Anger Co-rumination among same-sex adolescent peers

Co-rumination refers to an extensive discussion of problems, speculation about problems, and focus on negative emotions engaged in by two or more people. Research on emotion coping suggests that co-rumination, or dwelling on problems, leads to problematic emotional adjustment. Girls have been found to more frequently to co-ruminate about sadness with their peers and internalize symptoms more than boys. In my project, I looked at aspects of anger conversations that might reflect co-rumination and its aftermath. I hypothesized that as operational definitions of co-rumination, particular listener behaviors would be associated with speaker rumination behaviors more so than others. I also hypothesized that girls would show more co-ruminative behaviors than boys and that more co-rumination would predict higher post-co-rumination anger. Conversations between peers were videotaped, transcribed, and coded to measure co-rumination between adolescent participants (N = 108, mean age = 12.9) and their same-sex peers while discussing anger experiences. Strong associations were found between listener corumination eliciting behaviors and speaker rumination behaviors, but no gender difference was found in co-rumination, and co-rumination behaviors did not predict post-conversation anger. Findings are discussed in terms of the literature on gender differences in corumination and the implications of co-rumination for adolescent emotional distress.

ANGER CO-RUMINATION AMONG SAME-SEX ADOLESCENT PEERS by Kaul Lee A Senior Honors Thesis Submitted to the Faculty of The University of Utah In Partial Fulfillment of the Requirements for the Honors Degree in Bachelor of Science In The Department of Psychology Approved: ______________________________ Monisha Pasupathi Thesis Faculty Supervisor _____________________________ Bert Uchino Chair, Department of Psychology _______________________________ Jeanine Stefanucci Honors Faculty Advisor _____________________________ Sylvia D. Torti, PhD Dean, Honors College April 2019 Copyright © 2019 All Rights Reserved ii ABSTRACT Co-rumination refers to an extensive discussion of problems, speculation about problems, and focus on negative emotions engaged in by two or more people. Research on emotion coping suggests that co-rumination, or dwelling on problems, leads to problematic emotional adjustment. Girls have been found to more frequently to co-ruminate about sadness with their peers and internalize symptoms more than boys. In my project, I looked at aspects of anger conversations that might reflect co-rumination and its aftermath. I hypothesized that as operational definitions of co-rumination, particular listener behaviors would be associated with speaker rumination behaviors more so than others. I also hypothesized that girls would show more co-ruminative behaviors than boys and that more co-rumination would predict higher post-co-rumination anger. Conversations between peers were videotaped, transcribed, and coded to measure co-rumination between adolescent participants (N = 108, mean age = 12.9) and their same-sex peers while discussing anger experiences. Strong associations were found between listener corumination eliciting behaviors and speaker rumination behaviors, but no gender difference was found in co-rumination, and co-rumination behaviors did not predict post-conversation anger. Findings are discussed in terms of the literature on gender differences in corumination and the implications of co-rumination for adolescent emotional distress. iii TABLE OF CONTENTS ABSTRACT ii INTRODUCTION 1 METHOD 7 RESULTS 13 DISCUSSION 29 APPENDIX 39 REFERENCES 41 1 INTRODUCTION Adolescence is associated with the onset of exacerbated psychological problems, including depression (Birmaher et al., 1996) and antisocial behavior (Alink & Egeland, 2013). In adolescence, girls are more likely than boys to show symptoms of depression and eating disorders, and engage in suicidal ideation and suicide attempts, whereas boys are more likely to experience problems with anger and engage in high risk behaviors (World Health Organization, 2002). This suggests that girls are more prone to experiencing symptoms that are directed inwardly, such as anxiety and depression, while boys are more likely to act out, leading to aggression. Considering these findings, it makes sense that women would have a higher prevalence of most affective disorders compared to men (Riecher-Rossler, 2010). Identifying potential sources or predictors of affective dysfunction in women could potentially inform therapies or interventions to reduce the prevalence of these disorders. Ruminative coping is one of the variables associated with the emergence of gender differences in depression (Nolen-Hoeksema & Girgus, 1994). Past research supports that rumination, defined as internalizing and focusing excessively and continually on one’s distress or negative emotions, is associated with longer and more severe periods of negative mood. Rumination is maladaptive because it achieves no resolution (Nolen-Hoeksema, 1998). Early work investigated self-reported rumination as a more cognitive and internal phenomenon, and did not distinguish between different contexts in which people might engage in rumination. However, rumination occurs not only in individual context but can also take place within friendships and social 2 interactions, as by adolescence, individuals tend to discuss their problems and their negative emotions extensively with peers (Rose, 2002). This act of co-rumination within peers brings both positive and negative outcomes, such as increased relationship quality and internalization of symptoms of anxiety and depression (Rose, 2002). Although relationship quality between peers is important for youth’s well-being, co-rumination, may result in more emotional risks than benefits. Prior research has found that sadness rumination in general (Morrow & NolenHoeksema, 1990) and sadness co-rumination in particular (Stone et al., 2011) are highly associated with depression, which suggests that rumination exacerbates worry and depressed mood. Much of the literature on sadness rumination and sadness co-rumination found that girls reported higher levels of co-rumination/rumination than boys (Rose, 2002; Stone et al., 2011; Calmes & Roberts, 2008; Moreira et al., 2015; Lopez et al., 2009). Rose (2002) also found girls were more likely to co-ruminate about sadness and anxiety as they grew older, but this was not the case for boys. These findings are relevant to the idea that rumination accounts for why girls tend to experience depressive episodes more often than boys, as girls are more likely to ruminate and co-ruminate with their peers about negative events. While much research has focused on sadness and rumination, recent work suggests that rumination may operate differently for other emotions, such as anger. Unlike sadness rumination/co-rumination, people may be less likely to detach from their angry ruminations because anger is more likely than sadness to involve self-justification and blaming others (Tice & Baumeister, 1993). Because individuals feel more justified in their anger, this may lead individuals to perceive that they have a right to be angry more 3 than that they have a right to feel sad, which may intensify anger (Rusting & NolenHoeksema, 1998), and lead to relational and overt aggression (Peled & Moretti, 2007; Denson, 2012). Studies on children reveal no significant difference between boys and girls on self-reported anger, however, there were significant differences in the expression of anger, with boys reporting higher levels of aggression compared to girls (Buntaine & Costenbader, 1997). Concerning anger ruminations, there have been discrepant results on whether more women or more men ruminate. One study on Chinese and British college students showed that men were more likely than women to ruminate about anger (Maxwell et al., 2005). Within a clinical sample of adolescents, Peled and Moretti (2007) found that girls reported higher levels of angry rumination compared to boys. Still other studies report no significant differences in angry rumination between males and females (Sukhodolsky et al., 2001; Maxwell, 2004; Stimmel et al., 2006; Ottaviani et al., 2010; White & Turner, 2014). Taken together, research points to discrepant results in gender differences for anger rumination, with some studies showing notable gender differences and others showing no obvious differences. The different participant demographics of these studies may help explain the conflicting results. Most of the research described above studied angry ruminations within adult women and men (Maxwell et al., 2005; Sukhodolsky et al., 2001; Maxwell, 2004; Stimmel et al., 2006; Ottaviani et al., 2010; White & Turner, 2014), whereas Peled and Moretti (2007) studied angry rumination within adolescents. As adults tend to become more aware of the different coping mechanisms to deal with their problems, it is possible that they have better knowledge of the variety of options to choose from, whereas adolescents may perceive ruminating about anger as the most convenient coping 4 mechanism. It is also important to note that the studies on adults were on university students (Sukhodolsky et al., 2001; Rusting & Nolen-Hoeksema, 1998; Stimmel et al., 2006; White & Turner, 2014; Maxwell et al., 2005; Siewert et al., 2011), competitive college athletes (Maxwell, 2004), and adult individuals from a local community (Ottaviani et al., 2010). Peled and Moretti (2007) studied adolescents who had been referred to an assessment program for youth with severe conduct problems. The finding that adolescent girls reported higher levels of anger rumination compared to boys could be due to the use of a sample of adolescents who already have problems with anger, aggression, and depression. Differences in culture and personal characteristics of participants would very likely bring about different results. Understanding differences in anger rumination for different age groups could have important implications for identifying affective dysfunction early in life, which may allow for early treatment and better long-term outcomes. Another critical piece of information concerning these studies is that different scales, questionnaires, and in-lab experiments were used in different research studying anger rumination. The Anger Rumination Scale (ARS), which measures the tendency to focus on angry moods, recall past anger experiences, and think about the causes and the consequences of angry events (Sukhodolsky et al., 2001) was used in several studies (Sukhodolsky et al., 2001; Maxwell, 2004; Maxwell et al., 2005; White & Turner, 2014). Peled and Moretti (2007) used the ARS along with other rumination scales and created a new item to use for their study. The procedure was much more complex for other studies as they utilized electronic daily diaries to ask individuals how much they ruminated about angry events (Siewert et al., 2011), an anger recall task (Ottaviani et al., 2010), 5 anger/mood inductions (Stimmel et al., 2006), and choice or no-choice response tasks (Rusting and Nolen-Hoeksema, 1998). The procedure used by Rusting and NolenHoeksema (1998) created different conditions (anger or neutral) and participants were asked to focus on their mood, distract themselves, or choose to do either. It is very likely that differences could arise from the use of different methods, as answering questionnaires and actively participating in the experiment could produce variation in the way mood is expressed. Moreover, none of these studies have examined co-rumination about anger—that is, ruminating in the context of social interactions with others. Given the prior work on co-rumination in sadness relating to affective dysfunction, it is interesting to consider the effects of social influences on anger rumination. It is important to note that the majority of researchers looked at sadness corumination and thus utilized the Co-Rumination Questionnaire, which was developed to assess the extent to which participants typically co-ruminate with close same-sex friends (Rose, 2002). Interestingly, researchers studying anger in the context of rumination have only examined solitary rumination and utilized the Anger Rumination Scale, which did not involve discussions about anger with another individual. Studies on sadness rumination/co-rumination have yielded results that strongly support the idea that there exists a significant gender difference in how much individuals ruminate and co-ruminate about sadness. In anger, however, most rumination studies report no obvious gender differences in how much men and women ruminate about anger. This summary of prior literature on rumination poses three interesting questions. First, if adolescents are likely to co-ruminate with their peers about sadness, would they also be likely to co-ruminate about anger when they are given the chance? Second, if adolescents do co-ruminate about 6 anger with one another, do gender differences exist in terms of how much individuals coruminate? And third, does co-ruminating about anger lead to sustained or increased anger? To address these questions, the current study utilized a sample of adolescents in a two-partner conversation paradigm. One partner discussed an event that recently made them angry and the other partner listened and responded. I looked at both speech from the target child and the responses from the listener, identifying aspects of anger conversations that might reflect co-rumination and its aftermath (post-conversation emotion ratings). The sample consisted of equal numbers of girls and boys, allowing me to examine whether girls were more likely to co-ruminate with their peers than boys. I hypothesized that, as indicators of co-rumination, particular friend responses such as support/agree/validation, information/co-narration, related experience, and emotion analysis and exploration would relate to speaker rumination (narrator’s emotion analysis, emotion exploration, and anger expression) more so than other friend responses (minimizing speaker’s account, backchanneling, asking questions, and giving advice). Rose and colleagues (2014) reported that the global co-rumination score was positively associated with the response types: support/agree, information, and related experience. These three types of response were also positively related to dwelling on negative affect and mutual encouragement of problem talk. The current study also included the listener response type of asking for or providing emotion exploration and analysis, as exploring one’s emotions and speculating about the causes and the consequences of the event would very likely elicit more problem talk. 7 The fact that adolescent girls are more likely to turn to rumination as a way of coping with emotion, I hypothesized that co-rumination about anger between girls would be more extensive compared to boys. Specifically, I anticipated that more content in the discussions between girls compared to boys would involve speech and response that provoked further co-rumination. I also anticipated that the more engaged adolescents are in co-ruminating about anger with their peers, the more anger and distress they would feel afterwards. Studies support that ruminating about one’s anger is positively associated with anger (Rusting & Nolen-Hoeksema, 1998; Sukhodolsky et al., 2001) and may account for the maintenance and intensification of anger (Sukhodolsky et al., 2001). METHOD Participants Data were collected from 108 participants (51% female, 94% Caucasian) and their same-sex peers, however data were missing for one of the dyads, thus 107 dyads were studied. The participants were between the ages 12 and 14 (M=12.9, SD=.64), and approximately 78% came from middle-class families. Both participants and peers gave written informed assent (combined with parental consent) and were individually compensated $50 for their participation in the study. Measures Speaker behaviors. A total of nine (A-I) speech categories were developed for this study to code for each of the target child’s turns in the conversation: (A) irrelevant, distraction, off-topic (statements not focused on the main problem/event, e.g., “I’m hungry”), (B) event detail (any statement that reveals factual information about the 8 problem/event), (C) anger expression (any statement or action in which the narrator expresses his/her anger, e.g., “I was so mad,” “I just want to (pretends slap and makes hitting noise)”), (D) positive emotion (emotion expressions of contentment, e.g., “That made me happy, however,” “I loved that he got in trouble”), (E) emotion exploration and/or analysis (information that explains why the narrator felt a certain way, or considers what the consequences of feeling that way might be, e.g., “I found out that she was talking behind my back,” “I feel like my friends get mad at me back for being angry,” also, non-anger related emotion expressions such as sadness, fear, anxiety, and distress were included in this category), (F) agreement (statements in which there is agreement with the friend, e.g., “I think that too”), (G) disagreement (statements in which there is disagreement with the friend, e.g., “That’s not what happened”), (H) invite listener participation (statements that ask for or hint at wanting further discussion with friend, e.g., “Do you have anything to say?” “What do you think?”), and (I) other (turns that do not fit in any of the prior categories, including unfinished statements, statements that have no substantive meaning, e.g., “well,” “um,” and time stamps). Speaker’s anger expression and emotion exploration and/or analysis represented rumination. Listener behaviors. A total of twelve (J-U) response categories were developed (several categories have been based on Leaper, Carson, Baker, Holiday, & Myers, 1995) to code for each of the friend’s turns in the conversation: (J) irrelevant, distraction, offtopic (e.g., “Are we done talking?” among other things mentioned in label A), (K) minimization/non-support (friend minimizes the narrator’s problem, e.g., “It wasn’t that serious”), (L) backchannel (response that conveys the friend is paying attention and acknowledges what the target child says, e.g., “uh-huh,” “yeah”), (M) question (questions 9 about the event, e.g., “When did that happen?” “Who were you with?”), (N) information and co-narration (friend adds new factual information about the event or person, or conarrates the story with the target child), (O) encourage discussion (statements or questions that ask for further discussion, e.g., “Do you want to talk more about it?” “you were saying?” “continue”), (P) support, agreement, validation (statements that indicate that the listener understands and supports the narrator’s view of the event/problem, e.g., “I think you did the right thing,” “I would be mad too if that happened to me”), (Q) same experience (friend talks about a similar situation without taking attention away from the target child’s main focus of the problem, e.g., “I had that happen to me too and it was annoying”), (R) advice (friend gives advice on how to handle a certain problem or offers a way to fix the problem, e.g., “Try not to take it too personally,” “You should tell her what really happened”), (S) ask for emotion exploration (e.g., “How did you feel when she said that?”), (T) ask for or provide emotion analysis (friend asks for why the narrator felt a certain way or speculates the causes and consequences about the event, e.g., “Why were you angry?” “She’s always rude to us when she’s with her friends”), and (U) other (turns that do not fit in any of the categories above, including unfinished statements, statements that have no substantive meaning, and time stamps). Adding information/conarrating, giving support, talking about same experience, asking for emotion exploration, and providing emotion analysis were classified as behaviors that elicit co-rumination. Coding scheme. To calculate the proportions of occurrences for all behaviors, all turns with a ‘1’ were added separately for each behavior and divided by the number of total turns for each conversation (e.g., if the listener/friend had talked about an off-topic 10 content in a total of 5 times out of 25 turns, then proportion for friend’s offtopic/irrelevant content = 5/25 or 20%). Reliability. Coding reliability was assessed by using intraclass correlation. Eighteen percent of the transcribed conversations was coded for reliability between myself (primary coder) and 3 other coders. Reliability was generally found to be above .70 in speaker behaviors, with highest reliability of .88 in speaker positive emotion, and generally above .40 in listener behaviors, with highest reliability of .97 for listener same experience (refer to tables 8 and 9 in Appendix). Co-rumination behaviors. Speaker rumination behaviors of anger expression and emotion exploration and/or analysis, and listener behaviors of support/agree/validation, information/co-narration, same experience, asking for emotion exploration, and asking for or providing emotion analysis were used as indicators of co-rumination in conversations. Pre-/post-conversation emotion. Pre- and post-conversation emotion was measured and aggregated by averaging the mad and anger ratings reported by the participant. Unhappy and worked-up ratings reported by the participant were measured independently. Procedure In a previous larger research project, participants were asked to talk with their same-sex peers about a time that made them angry. This took place inside an observation room with two chairs and a one-way mirror in a university laboratory. Before the task began, the target child/speaker (participant) was asked to think of a time when he/she was angry at another person that he/she could discuss with the friend/listener. The participants 11 rated the intensity of their anger emotions when recalling the time using a scale from 1 to 7 (1 = Not at all mad/angry/unhappy/worked up, 7 = The most mad/angry/unhappy/worked up I’ve ever felt), what they learned from the experience, and their perceived importance of the experience. Then, the target child and his/her friend were instructed to go inside the room to discuss the time that made the target child angry. After they had finished talking, the target child was asked to rate how mad/angry/unhappy/worked up he/she felt when talking to their friend about the time on a scale from 1 to 7 (1 = Not at all mad/angry/unhappy/worked up, 7 = The most mad/angry/unhappy/worked up I’ve ever felt) and how he/she thought their friend responded to him/her during the conversation. Conversations between adolescents have been videotaped and transcribed verbatim onto a word document by former research assistants in the laboratory. The word documents containing the transcribed conversations were used in the current study to code for anger co-rumination in each of the dyads. We developed a coding scheme using categories of different target and friend behaviors to obtain the proportions of the occurrences in the different types of behavior in each anger conversation. To establish inter-rater reliability among myself and three undergraduate research assistants in the laboratory, meetings were held to discuss the coding manual, practice coding the first few conversations, and to address any questions and disagreements with the coding scheme. Coding Coder training. Prior to beginning coding, each of the three research assistants helping with this study coded the first five conversations as practice. Then, meetings 12 were held with each coder to go over how they coded the conversations and to make clarifications about the coding manual if their codings did not align with the coding keys that were created following the manual. Another meeting was held with all three research assistants and me to code conversation 6 (dyad 6). This gave each coder, including myself, an opportunity to converse with each other about coding and come to an agreement about the rules. In a lab meeting with all the research assistants and professors overseeing this project, we shared and discussed our coding on conversation 6. Comments and feedback received from this meeting resulted in additional speech and response categories to the coding manual. The first six conversations were used in the coder training process. Coding manual. The three research assistants helping with this study were told to follow specific guidelines under the coding manual which included: disregarding any turns that came before the speaker’s introduction of his/her anger event, counting and coding each turn (target’s speech and friend’s response) beginning with the speakers’ introduction of the problem and ending with listeners’ last response before the “wave” (which was used as a signal to the researchers to indicate that they had finished talking), giving a score of 1 under each behavior that is present within each turn, and giving a score of 0 for any behavior that is absent within each turn. For example, if the speaker (participant) had talked about the event, expressed her anger, and analyzed her emotion all in one turn, we gave a score of 1 for each of those behaviors (B, C, and D) and gave a 0 for the rest (A, E, F, G, H, and I). The same rule applied when coding for listener behaviors. 13 In order to obtain comparable proportions of the occurrences in the different behaviors between speakers and listeners, an equal number of total turns in the conversation was recorded for both speaker and listener in each dyad. In the transcribed conversations, oftentimes pauses were noted and were followed by the same speaker who had been speaking before the pause. This resulted in uneven number of turns. To avoid having an unequal number of total turns between the speaker and the listener, we counted the same speaker’s/listener’s consecutive turns as one turn. RESULTS Preliminary Analyses In the next two sub-sections, I looked at intercorrelations of speaker and listener behaviors to gain some sense of the validity of the coding scheme. No corrections were made for multiple comparisons. Relations within speaker behaviors A test of Pearson’s correlation was computed to assess the relationship between each speaker response variables (see Table 1). This was done to gain an understanding of what speakers were likely to do in relation to their other behaviors during conversations. A negative association was found between speaker off topic and event detail (r=-.31, p<.01), and speaker off topic and agreement (r=-.21, p<.05). The more off-topic content the speaker introduced in the conversation, the lower the presence of information regarding event detail and agreement with the listener’s perspective. A negative association was also found between inviting listener participation and agreement with 14 listener (r=-.24, p<.05). Interestingly, when the speaker agreed less with the listener’s perspective, he/she was more likely to invite listener’s participation in the conversation. A positive association was found between off topic and inviting listener participation (r=.22, p<.05). When the speaker talked about off-topic content, he/she was more likely to invite the listener to participate. There was a positive association between inviting listener participation and event detail (r=.26, p<.01), inviting listener participation and anger expression (r=.50, p<.001), inviting listener participation and positive emotion (r=.39, p<.001), and inviting listener participation and emotion exploration or analysis (r=.29, p<.01). When the speaker invited the listener to participate in the discussion, the speaker was more likely to talk about event details, express anger, express positive emotion, and explore or analyze his/her emotions. There was a positive association between anger expression and event detail (r=.24, p<.05), anger expression and positive emotion (r=.29, p<.01), and anger expression and emotion exploration or analysis (r=.53, p<.001). Unsurprisingly, the more anger the speaker expressed during the conversation, the higher the presence of information about the event and more emotion exploration and analysis. Interestingly, more anger expression by the speaker was also related to more positive emotion expression. Relationships mentioned above were statistically significant at either .05, .01, or .001 level. No other relationships were statistically significant (p>.05). Disagreeing with the listener had no relationships with any of the other behaviors. 15 16 Relations within listener behaviors A test of Pearson’s correlation was computed to assess the relationship between each listener response variables (see Table 2). This was done to gain an understanding of what listeners were likely to do in relation to their other behaviors during conversations. A negative association was found between listener off topic and backchanneling (r=-.38, p<.001), listener off topic and asking questions (r=-.28, p<.01), off topic and giving information or co-narrating (r=-.22, p<.05), and off topic and support/agree/validation (r=-.21, p<.05). When there was more off-topic content by the listener, there was less response from the listener that conveyed the speaker that the listener was listening (backchanneling). More off-topic content by the listener was also related to asking less questions about the event, giving less information or less co-narrating in the discussion, and supporting or validating the speaker less. There was a negative association between minimizing speaker’s anger and backchanneling (r=-.22, p<.05). When the listener minimized the speaker’s story or anger, the listener was less likely to respond in ways that indicated that he/she was listening to the speaker. There was also a negative association between asking questions and support/agree/validation (r=-.30, p<.01). The more questions asked by the listener, the less support and validation by the listener was shown in the conversation. There was a positive association between backchanneling and giving information or co-narrating (r=.26, p<.01), backchanneling and encouraging discussion (r=.25, p<.05), backchanneling and support/agree/validation (r=.47, p<.001), backchanneling and talking about same experience (r=.26, p<.01), and backchanneling and asking for or providing emotion analysis (r=.33, p<.01). When the listener actively conveyed to the 17 speaker that he/she was listening to the speaker, the listener gave more information about the event, encouraged further discussion, supported and validated the speaker’s accounts, talked about same experience, and asked for or provided speaker’s emotion analysis. A positive association was found between support/agree/validation and giving information or co-narrating (r=.29, p<.01), support/agree/validation and talking about same experience (r=.52, p<.001), and support/agree/validation and asking for or providing emotion analysis (r=.56, p<.001). More support from the listener was related to the listener giving more information about the event, sharing same experience, and asking for or providing emotion analysis. A positive association was found between asking for or providing emotion analysis and talking about same experience (r=.40, p<.001), and asking for or providing emotion analysis and giving information or co-narrating (r=.22, p<.05). Asking for emotion analysis and emotion analysis by the listener was related to more sharing of same experience and giving more information about the event. There was also a positive association between giving advice and minimizing the speaker’s anger (r=.30, p<.01). When the listener minimized the speaker’s anger, the listener was more likely to give advice to the speaker. Relationships mentioned above were statistically significant at either .05, .01, or .001 level. No other relationships were statistically significant (p>.05). There was no significant relationship between asking for speaker’s emotion (emotion exploration) and any of the other behaviors. 18 19 Relations between listener co-rumination and speaker rumination behaviors Pearson’s correlation was used to examine relations between speaker variables and listener variables (see Table 3). A negative association was found between speaker off-topic content and listener backchanneling (r=-.34, p<.001), and speaker off-topic content and listener asking questions (r=-.26, p<.01). As the speaker introduced more offtopic content, the listener was less likely to backchannel and ask questions about the event. There was a negative association between speaker event detail and listener offtopic content (r=-.36, p<.001). More event detail by the speaker was related to less offtopic content by the listener. There was a negative association between listener asking questions and speaker anger expression (r=-.24, p<.05) and listener asking questions and speaker inviting listener’s participation (r=-.25, p<.05). More questions asked by the listener was related to less speaker anger expression and inviting less participation from the listener. A negative association was found between speaker agreement and listener encouraging discussion (r=-.19, p<.05). When the speaker had agreed with the listener, the listener was less likely to encourage further discussion of the event. There was a strong positive association between speaker off topic and listener off topic (r=.85, p<.001). Off-topic or irrelevant content in the discussions by either the speaker or the listener was related to more off-topic content by the other party. There was a positive association between speaker event detail and listener backchanneling (r=.22, p<.05), and speaker event detail and listener asking questions (r=.27, p<.01). When the speaker gave more information about the event, there was more backchanneling and asking questions by the listener. A positive association was found between speaker anger expression and listener backchanneling (r=.31, p<.01), listener giving information (r=.32, 20 p<.01), listener encouraging discussion (r=.20, p<.05), listener support (r=.48, p<.001), listener same experience (r=.25, p<.05), and listener emotion analysis (r=.40, p<.001). Listeners conveying to speakers that they are being heard, adding information or conarrating with the speaker, encouraging further discussion, giving support, talking about a same experience, and asking for or providing emotion analysis were related to more anger expression by the speaker. There was a positive association between speaker positive emotion and listener backchanneling (r=.26, p<.01), listener encouraging discussion (r=.45, p<.001), listener support (r=.24, p<.05), and listener same experience (r=.26, p<.01). Speakers were more likely to express positive emotion when listeners showed that they were listening to the speaker, encouraged further discussion, showed support, and talked about a same experience. There was a positive association between speaker emotion exploration or analysis and listener backchanneling (r=.33, p<.01), listener support (r=.43, p<.001), listener same experience (r=.29, p<.01), and listener emotion analysis (r=.49, p<.001). More backchanneling, support, talking about a same experience, and asking for or providing emotion analysis by the listener was related to more emotion exploration or analysis by the speaker. A positive association was found between speaker disagreement and listener minimization (r=.31, p<.01), listener advice (r=.36, p<.001), and listener emotion exploration (r=.27, p<.01). Minimization of speaker’s anger or event, giving advice, and asking for emotion exploration by the listener was related to more disagreement by the speaker. A positive association was found between speaker inviting listener’s participation and listener backchanneling (r=.29, p<.01), and listener encouraging discussion (r=.28, p<.01). When the speaker invited the listener to participate in the discussion, the listener was more likely to 21 backchannel and encourage further discussion from the speaker. There was also a positive association between speaker agreement and listener minimization (r=.23, p<.05). Interestingly, when the listener minimized the speaker’s anger or situation, the speaker was more likely to agree with the listener. Supporting my hypothesis, listener showing support, talking about a same experience, and analyzing speaker’s emotion were related to both speaker rumination behaviors (anger expression and emotion analysis). Adding information by the listener was related to only speaker’s anger expression. However, asking about speaker’s emotion was not related to either speaker rumination behaviors. No other relationships were statistically significant (p>.05). 22 23 Gender difference in co-rumination behaviors To determine whether girls co-ruminated about anger more than boys, an independent samples t-test was used by comparing co-rumination behaviors in the two genders (see Table 4). In speaker behaviors, there was a significant difference in the scores for off-topic content between girls (M=.28, SD=.26) and boys (M=.22, SD=.19); t(105)=-1.471, p<.05, in the scores for positive emotion between girls (M=.04, SD=.10) and boys (M=.01, SD=.02); t(105)=-2.584, p<.001, and in the scores for disagreement between girls (M=.02, SD=.05) and boys (M=.01, SD=.03); t(105)=-2.092, p<.001. In each case, girls had significantly higher scores compared to boys. In listener behaviors, there was a significant difference in the scores for off-topic content between girls (M=.29, SD=.27) and boys (M=.24, SD=.21); t(105)=-1.021, p<.05, in the scores for encouragement of discussion between girls (M=.04, SD=.08) and boys (M=.02, SD=.04); t(105)=-1.789, p<.01, and in the scores for emotion analysis between girls (M=.10, SD=.17) and boys (M=.05, SD=.10); t(105)=-1.636, p<.05. In each case, girls had significantly higher scores compared to boys. Other behaviors were not found to have significant gender differences. These results suggest that regardless of whether they were the speaker or the listener, girls introduced off-topic content in the discussions more than boys. In addition, girls tended to express more positive emotion during the discussions and disagree with the listener when they were speakers. Girls tended to encourage further discussion and ask for or provide emotion analysis when they were listeners. The results also indicate that gender differences were present in only 3 out of 8 speaker behaviors and 3 out of 11 listener behaviors, suggesting that girls and boys co-ruminated about anger to a similar 24 extent during conversations. This did not support my hypothesis that girls would show more co-rumination behaviors than boys. Table 4 One-Tailed Independent Samples T-Test Between Behavior and Gender 25 Co-rumination effect on post-conversation emotion A test of hierarchical regression was used to determine whether more corumination predicted higher post-conversation unhappy, worked-up, and anger emotions while controlling for pre-conversation emotions. Step 1 controlled for pre-conversation emotion ratings. Step 2 showed the effect of all co-rumination behaviors and emotion ratings on post conversation emotion after controlling for pre-conversation emotion ratings. Only pre-conversation unhappy ratings predicted post-conversation unhappy ratings (p<.01), and only pre-conversation worked-up ratings predicted post-conversation worked-up emotion ratings (p<.05) (see Table 6 and 7). Neither co-rumination behaviors nor pre-conversation mad & angry ratings predicted post-conversation mad & angry ratings (p>.08) (see Table 5). The results did not support my hypothesis that corumination behaviors would predict increased levels of anger after conversations. 26 27 28 29 DISCUSSION This research hypothesized that 1) certain listener behaviors (adding information, showing support, talking about a similar experience, asking for emotion exploration, and providing speaker’s emotion analysis) would be associated with speaker rumination behaviors (anger expression and emotion analysis), 2) girls would show more corumination behaviors than boys, and 3) more co-rumination would lead to sustained or increased levels of anger post conversation. The results only supported my first hypothesis and showed positive relationships between 4 out of 5 listener co-rumination behaviors and speaker rumination behaviors. Relations between speaker and listener behaviors The results supported my hypothesis that listener’s co-rumination eliciting behaviors such as showing support, giving information about the event, talking about a similar experience, and asking for or providing emotion analysis for the speaker, are related to speaker rumination behaviors, anger expression and emotion exploration or analysis. It was also anticipated that listeners asking about the speakers’ emotion would be related to the speaker rumination variables stated above, however, data showed no significant relationships between these behaviors. Listeners’ talk about their similar experience, emotion analysis, and support were positively associated with speakers’ expressions of anger and their emotion exploration and analysis during conversations. Listeners’ additional information about the event was associated with speakers’ anger expression, but not speaker’s emotion exploration or analysis. Co-ruminating listeners seem to have responded in engaged, and supportive ways, which elicited rumination from speakers (Rose, 2014). Due to the speaker’s 30 perception that he/she and his/her peer shared an understanding of the experience, discussions where listeners talked about their similar experience may have created a space for speakers to feel unjudged to further discuss their emotions and freely express anger about the event. Since listeners who talked about a similar experience also showed greater support to speakers, speakers may have felt more justified to express anger and talk further about their emotions to their peer. Furthermore, listeners’ additional information about the event, whether it was factual or opinionated, may have proven to the speaker that his/her peer acknowledges where they are coming from with regards to the event. Knowing that listeners understood their side of the stories, speakers may have further felt that it was righteous to express anger. Because anger is likely to involve selfjustification and blaming others in solitary rumination alone (Tice & Baumeister, 1993), a peer’s understanding and support of one’s account and anger may help them to further feel justified in their anger. The same listener behaviors are associated with the speaker’s expression of both negative (anger) and positive emotions The majority of listener behaviors resulted in anger expression by the speaker, and several of these behaviors were positively associated with speaker’s positive emotion expression as well. Along with listener co-rumination behaviors stated above, listeners’ encouragement of discussion and backchanneling, or showing the speaker that they are being heard, was strongly associated with speaker’s anger expression. By asking speakers to speak more about the anger event and actively listening to their story, it is likely that listeners provided a space for speakers to feel that they are welcome and even encouraged 31 to talk freely about the event. This type of space that encouraged communication may also have served to justify the speaker’s anger. When listeners responded in ways that showed the speakers that they were listening to their stories, encouraged further discussion, showed support, and talked about their similar experience, speakers were more likely to express positive emotion during conversations. Speakers’ positive emotion may have resulted from the fact that they were accompanied by a close friend who seemed willing to listen to their stories, support them, and contribute in the discussions. Research supports that perceived autonomy support from a close friend is positively associated with experience of positive affect and expression of both positive and negative affect (Deci et al., 2006). Agreement and disagreement with listener’s perspective. It is interesting to note that when listeners minimized the speaker’s accounts of their stories, speakers were likely to both agree and disagree with the listener’s input in the discussions. This can be taken to mean that speakers sometimes felt that their experiences were thus inadequate, as expressed by listeners and may be inclined to agree with them, as speakers feel that listeners may hold an objective view of the event unlike their own which may have been subject to their negative emotions. Conversely, speakers may sometimes feel that listeners do not completely understand their experience of the event when listeners minimize their accounts, and may want to defend their accounts by disagreeing with and correcting listeners’ perspectives throughout the conversation. 32 Gender differences in co-rumination behaviors The results did not support my hypothesis that co-rumination about anger between girls would be more extensive compared to boys. Significant gender differences were found in 3 out of 8 speaker behaviors that were not classified as speaker rumination behaviors: off-topic content, positive emotion expression, and disagreement with listener. In listener behaviors, significant gender differences were found in 3 out of 11 behaviors: off-topic content, encouraging further discussion, and asking for or providing emotion analysis, with emotion analysis being one of the co-rumination eliciting behaviors by the listener. In each of the behaviors that were found to have gender difference, girls showed greater frequencies than boys. First, when they were speakers, girls expressed more positive emotion than boys during discussions. This could suggest that girls express a wider range of emotions while speaking with one another than boys (Kuebli et al., 1995). Furthermore, girls tended to disagree more with listeners during the conversation. It may be that girls hold a wider range of ways to explain a single event and therefore, are more likely to come into disagreement with each other concerning different aspects of the event that may influence their ways of interpreting the event as a whole. This is not to say that boys may be less thorough and creative when it comes to interpreting events, but rather, that girls may greatly consider secondary details as well as primary details when making sense of events. Studies (Fivush & Baker-Ward, 2005; Bohanek & Fivush, 2010) report that females may use more cognitive processing words in narratives than males, which could suggest that there is more meaning-making about events in conversations between women. 33 When they were listeners, girls tended to encourage further discussion with speakers and ask for the speaker’s emotion analysis or analyze their emotions. And regardless of whether they were a speaker or a listener, girls had introduced or talked about an irrelevant topic more during the conversations. These relationships suggest that, generally speaking, girls tend to talk more with each other regardless of what the topic may be. Co-Rumination and Post-Co-Rumination Emotion Ratings The results did not support my hypothesis that co-rumination behaviors (speaker anger expression, speaker emotion exploration or analysis, listener additional information, listener support, listener similar experience, and listener emotion exploration and analysis) would predict higher ratings of anger emotions (anger & mad, unhappy, and worked-up) post conversation. I found that none of the co-rumination behaviors predicted unhappy ratings post conversation, but pre-conversation unhappy ratings did predict postconversation unhappy ratings at a significant level. The same result was also obtained for worked-up ratings, with pre-conversation worked-up ratings as the only predictor in postconversation worked-up ratings. For anger and mad ratings, none of the co-rumination behaviors predicted anger, and neither did pre-conversation anger and mad ratings. This was surprising as I expected to see at least pre-conversation anger and mad ratings to affect post-conversation anger. The results do not align with past research finding that ruminating about anger leads to sustained/increased anger (Rusting & Nolen-Hoeksema, 1998; Sukhodolsky et al., 2001). The use of different contexts (solitary rumination about anger vs. co-rumination about anger) may explain why I obtained different results with regards to anger emotions post co-rumination. Perceived support, and other input in the 34 discussion from the listener may have had an effect in the way adolescents feel after coruminating. Relations Within Speaker Behaviors Strong positive relationships were found between speaker behaviors of expressing anger and talking about the details of the event, and expressing anger and exploring or analyzing one’s emotion. These relationships support the general notion under corumination literature that continuous discussion of an upsetting event or emotions is closely linked to feeling negative emotions (Nolen-Hoeksema, 1994; Nolen-Hoeksema, 1998; Rusting & Nolen-Hoeksema, 1998; Rose, 2002). Relations Within Listener Behaviors Listeners who showed more support, agreed with, and validated the speaker’s experience of the event talked more about their own experiences similar to that of the speaker’s and further analyzed the speaker’s emotions. Having experienced a similar situation in which they were angered or hurt, listeners may have found it natural to empathize with the speakers and thus show greater support and understanding of their emotions (Batson et al., 1996; Hodges et al., 2010). Implications of This Study This study revealed that most listener behaviors that were hypothesized to elicit further rumination by speakers—showing support, giving information about the event, talking about their similar experience, and asking for or providing emotion analysis—do indeed draw out anger expression and emotion exploration and analysis from speakers, which reinforces co-rumination between adolescents. However, this study was not able to conclude that co-ruminating about anger led to exacerbated feelings of anger. Instead, 35 pre-conversation emotions of unhappiness and feelings of being worked-up predicted those emotions post conversation. Therefore, the data do not support the idea that discussions of anger significantly lead to more anger in adolescents. Perhaps anger corumination plays an interesting role in the way adolescents think and feel after having conversations. Rather than leading to only negative emotions, co-ruminating may induce different emotions and different ways of thinking about the event as well. The notion that rumination is not simply a single construct, but instead a behavior with different components that lead to different outcomes was introduced by Trapnell and Campbell (1999). Further research on these components by Verhaeghen et al. (2014) found that reflective self-focus, which is understood as a more active, positive type of rumination that seeks to understand and alleviate unpleasant feelings, had no effect on dysphoria (a psychological state of sadness); whereas ruminative self-reflection, a more passive, negative type of rumination during which individuals dwelled on negative aspects, was positively associated with dysphoria. In connection to this research, these findings could suggest that even in co-rumination, some adolescents may have co-ruminated by reflecting on the event and themselves (speculating about causes and consequences of the event, and trying to understand parts of the problem), and others may have co-ruminated by brooding, or dwelling on negative aspects of themselves or the event, leading to different emotional outcomes. Past literature on rumination and co-rumination generally reported negative consequences of the act (Nolen-Hoeksema & Girgus, 1994; Rose, 2002; Nolen-Hoeksema, 1998; Morrow & Nolen-Hoeksema, 1990; Stone et al., 2011) and did not consider different aspects of rumination that may potentially be helpful in 36 learning from upsetting experiences. Future research could look at whether reflection and brooding in the context of anger discussions result in different emotions. Although this study did not find any significant gender difference in the ways adolescents co-ruminate about anger, girls were shown to discuss irrelevant topics, express positive emotion, disagree with listeners, encourage discussion, and talk about their emotions more than boys. This suggests that even though girls may not talk more about anger than boys, they seem to enjoy engaging in conversations with their peers in general more than boys. It should be taken into consideration that past research reports that girls talk more about emotions, particularly sadness, than boys (Adams et al., 1995). Adams (1995) further explains that this gender difference may be due to differences in the way parents talk to children about emotions, with parents talking more about emotions, especially sadness, with girls and talking less about emotions with boys, but anger being talked about most when emotions are discussed with boys. However, research by Fivush et al. (2000) also found that parents talked about sadness more with girls than with boys, but did not find gender differences in the emotions girls and boys discussed and in how much they talked. Although it is generally assumed that girls talk more than boys and talk more about sadness, which tends to explain why girls ruminate more about sadness, there does not seem to be a clear perspective on whether boys talk more about anger than girls. Perhaps boys are less likely to talk about all emotions in general due to having less discussions with peers, but more research with focus on male friendships and discussions of emotions is needed to confirm this possibility. 37 Limitations One significant limitation of this study is that coder reliability was not perfect in many speaker and listener behaviors. Although I obtained useable reliability scores on most behaviors, the relatively low reliability limits how much this study can conclude about adolescents’ behaviors while they converse with one another. While this study found several reasonable results regarding speaker and listener behaviors, there is still a possibility that higher reliability between coders would yield different results in the ways adolescents talk about anger. Also, I examined a high number of correlations but did not correct for all of these statistical tests. This may have inflated the chances of type 1 error. Another limitation of this study is that I only looked at pre- and post-conversation emotion ratings of anger, unhappy, and worked-up and did not find that co-rumination behaviors predicted those emotions. Looking at pre- and post-conversation positive emotion ratings and whether they predict or are predicted by anger co-rumination behaviors would be necessary to determine whether adolescents feel better or the same as they did before talking with their peer, regardless of topic. Last, the sample was obtained from a mid-sized city in western United States with a heavy religious influence. Particular religious customs, values, and family dynamics may play an important role in the way children are raised in homes which may likely affect their social interactions with others. As well, most adolescents were Caucasian and from middle-class families and thus this limits generalizability on a larger population of adolescents. 38 Conclusion This research examined co-rumination behaviors between same-sex adolescent peers in the context of an anger emotion. One-hundred and seven speaker and listener dyads were studied. Coded conversations between adolescents were used to obtain proportions of different behaviors by speakers and listeners that occurred in conversations. I found that showing support, giving more information about the event, talking about a similar experience, and analyzing speakers’ emotion by listeners were related to more anger expressions and emotion analysis by speakers, thus reinforcing corumination. Except for listener emotion analysis, gender differences were not found in these behaviors, suggesting that girls and boys co-ruminate about anger to a similar extent. However, gender differences found in other behaviors that were not classified as co-rumination behaviors indicates that there may be more conversations between girls than between boys. Co-rumination behaviors also did not predict exacerbated feelings of anger post co-rumination, which suggests that certain aspects of co-rumination may lead to different emotions and thought processes. This finding could help bring to awareness that co-rumination may not bring only negative outcomes for those whose goals are to make sense of upsetting experiences. 39 APPENDIX 40 41 REFERNCES Adams, S., Kuebli, J., Boyle, P. A., & Fivush, R. (1995). Gender differences in parentchild conversations about past emotions: A longitudinal investigation. Sex Roles: A Journal of Research, 33(5–6), 309–323. doi:10.1007/BF01954572 Alink, L. R., & Egeland, B. (2013). The roles of antisocial history and emerging adulthood developmental adaptation in predicting adult antisocial behavior. Aggressive Behavior, 39(2), 131-140. doi:10.1002/ab.21464 Batson, C. D., Sympson, S. C., Hindman, J. L., Decruz, P., Todd, R. M., Weeks, J. L., et al. (1996). “I’ve been there, too”: Effect on empathy of prior experience with a need. Personality and Social Psychology Bulletin, 22, 474-482. Bohanek, J. G., & Fivush, R. (2010). Personal narratives, well-being, and gender in adolescence. Cognitive development, 25(4), 368–379. doi:10.1016/j.cogdev.2010.08.003 Birmaher, B., Ryan, N. D., Williamson, D. E., Brent, D. A., & Kaufman, J. (1996). Childhood and adolescent depression: A review of the past 10 years. Part II. Journal of the American Academy of Child & Adolescent Psychiatry, 35(12), 1575-1583. doi:10.1097/00004583-199612000-00008 Buntaine, R. L., & Costenbader, V. K. (1997). Self-reported differences in the experience and expression of anger between girls and boys. Sex Roles, 36(9/10), 625-637. doi:10.1023/a:1025670008765 Calmes, C. A., & Roberts, J. E. (2008). Rumination in interpersonal relationships: Does co-rumination explain gender differences in emotional distress and relationship 42 satisfaction among college students? Cognitive Therapy and Research, 32(4), 577-590. doi:10.1007/s10608-008-9200-3 Deci, E. L., La Guardia, J. G., Moller, A. C., Scheiner, M. J., & Ryan, R. M. (2006). On the benefits of giving as well as receiving autonomy support: Mutuality in close friendships. Personality and Social Psychology Bulletin, 32(3), 313–327. doi:10.1177/0146167205282148 Denson, T. F. (2012). The multiple systems model of angry rumination. Personality and Social Psychology Review, 17(2), 103-123. doi:10.1177/1088868312467086 Fivush, R., Brotman, M. A., Buckner, J. P., & Goodman, S. H. (2000). Gender differences in parent–child emotion narratives. Sex Roles: A Journal of Research, 42(3–4), 233–253. doi:10.1023/A:1007091207068 Fivush, R., & Baker-Ward, L. (2005). The search for meaning: Developmental perspectives on internal state language in autobiographical memory. Journal of Cognition and Development, 6(4), 455-462. Hodges, S. D., Kiel, K. J., Kramer, A. D. I., Veach, D., & Villanueva, B. R. (2010). Giving birth to empathy: The effects of similar experience on empathic accuracy, empathic concern, and perceived empathy. Personality and Social Psychology Bulletin, 36(3), 398–409. doi:10.1177/0146167209350326 Kuebli, J., Butler, S., & Fivush, R. (1995). Mother-child talk about past emotions: Relations of maternal language and child gender over time. Cognition and Emotion, 9(2–3), 265–283. doi:10.1080/02699939508409011 43 Leaper, C., Carson, M., Baker, C., Holiday, H., & Myers, S. (1995). Self-disclosure and listener verbal support in same-gender and cross- gender friends’ conversations. Sex Roles, 33, 387–404. doi:10.1007/ BF01954575 Lopez, C. M., Driscoll, K. A., & Kistner, J. A. (2009). Sex differences and response styles: Subtypes of rumination and associations with depressive symptoms. Journal of Clinical Child & Adolescent Psychology, 38(1), 27-35. doi:10.1080/15374410802575412 Maxwell, J. P. (2004). Anger rumination: An antecedent of athlete aggression? Psychology of Sport and Exercise, 5(3), 279-289. doi:10.1016/s14690292(03)00007-4 Maxwell, J. P., Sukhodolsky, D. G., Chow, C. C., & Wong, C. F. (2005). Anger rumination in Hong Kong and Great Britain: Validation of the scale and a crosscultural comparison. Personality and Individual Differences, 39(6), 1147-1157. doi:10.1016/j.paid.2005.03.022 Moreira, J. F., Miernicki, M. E., & Telzer, E. H. (2015). Relationship quality buffers association between co-rumination and depressive symptoms among first year college students. Journal of Youth and Adolescence, 45(3), 484-493. doi:10.1007/s10964-015-0396-8 Morrow, J., & Nolen-Hoeksema, S. (1990). Effects of responses to depression on the remediation of depressive affect. Journal of Personality and Social Psychology, 58(3), 519-527. doi:10.1037//0022-3514.58.3.519 44 Nolen-Hoeksema, S., & Girgus, J. S. (1994). The emergence of gender differences in depression during adolescence. Psychological Bulletin, 115(3), 424-443. doi:10.1037/0033-2909.115.3.424 Nolen-Hoeksema, S. (1998). The other end of the continuum: The costs of rumination. Psychological Inquiry, 9(3), 216-219. doi:10.1207/s15327965pli0903_5 Ottaviani, C., Shapiro, D., & Fitzgerald, L. (2010). Rumination in the laboratory: What happens when you go back to everyday life? Psychophysiology, 48(4), 453-461. doi:10.1111/j.1469-8986.2010.01122.x Peled, M., & Moretti, M. M. (2007). Rumination on anger and sadness in adolescence: Fueling of fury and deepening of despair. Journal of Clinical Child & Adolescent Psychology, 36(1), 66-75. doi:10.1207/s15374424jccp3601_7 Riecher-Rossler, A. (2010). Prospects for the classification of mental disorders in women. European Psychiatry, 25(4), 189-196. Rose, A. J. (2002). Co-rumination in the friendships of girls and boys. Child Development, 73(6), 1830-1843. doi:10.1111/1467-8624.00509 Rose, A. J., Schwartz-Mette, R. A., Glick, G. C., Smith, R. L., & Luebbe, A. M. (2014). An observational study of co-rumination in adolescent friendships. Developmental Psychology, 50(9), 2199-2209. doi:10.1037/a0037465 Rusting, C. L., & Nolen-Hoeksema, S. (1998). Regulating responses to anger: Effects of rumination and distraction on angry mood. Journal of Personality and Social Psychology, 74(3), 790-803. doi:10.1037//0022-3514.74.3.790 45 Siewert, K., Kubiak, T., Jonas, C., & Weber, H. (2011). Trait anger moderates the impact of anger-associated rumination on social well-being. Personality and Individual Differences, 51(6), 769-774. doi:10.1016/j.paid.2011.06.030 Stimmel, T., Crayton, C., Rice, T., & Raffeld, P. M. (2006). Pain perception as a function of self-focused rumination. Perceptual and Motor Skills, 103(5), 21. doi:10.2466/pms.103.5.21-28 Stone, L. B., Hankin, B. L., Gibb, B. E., & Abela, J. R. (2011). Co-rumination predicts the onset of depressive disorders during adolescence. Journal of Abnormal Psychology, 120(3), 752-757. doi:10.1037/a0023384 Sukhodolsky, D. G., Golub, A., & Cromwell, E. N. (2001). Development and validation of the anger rumination scale. Personality and Individual Differences, 31(5), 689700. doi:10.1016/s0191-8869(00)00171-9 Tice, D. M., & Baumeister, R. F. (1993). Controlling anger: Self-induced emotion change. In D. M. Wegner & J. W. Pennebaker (Eds.), Hand-book of mental control (pp. 393-409). Englewood Cliffs, NJ: Prentice Hall. Trapnell, P. D., & Campbell, J. D. (1999). Private self-consciousness and the five-factor model of personality distinguishing rumination from reflection. Journal of Personality and Social Psychology, 76, 284-304. doi:10.1037/0022-3514.76.2.284 Verhaeghen, P., Joormann, J., & Aikman, S. N. (2014). Creativity, mood, and the examined life: Self-reflective rumination boosts creativity, brooding breeds dysphoria. Psychology of Aesthetics, Creativity, and the Arts, 8(2), 211-218. doi: 10.1037/a0035594 46 White, B. A., & Turner, K. A. (2014). Anger rumination and effortful control: Mediation effects on reactive but not proactive aggression. Personality and Individual Differences, 56, 186-189. doi:10.1016/j.paid.2013.08.012 World Health Organization (2002). Gender and mental health. Gender and Health. Retrieved from http://www.who.int/gender/other_health/genderMH.pdf