problems,
and would predict maladaptive behavior. Results
revealed that shame, resentment, and interpersonal mistrust were related, and were also related to
avoidance, distance, and hostility in relationships. Resentment and shame predicted self-abasement,
self-defeating behavior and passive aggression.
Clients diagnosed with avoidant personality disorder reported
significantly more shame, mistrust and resentment, as well as
interpersonal problems related to coldness and vindictiveness than those without the diagnosis.
Introduction
The association of shyness, shame and
internal attributions for negative social outcomes has been reported frequently in the clinical and research
literature (Arkin et!al 1980; Buss, 1980; Girodo
et!al 1981; Henderson, 1992;
Henderson, 2002; Teglasi, 1982). Chronically shy and socially anxious individuals report shaming as a means of
discipline in their families (Bruch and Heimberg,
1994) and shame and self-blaming attributions have been reported by
adolescents and adults (Henderson, 1992;
Henderson, 1994; Henderson and Zimbardo, 2001a; Zimbardo, 1977).
Children probably become sensitized to
frequent criticism, developing an increasing emotional sensitivity to others in the anticipation of painful
emotional states, while this vigilance promotes
emotional reactivity and instability. Vigilance also promotes the
internalization of parental blaming behavior,
with the result that children learn to blame others as well as
themselves (Benjamin, 1993).
Clinical observation and self-reports
demonstrate that chronically shy individuals blame others as well as themselves, seeing others as
dangerous, rejecting and unreliable (Henderson, 1994; Henderson, 1997;
Henderson and Zimbardo, 1998; Henderson and Zimbardo, 2001b). Furthermore, because shame is a painful emotion, the
practice of externalizing blame is thought to
dampen the pain of self-blame in the short run and to protect one’s
self-esteem (Lewis, 1971; Lewis, 1979). It seems
likely that at least some of the motivation to blame others is related
to reducing negative emotion. Blaming others has negative interpersonal as well as
intrapersonal consequences and is associated
with hostility and resentment toward the self and others (Bartholomew
and Horowitz, 1991; Henderson and Zimbardo, 2002; Tangney and Fischer,
1995; Tangney et!al 1992; Tennen and Affleck, 1990).
Furthermore, shame has been found to be negatively correlated with measures of empathy (Feshbach and Lipian, 1987;
Tangney, 1991), which would modulate anger and reduce potentially
maladaptive behavior. Beyond the interpersonal consequences of
hostility are health problems such as high blood pressure and
increased risk for hypertension (Gentry
et!al
1982).
Both interpersonal sensitivity and a
sense of threat from others were documented in an MMPI study of Shyness clinic clients (Henderson, 1997).
A significant majority (67%) of clinic clients
also received a diagnosis of avoidant personality disorder (APD) in a
co-morbidity study, a diagnosis that implies extreme sensitivity to
criticism. A diagnosis of passive-aggressive
personality disorder was also present in 15% of clients (St. Lorant
et!al 2000). Similar findings have been
shown in studies of social anxiety and social anxiety disorder (Alden
and Wallace, 1995; Cloitre et!al 1992; Erwin et!al 2003;
Herbert et!al 1992; Leary and Atherton, 1986). Because DSM criteria for avoidant personality disorder
(APD) include a belief that one is inadequate
and vulnerable to others in social situations, both shame and anger
were expected to be more elevated in individuals
with APD than those with generalized social anxiety disorder without APD (Greene, 1991; Henderson, 1997). Shame and
self-blame have been reduced with specific
instructions to restructure self-blaming attributions (Henderson and
Zimbardo, 2001). Therefore it seemed useful to
measure the associations of shame and resentment to see both how they predicted interpersonal difficulties, and
to provide a rationale for specific restructuring techniques for tendencies to blame others (Blalock
et!al 1997).
In order to assess the presence of
shame and anger in shyness clinic clients, and their associations with avoidance and aggressive tendencies,
we utilized the test results of shyness clinic clients. Our hypotheses
were: 1) that shame and anger would be positively correlated in our sample and positively correlated with negative
thoughts about others; 2) that patients
diagnosed with APD would evidence more shame, more negative thoughts
about others, anger, and interpersonal mistrust than those without APD;
and 3) that anger, in addition to shame, would
predict self-abasement, self-defeating behavior patterns and passive
aggression.
Method
Sample: Data in this study included data
from a co-morbidity study of 114 clients at the Palo Alto Shyness
Clinic (St. Lorant et!al 2000). The present study includes data on
these 114!clients, plus baseline assessment data on clients entering
the program since those data were collected. Clients are interviewed
by licensed clinicians using The Anxiety Disorders Interview Schedule
for DSM-IV (DiNardo et!al 1994). 94% meet criteria for generalized
anxiety disorder, and 70% for avoidant personality disorder according
to The Millon Clinical Multiaxial Inventory (MCMI; Millon, 1987;
Millon, 1998). Clients in the co-morbidity study were also given the
Minnesota Multiphasic Personality Inventory (MMPI; Butcher, 1989;
Greene, 1991). Table 1 includes the age and gender data for our
current participants. Sub-sample sizes vary for each analysis.
Materials: Each participant was administered standard assessment
instruments as part of the initial evaluation. The Millon Clinical
Multiaxial Inventory (MCMI; Millon, 1987), with the NCS Interpretive
Scoring System, was used to ensure accuracy and standardization (Millon,
1987). Degrees of avoidance, self-abasement, self-defeating behavior
and passive aggression were measured by elevations in scores on their
respective scales. Shame was measured by two instruments: the Personal
Feelings Questionnaire (PFQ-2; Harder et!al 1993), and the Test of
Self-conscious Affect (TOSCA; Tangney, 1995). The shame subscale of
the PFQ is purported to measure trait shame, the tendency to endorse
the frequent experience of shame, in contrast to guilt, with
test-retest reliability of 0.85 at two weeks and 0.78 at five weeks.
The TOSCAS, the shame sub-scale of a scenario-based measure of shame
and guilt is a revision of the Self Conscious Affect and Attribution
Inventory (SCAAI; Tangney, 1990), on which estimates of internal
consistency (Cronbach’s alpha) ranged from 0.72 to 0.82, and
test-retest reliabilities over a 1 to 5 week period were 0.79.
Internal consistency for the TOSCA shame subscale was 0.76 (Tangney,
1991). Anger was measured by the anger/in scale of the State-Trait
Anger Expression Inventory (STAXI; Spielberger, 1988; Spielberger,
1996). Because of the specificity in the STAXI’s subscales,
respondents can report how frequently they experience anger in
contrast to how much they express it (Spielberger et!al 1995). The
anger/in scale is an eight-item scale that indicates feelings of anger
that are suppressed. Internal consistency is 0.84 for males and is
0.81 for females.
Blaming others was
measured by the externalization scale of the TOSCA (TOSCAE) and our
Estimations of Others Scale (EOS) (Henderson and Horowitz, 1998). The
Estimations of Others scale is a 12-item scale based on statements of
clients during treatment and measures the tendency to report negative
thoughts about others. We have gathered normative data on college
students (see Table 2) and our clinic patients. The scale shows good
internal consistency, Chronbach’s Alpha being 0.90. Interpersonal
problems were measured by the 64-item Inventory of Interpersonal
Problems (Horowitz et!al 1988; Horowitz et!al 2000), an assessment of
self-reported interpersonal problems which include interpersonal
mistrust and hostility. The scale has good psychometric properties and
demonstrates high internal consistency with an alpha of 0.96) and
test-retest reliability of 0.78.
Shyness was also measured
by the Henderson/Zimbardo Shyness Questionnaire (ShyQ), which includes
items related to shame, negative attributions, and resentment (Henderson
and Zimbardo, 2000, 2002; Bortnik et!al 2002). Internal consistency
for six samples is 0.92, and testretest (2 weeks) reliability is 0.87.
Criterion validity as measured by its correlation with the 20-item
Revised Cheek and Buss Scale is between 0.60 and 0.67 in two college
student samples, and 0.74 in a shyness clinic sample (Cheek and
Melchior, 1990; Henderson and Zimbardo, 2002).
Results
The first hypothesis, that
shame and anger would be positively correlated in our sample and
associated with the externalization of blame, was tested using Pearson
correlation analyses. Table 3 contains the correlations of shame and
suppressed anger as measured by their respective subscales. As
expected, the PFQ shame scale, which measures the trait-tendency to
experience the emotion of shame, and the TOSCA shame scale, the
scenario based measure, were both significantly correlated with
suppressed anger (p!<!0.0001). Both were also correlated with negative
thoughts about others (p!<!0.0001), and the externalizing scale of the
TOSCA (p!<!0.02, p!<!0.0001). They were also correlated with
interpersonal coldness (IIPDE, PFQ shame, p!<!0.01; IIPDE, TOSCA shame,
p!<!0.0001) and hostile vindictiveness (IIPBC, PFQ shame, p!<!0.001;
IIPBC, TOSCA shame, p!<!0.0001). The second hypothesis, that those
diagnosed with avoidant personality disorder (APD) would score
significantly higher than the rest of the sample in shame, anger and
the externalization of blame, was tested by Univariate ANOVAS. Table 4
presents the means of those with and without the diagnosis of APD.
Those diagnosed with APD were higher in both measures of shame
F!(1,!181)!=!10.3, p!<!0.001; F!(1,!179)!=!6.7, p!<!0.01. The
APD group reported significantly more negative automatic thoughts
about others F!(1,!76)!=!4.07, p!<!0.05, more coldness, mistrust and
distance in relationships (IIPDE) F!(1,!103)!=!11.93, p!<!0.001, and
more hostile vindictiveness (IIPBC) F!(1,!103)!=!6.42, p!<!0.02,
according to the Inventory of Interpersonal Problems. Clients
diagnosed with APD also scored at the 83rd percentile on the Anger-in
Scale of the STAXI while those without APD scored at the 73rd
percentile. They did not differ, however, on the externalization scale
of the Tosca (TOSCAE).
The third hypothesis, that
shame would predict self-defeating behavior and passive aggression,
was tested using multiple regression (see Table 5). Self-abasement was
predicted by shame (PFQ!t, (4, 72)!=!3.54, p!<!0.01; TOSCAS, t, (4,
72)!=!3.51, p!<!0.01). Self-defeating behavior was also significantly
predicted by shame (PFQ!t, (1, 79)!=!3.91, p!<!0.0001; TOSCAS, t,
(1,!79)!=!3.37, p!<!0.01; F!(2,!79)!=!20.70, p!<!0.001). Shame also
predicted elevation on the Passive aggression scale, t,
(1,!73)!=!4.99, p!<!0.001; t, (1,!73)!=!2.18, p!<!0.05,
F!(2,!73)!=!17.47, p!<!0.001). Interestingly, shame also predicted
elevations on the MMPI subscale, Work Interference (WRK), (PFQ, t,
(1,!76)!=!3.34, p!<!0.01, TOSCAS, t, (1,!76)!=!4.71, p!<!0.001), not
surprising, given the tendency to passive aggression.
Exploratory Analyses:
Shame is also likely to affect treatment compliance and success.
Consequently, we examined the relationship of shame and anger to
negative attitudes toward treatment from the earlier co-morbidity
sample (St. Lorant et!al 2000 measured by the MMPI sub-scale, Negative
Treatment Indicators (TRT). Shame on the TOSCA, but not the PFQ,
predicted elevations on Negative Treatment Indicators (TOSCAS, t,
(1,!77)!=!5.07, p!<!0.001). Interestingly, the only negative
predictors of self-reported goal attainment in this sample were higher
scores on interpersonal avoidance (IIP) and externalizing tendencies (TOSCAE)
at posttest (R2!=!0.52, p!<!0.0001).
Discussion
Results were consistent
with our first hypothesis, that shame, anger and the externalization
of blame would be elevated in relation to normative samples, and
correlated with each other. Shame and resentment were also correlated
with the externalization of blame, consistent with findings of earlier
clinicians and researchers (Lewis, 1971; Scheff, 1987; Tangney et!al
1992; Wurmser, 1981). Our results are also consistent with recent
results for social anxiety disorder (Erwin et!al 2003). Patients
diagnosed with APD were significantly higher in shame and anger,
consistent with our second hypothesis, but not in the externalization
of blame as measured by the TOSCA, contrary to expectation. However,
those with APD were significantly higher in negative thoughts about
others, suggesting that these individuals may not be aware of the
implications of their negative thoughts about others, nor of their the
interpersonal implications, seeing such thoughts only as related to
their own vulnerability and natural caution.
Chronically shy
individuals sometimes are not aware of anger and tend not to express
it openly. Clinical observation has shown that effective treatment
goals for many shy clients include helping them to articulate and
express anger constructively (Henderson, 1992). The ways in which shy
individuals behave when they feel resentful or angry represent a
fertile domain for further research.
As we predicted in our
third hypothesis, shame was a significant predictor of selfabasement,
self-defeating behavior and passive aggression. This finding provides
further evidence that shame plays an important role in the formation
of maladaptive behavior within the chronically shy population. Further
studies that differentiate fear-related vs. anger-related social
avoidance are needed.
The elevations on the Work
Interference Scale, as predicted by shame and the externalization of
blame, point to the significance of problems in occupational
functioning among shy individuals. Treatment interventions concerned
with assertiveness and the appropriate handling of anger may be
particularly important in this domain. Assertiveness training, however,
will not likely be sufficient without specific restructuring of
negative automatic thoughts about others as well as oneself.
Negative Treatment
Indicators were also predicted by shame, resentment, and the
externalization of blame. Our clinical experience has been that
alerting patients to expect issues around shame and anger to be
occasioned in treatment, developing plans to handle the negative
feelings and to planning avenues for perception checks, in addition to
systematic hypothesis testing, appears to ameliorate attrition and
demoralization. Following such discussions at initial evaluation,
patients often speak to the group therapist privately if they do not
wish to address the emotions in the group.
We have been addressing
negative thoughts about others, interpersonal mistrust, and tendencies
to blame others in treatment at our Shyness Clinic, with a recent
study revealing that negative thoughts about others was a significant
negative predictor of self-reported goal attainment (Bortnik et!al
2002).
Limitations of this study
are its correlation nature, which does not allow us to assess
causality, and the use of self-report data and clinical interviews
without behavioral observations by contemporaries outside the groups.
Findings are consistent with clinical observation, but more
fine-grained analyses of interaction patterns are needed in this
population. Experimental studies of dyadic interactions with shy and
non-shy college students are in progress (Henderson et al 2002).
|
