Biobehavioral Responses to Interpersonal Conflict During Anger Expression Among Anger-In and Anger-Out Men Sonia Suchday, Ph.D. Uniformed Services University of the Health Sciences
Kevin T. Larkin, Ph.D. West Virginia University
ABSTRACT To examine whether typical modes of anger expression (i.e., anger-in, anger-out) were related to cardiovascular, affective, behavioral, and cognitive responses to interpersonal conflict, 20 anger-in and 20 anger-out undergraduate men participated in 2 role plays, one in which they were instructed to exhibit their anger overtly and the other in which they inhibited their anger. Results showed that anger-in individuals used significantly more repression self-statements than anger-out individuals across both role play interactions (p < .01). Anger-out persons showed exaggerated diastolic blood pressure response in contrast to anger-in participants, but only during the exhibited anger role play (p < .04). When the anger exhibition role play followed anger inhibition, diastolic blood pressure responses were more intense (p < .05), and heart rate recovery was significantly slower (p < .03) among anger-out participants in contrast to anger-in participants. These findings indicate that modes of anger expression (trait) and contextual demands of the interaction (state) interact in complex ways to influence biobehavioral reactions to anger provocation. (Ann Behav Med
assumptions behind this hypothesis are true, the magnitude of behaviorally elicited cardiovascular response could serve as a physiological pathway through which any set of psychosocial responses that evoke differential cardiovascular reactions might exact their influence on the cardiovascular system. Such a model has been proposed, for example, to explain how the coronary-prone behavior pattern leads to heart disease (5) or how cynically hostile attitudes can lead to atherosclerosis (6). Based on these theoretical models, considerable research that examines whether different modes of anger expression (e.g., overt anger exhibition, anger inhibition) are related to exaggerated cardiovascular responses to stress (7–16) has emerged. Some researchers have observed exaggerated cardiovascular responses to stress among persons who inhibit anger expression compared to those who express anger overtly (13–15), and others have reported exaggerated cardiovascular reactions among anger expressive persons (2,11–12,16). In an effort to explain these contradictory findings, Engebretson, Matthews, and Scheier (7) proposed a “matching hypothesis” that stated that when an individual used a mode of anger expression (e.g., anger inhibition) that was different from his or her typical method of dealing with anger (e.g., anger exhibition), he or she evidenced exaggerated cardiovascular responses to those situations. They found evidence for this hypothesis in an investigation of young men differing in anger expression styles. In an attempt to replicate these findings, however, Lai and Linden (8) found no evidence for differential cardiovascular response to stress when male participants used an atypical versus typical anger expression coping style. Although many studies have examined the relation between anger expression and cardiovascular response to stress, no consistent or lawful relation has been observed. In addition to exaggerated cardiovascular responses to stress, mode of anger expression has been shown to affect cardiovascular recovery following provocation (17). Although some studies have shown that cardiovascular recovery following provocation is impeded by a lack of opportunity to react to the provocation (7,17), other research has revealed that antagonistic hostility (typically overtly expressed), but not neurotic hostility, impedes recovery following provocation in the laboratory (7,9,18). Like the evidence associating exhibited anger and anger inhibition with increased cardiovascular reactivity to provocation, both forms of anger expression have been associated with delayed recovery following provocation. A longitudinal study by Everson, Goldberg, Kaplan, Julkunen, and Salonen (19) sheds some light on the issue. According to these authors, both extremes of anger-in and anger-out
2001, 23(4):282–290)
INTRODUCTION Although there is ample evidence to suggest that both the subjective experience of anger and the overt expression of anger are related to cardiovascular health (e.g., 1–3), there is less certainty regarding the physiological mechanisms by which these psychosocial factors evoke atherosclerotic lesions. One model, originally coined the “reactivity hypothesis” (4), stated that persons who exhibited exaggerated cardiovascular reactions to mental challenges had the highest risk for subsequent cardiovascular disease in contrast to low-reactive counterparts. Provided that the
Portions of this article were presented at the 16th Annual Scientific Sessions of the Society of Behavioral Medicine, San Diego, March 1995. We thank Scott Schauss, Ty Callahan, Susan McClain, and Elizabeth Semenchuk for their assistance in conducting this project. Reprint Address: S. Suchday, Ph.D., Clinical Assistant Professor, Department of Medical and Clinical Psychology, Uniformed Services University of the Health Sciences, 4301 Jones Bridge Road, Bethesda, MD 20814. © 2001 by The Society of Behavioral Medicine.
282
Volume 23, Number 4, 2001 are positively correlated with the development of hypertension over a 4-year period. Congruent with Harburg, Blakelock, and Roeper’s (20) suggestion, Everson et al. (19) proposed that both the extreme exhibition of anger and intense anger suppression lead to a continued feeling of resentment. The overt expression of anger may lead to continued conflict and resentment, but anger inhibition may also lead to brooding, thus maintaining a sense of resentment following anger-provocation. Over time, this resentful experience of anger initially leads to elevated blood pressure responses, delayed cardiovascular recovery from stress, and eventually the development of hypertension. Research relating modes of anger expression with cardiovascular reactivity to stress or recovery from stress has focused predominantly on physiological measures, excluding any affective, behavioral, or cognitive responses that may be associated with them. A complete understanding of concomitant affective, behavioral, and cognitive responses may elucidate important relations across measurement domains that may lead toward a better understanding of the complete stress response. For example, previous research has shown that measures of affect are as powerful as trait measures of anger expression in predicting behavior in various anger provoking situations (21). Furthermore, measures of anger expression obtained via behavioral observation may contribute to a more comprehensive understanding of a person’s response to anger-eliciting situations than relying solely on physiological measures. The purpose of this study was to examine the affective, behavioral, cognitive, and cardiovascular correlates of anger expression among men classified as possessing two different typical modes of anger expression (i.e., anger-in and anger-out) in an attempt to clarify previous contradictory findings. Interpersonal tasks permit concomitant recording of behavioral responding of study participants and, because most anger is expressed in an interpersonal context, possess greater ecological validity for studies of anger expression than noninterpersonal tasks. In contrast to the two previous studies in which standard laboratory stressors were employed, this study used a set of interpersonal conflict stressors based on evidence that anger-evoking tasks consistently result in differences in cardiovascular reactivity between hostile and nonhostile participants in contrast to no-harassment or nonthreatening tasks (22). Moreover, in both Engebretson et al. (7) and Lai and Linden’s (8) studies, opportunities to express anger occurred only after the provocation where participants were asked to rate the confederate positively or negatively. Situations in daily life rarely occur in such a coordinated fashion. Finally, our project investigated the effects of two forms of anger expression (i.e., anger inhibition and anger exhibition) occurring in the same experimental session to mimic the occurrence of anger expression in the “real world.” Rarely can daily life experiences be grouped into neat categories of anger expression. Typically, a combination of different modes of anger expression (anger exhibition and anger inhibition) occurs within a fairly short time period. Hence, the study design comprised exhibited anger and inhibited anger role plays in quick succession. Because the order of task presentation could affect measures of
Anger Expression
283
anger expression, half the participants were instructed to exhibit anger first, and the remaining participants were instructed to inhibit anger first. Based on the findings of Engebretson et al. (7), it was hypothesized that anger-in men would demonstrate higher cardiovascular reactivity during an interpersonal provocation and delayed cardiovascular recovery following the interaction when they were required to exhibit their anger overtly compared to a situation where they inhibited their anger. In contrast, anger-out men would demonstrate higher cardiovascular reactivity during an interpersonal provocation and delayed cardiovascular recovery following the interaction when they were required to inhibit their anger compared to a situation where they expressed their anger overtly. Both anger-in and anger-out men were hypothesized to demonstrate significantly greater verbal and nonverbal negative behaviors, affect, and cognitive self-statements reflecting perceptions of being victimized and being righteously angry when asked to engaged in behaviors that were contrary to their stated preferred mode of functioning. METHOD Participants Two hundred eighty-two White male participants between the ages of 18 and 35 were recruited from undergraduate psychology classes at West Virginia University. Women were not recruited due to the uncertain effects of the menstrual cycle and reproductive hormones on measures of cardiovascular function (23). Participants were selected on the basis of responses to the Anger Expression Inventory (AXI) (24). Participants scoring within the top tertile on anger-in and below the sample mean on anger-out were classified as anger-in participants, and those scoring within the top tertile on anger-out and below the sample mean on anger-in were classified as anger-out participants. Participants were excluded from the experiment if they were taking any medication that affected the cardiovascular system, including beta blockers, anxiolytics, and sympathomimetics. Based on these criteria, 76 participants were classified as anger-in and 60 participants were classified as anger-out. From this sample, 20 anger-in and 20 anger-out men were randomly selected to participate in the study. Demographic data on these 40 participants, presented in Table 1, show that there were no significant differences on height; weight; and measures of frequency of activity, caffeine use, and alcohol intake between groups. Power analyses revealed that a sample of 20 participants per cell would be sufficient to detect group differences. Cell sizes were similar to cell sizes in two previous studies that reported significant differences between preferred styles of anger expression during opportunities to exhibit or inhibit anger (7,8). Engebretson et al. (7) included 19 participants per cell, and Lai and Linden (8) included cells comprising from 10 to 14 male participants. Assessment of Anger Expression Anger expression style was measured using Spielberger’s AXI (24). This inventory was originally designed to measure an-
284
Suchday and Larkin
Annals of Behavioral Medicine
TABLE 1 Demographic Characteristics of Anger-In and Anger-Out Participants
Demographic Variables Age (years) Height (in.) Weight (lb) Number of hours of aerobic activity/week Number of hours of physical activity/week Caffeine intake/Cups per day Alcohol intake/Drinks per week
Anger-Ina
Anger-Outa
M
SD
M
SD
21.2 69.7 181.2 3.0
3.5 3.8 35.7 3.1
19.5 70.3 174.6 4.0
1.7 2.7 27.0 4.6
4.8
5.8
5.7
3.5
4.2 12.2
4.2 11.8
3.4 13.11
2.5 10.0
structed not to reach a compromise during either vignette. During one scenario, participants were asked to exhibit their anger toward the confederate; during the other scenario, participants were asked to inhibit their anger. The instructions to exhibit anger were as follows: You get mad at your neighbor. Express your anger and irritation toward your neighbor and let him realize you are very mad with him and that you will not tolerate his inconsiderate attitude. SHOW HIM HOW ANGRY YOU ARE!
Note. No significant differences were found on any demographic variable. an = 20.
The instructions to inhibit anger were as follows: You get mad at your roommate. But do not let your roommate realize you are very angry with him because of his inconsiderate attitude. Do not let your feelings show while you convince him to clean the apartment. DO NOT LET HIM SEE HOW ANGRY YOU ARE!
ger expression as a bipolar dimension with anger-in as one dimension and anger-out as the other dimension. Reliability analysis, however, revealed that measures of anger-in and anger-out were independent. The final version of the scale consisted of 24 statements that defined an individual’s response when he or she was angry. Responses to items are given in terms of how frequently they engage in each of the stated behaviors on a 4-point scale, ranging from 1 (almost never) to 4 (almost always). Alpha coefficients for the anger-in and anger-out scale ranged from .73 to .84 using male and female participants, attesting to the instrument’s reliability. The convergent and divergent validities of these scales have also been found to be adequate (24). The selection criteria used in this study differed from previous studies, where anger-in scores on the AXI were subtracted from anger-out and a total anger expression score was used to classify participants as anger-in or anger-out (7), or where endorsement of more items on one scale than the other resulted in being assigned to the more frequently endorsed category (8). Because this study used more stringent criteria compared to these two studies, it is clear that all participants categorized as anger-in and anger-out in previous studies would have also been categorized as such using our criteria. Experimental Tasks In this investigation, participants engaged in two role plays that have been used in previous investigations to elicit anger (25–27). In each role play, participants were asked to reach a solution with a White male confederate introduced as “another participant.” In one vignette, the confederate played the role of a neighbor who refused to turn down loud music in his apartment. In the other vignette, the confederate played the part of an inconsiderate roommate who refused to clean the apartment despite repeated requests. The confederate was provided with a set of structured response prompts, which became increasingly “nasty” with the participants’ continued attempts to persuade him to turn his music down or clean the apartment (see 27). The same confederate was used during both role plays and was in-
The two different instructional sets (i.e., exhibited anger and inhibited anger) were presented in a counterbalanced order so that half the participants in each group received instructions to exhibit anger first and half received instructions to inhibit anger first. Similarly, half of the participants role played the exhibited anger with the “noise” scenario and inhibited anger with the “mess” scenario, and the remaining participants role played the “mess” scene while exhibiting anger and the “noise” scene while inhibiting anger expression. Both interactions with the confederate were videotaped using a JVC camcorder. Assessment of Cardiovascular Responses Cardiovascular responses were assessed throughout the laboratory session, during baseline as well as role play periods and during recovery. Heart rate (HR) was measured using three surface electrodes attached to the collarbone and lower chest of each participant; wave forms were transmitted to an EKG preamplifier and charted on a Grass Model 7 polygraph. An IBS automated sphygmomanometer was used to measure systolic blood pressure (SBP) and diastolic blood pressure (DBP). This apparatus employs an occluding cuff positioned over the brachial artery of the participant’s nondominant arm. Cuff inflation and deflation was controlled by a microprocessor; resulting SBPs and DBPs were digitally displayed and recorded by a laboratory assistant. Assessment of Affective, Behavioral, and Cognitive Responses State anger. Anger was assessed using the state form of the State and Trait Anger Scale (STAS) (24), which measures emotional responsivity to the perception of insult, injustice, or frustration in the environment. Awareness of physiological responses. The participants’ awareness of physiological responses occurring during interaction role plays was assessed using the Physiological Sensations
Volume 23, Number 4, 2001 Checklist (PSC) (28). Participants were asked to estimate their HR increase or decrease (in beats per minute) during the preceding task. Participants were also asked to rate the extent of arousal and anxiety experienced during the preceding role play on a 10-point Likert scale ranging from 1 (no arousal/anxiety) to 10 (most arousal/anxiety). Assessment of behavioral responses. Behavioral responses to both interpersonal tasks were coded using a modified version of the Marital Interaction Coding System (29). Based on previous work by Weider and Weiss (30), individual response codes were grouped into five categories: positive and negative nonverbal motor behaviors, positive and negative verbal content of the interaction, and problem solving. Each role play was divided into 30-sec segments and the frequency with which each category was observed during each segment was coded by two independent raters who were trained to 80% agreement using a criterion tape. Each rater then coded 60% of the data; 20% of data coded by both raters was used to assess maintenance of reliability defined as number of agreements divided by number of agreements plus disagreements. Agreement ratios ranged from .93 to 1.00 across all behavioral coding categories. Two anger-in and two anger-out participants’ behavioral data were lost due to video equipment failure. Cognitive responses. Cognitive responses were assessed using the Anger Cognitions Inventory (ACI) (31), which is a 27-item, 4-point Likert scale inventory assessing typical cognitions associated with anger. Participants were instructed to respond to each item, rating the frequency with which each self-statement occurred during the immediately preceding role play situation. The ACI is composed of four factors: Rational Coping, Repression/Distancing, Victimization, and Self-Righteousness. Items on the Rational Coping factor include self-statements that focus on the rational justification of anger (e.g., “I need to reason this through”). Items on the Repression/Distancing factor include self-statements concerning cessation of thinking about anger (e.g., “I’ll put it out of my mind”). Items on the Victimization factor include self-statements reflecting feelings of being unfairly treated and victimized (e.g., “Why should this happen to me?”). Items on the Self-Righteousness factor reflected cognitions that focused on self-statements that justified anger (e.g., “I have the right to be angry”). Procedure Participants were scheduled for a single 1-hr laboratory session. The participants were also requested to refrain from consuming any alcohol, nicotine, or caffeine for 2 hr prior to the laboratory session. This was confirmed via self-report prior to beginning the laboratory session. After the study was explained briefly and informed consent was obtained, electrodes were attached as explained previously, and the participant was introduced to a male confederate who was described as “another participant” who was recruited from the participant pool for this portion of the study. The confederate then
Anger Expression
285
left the room, and the participant was asked to complete the state form of the STAS to determine a resting level of anger. The participant was then asked to relax for a 20-min period; the last 6 min were recorded as the initial rest period. HR was measured continuously, and blood pressure was measured at 2-min intervals. Following the adaptation and initial rest period, the participant was presented with a description of the first vignette and instructions to inhibit or exhibit his anger. The confederate then entered the room, and the role play began with a standard prompt delivered by the confederate. The role play lasted for 3 min; blood pressure was taken each minute, HR was measured continuously, and the entire interaction was taped for coding behavioral responses. The experimenter then returned and requested that the participant complete the STAS, the ACI, and the PSC based on his experience during the preceding role play. The confederate left the room during this period. The completion of these questionnaires, which took approximately 5 min, was followed by a second rest period, during which cardiovascular measures were obtained (2 min recovery and 6 min rest). During the second scenario, participants who exhibited their anger during the first interaction were asked to inhibit it, and participants who inhibited their anger during the first interaction were asked to exhibit it openly. The participants were then debriefed and informed about the purpose of the study, and all recording apparatuses were detached. Data Analysis The primary analytic method used in this investigation was a 2 × 2 × 2 (Group [anger-in, anger-out] × Order [inhibited anger role play first or second] × Role Play [exhibited anger scene, inhibited anger scene]) mixed factors design. For all three cardiovascular indexes (i.e., HR, SBP, and DBP), analyses of covariance were used, adjusting for respective pretask resting levels. The confidence level adopted for all analyses was .05 and when significant, interactions were followed up using F tests for simple main effects. Because the purpose of this article is to compare responses between anger-in and anger-out males to the experience of anger, only main effects for group and Group × Role Play, Group × Order, and Group × Order × Role Play interactions are reported. The main effects for order and role play and the Order × Role Play interaction effects for cardiovascular responses have been reported previously in a methodological article describing the development and validation of the social confrontation paradigm used in this and other studies (27) and consequently are not discussed in this study. RESULTS Cardiovascular Measures Baseline measures. Prior to conducting analyses on the cardiovascular responses during role play scenes, analyses of variance (ANOVAs) were conducted on each baseline cardiovascular measure. There were no significant main effects or
286
Suchday and Larkin
interactions between the two groups on any of the three baseline cardiovascular measures. Task measures. The analysis of DBP reactivity revealed a significant Group × Role Play interaction, F(1, 35) = 4.35, p < .05 (see Figure 1). Although no significant difference was observed among adjusted mean DBP response of anger-in participants during exhibited anger and anger inhibition scenes, a significant difference was observed for anger-out individuals during the exhibited anger and inhibited anger role plays, F(1, 18) = 5.09, p < .04. Anger-out participants evidenced greater mean adjusted DBP reactivity during the exhibited anger role play than the inhibited anger role play.
Annals of Behavioral Medicine There was also a significant Group × Order interaction, F(1, 35) = 4.51, p < .05. No significant differences between the anger-in and the anger-out groups were observed when anger inhibition occurred after anger exhibition; however, there was a significant difference between the two groups’ mean DBP responses when inhibition occurred prior to exhibition, F(1, 18) = 4.48, p < .05. Anger-out individuals evidenced higher adjusted mean DBPs than anger-in individuals. The main effect for group and the Group × Order × Role Play interaction was not significant. No significant group main effects or interactions with group were observed for analyses of SBP and HR reactivity. Recovery measures. ANOVAs were also computed on cardiovascular measures obtained during the 2-min recovery period following each role play. Although there were no significant effects on recovery for SBP and DBP, there was a significant Group × Order interaction for HR recovery, F(1, 36) = 4.39, p < .05. Mean HR during the recovery period among anger-out individuals was significantly lower if exhibited anger was the first role play in comparison to inhibited anger being the first role play, F(1, 18) = 5.35, p < .03. The difference between HR recovery during the two task presentation orders was not significant among anger-in individuals, F(1, 18) = .86. Affective, Behavioral, and Cognitive Responses Multivariate analysis of affective responses to role plays (i.e., state anger, awareness of arousal, number of heartbeats counted, and self-reported anxiety) yielded no significant main effects for group or Group × Order, Group × Role Play, or Group × Order × Role Play interactions (see Table 2). Multivariate analysis of behavioral parameters also revealed no significant main effects for group. Group × Order and Order × Role Play interactions were not significant (see Table 3).
TABLE 2 Affective Responses for Anger-In and Anger-Out Groups During Anger Exhibition and Anger Inhibition Anger-In
FIGURE 1 (a) Systolic blood pressure (SBP; in mm Hg), (b) diastolic blood pressure (DBP; in mm Hg), and (c) heart rate (HR; in bpm) during baseline, interaction, and recovery periods for anger-in and anger-out men during anger exhibition and anger inhibition conditions. DBP response for anger-out participants was significantly greater during anger exhibition than anger inhibition (p < .04), a difference not observed for anger-in participants. Anger-out participants had higher mean DBP responses (p < .05) and slower HR recovery (p < .03) than anger-in participants when inhibition occurred prior to exhibition.
State anger Anger exhibition Anger inhibition Heartbeat change Anger exhibition Anger inhibition Arousal Anger exhibition Anger inhibition Anxiety Anger exhibition Anger inhibition
Anger-Out
M
SD
M
SD
21.58 17.58
8.90 7.39
20.28 16.52
8.56 5.90
+25.6 +19.3
22.91 24.53
+24.6 +15.6
25.79 22.62
6.05 5.20
2.19 1.61
6.15 5.05
2.11 2.04
6.10 5.50
2.51 1.93
6.10 5.30
2.05 2.08
Note. No group differences were observed on self-report measures of affect. Anger expression resulted in significantly higher affective responses across all variables than anger suppression.
Volume 23, Number 4, 2001
Anger Expression
A multivariate analysis of variance conducted on the four subscales of the ACI revealed a main effect for group, F(4, 30) = 3.80, p < .02 (see Table 4). Univariate ANOVAs revealed that anger-in individuals endorsed significantly more items on the Repression subscale for both role plays (M = 4.7) in contrast to anger-out participants (M = 2.7), F(1, 33) = 13.36, p < .01. There also was a significant Group × Role Play interaction at the multivariate level, F(4, 30) = 2.95, p < .04; follow-up univariate ANOVAs revealed a significant interaction for endorsement of items on the Rational Coping subscale, F(1, 34) = 4.21, p < .05. Anger-out participants endorsed significantly TABLE 3 Behavioral Responses for Anger-In and Anger-Out Groups During Anger Exhibition and Anger Inhibition Anger-In
Problem-solving behaviors Anger exhibition Anger inhibition Negative verbal behaviors Anger exhibition Anger inhibition Positive verbal behaviors Anger exhibition Anger inhibition Negative nonverbal behaviors Anger exhibition Anger inhibition Positive nonverbal behaviors Anger exhibition Anger inhibition
Anger-Out
M
SD
M
SD
7.70 5.90
2.74 3.35
8.41 7.65
3.08 4.41
6.95 12.20
3.22 5.84
8.06 13.77
3.54 6.08
0.90 0.55
0.97 0.89
0.59 0.29
1.00 0.77
1.35 1.05
1.76 1.91
1.59 0.88
2.43 1.76
2.60 2.20
2.66 3.35
2.18 1.88
1.59 1.65
Note. No group differences were observed on self-report measures of affect. Anger expression resulted in significantly more negative verbal and fewer problem solving behaviors than anger suppression.
TABLE 4 Anger Cognitions for Anger-In and Anger-Out Groups During Anger Exhibition and Anger Inhibition Anger-In
Repression Anger exhibition* Anger inhibition* Rational Coping Anger exhibition* Anger inhibition Victimization Anger exhibition Anger inhibition Self-Righteousness Anger exhibition Anger inhibition *p < .05.
Anger-Out
M
SD
M
SD
4.1 5.2
2.30 1.73
2.6 2.9
1.81 1.79
7.2 7.6
1.90 1.54
6.0 7.6
2.27 1.92
8.8 7.4
3.52 2.36
8.2 8.4
2.59 2.82
6.1 5.2
2.49 2.01
7.0 6.1
2.35 1.91
287
fewer rational coping cognitions during the exhibited anger role play when compared to their endorsement of cognitions during the anger inhibition role play, F(1, 18) = 9.40, p < .01. No significant differences were found between the anger-in participants’ exhibited anger and inhibited anger rational coping scores. No other main effects for group or interactions with group were significant. DISCUSSION The results of this study support the hypothesis that exaggerated cardiovascular responding is a function of both typical styles of anger expression and the environmental demands of the experimental manipulation. In contrast to previous research (7,8), individuals in this study did not uniformly react more when they were forced to engage in behaviors opposite of what they normally would do in confrontational situations. However, some findings pertaining to anger expression and cardiovascular response to stress uncovered in this study provided partial corroboration of Engebretson’s findings and indicate that our understanding of the “matching hypothesis” is far from complete. Two interesting order effects emerged regarding DBP response and HR recovery to the interpersonal tasks. First, anger-out participants exhibited greater DBP responses to both tasks than anger-in participants but only when inhibited anger instructions were delivered first. This suggests that there is something peculiar to the order of instructions that elicited greater DBP responding among anger-out men. Perhaps the instructions to inhibit anger, which stood in contrast to the anger-out men’s natural preference to express their anger overtly, elicited physiological arousal, and then when provided the opportunity to exhibit anger in the second vignette, their exaggerated response was maintained. These findings appear to be consistent with the “suppressed hostility” hypothesis of psychosomatic disorders initially proposed by Alexander (32), which suggested that anger, when suppressed, can result in an explosive expression over time. Because the exaggerated response profiles were not observed when tasks were presented in the opposite order, it seems likely that the opportunity to exhibit anger following inhibition of anger, not simply inhibition of anger per se, leads to exaggerated cardiovascular responding. No significant differences in reactivity between anger-in and anger-out participants were observed when anger-out participants were permitted to express their anger during the first interaction. At least for anger-out persons, anger inhibition followed by anger exhibition seems to be associated with exaggerated responding. In this regard, this study partially replicates the findings of Engebretson et al. (7). Significant group differences in HR recovery provide additional support for this contention. For anger-out participants, the inhibition first/exhibition second order resulted in delayed HR recovery when compared with anger-out participants who received the alternate order of task presentation. Interestingly, the opposite pattern was observed for anger-in participants, although for anger-in individuals, the pattern was not significant. Those who expressed anger first and suppressed anger second exhibited delayed HR recoveries in contrast to the alternate or-
288
Suchday and Larkin
der. In this regard, HR recovery data also partially supported the “matching hypothesis” in that delayed HR recoveries were only observed in conditions in which participants were requested to engage first in an anger expression strategy opposite of their preferred mode of expression. It is important to note that group differences were observed in participants’ DBP response to provocation rather than SBP or HR. This differential responsivity of DBP and SBP/HR may suggest that the physiological pathway linking anger expression and the development of coronary heart disease may occur via vascular responding rather than myocardial responding. This is not to state, however, that myocardial responses are not implicated at all in the expression of anger. Based on the evidence portraying delayed HR recovery in situations where participants engaged in modes of anger expression different from their preferred mode, it is clear that myocardial effects were observed, but during recovery periods and not during confrontation. In brief, it appears that anger expression may be related to an initial vascular response that may be followed by a compensatory myocardial response during recovery. Although it is unknown which of these vascular or myocardial changes is related to risk for coronary heart disease, it is clear that anger expression involves a complex physiological response consisting of both vascular and myocardial components. Regarding the final hypothesis of the study, measures of state anger and self-reported anxiety and arousal revealed that anger-in and anger-out individuals were not differentially susceptible to the experience of anger, anxiety associated with its expression, or the arousal that stems from its expression in response to both laboratory-induced interpersonal stressors based on their self report. Apparently the laboratory procedure was successful in arousing affective distress consisting of anger, anxiety, and arousal among all individuals regardless of their typical modes of coping with anger or at least in eliciting similar self-reports on these measures among all study participants. The magnitude of affective responses observed in the present study is equivalent to responses reported previously (33). Most important, this finding confirmed that individuals who typically employed anger-in and anger-out strategies for expressing anger experienced anger equivalently; they simply chose to express it differently. This is similar to Everson et al.’s (19) suggestion that both extreme forms of inhibition and exhibition of anger may lead to a resentful festering of anger that may ultimately be harmful. Unusually, anger-out individuals did not differ from anger-in individuals regarding the behavioral exhibition of anger across both experimental situations. Most previous studies that have examined behavioral correlates of anger exhibition have typically relied on participants’ self-reports of anger expressive behaviors rather than actual observation-based indexes of their behavior (e.g., 34). This is surprising because the measurement of anger expression can be confirmed more easily than many of its internal correlates (e.g., hostility, the experience of anger). To examine this, our study employed a measure of actual behaviors elicited during confrontation. Anger-in and anger-out participants exhibited no behavioral differences during either scenario, lending support to the hypothesis that both groups were capable
Annals of Behavioral Medicine of exhibiting and inhibiting anger when instructed to do so. In this regard, the findings of this study are consistent with Porter, Stone, and Schwartz’s (21) findings that the contextual factors of the experiment may have played a significant role in anger exhibition that differed from the role of preferred modes of anger expression. In our study, both anger-in and anger-out participants showed that they possessed comparable behavioral skills in exhibiting and inhibiting anger when directly instructed to do so. It is important to note that our participants were not instructed to “use the behavioral strategy you normally use” in dealing with the confrontations. If these instructions were provided, it is likely differences in behavioral responding may have been observed. Also, it is important to note that participants in a 3-min role play situation may not have had enough time to exhibit the full repertoire of behaviors that accompany anger exhibition or inhibition. Hence, the lack of positive findings with regard to behavioral data may reflect inadequacy of the study procedure in eliciting the full range of behaviors. Regarding cognitive variables, a significant group difference in repression-related self-statements was observed; anger-in individuals endorsed more repression-related cognitions when compared to anger-out individuals across both role plays. This finding is consistent with what has been commonly described as the anger-in individuals’ typical coping style—repressing or distancing oneself from the experience and expression of anger. Anger-in and anger-out individuals differentially endorsed rational coping statements during the two role play situations as well. Anger-out participants utilized fewer rational coping statements during anger exhibition in contrast to anger inhibition. No differences in rational coping statements were observed between anger exhibition and anger inhibition conditions among anger-in participants. Anger-out persons, it seems, may represent a group of individuals who employ fewer rational coping strategies in contrast to anger-in individuals, but only during anger exhibition. This study is limited by its reliance on a laboratory methodology that attempted to measure what might happen in a real-life situation. Consequently, it is quite possible that participants’ reactions observed in this study may not represent an accurate depiction of their reactions in a real-life situation. Therefore, generalizations from these findings to the natural environment must be made with caution. Likewise, because this experiment was conducted on male participants, generalization of findings to women cannot be made. Given the sex role stereotypic expectations surrounding anger and its exhibition, women may present qualitatively and quantitatively different patterns of responding to anger-eliciting situations than men. Inasmuch as risk for coronary heart disease is related to behaviorally elicited cardiovascular responding (e.g., 4), the specific components of anger exhibition require further exploration in reaching a better understanding regarding behavioral risk factors for cardiovascular disease. Anger exhibition may have more harmful health and social consequences than those associated with the experience of anger or rage. Perhaps one of the most interesting contributions of this study lies in the unexpected discovery that not expressing anger
Volume 23, Number 4, 2001 in an appropriate and timely fashion exacerbated negative physiological consequences, particularly for anger-out participants. Inhibition of anger itself was related to extreme patterns of reactivity only if it occurred initially. The experience of anger inhibition resulted in sustained levels of arousal as evidenced by slower HR recovery among anger-out participants. For anger-out males, inhibition of anger was not an effective means of dealing with anger because it led to the maintenance of increased physiological arousal during anger exhibition. Future research is needed to examine the relation between passage of time after anger suppression and intensity of anger exhibition. Over time, problem-solving efforts could alter the meaning attached to a situation, leading to decreased possibility of intense anger exhibition. If this is so, teaching individuals to leave the situation or to “count to 10” might be beneficial. However, if the passage of time leads to a festering of anger and increased intensity of anger expression, as suggested by these findings, then teaching individuals to assertively express their feelings immediately may be more clinically useful. This issue, if empirically explored, may provide clinically useful information pertinent to coping with anger for a number of patient populations, including cardiac and hypertensive patients. This study has shown that examination of modes of anger expression and its cardiovascular, affective, behavioral, and cognitive dimensions is worthy of continued exploration. It raises important questions surrounding the “age-old” issue concerning the importance of learned behavioral patterns versus the more proximal environmental demands as determinants of behavior. The findings generally supported the hypothesis that environmental demands guide the extent of cardiovascular reactions, affective and cognitive repertoires, and behavioral choices and that previously learned patterns of behavior (at least those that were self-reported) exerted only a minor influence on these reactions. Additional empirical work on the impact of anger expression in predicting and causing subsequent health problems will predictably provide us with clues regarding our understanding of psychosocial risk factors for cardiovascular disease. REFERENCES (1) Kawachi I, Sparrow D, Spiro A, Vokonas P, Weiss ST: A prospective study of anger and coronary heart disease. The Normative Aging Study. Circulation. 1996, 94:2090–2095. (2) Miller TQ, Smith TW, Turner CW, Guijarro ML, Hallet AJ: A meta-analytic review of research on hostility and physical health. Psychological Bulletin. 1996, 119:322–348. (3) Moller J: Do episodes of anger trigger myocardial infarction? A case-crossover analysis in the Stockholm Heart Epidemiology Program. Psychosomatic Medicine. 1999, 61:842–849. (4) Krantz DS, Manuck SB: Acute psychophysiological reactivity and risk of cardiovascular disease: A review and methodological critique. Psychological Bulletin.1984, 96:435–464. (5) Smith TW, Houston BK, Stucky RJ: Type A behavior, irritability and cardiovascular reactivity. Motivation & Emotion. 1984, 8:221–230. (6) Smith TW: Hostility and health: Current status of a psychosomatic hypothesis. Health Psychology. 1992, 11:139–150. (7) Engebretson TO, Matthews KA, Scheier MF: Relations between anger expression and cardiovascular reactivity: Recon-
Anger Expression
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
(24) (25)
289
ciling inconsistent findings through a matching hypothesis. Journal of Personality and Social Psychology. 1989, 57:513–521. Lai JY, Linden W: Gender, anger expression style, and opportunity for anger release determine cardiovascular reaction to and recovery from anger provocation. Psychosomatic Medicine. 1992, 54:297–310. Faber SD, Burns JW: Anger management style, degree of expressed anger, and gender influence cardiovascular recovery from interpersonal harassment. Journal of Behavioral Medicine. 1996, 19:31–53. Miller SB, Dolgoy L, Friese M, Sita A: Dimensions of hostility and cardiovascular response to interpersonal stress. Journal of Psychosomatic Research. 1996, 41:81–95. Siegman AW: Cardiovascular consequences of expressing, experiencing, and repressing anger. Journal of Behavioral Medicine. 1993, 16:539–569. Siegman AW, Snow CW: The outward expression of anger, the inward experience of anger and cardiovascular reactivity: The role of vocal expression. Journal of Behavioral Medicine. 1997, 20:29–45. Vogele C, Steptoe A: Emotional coping and tonic blood pressure as determinants of cardiovascular responses to mental stress. Journal of Hypertension. 1992, 10:1079–1089. Vogele C, Jarvis A, Cheeseman K: Anger suppression, reactivity, and hypertension risk: Gender makes a difference. Annals of Behavioral Medicine. 1997, 19:61–69. Gallacher JE, Yarnell JW, Sweetnam PM, Elwood PC, Stansfeld SA: Anger and incident heart disease in the Caerphilly study. Psychosomatic Medicine. 1999, 61:446–453. Angerer P, Siebert U, Kothny W, Muhlbauer D, Mudra H: Impact of social support, cynical hostility, and anger expression on progression of coronary atherosclerosis. Journal of the American College of Cardiology. 2000, 36:1781–1788. Brosschot JF, Thayer JF: Anger inhibition, cardiovascular recovery, and vagal function: A model of the link between hostility and cardiovascular disease. Annals of Behavioral Medicine. 1998, 20:326–332. Suarez EC, Williams RB: The relationships between dimensions of hostility and cardiovascular reactivity as a function of task characteristics. Psychosomatic Medicine. 1990, 52:558–570. Everson SA, Goldberg DE, Kaplan GA, Julkunen J, Salonen JT: Anger expression and incident hypertension. Psychosomatic Medicine. 1998, 60:730–735. Harburg E, Blakelock EH, Roeper RJ: Resentful and reflective coping with arbitrary authority and blood pressure: Detroit. Psychosomatic Medicine. 1979, 41:189–202. Porter LS, Stone AA, Schwartz JE: Anger expression and ambulatory blood pressure: A comparison of state and trait measures. Psychosomatic Medicine. 1999, 61:454–463. Suls J, Wan CK: The relationship between trait hostility and cardiovascular reactivity: A quantitative review and analysis. Psychophysiology. 1993, 30:615–626. Stoney CM: The role of reproductive hormones in cardiovascular and neuroendocrine function during behavioral stress. In Turner J, Sherwood A, Light KC (eds), Individual Differences in Cardiovascular Response to Stress. New York: Plenum, 1992, 147–163. Spielberger CD: State Trait Anger Expression Inventory. Odessa, FL: Psychological Assessment Resources, 1988. Semenchuk EM, Larkin KT: Behavioral and cardiovascular responses to interpersonal challenges among male offspring of essential hypertensives. Health Psychology. 1993, 12:416–419.
290
Suchday and Larkin
(26) Larkin KT, Frazer NL, Semenchuk EM: Physiological, affective, and behavioral responses to interpersonal conflict among males from families with different levels of cohesion and adaptability. Journal of Psychopathology and Behavioral Assessment. 1997, 18:239–254. (27) Larkin KT, Semenchuk EM, Frazer NL, Suchday S, Taylor RL: Cardiovascular and behavioral responses to social confrontation: Measuring real-life stress in the laboratory. Annals of Behavioral Medicine. 1998, 20:294–301. (28) Sturges LV, Goetsch VL: Physiological reactivity and heart beat awareness in anxiety sensitivity. Journal of Anxiety Disorders. 1996, 10:283–294. (29) Hops H, Mills TA, Patterson GR, Weiss RL: Marital Interaction Coding System. Eugene, OR: Oregon Research Institute, 1972.
Annals of Behavioral Medicine (30) Weider GB, Weiss RL: Generalizability theory and the coding of marital interactions. Journal of Consulting and Clinical Psychology. 1980, 48:469–477. (31) Suchday S, Desiderato O: Factor analysis of an anger cognitions inventory. Unpublished Master’s Thesis, Connecticut College, 1991. (32) Alexander F: Emotional factors in essential hypertension: Presentation of a tentative hypothesis. Psychosomatic Medicine. 1939, 1:173–179. (33) Pope MK, Smith TW, Rhodewalt F: Cognitive, behavioral, and affective correlates of the Cook Medley Hostility Scale. Journal of Personality Assessment. 1990, 54:501–514. (34) Patterson GR: A microsocial analysis of anger and irritable behavior. In Chesney M, Rosenman RH (eds), Anger and Hostility in Cardiovascular and Behavioral Disorders. New York: Hemisphere McGraw-Hill, 1985, 83–100.