- Scoring Instructions for the Clinical Anger Scale (CAS):
Abstract - The
purpose of the present investigation was to develop and validate an objective self-report
instrument, the Clinical Anger Scale (CAS), designed to measure the syndrome of clinical
anger. Several specific analyses were conducted to examine the psychometric properties of
the Clinical Anger Scale (CAS). Factor analysis was conducted to examine the factorial
validity of the instrument, and reliability coefficients were computed to examine the
internal consistency and stability of the CAS. Also, in addition to providing evidence for
the convergent and divergent validity of the CAS, an ancillary purpose of the present
study was to provide preliminary evidence for its validity by examining some personality,
psychopathological symptomology, behavioral, and family environmental correlates of
clinical anger. Factor analysis of the Clinical Anger Scale confirmed essentially a
unidimensional item structure; reliability analyses also demonstrated adequate alphas
(i.e., internal consistency) and test-retest coefficients (i.e., stability) for the CAS;
and other results indicated that the CAS was unrelated to social desirability influences.
Additional findings indicated that clinical anger was positively associated with several
anger-related concepts (e.g., trait anger, state anger, anger-in, anger-out,
anger-control). Other results showed that the Clinical Anger Scale was related in
predictable ways to men's and women's psychological symptoms, personality traits, and
early family environments. These results are discussed in terms of the need to distinguish
and to investigate the concept of clinical anger and its therapeutic treatment.
- Method
- An
objective self-report instrument--the Clinical Anger Scale (CAS)--was designed to measure
the psychological symptoms presumed to have relevance in the understanding and treatment
of clinical anger. Twenty-one sets of statements were prepared for this purpose. In
writing these groups of items, the format from one of Beck's early instruments was used to
design the Clinical Anger Scale (Beck et al., 1961; Beck, 1963, 1967). The following
symptoms of anger were measured by the CAS items: anger now, anger about the future, anger
about failure, anger about things, angry-hostile feelings, annoying others, angry about
self, angry misery, wanting to hurt others, shouting at people, irritated now, social
interference, decision interference, alienating others, work interference, sleep
interference, fatigue, appetite interference, health interference, thinking interference,
and sexual interference. Subjects were asked to read each of the 21 groups of statements
(4 statements per group) and to select the single statement that best described how they
felt (e.g., item 1: A = I do not feel angry, B = I feel angry, C = I am angry most of the
time now, and D = I am so angry all the time that I can't stand it). The four statements
in each cluster varied in symptom intensity, with more intense clinical anger being
associated with statement "D." Each cluster of statements was scored on a
4-point Likert scale, with A = 0, B = 1, C = 2, and D = 3. Subjects' responses on the CAS
were summed so that higher scores corresponded to greater clinical anger (21 items; range
0 - 63).
- A
scoring procedure similar to Beck's (Beck et al., 1996). is used with the
Clinical Anger Scale (CAS)--where a clinical anger score in a particular range
is labeled in a manner similar to Beck's procedure. That is, clinical
interpretation of the CAS scores is accomplished through the following
interpretive ranges: 0-13 - minimal clinical anger; 14-19 - mild
clinical anger; 20-28 - moderate clinical anger; and 29-63 - severe clinical
anger.
-
- Results
-
These results are presented in several major sections. The first section presents the
psychometric analyses of the Clinical Anger Scale. Included in this section are the factor
analysis results, the reliability results, and other scale validity results. Section two
then presents the gender norms and the ANOVA analyses conducted to examine the effect of
gender on the CAS. The third section reports the research evidence for the convergent
validity of both the CAS. This section presents the correlations between the CAS and
Speilberger's anger-related instruments. Section fourth includes the results of the
analyses conducted to examine the relationship between the CAS and the measures of
psychological symptoms, personality traits, and unhealthy behaviors (i.e., acting out and
neuroticism indexes). The fifth and final section describes the relationship between the
CAS and the measure of early family atmosphere, the Family Environment Scale.
-
- Factor Analysis Results
- To
examine the psychometric properties of the Clinical Anger Scale, a series of factor
analyses (principal axis with varimax rotation) were conducted for males and females
separately and in combination (using Sample IV). The results are shown in Table 1. An
inspection of Table 1 indicates that for the combined group of both males and females, all
of the CAS statements (except for item 3) loaded above |.30| on a single factor solution
(the eigenvalue for Solution I was 9.53 with 45.4% of the variance being explained). No
other factor solution had an eigenvalue greater than 1 (see Table 1). The CAS items were
then analyzed for males and females separately. The resulting factor loadings are also
shown in Table 1. Again, for both the male and the female analyses, only one factor
solution with an eigenvalue greater than 1 was found (for males, the eigenvalue for
Solution I was 11.33 and it accounted for 54% of the variance; for females, the eigenvalue
for Solution I was 8.71 and it accounted for 41.5% of the variance). Although neither the
male nor the female analyses produced more than one factor solution with an eigenvalue
greater than 1, it is apparent from Table 1 that some of the secondary solutions were
associated with the attitudinal, physiological, and performance manifestations of clinical
anger.
-
- Reliability
- The
internal consistency of the 21 items on the Clinical Anger Scale was analyzed by means of
Cronbach alpha, and yielded reliability coefficients of .94 (males and females together),
.95 (males only), and .92 (females only). The item-total correlations for these alphas are
presented in Table 1. All the item-total correlations exceeded |.30|, except for item 3
(anger about failure) which had item-total coefficients of .13, .19, and .11,
respectively, for the total sample, males only, and females only. [Although the item-total
coefficient for item 3 was low, it was decided nonetheless to retain this item in the
computation of the total CAS score, pending the results of additional investigations on
other older samples.] In addition to conducting internal reliability analyses, test-retest
analyses were also performed (see Table 2). The correlations between the two
administrations of the CAS were .85 (males), .77 (females), and .78 (both males and
females).
-
- Social Desirability
-
Finally, to determine whether people's scores on the Clinical Anger Scale were
contaminated by some type of response bias, the CAS was correlated with a measure of
social desirability and with the EPI Lie Scale. The results, shown in Table 2, indicate
that the CAS was independent of the tendency to respond in a socially desirable fashion
and was largely independent of the EPI Lie scores (the only exception was the Lie scale
for females, but this correlation only accounted for 4% of the overall variability in the
scores). In brief, this information indicates that the CAS was largely unifactorial in
nature, highly reliable, and essentially uncontaminated by social desirable and lying
tendencies.
-
- Gender Effects and Norms for the CAS
-
According to social stereotypes about gender, anger is an affect that characterizes men
more than women. A series of ANOVAs for Samples I, II, III, IV, and V was thus conducted
on the CAS to examine whether men and women would report different degrees of clinical
anger. In these analyses, gender was treated as the independent variable and the CAS was
regarded as dependent variable. The results are presented in Table 2, and reveal no
evidence that males and females differ in terms of the syndrome of clinical anger (all ps
> .05). This table also presents normative data (i.e., means and standard deviations)
for males and females in each of the major samples (Samples I-V).
-
- Convergent Validity Findings for the CAS
-
Preliminary evidence for the validity of the Clinical Anger Scale was determined by
examining the correlations between the CAS and the scores on Spielberger's anger-related
instruments. These correlations are shown in the bottom half of Table 2. As expected, the
scores on the Clinical Anger Scale were positively and strongly correlated with the two
subscales on the State-Trait Anger Scale. Moreover, the Clinical Anger Scale was
positively correlated with the subscales on the Anger Expression Scale, although the
relationships were not always as strong nor as significant as for the State-Trait Anger
Scale. These findings thus provide support for the convergent validity of the Clinical
Anger Scale. Additional Validity Findings for the CAS This section presents the results of
analyses conducted to examine the relationship between the CAS and the measures of
psychological symptoms, personality traits, and other unhealthy behaviors (i.e., acting
out and neuroticism indexes). The results are shown in Table 3.
-
- Psychological Symptoms and the CAS.
- As
an inspection of Table 3 indicates, the Clinical Anger Scale was positively correlated
with the full range of psychological symptoms measured by the SCL-90-R. Also, as one might
expect, males and females who reported greater clinical anger reported an elevated number
of psychological symptoms associated with hostility. In brief, clinical anger was
positively associated with a broad array of psychopathological symptoms.
-
- Personality Traits and the CAS.
- The
Clinical Anger Scale was also correlated with two personality instruments, the Eysenck
Personality Inventory and the Goldberg Big-5 Scale. As can be seen in Table 3 (for the
combined sample of males and females), clinical anger was positively correlated with
Eysenck's neuroticism scale and negatively correlated with Eysenck's extraversion scale,
and negatively correlated with the Big-5 measure of extraversion,
pleasantness-agreeableness, and emotional stability. Thus, feelings of clinical anger were
associated in a predictable pattern with measures of dispositional personality attributes.
-
- Unhealthy Behaviors and the CAS.
-
Table 3 (Snell et al., 1995) also shows the correlations between the Clinical Anger Scale
and the measures of acting-out behaviors, neurotic behaviors, and interpersonal
defensiveness. As expected, among the combined sample of males and females, clinical anger
was positively associated with all three measures of psychologically unhealthy behaviors
(although the pattern of results did vary somewhat among males and females). Thus, people
who were characterized by more intense clinical angry reported engaging in a greater
number of misdirected and inappropriate behaviors (e.g., lying, fighting, thefts, drug
use); as being more uncertain and dissatisfied with their social and personal life; and as
acting in a more suspicious and defensive manner about the intentions of others.
-
- Family Environments and the CAS.
- The
relationship between the Clinical Anger Scale and a measure of early family environment,
as assessed by the Family Environment Scale, was also examined. These correlations were
computed for both males and females separately, but only the results for the combined
sample will be interpreted (see Table 3). An inspection of this table reveals that
clinical anger was negatively associated with a family history of cohesion,
expressiveness, and shared recreational activities, but was positively related to an
earlier history of family conflict and exaggerated family control (but see the gender
specific correlations). No other correlations were statistically significant.
- Discussion
- The
need for a reliable and valid instrument capable of assessing the symptoms of clinical
anger led to the present research on the construction and preliminary validation of the
Clinical Anger Scale. The Clinical Anger Scale was specifically designed to measure the
array of psychological, physiological, affective, cognitive, motoric, and behavioral
symptoms constituting clinical anger. Preliminary evidence for the validity of the CAS was
demonstrated in a series of analyses showing that clinical anger was related in a
systematic and interpretable manner with measures of state anger, trait anger, anger
control, and anger expressed inwardly and outwardly. Moreover, other findings revealed
that men's and women's feelings of clinical anger were predictably associated with a
number of distinct personality characteristics, psychopathological symptoms, and
inappropriate as well as problematic interpersonal behaviors.
Snell,
W. E., Jr., Gum, S., Shuck, R. L., Mosley, J. A., & Hite, T. L. (1995).
The Clinical Anger Scale: Preliminary reliability and validity. Journal of
Clinical
Psychology, 51, 215-226.
Beck,
A. T., Brown, G., & Steer, R. A. (1996). Beck Depression
Inventory II
manual. San Antonio, TX: The Psychological Corporation.
- Explicit written permission is need from
Dr. William E. Snell, Jr.,
in order for individuals to use the Clinical Anger Scale (CAS).
Address internet e-mail to: wesnell@semo.edu.
