Gender, Anger Expression Style, and Opportunity for Anger Release Determine Cardiovascular Reaction to and Recovery from Anger Provocation JOSANNA Y. LAI AND WOLFGANG LINDEN This study represents an extension of Hokanson's research, which showed that for men anger release after provocation tends to accelerate cardiovascular recovery. The objective of this study was to investigate how gender and habitual style for anger-in or anger-out behavior modulate the effect of anger provocation and release. Male and female subjects (N = 105) were classified as anger-in/anger-out only when a double criterion (i.e., self-report and peer evaluation) was satisfied. Following a state anger rating at pre-test. subjects were harassed during the performance of a 12-minute math task. After task completion, subjects were randomly assigned to one of two 10-minute recovery protocols a) having an opportunity to release negative affect, and b) not having such an opportunity. All groups (including the anger-ins) that had an opportunity to express negative affect did in fact express similar levels of anger. Men reacted more strongly to the math task performed under anger provocation on all cardiovascular indices. Anger expression style as a trait-type disposition was important for the recovery process in women whereas the situational manipulation (i.e., the opportunity to release anger) had specific effects on the recovery process of men. Opportunity to release anger facilitated heart rate recovery (and to a lesser degree diastolic pressure recovery) in men but not in women. Women with anger-in tendencies on the other hand displayed better systolic pressure recovery than female anger-outs whereas no such effects were observed in men.
INTRODUCTION Anger and hostility have long been associated with coronary heart disease and essential hypertension (1-4). According to the definitions proposed by Spielberger et al. (5), anger, as opposed to hostility, can be more easily instigated and studied among subjects in a standard laboratory situation. Of the many different dimen-
From the Department of Psychology, The University of British Columbia, Vancouver, Canada. Address reprint requests to: Wolfgang Linden, Ph.D., Department of Psychology, The University of British Columbia, Vancouver, B.C., Canada V6T1Z4. Received for publication February 25, 1991; revision received November 22, 1991
Psychosomatic Medicine 54:297-310 (1992) 0033-31 74/92/5403-0297S03 00/0 Copyright (D 1992 by the American Psychosc
sions of anger, it appears that expression style is particularly important for the study of anger and its link to heart disease and essential hypertension. Evidence is available that anger suppressors had higher all-cause mortality, including heart disease, than non-anger suppressors (6), that consideration of emotional expressiveness increased the predictive validity of Type A Behavior (7-9), and that elevated blood pressure is associated with anger-suppression (10-13). However, it has also been shown that both anger-suppressors and anger-expressors tended to have significantly higher blood pressure than those who. used a reflective response (11), thus supporting a suggested U-shape relationship between blood pressure and 297
). Y. LAI AND W. LINDEN
amount of anger/hostility expressed (14, 15). An intriguing protocol for the study of anger expression behavior and cardiovascular consequences is that pioneered by Hokanson. This protocol involved serial subtraction, where repeated interruptions and harassment were used to provoke anger. Subsequent anger retention versus anger release effects were then studied during recovery periods (16-21). Reliable and rapid recovery from task-induced blood pressure increases were obtained following counter-aggression toward the harasser by the experimental subjects (17-19, 21). Research also showed that situational variables modulated the outcome. With a high status frustrator (i.e., a visiting professor as opposed to an undergraduate assistant), frustrated undergraduates who were not given a chance to counter-aggress still showed rapid recovery (21). Overt counter-aggression via application of electric shocks and/or verbal disapproval was more powerful than covert modes (i.e., fantasy and unharmful signaling) in accelerating recovery from anger arousal (16). Consistent gender differences emerged in the Hokanson paradigm. Counteraggression did not lead to accelerated recovery in women whereas it did in men (22). In another study, Hokanson and Edelman (23) found more rapid systolic recovery with counter-aggression in men than was found in women. A revealing contribution to an understanding of these sex differences was made by Hokanson et al. (24). These researchers found that women, who initially obtained faster recovery subsequent to provocation through friendly counter-responses, could be conditioned through avoidance learning to achieve the same physiologi298
cal effect following aggressive counterresponses. Men could with equal success be conditioned through avoidance learning to have faster recovery when initial aggressive counter-responses were replaced by friendly counter-responses. Next, the conditioning patterns for both genders were successfully unlearned. This study neatly demonstrated that social-behavioral learning can modify the cardiovascular recovery in counteraggression protocols. Only recently have individual differences, i.e., the role of anger expression style, become the focus of research in Hokanson-type protocols (25). Engebretson and his coworkers (25) attempted to further clarify the linkage between anger expression/suppression and cardiovascular consequences through a matching hypothesis. It was predicted that when anger-suppressors (anger-ins) and anger-expressors (anger-outs) were harassed, those who were allowed to express their anger consistent with their anger direction style (i.e., matched subjects) would display more rapid cardiovascular recovery than mismatched subjects. These researchers used male subjects and found partial support for the matching hypothesis in that anger-ins writing positive evaluations and anger-outs writing negative evaluations of their harassers had more rapid systolic blood pressure recovery than the mismatched groups. Although the Engebretson et al. study (25) appears quite similar to the current one because both test the interaction of a (presumably trait-type) behavioral style and situational opportunity, there are many differences in these two designs that will defy direct comparison. The current study differed in that we attempted to integrate additional findings from the literature on anger expression. Psychosomatic Medicine 54:297-310 (1992)
ANGER RELEASE
We were concerned that self-reported an- opportunity to express their anger after ger expression style was not an accurate provocation would do so or not; c) to deindex of a person's acute anger behavior termine which subgroup and which situbecause the self-report of emotional states ational factors would be associated with is confounded by response sets (26). We the most rapid cardiovascular recovery; also wanted to test whether experimental d) to replicate Hokanson's observation instruction could lead habitual anger-in that women do not benefit (i.e., show responders to act against their natural rapid blood pressure recovery) from retendency and express their anger in the leasing angry feelings. Concerns about the laboratory because research suggests that validity of self-reported anger expression physiological responses following coun- style were addressed by using both a selfter-aggression can be unlearned or situa- report and peer-validation procedure. tionally modified (24). Because there are gender-specific effects of counter-aggression on cardiovascular recovery (23), it METHOD was also intended to jointly examine the roles of gender, anger expression style, Overview and an opportunity for anger expression Female and male undergraduates whose anger in a standard laboratory situation. An ad- expression styles were previously assessed and ditional concern for this study was the cross-validated by both self-report and peer ratings operationalization of an anger-retention on the Spielberger Anger Expression Scale (27) were condition. We believed that the most di- asked to work on two tasks. The first task involved arithmetic with interruptions during which rect operationalization of an anger-reten- mental unreasonable criticisms of the subject's performance tion condition would be to enforce anger were delivered on a fixed schedule. For the second retention by not permitting a possibly task, half the subjects were randomly assigned to compensatory behavior, such as writing a evaluate the experimenter and to express their emopositive evaluation of the frustrator (cf. tion in writing, and the other half were randomly assigned to copy a written neutral paragraph. Thus, 25). the study design represents a two (Female-Male) x Therefore, the present study was de- two (Anger ins-Anger outs) x two (Opportunity-No signed to examine the interactions of gen- opportunity for emotional ventilation) factorial deder and anger expression style (anger-in sign. vs. anger-out) in an anger-inducing situation followed by opportunity/no opportunity for emotional ventilation on the Measurement of Anger Expression part of the frustrated subject. Those anStyle ger-ins and anger-outs who were assigned The 24-item Anger Expression Scale (27) was used to the opportunity for anger release were to assess subjects' usual way of handling anger. The encouraged to express their feelings about items tap two relatively independent dimensions: the harrasser and the protocol was to fa- anger-in vs. anger-out (27). Anger-in refers to how often angry feelings are experienced but not excilitate anger expression. More specifically, the objectives were: pressed. Typical anger-in items include: "I control temper"; "I withdraw from people"; and "I tend a) to assess cardiovascular responses to my to harbor grudges that I don't tell anyone about." harassment as a function of the subject's Anger-out, on the other hand, refers to aggressive gender and anger expression style; b) to physical or verbal behavior when angry. Typical investigate whether anger-ins given an items include: "I express my anger"; "I say nasty Psychosomatic Medicine 54:297-310 (1992)
299
). Y. LAI AND W. LINDEN things"; and "I strike out at whatever infuriates me." The internal consistency for males and females ranged from 0.77 to 0.80, respectively (27).
Subjects Inclusion criteria were: a) no established hypertension, that is
INTRODUCTION Anger and hostility have long been associated with coronary heart disease and essential hypertension (1-4). According to the definitions proposed by Spielberger et al. (5), anger, as opposed to hostility, can be more easily instigated and studied among subjects in a standard laboratory situation. Of the many different dimen-
From the Department of Psychology, The University of British Columbia, Vancouver, Canada. Address reprint requests to: Wolfgang Linden, Ph.D., Department of Psychology, The University of British Columbia, Vancouver, B.C., Canada V6T1Z4. Received for publication February 25, 1991; revision received November 22, 1991
Psychosomatic Medicine 54:297-310 (1992) 0033-31 74/92/5403-0297S03 00/0 Copyright (D 1992 by the American Psychosc
sions of anger, it appears that expression style is particularly important for the study of anger and its link to heart disease and essential hypertension. Evidence is available that anger suppressors had higher all-cause mortality, including heart disease, than non-anger suppressors (6), that consideration of emotional expressiveness increased the predictive validity of Type A Behavior (7-9), and that elevated blood pressure is associated with anger-suppression (10-13). However, it has also been shown that both anger-suppressors and anger-expressors tended to have significantly higher blood pressure than those who. used a reflective response (11), thus supporting a suggested U-shape relationship between blood pressure and 297
). Y. LAI AND W. LINDEN
amount of anger/hostility expressed (14, 15). An intriguing protocol for the study of anger expression behavior and cardiovascular consequences is that pioneered by Hokanson. This protocol involved serial subtraction, where repeated interruptions and harassment were used to provoke anger. Subsequent anger retention versus anger release effects were then studied during recovery periods (16-21). Reliable and rapid recovery from task-induced blood pressure increases were obtained following counter-aggression toward the harasser by the experimental subjects (17-19, 21). Research also showed that situational variables modulated the outcome. With a high status frustrator (i.e., a visiting professor as opposed to an undergraduate assistant), frustrated undergraduates who were not given a chance to counter-aggress still showed rapid recovery (21). Overt counter-aggression via application of electric shocks and/or verbal disapproval was more powerful than covert modes (i.e., fantasy and unharmful signaling) in accelerating recovery from anger arousal (16). Consistent gender differences emerged in the Hokanson paradigm. Counteraggression did not lead to accelerated recovery in women whereas it did in men (22). In another study, Hokanson and Edelman (23) found more rapid systolic recovery with counter-aggression in men than was found in women. A revealing contribution to an understanding of these sex differences was made by Hokanson et al. (24). These researchers found that women, who initially obtained faster recovery subsequent to provocation through friendly counter-responses, could be conditioned through avoidance learning to achieve the same physiologi298
cal effect following aggressive counterresponses. Men could with equal success be conditioned through avoidance learning to have faster recovery when initial aggressive counter-responses were replaced by friendly counter-responses. Next, the conditioning patterns for both genders were successfully unlearned. This study neatly demonstrated that social-behavioral learning can modify the cardiovascular recovery in counteraggression protocols. Only recently have individual differences, i.e., the role of anger expression style, become the focus of research in Hokanson-type protocols (25). Engebretson and his coworkers (25) attempted to further clarify the linkage between anger expression/suppression and cardiovascular consequences through a matching hypothesis. It was predicted that when anger-suppressors (anger-ins) and anger-expressors (anger-outs) were harassed, those who were allowed to express their anger consistent with their anger direction style (i.e., matched subjects) would display more rapid cardiovascular recovery than mismatched subjects. These researchers used male subjects and found partial support for the matching hypothesis in that anger-ins writing positive evaluations and anger-outs writing negative evaluations of their harassers had more rapid systolic blood pressure recovery than the mismatched groups. Although the Engebretson et al. study (25) appears quite similar to the current one because both test the interaction of a (presumably trait-type) behavioral style and situational opportunity, there are many differences in these two designs that will defy direct comparison. The current study differed in that we attempted to integrate additional findings from the literature on anger expression. Psychosomatic Medicine 54:297-310 (1992)
ANGER RELEASE
We were concerned that self-reported an- opportunity to express their anger after ger expression style was not an accurate provocation would do so or not; c) to deindex of a person's acute anger behavior termine which subgroup and which situbecause the self-report of emotional states ational factors would be associated with is confounded by response sets (26). We the most rapid cardiovascular recovery; also wanted to test whether experimental d) to replicate Hokanson's observation instruction could lead habitual anger-in that women do not benefit (i.e., show responders to act against their natural rapid blood pressure recovery) from retendency and express their anger in the leasing angry feelings. Concerns about the laboratory because research suggests that validity of self-reported anger expression physiological responses following coun- style were addressed by using both a selfter-aggression can be unlearned or situa- report and peer-validation procedure. tionally modified (24). Because there are gender-specific effects of counter-aggression on cardiovascular recovery (23), it METHOD was also intended to jointly examine the roles of gender, anger expression style, Overview and an opportunity for anger expression Female and male undergraduates whose anger in a standard laboratory situation. An ad- expression styles were previously assessed and ditional concern for this study was the cross-validated by both self-report and peer ratings operationalization of an anger-retention on the Spielberger Anger Expression Scale (27) were condition. We believed that the most di- asked to work on two tasks. The first task involved arithmetic with interruptions during which rect operationalization of an anger-reten- mental unreasonable criticisms of the subject's performance tion condition would be to enforce anger were delivered on a fixed schedule. For the second retention by not permitting a possibly task, half the subjects were randomly assigned to compensatory behavior, such as writing a evaluate the experimenter and to express their emopositive evaluation of the frustrator (cf. tion in writing, and the other half were randomly assigned to copy a written neutral paragraph. Thus, 25). the study design represents a two (Female-Male) x Therefore, the present study was de- two (Anger ins-Anger outs) x two (Opportunity-No signed to examine the interactions of gen- opportunity for emotional ventilation) factorial deder and anger expression style (anger-in sign. vs. anger-out) in an anger-inducing situation followed by opportunity/no opportunity for emotional ventilation on the Measurement of Anger Expression part of the frustrated subject. Those anStyle ger-ins and anger-outs who were assigned The 24-item Anger Expression Scale (27) was used to the opportunity for anger release were to assess subjects' usual way of handling anger. The encouraged to express their feelings about items tap two relatively independent dimensions: the harrasser and the protocol was to fa- anger-in vs. anger-out (27). Anger-in refers to how often angry feelings are experienced but not excilitate anger expression. More specifically, the objectives were: pressed. Typical anger-in items include: "I control temper"; "I withdraw from people"; and "I tend a) to assess cardiovascular responses to my to harbor grudges that I don't tell anyone about." harassment as a function of the subject's Anger-out, on the other hand, refers to aggressive gender and anger expression style; b) to physical or verbal behavior when angry. Typical investigate whether anger-ins given an items include: "I express my anger"; "I say nasty Psychosomatic Medicine 54:297-310 (1992)
299
). Y. LAI AND W. LINDEN things"; and "I strike out at whatever infuriates me." The internal consistency for males and females ranged from 0.77 to 0.80, respectively (27).
Subjects Inclusion criteria were: a) no established hypertension, that is
