See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/13701414

Articulated thoughts of maritally violent and nonviolent men during anger

arousal. Journal of Consulting and Clinical Psychology, 66, 259-269

Article  in  Journal of Consulting and Clinical Psychology · May 1998

DOI: 10.1037//0022-006X.66.2.259 · Source: PubMed

CITATIONS READS
86 384

3 authors, including:

Christopher I Eckhardt Gerald C. Davison

Purdue University University of Southern California
84 PUBLICATIONS   3,453 CITATIONS    170 PUBLICATIONS   4,262 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Addressing Mental and Physical Health Needs Among LGBTQIA+ Community Members Post-Sexual Assault View project

Addressing Mental and Physical Health Needs Among LGBTQIA+ College Students Post-Sexual Assault View project

All content following this page was uploaded by Gerald C. Davison on 25 December 2014.

The user has requested enhancement of the downloaded file.

Journal of Consulting and Clinical Psychology Copyright 1998 by the American Psychological Association, Inc.
I9VR, Vol. 66, No. 2, 259-269 0022-006X/98/$3.00

Articulated Thoughts of Maritally Violent and Nonviolent Men During

Anger Arousal

Christopher I. Eckhardt Krista A. Barbour and Gerald C. Davison

Southern Methodist University University of Southern California

The cognitive correlates of anger arousal were investigated in community-based samples of maritally
violent (MV), maritally distressed-nonviolent (DNV), and maritally satisfied-nonviolent (SNV)
husbands. Participants performed the Articulated Thoughts in Simulated Situations (ATSS) paradigm
while listening to anger-arousing audiotapes. Trained raters coded for irrational beliefs, cognitive
biases, hostile attributiona] biases, and anger control statements. Results indicated that MV men
articulated significantly more irrational thoughts and cognitive biases than DNV and SNV men. MV
men articulated more hostile attributional biases than DNV and SNV men across all ATSS scenarios.
SNV men. however, articulated more anger control statements during ATSS anger arousal than MV
or DNV participants. Discriminant function analyses indicated that specific thoughts discriminated
between the groups and differentiated mildly from severely violent participants. ATSS cognitive
distortions (a) were not correlated with questionnaire measures of cognitive distortion, and (b) were
superior to questionnaire measures in discriminating between the groups. The findings are interpreted
in light of recent advances in understanding the relationship between information processing, anger,
and marital aggression.

Researchers have suggested that anger and hostility are dis-
criminating characteristics of maritally violent ( M V ) men using
both self-report and marital interaction research designs (for a
review, see Eckhardt, Barbour, & Stuart, 1997). Despite these
data, theories of marital violence have not adequately described
how anger is aroused and how it influences the enactment of
aggression in marriage. Cognitive models of emotion and psy-
chopathology, such as those proposed by Ellis (1962) and Beck
(1976), have gained clinical acceptance but have only recently
been applied to marital violence (e.g., Holtzworth-Munroe,
1992; O'Leary & Vivian, 1990). To assess the degree to which
various cognitive distortions and deficiencies characterize MV
husbands during laboratory anger arousal, we employed a think-
aloud cognitive assessment task, the Articulated Thoughts in
Simulated Situations paradigm (ATSS; Davison, Robins, &
Christopher I. Eckhardt, Department of Psychology, Southern Method-
ist University; Krista A. Barbour and Gerald C. Davison, Department of
Psychology, University of Southern California.
This research was supported by National Institute of Mental Health
Grant MH55028-01. Portions of this research were presented at the 1996
convention of the Association for Advancement of Behavior Therapy,
New York, New York.
We acknowledge the assistance of numerous individuals in the com-
pletion of mis research, including the psychology department of the
University of North Carolina—Wilmington, Howard Kassinove, John
Wilson, Heather Smith, Matthew Kandies, Lara McGill, Melanie Dye,
Stacey Reed, and Tapestry Theater Company. We also appreciate the
helpful comments of Ellen Dennehy.
Correspondence concerning this article should be addressed to
Christopher I. Eckhardt, Department of Psychology, Southern Methodist
University, Dallas, Texas 75275. Electronic mail may be sent to
eckhardt@mail.smu.edu.
Johnson, 1983). with samples of MV, maritally distressed-non-
violent (DNV), and maritally satisfied-nonviolent (SNV) men.
Although ample data indicate that anger differentiates MV
from nonviolent men, Margolin et al. (1988) and Burman et al.
(1993) have noted that the cognitive processes associated with
anger hyperarousal have not received sufficient research atten-
tion, which is somewhat surprising for three following reasons.
First, it is now commonplace for clinical researchers and com-
munity intervention programs to espouse cognitively oriented
anger management therapeutic interventions for maritally ag-
gressive individuals (Arias & O'Leary, 1988; Holtzworth-
Munroe & Stuart, 1994a). If one assumes that cognitive treat-
ments are best targeted to cognitive phenomena, it is important
to demonstrate that MV men exhibit cognitive distortions and
deficiencies under appropriate conditions. Second, question-
naire-based research investigating the relationship between
anger and cognition in clinical and nonclinical samples has indi-
cated that endorsement of Ellis's (1962) list of irrational beliefs
is associated with higher scores on various anger scales (e.g.,
Hogg & Deffenbacher, 1986; Kassinove & Eckhardt, 1994).
One would therefore expect MV men, who also show high levels
of trait anger, to demonstrate similar anger-related cognitive dis-
tortions. Third, a wealth of research has indicated that a variety
of attitudinal and cognitive distortions are associated with mari-
tal dissatisfaction (e.g., Bradbury, Beach, Fincham, & Nelson,
1996). Given that MV men also experience higher levels of
marital distress than nonviolent husbands (Holtzworth-Munroe
etal., 1992; Rosenbaum & O'Leary, 1981), it should follow that
MV men should also show distortions in cognitive processing.
Inspection of the marital violence research literature, however,
suggests that many of the preceding assumptions have yet to be
confirmed. Researchers examining the cognitive characteristics
of MV men have typically investigated global attitudes and
attributions. One consistent finding is that MV men are more

259
260 ECKHARDT. BARBOUR, AND DAVISON
likely to espouse favorable attitudes toward the use of aggres-
sion in marriage (e.g., Saunders, Lynch, Grayson, & Linz, 1987;
Stith & Farley, 1993). Some data suggest that men who believe
in conservative, traditional sex role stereotypes have higher mar-
ital aggression levels (e.g., Grossman, Stith, & Bender, 1990),
but other investigators have not found differential attitudes to-
ward women among MV and nonviolent men (e.g., Neidig,
Friedman, & Collins, 1986). Unfortunately, it is difficult to draw
firm conclusions about the predictive validity of these attitudinal
constructs because it is not clear whether MV men held these
attitudes before the onset of marital aggression, or if they
emerged as post hoc explanations for aggressive acts (Holtz-
worth-Munroe et al., 1997).
Tb guide our assessment of anger-related cognitive variables,
we adopted the cognitive taxonomy presented by Kendall and
colleagues (Ingram, Kendall, Smith, Donnell, & Ronan, 1987;
Kendall, 1992; Kendall & Dobson, 1993). In this framework, the
construct o f ' 'cognition'' is partitioned into cognitive processes,
cognitive structures, cognitive content, and cognitive products.
Cognitive processes refer to active thinking processes involved
in attending, storing, retrieving, and interpreting information.
These processes can be further differentiated into cognitive dis-
tortions, which represent illogical, faulty, or misguided pro-
cessing, and cognitive deficiencies, which represent "an insuf-
ficient amount of cognitive activity in situations wherein more
forethought would be beneficial" (Kendall, 1992, p. 5). Cogni-
tive structures refer to schematic templates that guide cognitive
processing, such as the depressive schema hypothesized by Beck
(1967) to contribute to depression. Cognitive content refers to
' 'information that is stored and organized in memory" (Kendall,
1992, p. 2), such as representations of specific life events. Cog-
nitive products refer to the end-stage cognitions that result from
prior structures, processes, and content, such as causal attribu-
tions. The present research focuses on cognitive processes (both
cognitive distortions and cognitive deficiencies) and cognitive
products (hostile attributional bias), although all four cognitive
operations are presumed to be interactive (Kendall, 1992).
Recently, Holtzworth-Munroe (1992) presented a cognitive
model of social skill deficits in MV men outlining the sequence
of information processing stages that can result in marital ag-
gression. During the first stage, decoding, the individual attends
to, perceives, and interprets social stimuli. However, "various
cognitive deficits, including unrealistic expectations, faulty attri-
butions, and irrational beliefs could result in the misconsrrual
of social stimuli" (Holtzworth-Munroe, 1992, p. 607). Thus,
automatic, overlearned information processing distortions may
alter the significance of incoming stimuli or result in other cogni-
tive products, such as hostile attributions, that may otherwise
disrupt accurate, goal-congruent processing of social stimuli.
During the second decision-making stage, the individual is con-
fronted with the task of strategically constructing a number
of potential responses to manage the demands of the specific
situation. After selecting the most appropriate response, the indi-
vidual decides if he has sufficient skill to enact that response and
judges the response's likely positive and negative consequences.
Finally, the third stage describes response enactment, during
which the individual puts the selected response into action and
monitors its impact. Affective variables such as anger and hostil-
ity serve as "transitory factors" that can disrupt the social
information processing sequence.
Using the social information processing heuristic, Holtz-
worth-Munroe and Hutchinson (1993) assessed the attributions
occurring during the decoding stage in MV, DNV, and SNV
men. Nine brief conflict vignettes were presented to participants,
who completed attribution questionnaires following the presen-
tation of each vignette. Results indicated that MV men were
more likely than comparison men to attribute negative intent on
the part of the wife as the primary causal agent of the conflict.
To assess the processes occurring during the decision-making
stage, Holtzworth-Munroe and Anglin (1991) had MV, DNV,
and SNV men read 22 marital conflict vignettes. When asked
what they would say or do in each situation, MV men offered
the fewest competent responses during scenes in which the wife
rejected, teased, or taunted the husband, or wanted something
from the husband. Similar findings were obtained by Dutton
and Browning (1988), who found that MV men were more
likely to choose coping responses that involved little construc-
tive reasoning, more verbal aggression, and more physical ag-
gression after imagining themselves the subject of a series of
videotapes depicting marital conflict.
Although the preceding research added valuable information
regarding the faulty attributions and deficient decision-making
skills of MV men, additional improvements are needed. With
the exception of the Holtzworth-Munroe and Anglin (1991)
study, most research in this area has utilized paper-and-pencil
cognitive assessment procedures. These measures assess partici-
pants' post hoc generalizations about how they believe they
usually think and do not provide ' 'on-line'' accessibility to more
automatic cognitive activity arising without conscious delibera-
tion or strategic filtering (Davison, Feldman, & Osborn, 1984).
Given the automaticity of emotion-relevant cognitions, it follows
that cognitions will be more accessible and reportable if they
are assessed concurrently with mood state activation (Persons &
Miranda, 1992).
Davison and associates (Davison et al., 1984; Davison et
al., 1983) developed a methodology to assess naturalistic and
representative on-line cognitive activity occurring during emo-
tion activation. The ATSS paradigm provides open-ended verbal
reporting of thoughts as they are experienced during an emo-
tional reaction. In the ATSS method, participants listen to an
audiotaped scenario and imagine that they are an active part of
the interaction. After a brief segment of the tape is played, they
are asked to articulate their thoughts for 30 seconds. The tape
then continues, followed by another pause for articulation, and
so on. Articulations are transcribed and subsequently rated along
relevant dimensions by trained coders. Research by Davison and
colleagues using nonclinical and clinical samples has supported
the validity of the ATSS procedure as an alternative to question-
naire-based cognitive assessment (for reviews of the ATSS pro-
cedure see Davison, Navarre, & Vogel, 1995, and Davison,
Vogel, & Coftman, 1997).
The present study assessed anger-related cognitive distortions,
cognitive deficiencies, and cognitive products hypothesized to
occur during the decoding and decision-making stages of Holtz-
worth-Munroe's (1992) social information processing model in
samples of MV, DNV, and SNV men. All participants initially
completed self-report measures of irrational beliefs (Ellis, 1994)
 ARTICULATED THOUGHTS AND MARITAL VIOLENCE
and cognitive biases (Beck, 1976). Participants then articulated
their thoughts during an anger-arousing ATSS procedure, and
these responses were later coded for irrational thinking, cogni-
tive biases, hostile attributional biases, and anger-controlling
statements. A number of predictions guided our data analyses.
First, we hypothesized that MV men would articulate more total
irrational beliefs, total cognitive biases, and hostile attributional
biases during ATSS anger arousal relative to both nonviolent
comparison groups. Second, we predicted that both nonviolent
comparison groups would articulate more anger-controlling
statements than MV men (suggesting a cognitive deficiency in
MV men). Third, we predicted that specific ATSS cognitive
distortions (the irrational beliefs demandingness, self/other rat-
ing, and low frustration tolerance, and the cognitive biases arbi-
trary inference, magnification, and dichotomous thinking)
would be stronger predictors of group membership than ques-
tionnaire measures of cognitive distortion. Fourth, we did not
expect a relationship to emerge between the questionnaire-as-
sessed cognitions and those articulated by MV men in the labo-
ratory due to the heightened accessibility of the latter cognitions
during anger arousal. Finally, we expected that MV men who
had engaged in severe forms of marital violence in the previous
year would be characterized by higher levels of cognitive distor-
tions and hostile attributions than MV men without episodes of
severe violence.
Method
Participants
The participants were 88 married men who responded to local newspa-
per and radio advertisements.1 The men were contacted by phone by
trained interviewers who administered a variety of screening measures
to determine group membership. All participants were paid $40.
Husbands' reports were used to determine group assignment, as recent
research has indicated that husbands' reports of wife-directed violence
are not strikingly dissimilar from the reports of their wives (see Jacobson
et al., 1994, p. 983) .2 Participants in the MV group consisted of 31 men
who reported at least four incidents of physical aggression against their
wives within the past 12 months on the Modified Conflict Tactics Scale
(MCTS; Pan, Neidig, & O'Leary, 1994a), which assesses the frequency
of psychological aggression (12 items; e.g., threatened to leave relation-
ship), mild physical aggression (5 items; e.g., pushed, grabbed, or
shoved spouse), and severe aggression (6 items; e.g., beat up spouse)
used during marital conflicts.^ The validity of the qualitative distinction
between mild and severe MCTS marital aggression is supported by data
presented by Pan, Neidig, and O'Leary (1994b). Inclusionary criteria
for the 23 DNV men were (a) a score of 95 or below on the Short
Marital Adjustment Test (SMAT; Locke & Wallace, 1959),4 a 15-item
screening measure that assesses both general marital happiness as well
as the degree to which couples agree on a variety of more specific
marital issues (e.g., finances, sex, recreation), and (b) no reported
incidents of physical aggression used against their wives at any point
in the marriage on the MCTS. Thirty-four men classified as SNV were
required to (a) score 115 or above on the SMAT, and (b) report no
incidents of physical aggression used against their wives at any point
in the marriage on the MCTS. Finally, potential participants answered
various demographic questions.
All participants were required to be between the ages of 19 and 49
and married for at least 1 year. There were no significant differences
between the groups on age, race-ethnicity, socioeconomic status, yearly
household income, length of marriage, number of children, years of
261
education, and frequency of weekly alcohol and drug usage. The average
participant was White, age 35, and was married for 8 years; had 14.5
years of education, two children, and a household income of $46,500;
and reported drinking 4.5 alcoholic beverages per week and using other
drugs less than once per week. We also asked participants if they were
involved in individual or marital counseling. Only I MV man and 2
DNV men indicated that they had been involved in marital counseling.
On the Dyadic Adjustment Scale (DAS; Spanier, 1976), MV and DNV
groups were well matched, and both groups' mean DAS scores were
significantly lower than SNV husbands' scores (MV: M = 86.9, SD =
20.2; DNV: M = 85.7, SD = 14.9; SNV: M = 122.0, SD = 12.2).
Materials
State Anger Scale. To provide a measure of participants' anger levels
at the time of the experiment, we administered the State Anger Scale
from the State-Trait Anger Expression Inventory (STAXI; Spielberger,
1988). This 10-item scale assesses the intensity of angry feelings at the
time of assessment. An alpha of .90 was reported in a large sample of
nonclinical men (Spielberger, 1988) and accumulated data point to the
scale's excellent validity (Spielberger, 1988; Spielberger, Reheiser, &
Sydeman, 1995).
Survey of Personal Beliefs. The 50-item Survey of Personal Beliefs
(SPB; Demaria, Kassinove, & Dill, 1989) was used to assess the en-
dorsement of irrational ideation. Items on the SPB assess the four hy-
pothesized core irrational beliefs of rational emotive behavior therapy
(Ellis, 1994; Ellis & Dryden, 1987) along a 6-point Likert scale. The
SPB yields subscale scores for awfulizing (e.g., "Some situations in
life are truly terrible"), low frustration tolerance (e.g., "I can't stand
some of the things that have been done by members of my family"),
demandingness (e.g., "My family should be reasonable in what they
ask of me"), self/other rating (e.g., "A person who sins or harms
others is a 'bad person' "), and a total rationality score. Higher scores
indicate greater rationality. Previous research has supported the reliabil-
ity of the SPB (Demaria et al., 1989; Kassinove, 1986). The items
make no mention of emotion, which addresses criticisms made by Smith
(1982) regarding the lack of discriminant validity used in the construc-
tion of other measures of irrational ideation that use affect-laden word-
ing. The SPB has been shown to correlate significantly with various
1
Recruitment began by placing advertisements in local newspapers
and on radio stations for married men willing to participate in a research
project about their opinions of their marriage and their spouse. Partici-
pants left their first names on a voice mail system and were later con-
tacted by trained interviewers who administered the screening measures
over the telephone.
2
Nevertheless, other researchers have indicated that MV men tend to
minimize their own violence frequency. The issue at hand is the degree
of confidence that our samples are actually representative of the popula-
tion from which they are purported to be drawn. This is less of an issue
for the MV husbands, who had to indicate four incidents of violence to
qualify. The DNV group (and to a lesser extent the SNV group), how-
ever, may include men who did not report on their own marital aggres-
sion. However, if this were true, group differences between violent and
nonviolent husbands would he harder to discover.
3
Analysis of the MV group's MCTS-assessed marital aggression indi-
cated that 67.7% (n = 21) engaged in only "mild" acts of physical
aggression and 32.3% (n = 10) reported at least one act of severe
violence.
* Although the widely used SMAT is adequate as a gross screening
measure of marital satisfaction (e.g., Smith, Vivian, & O'Leary, 1990),
the Dyadic Adjustment Scale (DAS; Spanier, 1976) was also adminis-
tered upon participants' arrival to the laboratory to provide a more
comprehensive assessment of marital quality and cohesion.
262 ECKHARDT, HARBOUR, AND DAVTSON
measures of negative emotion and psychopathology in nonclinical and
clinical samples from the United States and abroad (Kassinove & Eck-
hardt, 1994; Muran, Kassinove, Ross, & Muran, 1989).
Dysfunctional Attitudes Scale. The 40-item Dysfunctional Attitudes
Scale (DYS; Weissman & Beck, 1978) was originally designed to mea-
sure specific depressogenic cognitive biases. More recent findings have
indicated that the DYS assesses cognitive biases that are not necessarily
specific to depressive syndromes; nondepressed clinical samples also
scored higher than relevant comparison samples (Hollon, Kendall, &
Lumry, 1986). As such, the DYS represents the only available measure
of general cognitive bias. Items are rated on a 7-point Likert scale. A
total cognitive bias score is computed by summing across the 40 items
with higher scores reflective of greater cognitive bias.
Procedure
ATSS. Upon arrival at the laboratory, all men signed a consent form
acknowledging that the study was investigating negative emotions in
marriage and that the actual experience of negative emotions during the
experiment was likely. Following completion of an assessment packet,
participants were introduced to the ATSS procedure via tape-recorded
instructions, which stated that they would be asked to listen to several
taped scenarios, imagine that they were actually involved in each, and
when prompted by a tone, to talk out loud about their thoughts and
feelings into a microphone connected to a hidden tape recorder. The
anonymity of participants' responses was emphasized to encourage open
and frank thought articulation. When participants indicated that they
understood the procedure following a nonarousing practice scenario, the
digital-quality ATSS scenarios were played by means of a Sony Digital
Audio Tape (DAT) player and Pioneer large cup headphones. The experi-
menter left the laboratory room until the end of the ATSS procedure.
The ATSS procedure included three stimulus scenarios (control, over-
heard conversation, and jealousy), each divided into eight 30-s segments.
The taped characters were portrayed by paid professional actors. On the
basis of pilot data by Eckhardt and Kassinove (in press), we used three
stimulus scenarios, two of which were anger-inducing.s In the control
scenario, the participant was asked to imagine that he and his wife were
invited to another couple's house, where they all played a game. Al-
though the couple was somewhat competitive and annoying, the source
of this annoyance had nothing to do with the participant's wife. Tn the
anger-arousing overheard conversation scenario, the participant imagined
that he was overhearing a conversation between his wife and her friend,
who negatively evaluated the participant and expressed uncertainty about
remaining in the marriage. This scenario was scripted around the themes
of wife abandonment and wife-to-husband ridicule, which have been
previously shown to be associated with spousal violence (e.g., Daly &
Wilson, 1988;Holtzworth-Munroe&Hutchinson, 1993).In the jealousy
scenario, the participant imagined that he had arrived home early to find
his wife talking and laughing with a male acquaintance. The participant
decided to listen to their conversation, which consisted of subtly flirta-
tious dialogue. The purpose was to elicit anger by suggesting that the
wife may have been or might become amorous with another man, capi-
talizing on the theme of male sexual jealousy (Daly, Wilson, & Weghorst,
1983).
The three stimulus tapes were presented to participants successively
on a single tape. Because it was expected that participants might initially
be uncomfortable with the procedure, the control tape was always pre-
sented first. The presentation order of the two anger-arousing tapes was
counterbalanced. Following the ATSS procedure, the State Anger Scale
was again administered.
Debriefing. Following the ATSS procedure, participants were fully
debriefed for a period of 15 to 30 min, The "process" debriefing was
designed to fully explain the rationale of the study and explore the nature
of the men's affective reactions to the procedure. All participants were
given the opportunity to engage in relaxation training, and all MV partici-
pants were given the phone numbers to local spouse abuser support
groups and mental health centers. No participant was allowed to leave
the laboratory in emotional distress, and follow-up telephone calls were
made to MV participants to assess postexperimental reactions.fi
ATSS Coding Procedures
Each participant's tape containing his articulated thoughts for the
three stimulus situations was transcribed by undergraduate research as-
sistants. Two raters (advanced undergraduates) who were not aware of
the experimental conditions received 30 hr of training according to an
ATSS coding manual7 and on basic issues regarding cognitive theories
of emotion. Following training to criterion, the two raters read transcrip-
tions of participants' articulated thoughts while listening to their articu-
lated thoughts tape and independently rated each segment on a special-
ized coding sheet. Pearson product-moment coefficients were computed
to assess inter-rater agreement.
Ratings were made in four dependent variable categories: irrational
beliefs, cognitive biases, hostile attributional biases, and anger-control
strategies. Irrational belief and cognitive bias verbalizations were rated
on a 5-point scale (0-4) that assessed the degree to which each variable
was present in each segment, with 0 indicating not present at all and 4
indicating extremely present. Scores were calculated by averaging the
two coders' ratings on each variable for each segment, and summary
scores were calculated by totaling the averaged ratings across the eight
segments of each scenario. We assessed the hostile attributional bias
and anger-control statements variables by tallying the frequencies of each
verbalization across each scenario.
Irrational beliefs. The total irrational beliefs score, a composite of
the four specific irrational beliefs, had very high intercoder reliability
(r = .92). As outlined by Ellis and Dryden (1987), four core irrational
beliefs are presumed to contribute to a general style of irrational think-
ing: (a) awfulizing (an exaggerated rating of the "badness1" of an
aversive event evidenced by words such as "awful," "horrible," "terri-
ble," or "catastrophic"); (b) low frustration tolerance (asevere intoler-
ance for discomfort evidenced by phrases such as "can't stand," "can't
take," "can't tolerate"); (c) demandingness (rigid, absolutist beliefs
that events or people must be a certain way and that conditions such as
3
Transcripts of the ATSS stimulus tapes can be obtained from Christo-
pher I. Eckhardt.
6
Here is an additional note about debriefing. Because anger induction
is neither pleasant to administer nor receive, we were concerned about
participants (especially MV men) leaving our study angrier than when
they arrived. However, it should be kept in mind mat the entire duration
of anger arousal was about 14 min (half of which consisted of articula-
tion pauses). The process debriefing was structured for all participants to
assess (a) thoughts about the experiment, (b) any emotions experienced
during the ATSS procedure, (c) the intensity and subjective experience
of those emotions, (d) whether the men actually thought of their own
relationships during the imagined scenarios, and (e) whether they felt
more negative emotions at the current time than when they began the
experiment on a scale ranging from 1 (perfectly calm) to 10 (intensely
aroused). By the end of the debriefing session, 25 (81 %) of the 31 MV
participants indicated that they did not feel more negative emotions than
when they arrived. The 6 men who indicated that they did feel upset
indicated an emotionality level of 5 or below when queried. All MV
and DNV men were encouraged to seek counseling and appropriate
referrals to competent professionals were given. Finally, the follow-up
phone calls indicated that no participant felt any lingering unpleasantness
as a result of their participation.
1
A copy of the ATSS coding manual can be obtained from Christopher
I. Eckhardt.
 ARTICULATED THOUGHTS AND MARITAL VIOLENCE
success and approval are absolute necessities); and ( d ) self/other rating
(evaluations of the total value or worth of a human being on the basis
of a specific behavior or attribute in the basic form of "He/She/They/
I is/am/are a "). Interrater reliabilities ranged from .67 (awfuliz-
ing) to .94 (low frustration tolerance), with a mean of .85.
Cognitive biases. The total cognitive bias score, a composite of the
following six automatic thoughts, had good intercoder agreement (r =
.82). The following six individual cognitive biases outlined by Beck
(Beck, 1976; Beck, Rush, Shaw, & Emery, 1979) were assessed: (a)
arbitrary inference (making assumptions or drawing conclusions in the
absence of supporting evidence); (b) selective abstraction (understand-
ing an experience on the basis of a detail taken out of context, while
ignoring more salient aspects of the situation); (c) overgeneralization
(constructing a general rule from one or a few isolated incidents and
applying this rule universally); (d) magnification (overestimating the
significance of events and reacting incongruously to the presenting situa-
tion); (e) personalization (the tendency to engage in self-referent think-
ing even when presented with situations having little to do with the
self); and (f) dicholomous thinking (categorizing an event in one of
two extremes and thinking in all-or-none terms). Interrater agreement
ranged from .11 (selective abstraction) to .88 (magnification), with an
average of .72. Omitting the very low reliability for selective abstraction
results in r — .84.
Hostile attributional biases. This variable was coded by tallying the
number of statements wherein participants blamed the cause of an event
on the malicious and hostile intentions of another character. Hostile
attributional bias was considered a special case of the arbitrary inference
distortion in that the individual arrives at a conclusion in the absence of
confirming evidence (an incorrect conclusion regarding the protagonist's
intentionality). However, the hostile attributional bias involves not only
the misperception of causal intent but also the assumption of hostile
motivation (Epps & Kendall, 1995). For example, the statement "It's
all her fault!" would be classified as arbitrary inference because it is
not clear if this intentionality conclusion is valid. However, the statement
"She meant for this happen just to get back at me!" is classified as
hostile attributional bias because there is a question of both intent and
hostile motivation. Interrater agreement was very high (r = .92).
Anger-control statements. Finally, articulated usage of strategies for
anger control were rated by summing their frequency of occurrence.
These statements included expressing the desire to walk away from the
character or escape the situation in order to calm down, attempting to
actively change one's views about the situation or characters in order to
decrease negative emotionality, suggesting counseling or other external
mediation, initiating a request to calmly talk over the situation, or other
anger-control strategies. Interrater agreement was again very high
(r = .94).
Results
Validity o/ATSS Anger Induction
To assess whether the three groups of men were indeed an-
gered by the ATSS procedure, we conducted three dependent
samples t tests on STAXI State Anger scores gathered before
and immediately after the ATSS procedure. Results indicated
that men in all three participant groups reported significant in-
creases in state anger after the ATSS procedure: For MV, f(30)
= 3.82, p < .001; for DNV, t(22) = 3.55, p < .002; and for
SNV, f(32) = 3.23, p < .003. The three groups did not signifi-
cantly differ in post-ATSS state anger, F(2, 84) = 2.26, ns.
Thus, all three groups reported that the ATSS tapes were anger
inducing.
263
Group Differences on ATSS Cognitive Variables
Table 1 presents means and standard deviations for all ATSS
cognitive variables. As post hoc analyses revealed that there
were nonsignificant within-group differences between ATSS
jealousy scenario and overheard conversation scenario scores
for any cognitive variable, the total scores of these two scenarios
were averaged into a single anger scenario score. To examine
general group differences in articulated cognitive processes,
products, and deficiencies, we first present the results of four 3
X 2 (Group X Scenario) mixed ANOVAs on the ATSS variables
of total irrational beliefs, total cognitive biases, frequency of
hostile attributional biases, and frequency of anger-control state-
ments. Effect sizes are reported in the form of eta (77). Where
significant main effects coexisted with significant interaction
terms, only significant interactions were probed. Finally, direct
discriminant function analyses were undertaken with group
membership being predicted by the four irrational beliefs, six
cognitive biases, frequency of hostile attributional biases, and
frequency of anger-control statements.
Irrational beliefs. On the total irrational beliefs score, the
ANOVA indicated significant effects of group, F(2, 77) = 7.08,
p < .002,17 = .39; scenario, F(1, 77) = 27.57, p < .001, r\ =
.51; and the Group X Scenario interaction, F(2, 77) = 4.63, p
< .013, r) = .33. To explore this interaction, we evaluated
within-group differences between the control scenario and anger
scenario using dependent t tests. In addition, we examined group
differences within each scenario using univariate ANOVAs and
Tukey's (1977) honestly significant difference (HSD) test as
a control for Type I errors resulting from multiple pairwise
comparisons. Although MV and SNV husbands articulated a
significantly greater level of articulated irrational beliefs in the
anger tape versus the control scenario—for MV, ((27) = 4.91,
p < .001; for SNV, /(30) = 4.07, p < .001—DNV men articu-
lated a similar degree of irrationality across the two scenarios,
1(20) = .66, ns. During the control scenario, the ANOVA indi-
cated a significant effect of group, F(2, 79) = 4.23, p < .02.
The HSD test indicated that MV men were not significantly
different than DNV men during the control scenario, with both
M V and DNV men articulating greater levels of irrational beliefs
than SNV men. During the anger scenario, the univariate
ANOVA indicated a significant group effect, F(2, 77) = 7.42,
p < .01, with MV husbands scoring significantly higher than
DNV and SNV men. DNV did not differ significantly from
SNV men.
Cognitive biases. On the total cognitive biases score, the
ANOVA indicated significant effects of group, F(2, 77) =
15.32, p < .001, rj = .53; scenario, F(l, 77) = 121.92, p <
.001, rj = .78; and the Group X Scenario interaction, F(2, 77)
= 7.49, p < .001. All three groups significantly increased their
level of cognitive bias articulations from the control scenario to
the anger scenario: For MV, <(27) = 7.69, p < .001; for DNV,
((20) = 3.09, p < .006; for SNV, r(30) = 9.66, p < .001.
During the control scenario, the univariate ANOVA indicated a
significant effect of group, F(2, 78) = 9.06, p < .001. The
HSD test indicated that MV men were not significantly different
from DNV men, with both MV and DNV men articulating
greater levels of cognitive bias than SNV men. During the anger
scenario, the univariate ANOVA indicated a significant group
264 ECKHARDT, BARBOUR, AND DAVISON

Table 1
Means (and Standard Deviations) for ATSS Cognitive Variables

Control scenario Anger scenario

MV DNV SNV MV DNV SNV

Variable M SD M SD M SD M SD M SD M SD

Irrational beliefs
Awfulizing 0.03 0.18 0.07 0.32 0.03 0.12 0.19 0.71 0.03 0.12 0.14 0.28
LFT 0.43 1.29 0.45 0.94 0.05 0.27 1.67 3.10 0.76 1.04 0.98 1.29
Demanding 5 17 4.38 4.64 4.93 3.24 2.50 10.04 6.15 606 4.15 6.10 5.04
Rating 2.16 2.75 2.72 3.42 0.79 1.25 3.74 3.76 2.52 1.83 1.39 1.34
Total 7.79 6.14 7.89 7.15 4.1 1 3.31 15.55 10.20 9.23 4.78 8.60 5.75

Cognitive biases
Arbitrary inf 3.12 3.18 2.16 2.45 1.11 1.68 6.71 3.36 4.62 3.16 4.23 3.66
Selective abs 0.03 0.18 0.00 0.00 0.00 0.00 0.35 0.56 0.07 0.27 0.03 0.11
Overgen 1.02 1.28 1.16 1.65 0.94 1.78 1.67 1.46 2.36 3.49 1.91 2.15
Magnification 5.88 4.47 5.02 6.47 1.92 2.01 12.22 7.08 7.25 3.08 5.70 5.00
Personalization 0.46 0.97 0.30 0.59 0.21 0.44 3.03 4.12 0.82 1.16 2.37 3.02
Dichot think 3.98 4.28 5.25 4.53 1.79 2.01 10.28 4.72 6.72 3.70 6.20 3.46
Total 14.91 9.12 13.89 11.74 5.97 5.34 34.26 13.16 21.73 8.03 20.45 9.61

Hostile 0.59 1.18 0.36 0.73 0.22 0.70 1.14 0.87 0.36 0.51 0.25 0.46
Ang con 0.07 0.18 0.18 0.45 0.18 0.35 0.87 1.00 1.06 1.01 2.64 2.31

Note. ATSS = Articulated Thoughts in Simulated Situations; MV = maritally violent (n = 31); DNV =
maritally distressed-nonviolent (ti = 23); SNV = maritally satisfied-nonviolent (n = 34); LFT = low
frustration tolerance; Demanding = demandingness; Rating = self/other rating; Arbitrary inf = arbitrary
inference; Selective abs = selective abstraction; Overgen = Overgeneralization; Dichot think = dichotomous
thinking; Hostile = Hostile attributional bias; Ang con = Anger-control statements.

effect, F(2,78) = 14.49,p < .001, with MV husbands articulat-
ing a significantly greater degree of cognitive biases than DNV
and SNV men. DNV men did not differ significantly from
SNV men.
Hostile attributional biases. On the hostile attributional bias
variable, the ANOVA indicated a significant effect of group,
F(2, 79) = 9.52, p < .001, r, = .44. No significant effects were
found for the scenario effect, F(1, 79) = 2.56, ns, or the Group
X Scenario interaction, F(2, 79) = 2.35, ns. Across both sce-
nario conditions, MV men articulated a significantly greater
number of hostile attributional biases than DNV men, ((79) =
3.06. p < .003, and SNV men, f(79) = 4.19, p < .001. DNV
men did not differ from SNV men, ((79) = .78, ns.
Anger-control statements. For the frequency of articulated
anger-control statements variable, the ANOVA indicated a sig-
nificant effect of group, F(2, 79) = 10.36, p < .001, 77 = .46;
a significant effect of scenario, F(l, 77) = 58.15, p < .001,
17 = .65; and a significant Group X Scenario interaction, F(2,
79) = 9.79, p < .001. »j = .45. All three groups significantly
increased their level of anger-control statements from the control
scenario to the anger scenario; ForMV, ((28) = 4.24, p < .001;
for DNV, ( ( 2 1 ) = 3.67, p < .001; for SNV, ((30) = 6.09, p <
.001. During the control scenario, the ANOVA indicated no
effect of group, F(2, 79) = 0.96, ns. During the anger scenario,
the univariate ANOVA indicated a significant group effect, F(2,
79) = 10.50, p < .001, with SNV husbands scoring significantly
higher than MV and DNV men. MV men did not differ signifi-
cantly from DNV men.
Predicting group status from ATSS cognitive variables. Be-
cause our prior analyses clearly indicated that MV men were
characterized by excesses on aggregate cognitive variables dur-
ing anger arousal, in the present section we examine the role of
specific cognitive variables in the differentiation of participant
groups. To control for the increase in Type I error rates that
would have resulted from numerous univariate ANOVAs for
specific cognitive variables and to more directly answer the
research question, we used a direct discriminant function analy-
sis with specific ATSS anger scenario cognitive variables to
predict group status. In accordance with recommendations by
Pedhazur (1982) and Mueller and Cozad (1993), interpretation
of the functions will be made primarily using structure coeffi-
cients (pooled within-group correlations between predictors and
functions) rather than standardized discriminant function coef-
ficients, although both types of data are presented in Table 2.
The first of the two obtained functions explained 59% of the
variance in group membership, ^; 2 (22) = 63.76, p < .0001,
and the remaining function explained 15% of the variance in
group membership, x 2 (10) = 11.65, ns. Using predictor-func-
tion correlations above .30 as the guideline for determining
significance (Pedhazur, 1982), examination of correlations be-
tween the two functions and the set of cognitive predictor vari-
ables indicated that hostile attributions, magnification, dichoto-
mous thinking, self/other rating, arbitrary inference, and de-
mandingness were significantly correlated with Function 1,
whereas anger-control statements and personalization loaded sig-
nificantly on Function 2. Anger control statements were nega-
tively correlated with Function 1. Analysis of group centroids
indicates that high scores on Function 1 were characteristic of
MV participants (Function 1 centroid = 1.32; Function 2
centroid = .17), whereas higher scores on Function 2 were
 ARTICULATED THOUGHTS AND MARITAL VIOLENCE 265

Table 2
Results of Discriminant Function Analysis Using ATSS Cognitive
Variables to Predict Group Membership

Correlations with discriminant Standardized discriminant

functions function coefficients

Variable Function 1 Function 2 Function 1 Function 2

Hostile attributions .65 .38 .93 .06

Magnification .51 .11 .18 -.00
Dichotomous thinking .47 .17 .41 .05
Self/other rating J9 -.08 .13 .02
Arbitrary inference .36 .26 -.44 .21
Demandingness .34 .27 -.21 .15
Anger-control statements -.42 .76 -.61 .70
Personalization .14 .60 .37 .38
Awfulizing .07 .27 -.25 .12
Overgeneralization -.05 -.23 -.33 -.09
Low frustration tolerance .13 .20 -.16 .27

Function Eigenvalue Canonical R Wilks's A <tf

1 1.06 .72 .41 63.76 22 .001
2 0.18 .39 .85 11.65 10 .31

Note. Significant loadings in boldface type. ATSS = Articulated Thoughts in Simulated Situations.

more characteristic of SNV participants (Function 1 centroid =
— 1.01; Function 2 centroid = .31). DNV participants' group
centroid was low on both Function 1 (centroid = —.26) and
Function 2 (centroid = —.68). Overall classification results
indicate that 70% of cases were correctly classified according
to the two discriminant functions.
Self-Report Measures of Cognitive Distortion
Between-group one-way ANOVAs on the total scores of the
SPB and DYS were examined. Results indicated a significant
group effect on Total SPB score, F(2, 71) = 4.04, p < .02,
with MV men endorsing greater irrationality than SNV men,
( ( 7 1 ) = -2.68, p < .009, but not DNV men, r(71) = -.47,
ns. DNV men endorsed greater irrationality than SNV men,
f(71) = -2.03, p < .05. On the total score of the DYS, no
significant group effects were observed, F( 2, 79) = 2.72, ns.
To assess the differential ability of the questionnaire measures
of cognition to predict the group status of individual cases,
the SPB subtests of awfulizing, demandingness, low frustration
tolerance, and self/other rating, and the DYS total score were
entered into a discriminant function analysis. The analysis pro-
duced two nonsignificant functions, with Function 1 explaining
19.2% of the variance in group membership x 2 (10) = 14.24,
p < .16, and Function 2 explaining 2.7% of the variance in group
membership, x 2 ( 4 ) = 1.86, p < .76. As might be expected from
the lack of significance among the two functions, only 48.61%
of cases were accurately classified and no additional analyses
were undertaken.
Correlations Between ATSS and Questionnaire
Cognitive Variables
This lack of convergence led us to evaluate the degree of
correlation between the cognitions articulated by MV men dur-
ing the ATSS procedure and those assessed by the question-
naires. No significant correlations were obtained (range =
.01 -.26). However, because participants experienced anger dur-
ing the ATSS anger scenario but not during the completion of
the questionnaires, we also evaluated correlations between ATSS
control scenario articulations (no anger arousal) and question-
naire measures of cognitions within the MV sample. Once again,
no significant correlations were obtained. The overall pattern
of nonsignificance remained even when the three groups were
collapsed into one, although one significant correlation emerged
between ATSS anger scenario total irrational beliefs and SPB
total irrationality (r = -33, p < .01).
Differentiating Mildly Versus Severely Aggressive Men
To evaluate the ability of ATSS cognitive data to discriminate
between MV men who engaged in "mild" acts of physical
aggression and those who reported engaging in at least one act
of severe wife-directed violence, a discriminant function was
calculated using all ATSS cognitive variables as predictors. The
resulting function, although nonsignificant, x z (12) = 17.32,.
p < .14, explained 58% of the variance in the prediction of
mild versus severe aggression. Because the eigenvalue, canoni-
cal correlation, and amount of explained variance were of large
magnitude, the function was deemed interpretable. As can be
seen in Table 3, the ATSS anger scenario articulations of
demandingness, magnification, and awfulizing were signifi-
cantly correlated with the function. Frequency of anger-control
statements approached significance. Analysis of group centroids
indicated that high scores on the function were associated with
being classified a severe batterer (severely aggressive = 1.79;
mildly aggressive = —.72). Overall, the function correctly clas-
sified 92.9% of MV participants.
Using cognitive data gathered from the SPB and DYS, we
266 ECKHARDT, HARBOUR, AND DAVISON

Table 3
Results of Discriminant Function Analysis Using ATSS Cognitive Variables
to Predict Severe Versus Mild Aggression

Correlations with Standard discriminant

Predictor discriminant functions function coefficients

Demandingness .44 1.48

Magnification .40 .44
Awfulizing JO -.04
Anger-control statements -.29 -.29
Selective abstraction .24 .43
Overgeneralization -.21 -.51
Self/other rating .14 -.22
Arbitrary inference .10 -.30
Dichotomous thinking .04 .94
Hostile attributions .02 -.38
Low frustration tolerance .01 .22
Personalization .00 .58

Function Eigenvalue Canonical R Wilks's \ X2 df p

1 1.38 .76 .42 17.32 12 .14

Note. Significant loadings in boldface type. ATSS = Articulated Thoughts in Simulated Situations.

evaluated the ability of self-reported cognitive data to differenti-
ate between mildly and severely aggressive men. The single
nonsignificant function obtained from the questionnaire data
explained only 15.3% of the variance in the prediction of mild
versus severe violence, X 2 ( 5 ) = 3.25, p < 66. Because of the
small amount of variance explained by this function, no further
interpretation of these data will be made.
Discussion
The data indicated that articulated cognitive distortions, cog-
nitive deficiencies, and cognitive products differentiated MV
husbands from their nonviolent counterparts during anger induc-
tion. During anger arousal, aggregate irrational beliefs and cog-
nitive biases were articulated to a greater degree by MV men
than DNV and SNV men, as cognitive theorists (Beck, 1976;
Ellis, 1977) would predict. Additionally, discriminant function
analyses indicated that specific ATSS irrational beliefs and cog-
nitive biases differentiated (a) between MV men and their non-
violent counterparts, and (b) between severely violent and
mildly aggressive husbands. MV men also articulated more hos-
tile attributional biases, but fewer anger-controlling statements.
Although few notable differences in cognitive processing were
present between MV and DNV men in the absence of anger
arousal, MV men clearly differentiated themselves from the
other groups during anger-arousing scenarios. This points to a
substantial strength of the ATSS assessment method in that it is
designed for the measurement of cognition while respondents
are experiencing affective arousal. Because we controlled for
the effects of marital discord on aggression by including a com-
parison sample of DNV men, we are able to conclude that these
cognitive distortions are associated with the presence of violence
in the participants' relationships rather than levels of marital
discord.
How might cognitive distortions contribute to marital vio-
lence? A social information processing framework (Holtzworth-
Munroe, 1992) posits that MV men may selectively attend to,
actively misconstrue, or otherwise distort situations in ways that
increase the likelihood of marital anger and aggression. MV
participants in the present study, for example, were character-
ized by verbalizations reflective of a tendency to magnify the
importance of aversive situations, dichotomous thinking that
establishes rigid all-or-none ground rules for acceptable and
unacceptable behavior, inflammatory ratings of other individu-
als' worth, articulations characterized by absolutist demands
that people act appropriately, and arbitrary inferences estab-
lished in the absence of confirming evidence. In addition, we
found that specific cognitive distortion articulations distin-
guished between violent husbands who had engaged in more
severe forms of marital violence and those husbands who used
more mild forms of aggression. The 32% of our MV sample
who reported at least one act of severe violence were discrimi-
nated on the basis of high levels of ATSS demandingness, magni-
fication, and awfulizing, and low levels of anger-control state-
ments. If these cognitive distortions are assumed to be trait-
like, automatic processes that are applied in a wide variety of
potentially provocative situations, it follows that MV men will
experience anger arousal and perceive threat provocation at a
higher frequency than individuals who do not exhibit this pattern
of information processing.
Social information processing models also predict that defi-
ciencies in making effective and appropriate decisions can also
contribute to aggressive outcomes (Holtzworth-Munroe, 1992;
Holtzworth-Munroe & Anglin, 1991). MV men in the present
research did not spontaneously produce thought articulations
that function to de-escalate and control anger arousal. SNV
husbands articulated anger-controlling cognitions at a signifi-
cantly higher rate than MV and DNV men, which corresponds
to similar findings by Dutton and Browning (1988) and Holtz-
worth-Munroe and Anglin (1991). As reported by numerous
marital interaction researchers (e.g., Burman et al., 1993), MV
 ARTICULATED THOUGHTS AND MARITAL VIOLENCE
couples appear to be entrenched in a communication style during
marital conflicts that involves an escalating cascade of recipro-
cated anger, contempt, and belligerence. The present data sug-
gest that, in part, MV men are unable to extricate themselves
from this aversive communication pattern because of their in-
ability to spontaneously generate anger-controlling, prosocial
coping strategies. There are important clinical implications asso-
ciated with this finding. Whereas cognitively oriented therapies
are generally structured around the goal of identifying and modi-
fying distortions in information processing that affect emotions
and behavior, merely substituting rational beliefs for their irra-
tional counterparts, or objectively valid thoughts in place of
faulty automatic inferences, does not address the individual's
decision-making skill deficiencies in the area of anger coping
and conflict management. Skill-building techniques such as so-
cial skills training, problem-solving therapy, or interventions that
teach couples how to argue "better" (Gottman, 1994) may thus
be important treatment adjuncts for MV men.
Do the ATSS data support the primacy of cognition in the
enactment of marital aggression? Perhaps, but our data are cor-
relational and not without their limitations. First, it is still un-
known whether MV men possessed these cognitive processes
and products before engaging in violence against their wives,
or whether they emerged as a result of the violence. Second,
the present assessment scheme coded for the presence of hostile
attributional biases regardless of the individual they were di-
rected toward, which stands in contrast to previous studies of
MV men (e.g., Holtzworth-Munroe & Hutchinson, 1993) that
only assessed wife-directed attributional biases. In addition, al-
though the ATSS anger-arousing scenarios depicted conflict situ-
ations of direct relevance to the marital relationship, they did
not involve an actual husband-wife argument per se. Future
studies might examine the utility of including argument-based
scenarios. Finally, although the ATSS procedure provides an
optimal method for assessing affect-relevant cognitions as they
are occurring on-line, there is still an opportunity for respon-
dents to withhold or otherwise distort their responses. However,
the fact that respondents do not verbalize everything that is on
their minds during ATSS assessment does not compromise data
that are coded schematically (Kashima & Davison, 1990), as
they were in the present study. Indeed, the relatively large effect
sizes reported in the present research suggest that MV men were
minimally affected by social desirability and censoring issues.
Thus, given these limitations, the present data suggest that cogni-
tive variables may strongly influence affective arousal and play
a key role in the complex interaction of events that unfold during
relationship conflicts.
Because cognitions are so heavily connected to intense af-
fective states such as anger, it is critical that future researchers
continue to tailor cognitive assessment methods to address this
cognition-emotion association. A strength of the ATSS proce-
dure is its ability to tap into on-line, affect-relevant cognitive
processes that may not be accessible in the absence of affective
arousal. One may hypothesize that for participants in the present
study, the elicitation of anger during the ATSS procedure acti-
vated anger-relevant cognitive processes. In accordance with
a spreading activation perspective on cognition and emotion
(Anderson & Bower; 1973; Bower, 1981; Collins & Loftus,
1975), MV men are hypothesized to possess more associative
267
connections between cognitive operations (e.g., cognitive distor-
tions, memories, images) and angry affect that are activated and
articulated during think-aloud tasks such as the ATSS procedure.
This theoretical position, which Persons and Miranda (1992)
have recently modified into the mood-state hypothesis, may ex-
plain the lack of relationship found in the present study between
the questionnaire measures of cognitive distortion and ATSS
cognitive data. The SPB and DYS possess strong psychometric
properties but were nevertheless poor predictors of group mem-
bership and severity of marital violence in the discriminant func-
tion analyses. According to the mood-state hypothesis, one
might hypothesize that cognitions were more accessible, and
thus more reportable, during the ATSS procedure because the
individual's emotional state was manipulated at the time of as-
sessment. However, it should be kept in mind that anger induc-
tion data in the present study are limited because anger was
only assessed before and immediately after the ATSS tapes were
presented. Future studies should more precisely assess the pro-
cess of anger arousal in MV men and at what point this arousal
influences cognition.
The present data support a social information processing
model of marital aggression (Holtzworth-Munroe, 1992). MV
men were differentiated from DNV and SNV husbands on the
basis of articulated cognitive processes and products during
anger arousal, and these cognitions further distinguished mildly
aggressive from severely aggressive husbands. Future research
should address whether additional patterns of cognitive activity
distinguish between different subtypes of MV men (Holtzworth-
Munroe & Stuart, 1994b). For example, are batterers who pre-
sent with a psychopathic pattern of emotional (under) arousal
and behavior (generally aggressive/antisocial subtype) charac-
terized by a different set of distorted cognitive processes than
batterers who are aggressive while intensely angry and enraged
(dysphoric/borderline subtype)? In addition, researchers could
examine whether information processing mechanisms operating
at a preattentive level differentiate violent from nonviolent hus-
bands. Do MV men characterized by high levels of anger bias
their attention during preliminary information processing to-
ward anger-relevant cues at the expense of other contextual cues
that would otherwise not induce anger? It appears crucial to
continue to integrate findings from diverse theoretical perspec-
tives and experimental methodologies in order to arrive at a
comprehensive model of marital violence to aid in the treatment
of abusive men and the prevention of the suffering experienced
by their victims.
References
Anderson, J., & Bower, G. (1973). Human associative memory. Wash-
ington DC: Winston.
Arias, I., & O'Leary, K. D. (1988). Cognitive-behavioral treatment of
physical aggression in marriage. In N. Epstein, S. E. Schlesinger, &
W. Dryden (Eds.), Cognitive-behavioral therapy with families (pp.
118-150). New York: Brunner/Mazel.
Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical
aspects. New "fork: Harper & Row.
Beck, A. T. (1976). Cognitive therapy and the emotional disorders.
New "Vbrk: International Universities Press.
Beck, A. T., Rush, A. J., Shaw, B. E, & Emery, G. (1979). Cognitive
therapy of depression. New-York: Guilford Press.
268 ECKHARDT, BARBOUR, AND DAVISON
Bower, G. H. (1981). Mood and memory. American Psychologist, 36,
129-148.
Bradbury, T. N., Beach, S. R. H., Fincham, F. D., & Nelson, G. M.
(1996). Attributions and behavior in functional and dysfunctional
marriages. Journal of Consulting and Clinical Psychology, 64, 569-
576.
Burman, B., Margolin, G., & John, R. S. (1993). America's angriest
home videos: Behavioral contingencies observed in home reen-
actments of marital conflict. Journal of Consulting and Clinical Psy-
chology, 61, 28-39.
Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory
of semantic processing. Psychological Review, 82, 407-428.
Grossman, R., Stith, S., & Bender, M. (1990). Sex role egalitarianism
and marital violence. Sex Roles, 22, 293-303.
Daly, M., & Wilson, M. (1988). Homicide. New %rk: Aldine de Gruyter.
Daly, M., Wilson, M., & Weghorst, S. J. (1983). Male sexual jealousy.
Ethology and Sociobiology, 3, 11-27.
Davison, G. C, Feldman, P. M., & Osbom, C. E. (1984). Articulated
thoughts, irrational beliefs, and fear of negative evaluation. Cognitive
Therapy and Research, 8, 349-362.
Davison, G. C., Navarre, S. G., & Vogel, R. S. (1995). The articulated
thoughts in simulated situations paradigm: A think-aloud approach to
cognitive assessment. Current Directions in Psychological Science, 4,
29-33.
Davison, G. C., Robins, C., & Johnson, M. K. (1983). Articulated
thoughts during simulated situations: A paradigm for studying cogni-
tion in emotion and behavior. Cognitive Therapy and Research, 7,
17-40.
Davison, G. C., Vogel, R. S., & Coffman, S. G. (1997). Think-aloud
approaches to cognitive assessment and the Articulated Thoughts in
Simulated Situations paradigm. Journal of Consulting and Clinical
Psychology, 65, 950-958.
Demaria, T., Kassinove, H., & Dill, C. ( 1989). Psychometric properties
of the Survey of Personal Beliefs: A rational-emotive measure of
irrational thinking. Journal of Personality Assessment, 53, 329-341.
Dutton, D. G., & Browning, J. J. (1988). Concern for power, fear of
intimacy and wife abuse. In G. T. Hotaling, D. Finkelhor, J. T. Kirkpat-
rick, & M. Straus (Eds.), New directions in family violence research
(pp. 163-175). Beverly Hills, CA: Sage.
Eckhardt, C. I., Barbour, K. A., & Stuart, G. L. (1997). Anger and hostil-
ity in maritally violent men: Conceptual distinctions, measurement
issues, and literature review. Clinical Psychology Review, 17, 333-
358.
Eckhardt, C. I., & Kassinove, H. (in press). Articulated cognitive distor-
tions and cognitive deficiencies in maritally violent men. Journal of
Cognitive Psychotherapy.
Ellis, A. (1962). Reason and emotion in psychotherapy. New "York:
Citadel Press.
Ellis, A. (1977). Anger: How to live with and without it. New York:
Citadel Press.
Ellis, A, (1994J. Reason and emotion in psychotherapy (2nd ed.). New
York: Citadel Press.
Ellis, A., & Dryden, W. (1987). The practice of rational-emotive ther-
apy. New York: Springer.
Epps, J., & Kendall, P. C. (1995). Hostile attributional bias in adults.
Cognitive Therapy and Research, 19, 159-178.
Gottman, J. M. (1994). What predicts divorce? Hillsdale, NJ: Erlbaum.
Hogg, J. A., & Deffenbacher, J. L. (1986). Irrational beliefs, depression,
and anger among college students. Journal of College Student Person-
nel, 27, 349-353.
Hollon. S. D., Kendall, P. C., & Lumry, A. (1986). Specificity of de-
pressogenic cognitions in clinical depression. Journal of Abnormal
Psychology, 95, 52-59.
Holtzworth-Munroe, A. (1992). Social skill deficits in maritally violent
men: Interpreting the data using a social information processing
model. Clinical Psychology Review, 12, 605-618.
Holtzworth-Munroe, A., & Anglin, K. (1991). The competency of re-
sponses given by maritally violent men versus nonviolent men lo
problematic situations. Violence and Victims, 6, 257-268.
Holtzworth-Munroe, A., Bates, L., Smutzler, N., & Sandin, E. (1997).
A brief review of the research on husband violence. Part I: Maritally
violent versus nonviolent men. Aggression and Violent Behavior: A
Review Journal, 2, 65-99.
Holtzworth-Munroe, A., & Hutchinson, G. (1993). Attributing negative
intent to wife behavior: The attributions of maritally violent versus
nonviolent men. Journal of Abnormal Psychology. 102, 206-211.
Holtzworlh-Munroe, A., & Stuart, G. L. (I994a). An introduction to
marital violence treatment. In L. Vandecreek (Ed.), Innovations in
clinical practice: A source book (Vol. 13, pp. 5-14). Sarasota, FL:
Professional Resource Press.
Holtzworth-Munroe, A., & Stuart, G. L. (1994b). Typologies of male
batlerers: Three subtypes and the differences among them. Psychologi-
cal Bulletin, 116, 476-497.
Holtzworth-Munroe, A., Waltz, J., Jacobson, N. S., Monaco, V,, Fehren-
bach, P. A., & Gottman, J. M. (1992). Recruiting nonviolent men as
control subjects for research on marital violence: How easily can it
be done? Violence and Victims, 7, 79-88.
Ingram, R. E., Kendall, P. C., Smith, T. W, Donnell, C., & Ronan, K.
(1987). Cognitive specificity in emotional distress. Journal of Per-
sonality and Social Psychology, 53, 734-742.
Jacobson, N. S., Gottman, J. M., Waltz, J., Rushe, R., Babcock, J., &
Holtzworth-Munroe, A. (1994). Affect, verbal content, and psycho-
physiology in the arguments of couples with a violent husband. Jour-
nal of Consulting and Clinical Psychology, 62, 982-988.
Kashima, K., & Davison, G. C. (1990). Functional consistency in the
face of topographical change in articulated thoughts. Journal of Ratio-
nal Emotive & Cognitive Behavioral Therapy, 7, 131-139.
Kassinove, H. (1986). Self-reported affect and core irrational thinking:
A preliminary analysis. Journal of Rational Emotive Therapy, 4, 119-
130.
Kassinove, H., & Eckhardt, C.I. (1994). Irrational beliefs and self-
reported affect in Russia and America. Personality and Individual
Differences, 16. 133-142.
Kendall, P. C. (1992). Healthy thinking. Behavior Therapy, 23, 1-12.
Kendall, P. C., & Dobson, K. S. (1993). On the nature of cognition and
its role in psychopathology. In P. C. Kendall & K. S. Dobson (Eds.),
Psychopathology and cognition (pp. 3-19). San Diego: Academic
Press.
Locke, H. J., & Wallace, K. M. (1959). Short marital adjustment and
prediction tests: Their reliability and validity. Marriage and Family
Living, 21, 251-255.
Margolin, G., John, R. S., & Gleberman, L. (1988). Affective responses
to conflictual discussions in violent and nonviolent couples. Journal
of Consulting and Clinical Psychology, 56, 24-33.
Mueller, R. O., & Cozad, J. B. (1993). Standardized discriminant coef-
ficients: A rejoinder. Journal of Educational Statistics, 18, 108-114.
Muran, J. C., Kassinove, H., Ross, S., & Muran, E. (1989). Irrational
thinking and negative emotionality in college students and applicants
for mental health services. Journal of Clinical Psychology, 45, 188-
193.
Neidig, P. H, Friedman, D. H., & Collins, B. S. (1986). Attitudinal
characteristics of males who have engaged in spouse abuse. Journal
of Family Violence, I, 223-233.
O'Leary, K. D., & Vivian, D. (1990). Physical aggression in marriage.
In F. D. Fincham & T. N. Bradbury (Eds.), The psychology of mar-
riage (pp. 323-348). New York: Guilford Press.
Pan, H., Neidig. P., & O'Leary, K. D. (1994a). Male-female and ag-
gressor-victim differences in the factor structure of the Modified
 ARTICULATED THOUGHTS AND MARITAL VIOLENCE 269

Conflict Tactics Scale. Journal of Interpersonal Violence, 9, 366-
382.
Pan, K, Neidig, P., & O'Leary, K. D. (1994b). Predicting mild and
severe husband-to-wife physical aggression. Journal of Consulting
and Clinical Psychology, 62, 975-981.
Pedhazur, E. (1982). Multiple regression in behavioral research. New
York: Holt, Rinehart, & Winston.
Persons,!. B., & Miranda, J. (1992). Cognitive theories of vulnerability
to depression: Reconciling negative evidence. Cognitive Therapy and
Research, 16, 485-502.
Rosenbaum, A., & O'Leary, K. D. (1981). Marital violence: Character-
istics of abusive couples. Journal of Consulting and Clinical Psychol-
ogy, 49, 63-71.
Saunders, D. G., Lynch, A. B., Grayson, M., & Linz, D. (1987). The
inventory of beliefs about wife beating: The construction and initial
validation of a measure of beliefs and attitudes. Violence and Victims,
2, 39-57.
Smith, D. A., Vivian, D., & O'Leary, K. D. (1990). Longitudinal predic-
tion of marital discord from premarital expressions of effect. Journal
of Consulting and Clinical Psychology, 58, 790-798.
Smith, T. W. (1982). Irrational beliefs in the cause and treatment of
emotional distress: A critical review of the rational-emotive model.
Clinical Psychology Review, 2, 505-522.
Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for
assessing the quality of marriage and similar dyads. Journal of Mar-
riage and the Family, 38, 15-28.
Spielberger, C. D. (1988). Manual for the State-Trait Anger Expression
Inventory. Odessa, FL: Psychological Assessment Resources.
Spielberger, C. D., Reheiser, E. C., & Sydeman, S. J. (1995). Measuring
the experience, expression, and control of anger. In H. Kassinove
(Ed.), Anger disorders: Definition, diagnosis, and treatment (pp. 49-
67). Washington, DC: Taylor & Francis.
Stith, S. M., & Farley, S. C. (1993). A predictive model of male spousal
violence. Journal of Family Violence, 8, 183-201.
Tukey, J. W. (1977). Exploratory data analysis. Reading, MA: Addison-
Wesley.
Weissman, A. N., & Beck, A. T. (1978). Development and validation
of the Dysfunctional Attitudes Scale. Paper presented at the annual
convention of the American Educational Research Association, To-
ronto, Ontario, Canada.
Received October 21, 1996
Revision received March 14, 1997
Accepted August 15, 1997 •

Low Publication Prices for APA Members and Affiliates

Keeping you up-to-date. All APA Fellows, Members, Associates, and Student Affiliates
receive—as part of their annual dues—subscriptions to the American Psychologist and
APA Monitor. High School Teacher and International Affiliates receive subscriptions to
the APA Monitor, and they may subscribe to the American Psychologist at a significantly
reduced rate. In addition, all Members and Student Affiliates are eligible for savings of up
to 60% (plus a journal credit) on all other APA journals, as well as significant discounts on
subscriptions from cooperating societies and publishers (e.g., the American Association for
Counseling and Development, Academic Press, and Human Sciences Press).

Essential resources. APA members and affiliates receive special rates for purchases of
APA books, including the Publication Manual of the American Psychological Association,
and on dozens of new topical books each year.

Other benefits of membership. Membership in APA also provides eligibility for

competitive insurance plans, continuing education programs, reduced APA convention fees,
and specialty divisions.

More information. Write to American Psychological Association, Membership Services,

750 First Street, NE, Washington, DC 20002-4242.

View publication stats