Journal of Psychology and Christianity Copyright 2006 Christian Association for P.sychological Studies 2006, Vol. 25, No. 3, 195-204 ISSN 0733-4273 The ACE Model of Emotions: What Separates Righteous and Sinful Emotions? Rodney L. Bassett Catherine Luzadder Patricia Manley Rosemary Miano Christen Santore Julie Putnam Roberts Wesleyan College The purpose of this study was to test the assumption made by the ACE model that what separates func- tional from dysfunctional (righteous from sinful) emotional experiences are the cognitive and expressive ele- ments, not physiological activation. In this study, 36 students were asked to recall a past experience with functional or dysfunctional anger. Heart rate was monitored before and after the recollection. Following the recollection, students evaluated the experience along dimensions designed to tap the cognitive and expres- sive aspects of the emotional experience. Consistent with the ACE model of emotion, the results revealed meaningful differences between functional and dysftjnctional experiences within the cognitive and expres- sive domains but not within the physiological domain. For some Christians, particular emotions may same way of thinking about emotional experience seem more "righteous" than others. After all, the to the emotion of anger. Ellis makes a distinction biblical listings of the fruit of the Spirit include between healthy and unhealthy anger in his clini- emotions like love and joy. Those listings do not cal practice. He contends that many therapists try include emotions like anger and fear. But, the to suppress client's anger and hope that just talk- ACE model of emotion (Bassett & Hill, 1998) ing about the emotions will provide sufficient suggests such a view of emotions may be too therapeutic relief. However, in his clinical experi- limited. Instead, the ACE model proposes three ence, Ellis has found it more beneficial to channel basic ideas: (a) all emotions are morally angry feelings in a positive fashion. He uses ratio- ambidextrous (any emotion has the potential to nal-emotive methods to help hostile clients to be be righteous or sinful); (b) the essential elements "constructively assertive instead of defeatingly of all emotions are physiological activation [A], angry" (p. 251). So, clients can be helped to use cognition [C], and expression [E]; and (c) what the emotion of anger in a positive fashion. separates righteous from sinful (functional from Of course, if you can not automatically classify dysfunctional) emotional experiences is the cog- an emotion as always good or always bad, then nitive and expressive components of emotion, there need to be criteria that help one discern not physiological activation. when an emotion is in fact good or bad. Ellis Others seem to agree that it is unwise to con- (1976) argued that positive emotions uphold basic sistently categorize an emotion as functional or human values and negative emotions obstruct dysfunctional (good or bad). For example, those values. The basic human values that he Solomon and Stone (2002) argue that this tradi- identified were: "survival, happiness, social accep- tional polarization of emotions oversimplifies the tance, and intimate relations" (p. 240). Ellis complexity of the emotional experience. Accord- believed that a significant contributor to these ing to Solomon and Stone, all emotions have the basic human values is cognitive. He contended capacity to be positive or negative. Fear for that once people evaluate their emotions they are example, which is usually thought of as nega- subject to either rational or irrational beliefs that tive, "... might even be perceived as a good influence their actions. Rational beliefs are consis- emotion if it propels us to remove ourselves tent with human values while irrational beliefs from danger." (p. 419) undermine those values. Hence positive (or con- Going from a broad overview of emotions to a structive) emotions can be assessed by how peo- narrower perspective, Ellis (1976) applies the ple cognitively respond to their emotions and how their resulting actions impact the basic Portions of this article were presented at the annual human value system. meeting of the Christian Association for Psychological Studies, Dallas, TX. Correspondence should be Another way of thinking about good and bad addressed to Rod Bassett at RWC, 2301 Westside emotions has flowed out of Tangney's distinction Drive, Rochester, NY 14624. [email protected] between shame and guilt (Tangney, 1990; 1995). 195 196 THE ACE MODEL OF EMOTIONS: WHAT SEPARATES RIGHTEOUS AND SINEUL EMOTIONS? Shame and guilt are emotions that people experi- constructive and destructive emotional experi- ence when they believe they have acted in ways ences are the cognitive and expressive elements. that are socially or morally unacceptable. Both Of the three emotional elements, physiological emotions are negative self-relevant emotions. activation (A) refers to the bodily changes that However, shame is conceptualized as more global are under the control of the limbic system and in nature with the entire person feeling inade- the autonomic nervous system. With some emo- quate, under scrutiny, worthless, and powerless. tions these changes may have an alerting or In contrast, with guilt the emotional experience is arousing quality. With other emotions the conceptualized as more focused on the specific changes may be more calming or restorative. As behavior. The inadequacy of the behavior is not a result, all emotions have an "embodied" quali- generalized to the entire individual. The behavior ty. We literally "feel" emotions. Cognition (C) may be reprehensible, but the person is not. simply refers to all relevant conscious and sub- Tangney and her colleagues have suggested that conscious thoughts. Expression (E) refers to all tendencies toward experiencing shame- and guilt body movements and speech controlled by the can be dispositional (see Tangney, 1990). voluntary nervous system. Such movements may Obviously, shame is a "bad" emotion while be subtle (clenching of the teeth) or dramatic guilt is a "good" emotion. However, even more (striking an opponent). These movements serve interesting was the prediction of Tangney, Wagn- a communicative function, to self and others, for er, Hill-Barlow, Marschall, and Gramzow (1996) the different emotions. that anger would play a more problematic role Applying a Christian perspective to these ideas for shame-prone individuals than for guilt-prone about emotion, Bassett and Hill propose that two individuals. In a cross-sectional developmental dimensions are especially important for under- study of children, adolescents, college students, standing emotions: (a) reflective/nonreflective and non-college adults, two sets of age-appropri- processing, and (b) constaictive/destructive pro- ate measures were taken. The first set of mea- cessing. The reflective/nonreflective dimension is sures assessed dispositional tendencies toward what makes the ACE model a dual-process shame and guilt. The second set of measures model. The nonreflective end of this dimension assessed the extent to which anger played a con- involves subconscious processes. These process- structive or destructive role in participants' lives. es can be subconscious for several reasons. First, Across all age groups, shame proneness was the emotional experience may not have enough generally associated with destructive anger expe- significance for the individual to justify more riences (malevolent intentions, direct and indi- careful consideration. Second, the emotional rect aggression, and negative long-term experience may be automatic. Finally, the emo- consequences from the anger experience). In tional experience may be pushed into the uncon- contrast, guilt proneness tended to predict con- scious by defensive mechanisms. The reflective structive anger experiences (constructive inten- end of this dimension involves conscious pro- tions, corrective action and nonhostile communication, and more positive long-term cessing that will be more effortful and elabora- consequences from the anger experience). tive. Reflective processing is particularly important for the ACE model because the model Finally, a model that considers the issue of proposes that this is the locus of hutnan choice. good and bad emotions from a psychological Human choice introduces the element of human and Christian perspective is the ACE model responsibility. Thus, reflective emotional process- (Bassett & Hill, 1998). The ACE model assumes that the necessary and sufficient elements of an ing opens the door for the distinction between emotion are physiological activation, cognition, sinful and righteous emotional experiences. and expression. If one or more of these elements The second dimension that is particularly impor- is missing, then a persori will experience some- tant in the ACE model is constructive/destaictive thing that is different from an emotion. Also, processing. The concepaial basis for this dimen- these three elements are assumed to be interac- sion is the person and character of God. The tive; therefore, what happens with one element model assumes that everything that is consistent can impact the other elements. Further, the with the character of God is ultimately constmctive model argues that all emotions have the capacity in human experience. Everything that is inconsis- to be constructive or destructive (righteous or tent with God's character is ultimately destmctive sinful). According to the model what separates in human experience. BASSETT, LUZADDER, MANLEY, MIANO, SANTORE, AND PUTNAM 197 Orthogonally combining these two dimensions that during this process physiological rneasures produces four emotional quadrants. Emotional would be taken. Finally, all participants were experiences that are reflective and constructive, guaranteed that their recalled anger experiences the model identifies as "righteous." Emotional could be kept confidential. The group of partici- experiences that are reflective and destructive, pants included 9 freshmen, 8 sophomores, 6 the model identifies as "sin." Emotional experi- juniors, 3 seniors, and 10 graduate students. The ences that are nonreflective and constructive, the group was evenly split between males and model identifies as "natural." And, emotional females and the average age was 21.2 years. The experiences that are nonreflective and destruc- undergraduate students received a small atnount tive, the model identifies as "distorted." The of class extra credit for their participation. The model further proposes that of the three emo- graduate students were requested to "help out" tional elements (physiological activation, cogni- other students in the same graduate research tion, and expression), it is where the cognitive methods course who were collecting data for one and expressive elements fall along the reflec- of several research projects. tive/nonreflective and constructive/destructive Procedures and Materials dimensions that determine the moral location of Participants arrived individually at a campus an emotional experience. Physiological activation lab. They were greeted by one of the researchers is assumed to be morally neutral in emotional and given a five-page questionnaire. Participants experience. were asked to read the first page of the question- Thus, the ACE model proposes that there is no naire which gave a general review of the proce- definitive list of emotions that are spiritual and dures and served as a consent form. Included in emotions that are not. There is no definitive list of the review was a reminder that physiological emotions that are always functional and emotions measures would be taken during the study. If that are always not. Instead, Christians need to be participants agreed to continue to participate, discerning. Any emotion can please or displease then they signed the consent form and it was God. Or, to put this idea another way, no emo- detached from the questionnaire. At that point tion will always please God. the equipment for measuring heart-rate was The purpose of this study was to test the attached. The researcher then engaged the partic- assumption made by the ACE model that what ipant in reading and rereading a grocery list (the separates functional from dysfunctional (righ- same list for all participants); while a one-minute teous from sinful) emotional experiences are the baseline measure for heart-rate was established. cognitive and expressive elements, not physio- logical activation. In this study, students were Following the baseline for the physiological asked to recall a past experience with functional measure, the participant then proceeded to the or dysfunctional anger. Before and after the rec- second page of the questionnaire. On this page, ollection heart rate was monitored. Following participants provided some basic demographic the recollection, students evaluated the experi- information and responded to a modified version ence along dimensions designed to tap the cog- of the Positive and Negative Affect Schedule nitive and expressive aspects of the emotional (PANAS; Watson, Clark, & Tellegen, 1988). This experience. Consistent with the ACE model of scale is designed to assess emotional state. Partic- emotion, it was predicted there would be mean- ipants were asked to estimate the extent to which ingful differences between functional and dys- emotional words (e.g., interested, scared, proud, functional experiences within the cognitive and jittery, etc.) characterized how they felt at that expressive domains but not within the physio- moment. Individuals responded to each word on logical activation domain. a 3-point Likert-like scale (1 = very slightly or not at all, 5 = extremely). Given that we intended to Method then ask participants to recall a past anger expe- rience, we modified the PANAS by adding the Participants emotional words "happy" and "angry." Thirty-six individuals were recruited to partici- pate in the study from graduate and undergradu- Participants then moved to the third page of the ate classes (participation rate = 88%). All questionnaire. Each participant was asked to write participants were informed in advance that they about a past anger experience. Half of the partici- would be asked to recall and evaluate a past pants were randomly assigned to recall a func- anger experience. They were further informed tional anger experience and the other half were 198 THE ACE MODEL OF EMOTIONS: WHAT SEPARATES RIGHTEOUS AND SINFUL EMOTIONS? randomly assigned to recall a dysfunctional anger asked participants to estimate the emotional experience. The instructions clarifying these two intensity of the recalled anger experience. The different types of anger are presented below. remaining 12 items were designed to tap dimen- sions that the researchers thought might separate Try to think of an event where you functional from dysfunctional anger. All of the believe that your anger was a reason- items from this page of the questionnaire are able and correct response where, in presented in Table 1. retrospect, it might have been wrong not to get angry. Perhaps it was a sit- The final page of the questionnaire contained uation where your anger motivated the modified PANAS scale. This allowed us to you to do the right thing. For exam- have a post measure of affective state. Individuals ple: you and your best friend got into again responded to each word describing feeling a fight because your friend lied to states on a 5-point Likert-like scale (1 = very you. Because of your anger, you slightly or not at all, 5 = extremely). Once partici- were able to confront your friend in pants completed this last page, they were a constructive fashion, (functional) detached from the physiological monitoring equipment and thanked for their participation. Try to think of an event where you believe that your anger was not the Resiilts reasonable and correct response One preliminary, but important, issue involved where, in retrospect, it might have the extent to which participants were able to been wrong to get angry. Perhaps it recall anger experiences that fit the requested was a situation where your anger functional and dysfunctional categories of anger. motivated you to do the wrong thing. The mean "fit" score for the functional anger For example: you and your best group was 5.1 iSD = .73). The mean score for the friend got into a fight because your dysfunctional anger group was 5.0 iSD = .69). friend lied to you. Because of your Since the highest possible score was 6.0, both anger, you said some things to your groups seemed to report that their recalled expe- friend that were unnecessarily hurtful riences fit well the type of experience requested. and you later regretted saying those In fact the lowest "fit" rating given by any partici- things, {dysfunctional) pant was a "4" in both groups (above the mid- Participants were asked to recall an event by point of the scale). A one-way ANOVA indicated writing about the event. It was thought that writ- that there was not a difference between the aver- ing about the event would allow participants to age "fit" ratings for the functional and dysfunc- more graphically re-experience the event. In addi- tional anger groups [F(l, 34) = .05, ns]. tion, writing about the event protected the confi- An important analysis for this study involved dentiality of the participant. After providing a the comparison of dysfunctional and functional written account of the event, participants anger regarding level of arousal. The ACE detached that page from the questionnaire and model predicted no differences in arousal took it with them after the study was completed. between the two types of anger. The physiolog- Participants had been informed in advance that ical measures of arousal were the baseline aver- they would be free to keep their written recollec- ages for heart-rate and the treatment averages tions. As a result, the researchers were never for heart-rate (taken while participants were exposed to the recollection of the event. While recalling their anger experiences). Both of these participants were writing their recollections, a sec- measures were based on an average of heart- ond one-minute measure of heart-rate was taken. rate measures taken over a one minute period. Following the written recollection, participants These data were analyzed in two ways. First, a then moved to the next page of the question- comparison was done between the two anger naire. On this page, participants were requested groups on the baseline measure and then on to evaluate the recalled anger experience using the treatment measure. The comparison 14 Likert-like items. The first of these items was between the two groups on the baseline mea- essentially a manipulation check with partici- sure was nonsignificant [functional mean = 86.3 pants indicating the extent to which the recalled vs. dysfunctional mean = 79.6; F(l, 34) = 1.63, anger experience was a "good fit" for the type of ns, eta = .21]. The comparison between the two anger experience requested. The second item groups on the treatment measure was also BASSETT, LUZADDER, MANLEY, MIANO, SANTORE, AND PUTNAM 199 Table 1 Mean Responses to the Items Used to Evaluate Functional and Dysfunctional Anger Experiences The anger experience I recalled was a 5.1 5.0 .82 .04 good fit for the type of experience requested. As I recall the situation, the anger I A.6 5.2 .17 .23 experience at the time was very intense. I believe I relied upon God to help me 3.2 2.9 .53 .01 navigate this experience. Other people were negatively affected 3.3 4.6 .01 .43 • by this experience. My anger was triggered by a concern 2.7 3.2 .35 .16 about the well-being of others. I find that I still think about that event, 3.8 4.4 .24 .20 and those thoughts give me a sense of dissatisfaction. I would say the short-term outcome of 3.3 2.3 .03 .36 this experience was positive. I would say the long-term outcome of 4.1 3.4 .15 .08 this experience was positive I believe I have grown spiritually as a 3.7 3.3 .47 .12 result of this experience. I believe my reaction was appropriate, 4.6 4.0 .19 .22 given the circumstances. I'm afraid I said and did things during 2.7 3.5 .04 .34 that experience that I now regret. During the midst of the experience, 3.4 4.6 .02 .38 my feelings were out of control. In my anger, I said appropriate things 4.3 3.5 .06 .32 to the other person. While I was angry, I was still concerned 4.1 3.8 .68 .07 about the well-being of the person with whom I was angry. 'Shaded rows indicate the difference between the two types of anger met traditional levels of significance (p < .05). nonsignificant [functional mean = 93.7 vs. dys- Second, an analysis of covariance was conduct- functional mean = 85.0; 7<'(1,34) = 2.37, ns, eta - ed using multiple regression with the baseline .26]. Even if more liberal levels of alpha were measure of heart rate as the covariate. The criteri- adopted (e.g., p = .10), neither analysis would on variable in this analysis was the treatment have been statistically significant. If effect sizes measure of heart rate. A forward stepwise strate- are considered for these analyses (eta), then gy was then adopted for introducing the predic- according to the criteria presented in Rosenthal tor variables (baseline heart rate and the two and Rosnow (1984), the effect sizes for both anger groups). The first predictor variable analyses would be identified as "small." entered into the equation was baseline heart rate 200 THE ACE MODEL OF EMOTIONS: WHAT SEPARATES RIGHTEOUS AND SINFUL EMOTIONS? {beta = .88, adjusted R Square = .78, standard We also looked at the correlations within the error of the estimate = 8.07). The other predictor different evaluations for functional and dysfunc- variable, type of anger, did not meet the criterion tional anger. The significant correlations for each (set at alpha = .10) for being entered into the type of anger are reported in Table 2. It is worth equation. In fact, the type of anger only account- noting that the degrees of freedom for these ed for 2% (partial correlation = < -.15) of the vari- analyses were relatively small (n = 18 for each ability in treatment heart rates. Thus, both type of anger). Therefore, only rather large cor- analyses revealed no significant differences in relations (r > .45) reached traditional levels of heart-rate between the functional and dysfunc- significance. Of these significant relationships, tional anger groups. three especially stand out because the same pat- tern of relationship was found for functional and Participants were also asked to provide a sub- dysfunctional anger. With both types of anger: jective measure of arousal when they were asked (a) having an increased belief they had relied to rate the intensity of the anger experience after upon God to help them navigate the anger expe- they recalled the experience (see Table 1). On a rience predicted a greater sense of having grown 6-point scale, the functional anger group gave an spiritually from the experience, (b) the tendency average intensity rating of 4.6 while the average of participants to believe their reactions were intensity rating for the dysfunctional anger group appropriate predicted lower tendencies on the was 5.2. A one-way ANOVA revealed that the dif- part of participants to believe they did things ference between these average intensity ratings during the anger experience that they now was nonsignificant IF(1, 34) = 1.94, ns]. Again, the regret, and (c) as ratings of loss of self-control effect size was small (eta = .17). Thus, both objec- increased participants tended to indicate tive and subjective measures of arousal seemed to decreased levels of being concerned about the indicate that there were no differences between well-being of the other person during the anger the two types of anger. experience. However, the ACE model also suggests that Finally, we analyzed the emotional state ratings there will be differences between functional and given before and after the recall of the anger dysfunctional anger in the cognitive and expres- experience on the PANAS. Our version of the sive realms. So, in what ways were the two scale requested participants to respond to 22 types of anger different? Table 1 presents a sum- words capturing feelings/emotions were charac- mary of the evaluations of functional and dys- teristic of them at that point in time. To simplify functional anger provided by the participants. the analyses, five words were identified that Comparing the average responses for functional seemed to particularly speak to the experience of and dysfunctional anger using one-way anger (distressed, upset, hostile, irritable, and ANOVAs, four of the evaluations were signifi- angry). Cronbach's alpbas seemed to indicate cantly different. With dysfunctional anger, partic- these words had good internal reliability (pre- ipants were more likely to report that other measure, alpha = .85; postmeasure, alpha = .89). people were negatively affected by the experi- So, "pre" and "post" composite measures were ence [F{1, 34) = 7.56, p < .01]. With functional created by averaging the scores for these words anger, participants were more likely to indicate for each participant. A repeated measures one- that the short-term outcome of the experience way ANOVA indicated a significant difference was positive IF(.1, 34) = 5.04, p < .05]. With dys- between the composite measures, F(\, 34) = functional anger, participants were more likely 25.66, p < .001 (premeasure mean = 1.63, post- to share that they said and did things during the measure mean = 2.42). Thus, after the recall of experience that they now regret IFil, 34) = 4.38, an anger experience, participants reported p < .05]. Also with dysfunctional anger, partici- greater negative affect than before recalling that pants were more like to indicate their feelings experience. However, the average amount of were out of control during the experience [Fil, negative affect reported after recalling the experi- 34) = 5.65, p < .05]. Finally, there was a ence fell towards the middle of the 5 point marginally significant difference between the response scale. two types of anger. Participants recalling func- Discussion tional anger had more of a tendency to indicate they said appropriate things to the other person Participants were asked to recall a past anger [Fa, 54} = 5.76, .05>p< .10]. experience that was functional or dysfunctional. BASSETT, LUZADDER, MANLEY, MIANO, SANTORE, AND PUTNAM 201 Table 2 Significant Correlations between the Participant Evaluations of their Functional or Dysfi^nctional Anger Experiences msimimm mmLu mm>n asdig iisgfl,^ Eval. 1: As I recall the situation, the anger I experienced at the time was very intense. Eval. 2: I believe I relied upon God to help me navigate this experience. Eval. 3: Other people were negatively affected by this experience. Eval. 4: My anger was triggered by a concern about the well-being of others. Eval. 5: I fmd that I still think about that event, and those .57* -.63" thoughts give me a sense of dissatisfaction. Eval. 6: I would say the short-term outcome of this experience was positive. Eval. 7: I would say the long-term outcome of this -.60** .54* experience was positive. Eval. 8: I believe I have grown spiritually as a result .57* of this experience. .81** Eval. 9: I believe my reaction was appropriate, -.57* given the circumstances. Eval. 10: I'm afraid I said and did things during .47* that experience that I now regret. Eval. 11: During the midst of the experience, my feelings were out of control. Eval 12: In my anger, I said appropriate things to the other person. Eval. 13: While I was angry, I was still concerned about the well-being of the person with whom I was angry. Correlations in regular type = functional anger, correlations in italics and bold = dysfunctional anger. 'p<.05 "p<.01. Table 2 continues next page Before and after (or during) recalling the expe- It is true; the design of our study put us in an rience, heart-rate and emotional state were interesting position. Logically, it is difficult to assessed. In addition, participants were asked support the null hypothesis (in this project, the to evaluate the anger experience on several prediction of no differences in arousal for the dimensions. An important finding of this project two types of anger): The reason for this difficul- was that the levels of physiological arousal for ty is that several things can account for non- the functional and dysfunctional anger experi- significant differences between groups. So, let's ences were not significantly different. Equally consider some of those possibilities. One possi- important for the predictive validity of the bility is that the manipulations simply did not model was the existence of significant differ- work. That seems unlikely since the negatively ences between the two types of anger in the valenced words on the PANAS showed cognitive and expressive domains. Thus, the increased values after the manipulation were all combination of nonsignificant and significant negative in connotation. Another possibility is differences discovered in this study was consis- that the manipulations worked, but not as tent with the ACE Model. hoped. That also seems unlikely since the self- 202 THE ACE MODEL OE EMOTIONS: WHAT SEPARATES RIGHTEOUS AND SINEUL EMOTIONS? Table 2 (continued) Significant Correlations between the Participant Evaluations of their Functional or Dysfunctional Anger Experiences . '. ; • t it Eval. 1: As I recall the situation, the anger I experienced at the time was very intense. Eval. 2: I believe I relied upon God to help me navigate this experience. Eval. 3: Other people were negatively affected by this experience. Eval. 4: My anger was triggered by a concern about the well-being of others. Eval. 5: I find that I still think about that event, and those thoughts give me a sense of dissatisfaction. Eval. 6: I would say the short-term outcome of this experience was positive. Eval. 7: I would say the long-term outcome of this .48 experience was positive. Eval. 8: I believe I have grown spiritually as a result .12" of this experience. Eval. 9: I believe my reaction was appropriate, .66** given the circumstances. Eval. 10: I'm afraid I said and did things during that experience that 1 now regret. Eval. 11: During the midst of the experience, my feelings were out of control. Eval 12: In my anger, I said appropriate things to the other person. Eval. 13: While I was angry, I was still concerned about the well-being of the person with whom I was angry. Correlations in regular type = functional anger, correlations in italics and bold = dysfunctional anger. •p<05 "p<.m. Tabie 2 continues next page reported indications of how well the recalled comparisons between the premeasures for both anger experiences fit the requested anger experi- types of anger and the postmeasures for both ences were quite high for both types of anger. types of anger were not significantly different. Finally, another possibility is that the lack of sig- However, comparisons between the pre and nificant differences reflected insufficient statisti- postmeasures for functional anger were signifi- cal power for the analyses. Again, that seems cantly different [preM = 86.3 vs. postM = 93.7, unlikeily. With the self-ratings evaluating the Fil, 17) = 8.85, p < .01] as were the comparisons anger experience, four out of thirteen of the self- for the pre and postmeasures for dysfunctional ratings showed statistically significant differences anger [preM = 79.6 vs. postM = 85.0, F(.l, 17) = between the two types of anger (and there was 32.61, p < .001]. It is important to note that these one marginally significant difference). Concerns statistically significant comparisons were based about the possibility of low power among the on the same sample sizes as those of the non- evaluations would have been more convincing if significant analyses. Apparently, the heart-rate fewer of the evaluations had shown significant analyses also had sufficient statistical power. differences. With the heart-rate analyses the Thus, perhaps the most reasonable explanation BASSETT, LUZADDER, MANLEY, MIANO, SANTORE, AND PUTNAM 203 Table 2 (continued) Significant Correlations between tbe Participant Evaluations of their Functional or Dysfimctional Anger Experiences Eval. 1: As I recall the situation, the anger I experienced at the time was very intense. Eval. 2: I believe I relied upon God to help me navigate this experience. Eval. 3: Other people were negatively affected by this experience. Eval. 4: My anger was triggered by a concern about the well-being of others. Eval. 5: I find that I still think about that event, and those thoughts give me a sense of dissatisfaction. Eval. 6: I would say the short-term outcome of this experience was positive. Eval. 7: I would say the long-term outcome of this experience was positive. Eval. 8: I believe I have grown spiritually as a result of this experience. Eval. 9: I believe my reaction was appropriate, given the circumstances. Eval. 10: I'm afraid I said and did things during -.71" that experience that I now regret. Eval. 11: During the midst of the experience, my feelings were out of control. Eval 12: In my anger, I said appropriate things to the .74" -.56* other person. .54 Eval. 13: While I was angry, I was still concerned about -.55* the well-being of the person with whom I was angry. .49* Correlations in regular type = functional anger, correlations in italics and bold = dysfunctional anger. •p<.05 "p<.01. for the nonsignificant differences in arousal significant difference between the two types of between functional and dysfunctional anger is anger for long-term outcomes. One way to that there simply are no differences in arousal explain this lack of difference for the long-term between functional and dysfunctional anger. outcomes would be the perception that over the Of course, one of the wonderful things about long-term people who experienced dysfunctional research is that typically the researcher ends up anger perceived God as using that experience to with more questions to consider. For this study, help them to mature. Yet, the data do not seem to one of those questions is the issue of short-term fit that explanation. Perceiving that the long-term and long-term outcomes for the different types of outcome was positively connected with believing anger. Participants reported significantly more that God had used the experience to mature the positive short-term outcomes for functional than individual occurred only with the functional ver- dysfunctional anger. However, there was not a sion of anger. With dysfunctional anger, there was 204 THE ACE MODEL OF EMOTIONS: WHAT SEPARATES RIGHTEOUS AND SINFUL EMOTIONS? not such connection. So, what diminished the gap References between functional and dysfunctional anger expe- Bassett, R. L., & Hill, P. C. (1998). The ACE Model of riences regarding long-term outcomes? emotion: Living Jesus Christ while experiencing emo- Another interesting question is what might be tions. Journai ofPsychoiogy and Theology, 26, lil-lAd. the next step in this line of research? Several pos- Ellis, A. (1976). Healthy and unhealthy aggression. sibilities come to mind. For example, this study Humanitas, 12, 239-254. seems to confirm the ACE Model assumption that Rosenthal, R., & Rosnow, R. L. (1984). Essentials of levels of arousal does not differentiate functional behavioral research: Methods and data analysis. New from dysfunctional anger. However, does this York: McGraw-Hill Book Company. assumption also apply to other emotions? Are Solomon, R., & Stone, L. D. (2002). On "positive" and "negative" emotions. Journal for the Theory of levels of arousal similar for functional and dys- Social Behavior, 32, All-AiS. functional love, fear, etc.? Another direction might Tangney, J. P. (1990). Assessing individual differ- be to map out those factors that do differentiate ences in proneness to shame and guilt: Development functional and dysfunctional emotions. The ACE of the Self-Conscious Affect and Attribution Inventory. Model proposes that it is the "cognitive" and Journal of Personality and Social Psychology, 59, 102- "expressive" aspects of emotion that will separate 111. functional and dysfunctional versions of an emo- Tangney, J. P. (1995). Shame and guilt in interper- tion. Charting these differences across various sonal relationships. In J. P. Tangney & K. W. Eischer emotions might allow us to discern which differ- {"Eds^, .Self-conscious emotions: Shame, guilt, embar- entiating factors are common across most emo- rassment, and pride ipp. 114-139). New York: Guilford tions and which differentiating factors are unique Press. to a particular emotion. So, resolving quickly an Tangney, J. P., Wagner, P. E., Hill-Barlow, D., emotional state might be characteristic of func- Marschall, D. E., & Gramzow, R. (1996). Relation of tional "negative" emotions (like anger), but not of shame and guilt to constructive versus destructive functional "positive" emotions (like love). responses to anger across the lifespan. Journal of Per- sonality and Social Psychology, 70, 797-809. But it is probably good to end with a word of Watson, D., & Clark, L. A., & Tellegen, A. (1988). caution. The methodology of this study involved Development and validation of brief measures of posi- the recollection of anger experiences. Frankly, tive and negative affect: The PANAS Scales. Journal of part of the rational behind that methodology was Personality and Social Psychology, 54, 1063-1070. the researchers were hesitant to inflict anger on participants (particularly dysfunctional anger). Therefore, the researchers settled for rekindling Authors previous anger experiences from memory. There Rod Bassett is on the faculty of Roberts Wesleyan Col- may be important ways in which the original lege (RWC). The other authors were, at the time of this experience differed from the recalled experience. study, students at RWC.