فعال شدن مراقب داخلی: نقش طرح های حمایتی در افزایش حالت خوددلسوزی

Abstract Self-compassion, which involves treating one's own suffering with compassion, mirrors the interpersonal experience of giving support to others. In four experiments we examined the hypothesis that activating support-giving schemas can increase state self-compassion. In Experiments and , participants first recalled a negative event (Experiment 1) or experienced a lab-based test failure (Experiment 2), then were randomly assigned to recall an experience of giving support to versus having fun with another person, and finally completed a measure of state self-compassion. Experiments and examined the effects of actually giving support to another person (via written advice), compared to not giving support or simply reading about another's problem, and assessed effort invested in writing a self-comforting statement, operationalized as statement length (Experiment 3), and self-reported self-compassion (Experiment 4). As predicted, both forms of support-giving schema activation increased self-compassion. Alternative explanations involving affect, self-esteem, and awareness of others' problems were addressed. These results suggest that one way to increase compassion for the self is to give it to others.

Introduction Recent research suggests that self-compassion has numerous psychological benefits, but little work has experimentally examined the contextual factors that give rise to it. Because self-compassion mirrors the interpersonal process of giving support to others, we propose that activating support-giving schemas may increase compassion directed at the self. Self-compassion Self-compassion involves approaching one's own suffering with an attitude of kindness and nonjudgmental understanding (Neff, 2003a). Self-compassion is especially relevant in the context of negative events such as failure or rejection, when people are more likely to be self-critical. Neff (2003a) identified three components of self-compassion: self-kindness, or being understanding and patient with negative aspects of the self; common humanity, or recognizing that making mistakes is part of being human; and mindfulness, or taking a balanced, nonjudgmental approach to negative emotions. Although similar to self-esteem, self-compassion is conceptually and empirically distinct: unlike self-esteem, self-compassion is non-evaluative and helps people confront their weaknesses without being either self-deprecating or self-enhancing (Neff, Kirkpatrick, & Rude, 2007). Furthermore, self-compassion uniquely predicts a number of positive outcomes above and beyond self-esteem, such as more balanced reactions to stressful events (Leary, Tate, Adams, Allen, & Hancock, 2007), greater self-worth stability (Neff & Vonk, 2009), and lower narcissism (Neff, 2003b). Self-compassion is associated with positive psychological and social functioning. Self-compassionate people are lower in symptoms of anxiety and depression, even when controlling for self-esteem (Neff et al., 2007), and self-compassion-focused therapeutic interventions reduce shame and self-criticism (Gilbert & Procter, 2006). Rather than being a form of self-indulgence or complacency, self-compassion is associated with taking the initiative to make positive changes (Neff et al., 2007), engaging in constructive relationship-maintenance behaviors (Baker & McNulty, 2011), and pursuing mastery goals in academic settings (Neff, Hseih, & Dejitthirat, 2005). In addition, inducing self-compassion has been shown to increase self-improvement motivation (Breines & Chen, 2012). Given such wide-ranging benefits of self-compassion, it is important to identify factors that increase or decrease it. The role of support-giving schemas Although there is likely to be some degree of cross-situational stability in self-compassion, it may also be sensitive to the social context, so that self-compassion differs in various situations. To date, the overwhelming majority of research on self-compassion has assessed the construct as a stable trait, with little research examining how situational factors influence self-compassion levels in a given context. In particular, the experience of giving support to another person may in turn increase the likelihood that people will take a supportive attitude towards themselves while their support-giving schemas are activated. Although one might suspect that giving support to others would reduce self-focus, orienting people towards others' needs rather than their own, we propose that this relational orientation may in fact facilitate self-compassion. Self-compassion is an intrapersonal supportive exchange that mirrors interpersonal supportive exchanges and thus may be momentarily increased by the contextual activation of such exchanges. Much of the research on the influence of relationships on the self has focused on examining how others' behavior towards the self influences self-relevant processes and outcomes. For example, priming representations of critical versus accepting significant others led participants to make corresponding self-evaluations ( Baldwin, 1992, Baldwin, 1994 and Baldwin et al., 1990), and expecting to receive social support led participants to make less use of self-denigrating coping strategies ( Pierce & Lydon, 1998). One of the basic tenets of attachment theory states that representations of others' behavior towards the self shape internal working models of self, such as beliefs about being worthy of love (e.g., Bartholomew and Horowitz, 1991, Bowlby, 1982 and Hazan and Shaver, 1987). Consistent with attachment theory, research suggests that self-compassionate adults are more likely to have supportive rather than critical mothers ( Neff & McGeehee, 2010). Far less research, however, has explored the influence of one's own behavior towards others on the self. We propose that self-compassion is likely to be influenced by the priming of interpersonal schemas related to one's own behavior towards others—specifically those involving giving support. The way we treat ourselves in difficult times may be just as much due to the salience of interpersonal schemas related to giving support as they are to schemas related to receiving support. Cross-sectional and longitudinal research provides mixed evidence for a linkage between compassion for others and compassion for the self. Neff and Pommier (2012) found a positive association between trait self-compassion and some measures of other-focused concern (i.e., forgiveness and perspective taking) in both undergraduates and community adult samples, but in the undergraduate samples trait self-compassion was not significantly correlated with compassion for humanity, empathic concern, or altruism. Consistent with previous research showing that undergraduates tend to report being kinder to others than they are to themselves (Neff, 2003a and Pommier, 2011), this finding seemed to be driven by a sub-group of participants who were low in self-compassion but high in empathic concern (Neff & Pommier, 2012). Other studies, by contrast, have found a positive association between compassion for others and for the self among undergraduates: for example, a longitudinal study showed that incoming college freshman who held more compassionate goals towards their roommates also had higher levels of self-compassion (Crocker & Canevello, 2008). In addition, recent fMRI research found that self-compassion engages brain regions that are also involved in feeling compassion towards others (Longe et al., 2009), suggesting that caring for the self and caring for others may be neurologically linked. None of this research, however, addresses the question of whether the contextual activation of support-giving schemas may increase self-compassion in the moment, independent of trait levels of compassion or self-compassion. Furthermore, although some past research suggests that self-compassion and compassion for others may not be correlated at the trait level, we hypothesized that momentary activation of support-giving schemas might lead individuals to show greater state self-compassion. The present experiments Prior research has considered self-compassion primarily as a trait and has focused on assessing its correlations with various other traits and behaviors. An important next step is to examine how interpersonal processes may influence levels of state self-compassion. In four experiments, we tested the hypothesis that activating support-giving schemas can increase self-compassion. The first two experiments examined whether activating a support-giving schema by having participants think about giving support to a friend led to greater self-compassion for a recalled negative event (Experiment 1) and a lab-based academic failure (Experiment 2). Because research suggests that giving can increase positive mood (Dunn, Aknin, & Norton, 2008), and positive mood could in turn increase positive self-directed attitudes such as self-compassion, in both experiments we included a control condition in which participants were asked to think about having fun with a friend, an experience that is likely to be associated with positive mood. By having participants in this control condition write about a friend, we also controlled for any effects due to the activation of positive interpersonal schemas more generally. In Experiments and , support-giving schemas were activated by having participants actually give support (i.e., written suggestions to someone going through a romantic break-up, Experiment 3, or someone who recently got into a car accident, Experiment 4) compared to a control condition in which participants were not asked to give support (Experiments 3 and 4), or a control condition where they read about another person's problem without giving support (Experiment 4). The latter control condition allowed us to address the alternative explanation that self-compassion is increased simply by reminding people that others have problems, rather than because of something specific to giving support. As a dependent measure in Experiment 3, we assessed length of the self-comforting statement that participants wrote regarding a recalled personal negative event, as an index of the effort they expended towards being self-compassionate. Experiments 1, 2, and 4 used self-report measures of state self-compassion. Across experiments, we expected that participants in the support-giving schema condition (i.e., those who were asked to give support) would exhibit greater self-compassion compared to control participants. Experiment 1 Experiment 1 compared the effects of activating a support-giving schema with those of a positive interpersonal schema unrelated to support-giving. Participants recalled a time when they gave support to a friend (experimental condition), or had fun with a friend (control condition), and then reported their state self-compassion regarding a personal negative event that they wrote about before the manipulation. Method Participants Sixty-three undergraduates (61% female) participated for course credit. Five were excluded from the analyses for non-compliance with the manipulation instructions, leaving a final sample of 58. Fifty-seven percent of participants identified themselves as Asian-American, 19% as European-American, 14% as Latino/a, and 2% as African-American. The remainder identified as “Other” or did not indicate their ethnicity. Participants ranged in age from 18 to 29 (M = 20.6, SD = 2.4). Procedure Participants were told that the purpose of the experiment was to understand the relationship between personality and autobiographical memory. Participants first spent 3 min writing about a recent negative event that involved failure, rejection, or humiliation and that made them feel bad about themselves (Leary et al., 2007). They also rated how long ago the event occurred, how upsetting it was at the time, and how bad they still feel about themselves due to the event. Participants were then randomly assigned to one of the two conditions. In the support-giving schema condition, participants were asked to recall a time when they gave emotional support to a good friend and describe it for 2 min. Participants in the control condition were asked to do the same for a time when they had fun with a good friend, a task intended to activate a positive interpersonal schema unrelated to support-giving. After writing the event description participants filled out measures of self-esteem and affect. Following the stem “Right now, how much do you feel…”, state self-esteem was assessed with two items: good about myself and proud (α = 0.73, M = 2.97, SD = 0.98). Positive affect was assessed with six items: happy, content, and reverse-scored sad, upset, angry, and disappointed (α = 0.86; M = 3.78, SD = 0.88). 1 Self-esteem and affect ratings were made using 5-point scales (1 = Not at all, 5 = A lot). Next, state self-compassion was assessed using an adapted version of the self-compassion scale (SCS; Neff, 2003b), which asks participants to reflect on how they generally behave towards themselves in difficult times. The SCS has demonstrated convergent and discriminate validity (Neff, 2003b). For the present experiment, participants were asked to indicate how much they agreed with the statements regarding the negative event they had previously recalled (1 = Strongly Disagree, 7 = Strongly Agree). The scale was shortened to 16 items and reworded to reflect current feelings regarding the recalled negative event. For example, the item “Right now, I'm being understanding towards myself” was used in place of the original item “I try to be understanding and patient towards those aspects of my personality I don't like,” and “A lot of people have negative experiences, I'm not the only one” was used in place of “When things are going badly for me, I see the difficulties as part of life that everyone goes through.” Where applicable, we also removed stems such as “When I fail at something important to me” and “When I feel inadequate in some way…” because they do not make sense in a state context. Because the original measure was created for trait-level ratings, some items were less meaningful or even misleading for “in the moment” ratings — thus, we omitted or, when possible, reworded items like this for the revised scale. We included at least two items from each subscale to capture the full breadth of the original scale. 2 See Appendix A for the complete list of items. The adapted scale was significantly positively correlated with a measure of trait self-compassion that was included in a prescreening survey (r = 0.61, p < 0.001) and it was internally consistent (α = 0.76; M = 4.70, SD = 4.70). Finally, participants completed demographic questions and a suspicion probe and then were debriefed. Results and discussion Most of the negative events described involved academic or social experiences and occurred within the past 2–3 years. In response to the suspicion probe, no one guessed the hypothesis that thinking about giving support might increase self-compassion. Neither gender nor ethnicity interacted with condition to predict self-compassion in this or any of the subsequent studies, so these variables will not be discussed further. State self-compassion scores were marginally significantly higher in the experimental condition (M = 4.88) than in the control condition (M = 4.52), F(1,56) = 3.96, p = 0.05. Thus, recalling an experience of giving support to a friend led to greater compassion directed towards the self relative to recalling a fun experience with a friend. State self-esteem was marginally significantly higher in the experimental (M = 2.76) compared to control condition (M = 3.19), F(1,56) = 2.93, p = 0.09, and was marginally significantly correlated with state self-compassion (r = 0.24, p = .08). Positive affect was significantly lower in the experimental (M = 3.52) compared to the control condition (M = 4.05), F(1,56) = 5.52, p < 0.05. When we controlled for the unique effects of self-esteem and positive affect on state self-compassion, the effect of condition was significant (ps < 0.01). Finally, when controlling for participants' ratings of the negative event (e.g., how long ago it occurred, how upsetting it was), the effect of condition was significant (p < .05), and marginal when controlling for how bad participants still felt about themselves (p = .056). Experiment 2 Experiment 1's results suggest that the activation of support-giving schemas leads to greater self-compassion for a recalled negative event than the activation of positive interpersonal schemas more generally. In Experiment 2, we examined the effects of our support-giving schema manipulation in the context of a lab-based negative event—poor performance on a test. The purpose of using a lab-based event was to ensure that support-giving schemas increase self-compassion for negative experiences that occur in the moment, in addition to past events, and also to construct a more controlled setting where participants would have a similar experience. Method Participants Thirty-four undergraduates (75% females) participated for course credit. Two were excluded because they did not believe that the test was real, leaving a final sample of 32. Fifty-three percent of participants identified as Asian-American, 28% as European-American, 13% as Latino/a, and the remainder as “Other.” Participants ranged in age from 18 to 35 (M = 20.2, SD = 3.0). Procedure Participants were told that the purpose of the experiment was to understand the relationship between personality and test performance. Towards this aim, participants were first asked to complete a “cognitive” test. This test, which included some impossible questions, has been used in previous research primarily to assess cheating behavior, but was used in the present study to ensure that all participants would perform relatively poorly (Gillath et al., 2010 and Niiya et al., 2008). After taking the test participants indicated how well they felt they performed. Next, participants were randomly assigned to one of the two conditions used in Experiment 1 (i.e., giving support to a friend vs. having fun with a friend). Next, they filled out measures of self-esteem and affect. State self-esteem was assessed with the item proud (M = 2.44, SD = 1.24), and positive affect was assessed with the same items used in Experiment 1 (α = 0.63; M = 4.75, SD = 0.77). Finally, participants filled out a state self-compassion scale similar to the scale used in Experiment 1 but some items were tailored to be relevant to the test failure (e.g., “A lot of people have difficulty with tests like this” instead of “A lot of people have negative experiences, I'm not the only one”). Other items remained the same or similar (e.g., “I'm trying to be understanding towards myself”). The scale was shortened to 8 items because not all of the previous items made sense in the context of the test failure (see Appendix B). The adapted scale was significantly positively correlated with a measure of trait self-compassion that was included in a prescreening survey (r = 0.43, p < 0.05) and it was adequately internally consistent (α = 0.62; M = 5.00, SD = 0.66). Finally, participants completed demographic questions and the same suspicion probe used in Experiment 1, and then were debriefed. Results and discussion All participants performed poorly on the test (M = 3.47 correct out of 12 questions, SD = 1.29), and almost all felt that they did poorly on the exam (M = 2.13, SD = 1.13) in response to the question “How well do you think you did?” which was rated on a 1–7 scale (1 = very poorly, 7 = very well). Two participants felt they performed slightly well (i.e., gave a rating of 5). Excluding them did not change the results. As expected, state self-compassion scores were significantly higher in the experimental condition (M = 5.27) than in the control condition (M = 4.76), F(1,30) = 5.44, p < 0.05. This finding conceptually replicates Experiment 1's key finding, but extends it to a recent, lab-based negative event. State self-esteem did not differ significantly between the experimental (M = 2.67) and control conditions (M = 2.24; p > 0.3), nor did positive affect (Ms = 4.74 and 4.75, respectively). When we controlled for the unique effects of self-esteem and positive affect on state self-compassion, the effect of condition remained significant (ps < 0.05). Controlling for perceived test performance, actual test performance, and college GPA also did not change the results. Experiment 3 The results of Experiments and indicate that participants who thought about a time when they gave support to a friend reported greater state self-compassion for a recalled negative event as well as for a lab-based test failure, compared to participants who recalled having fun with a friend, even though thinking about having fun led to greater positive affect and marginally greater state self-esteem in Experiment 1. In Experiment 3, we used a different method of activating support-giving schemas to test our central hypothesis. Specifically, we examined whether actually giving support to another person might also lead to greater self-compassion. In this experiment the target was a stranger rather than a friend, allowing us to examine whether the effects from Experiments and generalize beyond close relationship contexts. Further extending Experiments and , both of which used a positive interpersonal schema control condition (i.e., recalling a fun experience with a friend), we used a neutral control condition in this study. Using such a control condition allowed us to speak to whether the activation of support-giving schemas increases self-compassion relative to a neutral task. To assess state self-compassion, we used an open-ended measure where participants were instructed to give themselves comforting suggestions to make themselves feel better about a recalled negative event. We operationalized state self-compassion as the length of participants' self-comforting statements, which we also coded for qualitative evidence of self-compassion. Consistent with research using amount of cognitive responses as a measure of effortful processing (e.g., Chaiken and Maheswaran, 1994 and Petty and Cacioppo, 1979), we interpreted the length of these statements as an index of the effort they directed at comforting themselves. Using this measure allowed us to bolster our previous findings with a less explicit measure, thereby minimizing potential demand effects since participants would be unlikely to suspect that we were interested in the length of their responses (and in fact no participants guessed this).