Abstract
Previous studies demonstrated that people socially share both their positive and negative emotional experiences. This article reports two studies aimed to clarify and assess the motives underlying this human propensity to share emotions. A large number of motives were collected from 182 participants (Study 1). A content analysis resulted in 8 categories of motives; 72 representative items were retained to create a Social Sharing Motives Scale (SSMS). The final structure of the SSMS, tested with 719 respondents, includes 39 items organized into 7 factors, namely “clarification and meaning,” “rehearsing,” “venting,” “arousing empathy/attention,” “informing and/or warning,” “assistance/support and comfort/consolation,” and “advice and solutions” (Study 2). This scale will usefully contribute to the investigation of the role played by interpersonal processes in emotional regulation.
Keywords
Traditionally, psychologists have investigated emotions as intrapersonal processes. It was stressed that emotions develop in the physiology of the individual and resonate in the depths of this individual’s subjective life (e.g., Arnold, 1960; Frijda, 1986; Tomkins, 1995). However, the empirical literature of recent decades is replete with concepts emphasizing emotions as sustaining essential connections with interpersonal relationships. These concepts accent the fact that emotions are accompanied with verbal expression processes. What used to be considered as a private experience is generally put into words and communicated to members of the entourage. Thus, for instance, the study of emotional disclosure addresses how people respond to emotional upheavals and why translating emotional events into language increases physical and mental health (for reviews, see Pennebaker & Chung, 2007; Smyth & Pennebaker, 2008). Thus, while self-disclosure is defined as “an interaction between at least two individuals where one intends to deliberately divulge something to another” (Greene, Derlega, & Mathews, 2006, p. 411), emotional disclosure represents a specific form of self-disclosure focused at the verbal expression and communication of a personal emotional experience. An important scientific interest exists for the investigation of effects of written emotional disclosure on well-being (e.g., Pennebaker, 1997; Smyth & Pennebaker, 2008). Participants write about past stressful or traumatic events in their lives for short sessions (15–30 min) held on several consecutive days (for a meta-analytic study of effects, see Frattaroli, 2006). Co-rumination was particularly examined in the friendships of children and adolescents. It involves “extensively discussing and revisiting problems, speculating about problems, and focusing on negative feelings” (Rose, 2002, p. 1830). Co-rumination was related to positive friendship quality but also to elevated internalizing symptoms. Studies showed that such rehashing of one’s emotions is socially reinforced and perpetuated by target persons and that co-rumination predicts the onset of depressive disorders during adolescence (Calmes & Roberts, 2008; Stone, Hankin, Gibb, & Abela, 2011). Whereas co-rumination studies examine the expression and verbalization of negative emotions and feelings, capitalization studies investigate people’s propensity to share with close persons the positive emotional experience they just went through (for a review, see Gable & Reis, 2010). Capitalization occurs when one member of a relationship dyad experiences a personal event that positively affects himself or herself and then relates it to the other member of the dyad. Langston (1994) proposed the term “capitalization” after having observed that sharing the news of a positive event with others led the subject to experience more positive affect than could be attributed to the event itself. Capitalization studies later evidenced the positive effect that sharing a positive emotion can have on the interpersonal relationship itself (e.g., Gable, Reis, Impett, & Asher, 2004).
Besides these different concepts, studies on the social sharing of emotion were the first to evidence people’s propensity to share their emotional experiences. The social sharing of emotion is defined as a communication process involving the description of an emotion in a socially shared language by the person who experienced it to another one (Rimé, Mesquita, Philippot, & Boca, 1991). Abundant data showed that when people go through an emotional experience, they immediately feel the need to talk with members of their entourage, and they actually do so in almost all cases (for reviews, see Rimé, 2009; Rimé, Finkenauer, Luminet, Zech, & Philippot, 1998; Rimé, Philippot, Boca, & Mesquita, 1992). Overall, emotional episodes are subject to social sharing conversations in 80% to 95% of the cases, a figure that comes close to those reported for the sharing of positive emotions in recent capitalization studies (Gable & Reis, 2010, pp. 215–216). The social sharing of a given episode occurs most often repetitively, usually several times, with different people for a same emotional episode. The more intense the emotion is, the higher the propensity to talk about it (Luminet, Bouts, Delie, Manstead, & Rimé, 2000). The sharing process typically begins early after an emotional experience has occurred. In 60% of the cases indeed, the first sharing occurs on the actual day that the event occurred, as was also found for the specific case of positive emotions (Gable & Reis, 2010, p. 215). Across age-groups, targets of social sharing were consistently found to be intimates (i.e., parents, brothers, sisters, friends, or spouse/partner), whereas nonintimates hardly played some role in the sharing process, as was also observed for the specific case of positive emotions (Gable & Reis, 2010, p. 216). Communicating an emotional experience seems to be a universal response to an emotion. It is observed with approximately equal magnitude in Asian and Western countries (Singh-Manoux & Finkenauer, 2001). Episodes that involve fear, anger, or sadness are reported to others as often as those involving happiness or love (Rimé et al., 1992). However, emotional episodes that involve shame and guilt tend to be verbalized to a lesser degree (Finkenauer & Rimé, 1998; Rimé et al., 1998; Singh-Manoux & Finkenauer, 2001).
An important feature is that social sharing of an emotion reactivates the shared emotion in the sharing person. Thus, related mental images are reexperienced, body sensations are felt, and subjective feelings are aroused (Rimé, 2009; Rimé, Noël, & Philippot, 1991). In the case of negative emotions, emotional reactivation typically leaves the sharing person in an arousal state. Interestingly, despite these negative consequences, research on social sharing has shown that people are generally eager to discuss their emotional experiences, whether negatives or positives (for reviews, see Rimé et al., 1998, 2009). Then, why are people so eager to share their emotions? The studies that are presented in this article aimed to clarify and assess the motives underlying the universal propensity to share both pleasant and unpleasant emotions. First, we will briefly review the theoretical concepts that are pertinent to this topic, first those for positive emotions and then those for negative emotions.
Positive emotions result from circumstances that facilitate goal-attainment activities (Carver & Scheier, 1990, 2001), and they enhance a subject’s well-being by increasing his/her level of positive affect. Likewise, the social sharing of a past positive emotional experience is likely to elicit pleasurable emotional feelings. In two different studies, Langston (1994) confirmed that the communication of positive events to others was indeed associated with an enhancement of positive affect far beyond the benefits resulting from the valence of the positive events themselves. Gable, Reis, Impett, and Asher (2004) demonstrated that close relationships in which one’s partner typically responds enthusiastically to such a capitalization were associated with higher relationship well-being (e.g., intimacy and daily marital satisfaction). Thus, sharing positive emotions can enhance both the positive affect of individuals and the social bonds between them (Reis et al., 2010). Therefore, capitalization and social integration constitute two demonstrated motives underlying the sharing of positive emotions.
With regard to the question of why people share negative emotional experiences, Schachter (1959) first proposed an answer in the framework of his classic “stress and affiliation” studies. He found that the participants who became anxious at the prospect of being administered electric shocks expressed a preference for waiting in the company of other persons, whereas the control participants preferred to wait alone. Schachter hypothesized that individuals encountering stress attempted to reduce their anxiety by verbally interacting with others in the same situation and thus using others as a lens through which to evaluate their own emotional state. This social comparison motive (Festinger, 1954) is especially relevant when people lack objective standards or undergo a confusing experience, which are typical characteristics of negative emotional experiences.
Negative emotional episodes undermine a person’s knowledge base because these episodes disconfirm expectations and models of the world. Thus, such episodes represent a broad form of distress that a person is highly motivated to reduce (Epstein, 1973, 1990; Rimé, 2009). Although he favored a social comparison explanation for his “stress and affiliation” effect, Schachter (1959) also considered emotional support, or direct distress reduction through the presence of others, to be involved in the process. Since the observations of Bowlby (1969) on attachment, ample evidence has shown that both primate and human infants seek contact with others during periods of uncertainty and distress (e.g., Ainsworth, Blehar, Waters, & Wall, 1978; Sroufe & Waters, 1976). According to Shaver and Klinnert (1982), this early form of affiliation is perpetuated among adults and serves two distinct but related functions: direct anxiety reduction and increased cognitive clarity. This contact seeking would however depend on the quality of the attachment figure’s responses when proximity/help was sought during infancy and childhood and the expectations these responses elicited about the help the others can provide when distressed in adulthood—leading to interpersonal differences in attachment style and proximity seeking (Ainsworth, 1972; for review, Cassidy & Shaver, 2008; Mikulincer & Shaver, 2007; Shaver & Mikulincer, 2002). Thus, the generalized distress that negative emotions produce likely motivates adults (particularly secure and insecure anxious individuals but to a lesser extent insecure avoidant individuals—whose expectations regarding support from others are negative, for a review, see Shaver and Mikulincer, 2002) to search for emotional support and to turn to their attachment figures for this purpose.
Many arguments favor the search for cognitive clarity as the primary motive for the social sharing of negative emotional episodes. By disconfirming aspects of a person’s schemas, models, theories, or assumptions, negative episodes both elicit a state of emotional distress and generate a state of cognitive dissonance within an individual. Therefore, negative emotions are likely to stimulate cognitive efforts toward dissonance reduction (Festinger, 1957). This reasoning was anticipated by both Cantril (1950) and Kelly (1955) who viewed emotions as occurring in moments at which events “do not stick” with cognitive constructions and thus compel individuals to modify these constructions. More recently, Martin and Tesser (1989) argued that when progression toward a goal is blocked or when a discrepancy occurs between the current state of affairs and the expected situation, conditions for the development of cognitive activity are fulfilled. Similarly, Weick (1995) observed that when expectations are disconfirmed or when activities in progress are blocked, efforts to produce meaning emerge. In accordance with these theoretical propositions, a review of the previous empirical findings suggested that one of the most reliable predictors of the need to discuss an emotional episode is the extent of the cognitive needs that are aroused by a given episode (Rimé et al., 1998). Thus, when emotional experiences elicited a need to “put things in order with regard to what occurred”, to “find meaning in what occurred”, or to “understand what occurred”, these experiences were more likely to be subsequently shared. This finding suggests that as the extent to which an emotional episode creates a subjective sense of unfinished cognitive business increases, individuals are likely to feel more motivated to discuss their emotional experiences with others. The results of studies of “secret emotions” support this view. Memories of unshared emotional episodes were found to elicit feelings of unresolved cognitive business among the respondents more so than did memories of episodes that had been shared (Finkenauer & Rimé, 1998). Thus, a need to obtain cognitive clarity or to find meaning appears to constitute a third motive for the social sharing of negative emotion.
In sum, three major motives appear to lead people to share their negative emotional experiences: emotional comparison, emotional support, and cognitive clarity concern. Naturalistic investigations of the stress and affiliation effect have also supported such a conclusion. Kulik and colleagues (Kulik, Mahler, & Earnest, 1994; Kulik, Mahler, & Moore, 1996) examined affiliation toward roommates among hospital patients expecting to undergo major cardiac surgery. In addition to using real life-threatening health events, the authors also assessed actual interaction patterns rather than the mere expression of intentions. In one such study, Kulik, Mahler, and Earnest (1994) concluded that cognitive clarity most accurately accounted for the effects on verbal affiliation that were observed. However, in a subsequent study that examined cognitive clarity concerns, emotional comparison, and emotional support, Kulik, Mahler, and Moore (1996) found evidence for all three motives. These authors concluded that when stress and affiliation relationships were considered in more naturalistic situations, multiple reasons for interpersonal affiliation under threat emerged. This conclusion can likely be extended to the social sharing of emotion.
Another relevant source of information lies in the motives that people openly allege for engaging in sharing behavior. Three sets of data are available in this regard (for a review, see Rimé, 2007). The first set was obtained from a group of psychology students who were enrolled in an advanced class on emotion. These students first recalled a recent emotional experience that they had shared and then listed all of the possible reasons that they had engaged in sharing (Finkenauer & Rimé, 1996). In a second study, a pool of 200 answers was collected from non-psychology students who also referred to a recent emotional experience that they had shared. Their alleged motives for sharing were then grouped by judges using the smallest possible number of classifications (Delfosse, Nils, Lasserre, & Rimé, 2004). Finally, in a third study, 100 male and female participants were recruited in university libraries, and each participant was asked to list five different reasons that they had shared a recent emotional episode in their lives (Nils, Delfosse, & Rimé, 2005, cited by Rimé, 2007). These three studies manifested a striking consistency in the sources of motives that they evidenced (see Table 1). Together, these studies yielded a list of 12 motivational subtypes (see Table 2). Some of these motives are essentially self-oriented, including rehearsing an episode or venting about it, whereas other motives are more clearly other oriented, such as entertaining, informing, or warning the target. In contrast, all of the remaining motives in the list manifest considerable demands on the social targets with regard to emotion regulation. Social sharing partners are indeed expected to provide contributions that are as diverse as providing assistance and support, comfort and consolation, legitimization, clarification and meaning, and advice and solutions. Moreover, this long list of specific social solicitations is still augmented with less specific and more personally involving demands of sharing partners, such as providing attention, bonding, and eliciting empathy. Thus, the motives that are openly alleged for socially sharing emotions reveal an overabundance of social demands aimed at emotional regulation. Although these motives also involve cognitive regulation needs, such as the pursuit of clarification and meaning, they are overwhelmingly likely to meet socioaffective regulation needs such as the search for comfort/consolation.
Summary of the data from the three previous studies about alleged motives for social sharing of emotions (adapted from Rimé, 2007).
A list of motives alleged by respondents for sharing an emotion (from Rimé, 2007).
The current studies intended to examine the interrelationship between the major classes of motives for social sharing and to construct a reliable questionnaire for the assessment of the various motives evidenced. Such a questionnaire may be useful in many different regards. For example, this type of survey would facilitate an examination of variations in motives as a function of the type of emotion that is involved in a shared episode, aspects of emotional circumstances surrounding an episode, types of target persons, and personality traits or clinical diagnoses of sharing persons. The questionnaire could also facilitate the investigation of relationships between sharing motives and the actual responses of targets, in addition to the effects on the emotional recovery and general well-being of sharing individuals. Several existing scales already aim at assessing emotional disclosure: Emotional Self-Disclosure Scale (Snell, Miller, & Belk, 1988), Self-Disclosure Index (Miller, Berg, & Archer, 1983), Distress Disclosure Index (Kahn & Hessling, 2001), and Ambivalence Overemotional Expression (King & Emmons, 1990). Yet these scales all measure confiding/not confiding one’s emotions as a stable individual difference. However, the reasons why people talk about their emotional experiences are determined by both the characteristics of the emotional experience (e.g., valence and type of emotional episode) and the characteristics of the individual (e.g., gender and age), (Delfosse et al., 2004). Furthermore, these existing scales mostly assess the extent of disclosure of negative emotions, thus neglecting positive ones. Having a tool allowing to measure the alleged motives for social sharing would not only permit to better identify expectations in terms of emotion regulation but also to assess the impact of characteristics of both the individual and the event on these emotion regulation needs. The three studies that previously investigated the alleged motives for social sharing have already documented this phenomenon. Yet they simply aimed at exploring existing social sharing motives. The present study is in line with these studies while differing by its aims of creating a questionnaire and as a consequence by collecting data in a more exhaustive way. The present study was planned for the purpose of collecting from participants, and in participant’s own colloquial verbal formulations, a large number of motives, in order to create an assessment tool made of items directly inspired from these colloquial formulations. In addition, insofar as in these previous works the alleged motives for social sharing were mainly collected from students, it seemed important to collect data from a much more varied sample of respondents.
In the first study, we collected a large the broadest possible number of motives that the respondents could identify for having shared an emotional experience with others. The collected motives were then organized into categories and transformed into items. In the second study, the resulting questionnaire was tested on a large sample of respondents.
Study 1
Method
Procedure
A total of 240 people were contacted individually by a female investigator in university libraries, on campus, or through social networks on the Internet. These individuals were invited to participate in a university investigation of the memory of emotional events by completing a questionnaire. The contacted persons who accepted were then asked to recall a recent emotional event that happened less than 3 months before and that they had personally experienced and shared with other persons. Half of the participants were randomly selected to recall a positive emotional episode, whereas the other half were asked to recall a negative emotional experience. Individuals who declined to participate or were unable to recall an emotional event that they had experienced were thanked and dismissed. Those who retrieved a memory as requested then answered the study questionnaire. Confidentiality and anonymity were guaranteed. After indicating their age and gender, the participants were first asked to provide a short written description of the emotional episodes that they had recalled. This procedure, commonly used in studies about social sharing, helps participants to reimmerse themselves in the memory of the emotional situation and to experience a reactivation of the various emotional components before answering the study questionnaire (Rimé, Noël et al., 1991).
Participants
In total, 182 participants (97 females) whose ages ranged from 18 to 79 years (M = 30.16, SD = 12.08) completed the questionnaire, with 81 of them (43 females) in the positive emotion condition and 101 individuals (54 females) in the negative emotion condition. Nearly half of the participants were students (48.90%), 37.91% were employees, 9.89% were unemployed, and 3.30% were retired. A majority of the participants were living as couples (60.44%), 34.06% were single, 3.30% were living alone with children, and 2.20% were widows.
Measurements
The respondents rated the valence of these episodes on a 7-point scale (1 = not positive at all to 7 = very positive) and the intensity of the distress that the episode had elicited (0 = not upset at all to 10 = extremely upset). Subsequently, the participants responded to items that were intended to examine their sharing of these episodes: (1) with whom did they share their experiences (partner, spouse, family member, relative, or stranger); (2) how long after the events did they first share them (same day/same week/more than a week later); (3) the number of people to whom they had spoken (1 or 2/3 to 10/more than 10); and (4) the total number of times that they had discussed their experiences with someone (1 time or 2 times/3 to 10 times/more than 10 times). The questionnaire concluded with one question that was intended to collect a broad range of potential motives for sharing an emotion: “Please list the first 10 reasons that you can recall for discussing this episode with people around you.” A prompt reading “I talked about this event because I wanted to …” was then followed by a blank space in which the participant could freely formulate up to 10 social sharing motives. This procedure allowed us to collect a wide range of alleged motives for social sharing.
Results and discussion
Emotional episodes
The reported emotional episodes were rated as moderately upsetting, both in the positive valence condition (M = 6.04, SD = 1.09) and in the negative condition (M = 5.44, SD = 1.58), which did not differ significantly, F(1, 180) = 1.89. Positive episodes were related to personal instances of achievement (e.g., “finding a job”; 34.57%), leisure (e.g., “attending a concert”; 14.81%), or relationships (e.g., “falling in love”; 12.34%). Negative episodes primarily involved relationship problems (e.g., “breakup”; 15%), health (e.g., “partner being hospitalized”; 15%), or experiences of defeat (e.g., “exam failure”; 14%).
Social sharing
The episodes were first shared on the day that they occurred in 62.64% of cases, during the following week in 26.92% of instances, and more than 1 week later in 10.44% of the time, with no significant difference between the valence conditions, χ2(2, 182) < 1.00. The participants reported having first shared their experiences with their spouse or partner in 41.76% of cases, with another family member in 32.97% of instances, with a friend in 20.88% of instances, and with a relative in only 4.39% of cases. The episodes were shared with 3 to 10 people by a majority of the participants (54.4%), whereas 26.9% had shared their experiences with more than 10 people and only 18.7% reported having shared with only 1 or 2 persons. Positive episodes were shared with more people than negative episodes, χ2(2, 182) = 7.58, p < .05. The frequency of social sharing was generally high: 43.96% of the respondents shared their episodes 3 to 10 times, 32.42% shared their experiences more than 10 times, and only 23.63% shared only once or twice. Positively and negatively valenced episodes did not differ for this variable, χ2(2, 182) = 3.46. All of the above results were perfectly consistent with the previous findings regarding the basic parameters of the social sharing of emotions (for reviews, see Rimé et al., 1998, 2009).
Motives for social sharing
The respondents provided a total of 514 motives for sharing the episodes that they reported, with 308 motives in the positive episode condition and 206 motives in the negative episode condition. As these figures show, positive emotional episodes elicited a much larger number of motives from the participants than did negatively valenced episodes. This difference most likely reflects the heightened creativity and broadened perspective that is observed when positive emotional memories are retrieved or, more generally, when participants are exposed to a positive mood induction (Fredrickson, 1998, 2001; Isen, 2000). Of the 182 participants, 83 (n = 55 in the “negative episode” condition, and n = 28 in the “positive episode” condition) reported two or more reasons which after analysis appeared to belong to a same class of motives.
The 514 collected motives were submitted to a content analysis (Bardin, 1991) in order to organize them into categories of motives. This analysis was led by two independent judges who were uninformed of the categorization scheme proposed by Rimé (2007). The answers that were collected were initially submitted to a semantic analysis that aimed to group items with similar meaning (e.g., “the need to free myself” and “emotional release”). Subsequently, items with similar objectives were combined into the same category (e.g., “externalize my happiness” and “venting my good mood state”). Categories that were obtained by the two judges were then compared to examine discordance. Seventeen items were ultimately eliminated because their content appeared peculiar or irrelevant to social sharing. In addition, 10 other items were discarded because the judges could not agree on how to categorize them. This data-driven categorization resulted in eight classes of motives, which were very similar to the categorization proposed by Rimé (2007) (see Table 2). The titles for each of the categories resulting from the content analysis were adjusted in order to better correspond to those defined by Rimé (2007) (e.g., a category first entitled “informational social support” was entitled “advices and solutions” after the judges learnt the classification established by Rimé, 2007).
Table 3 lists these categories, the number and proportion of items within each category, and examples of such items. The table shows that the answers of the respondents generally corresponded to categories proposed by Rimé (2007). However, the items within the categories of “assistance and support” and “comfort/consolation” could not be meaningfully distinguished from one another. Thus, it was determined that these categories should be merged into a single category labeled “assistance, support, and comfort.” A merging also occurred for “arousing empathy” and “gaining attention” which were merged into a category labeled “arousing empathy/attention.” Encompassing nearly 28% of the items, “venting” accounted for the largest proportion of answers, followed by “informing and/or warning,” “advice and solutions,” “assistance, support, and comfort,” and “arousing empathy/attention,” each of which accounted for 10% to 20% of the collected answers. Less than 10% of the items were categorized as “rehearsing,” “bonding,” or “clarification and meaning.” Finally, two categories, “legitimization” and “entertaining,” which were present in the categorization proposed by Rimé (2007), were not represented at all in the answers of the respondents. Overall, the classes of alleged motives that result from the present study confirm with a larger and more varied sample than those that emerged from the three previous studies. In addition, the collected motives show strong links between the reasons why people talk about their emotional experiences and the strategies they initiate in order to regulate their emotions (e.g., searching for meaning).
Results of the content analysis: distribution of the 487 collected social sharing motives in the various categories.
Venting, the most frequently observed motive in these data, is consistent with the common belief that discussing an emotional experience will reduce or even eliminate its emotional load (Nils & Rimé, 2012; Rimé, 2009; Zech & Rimé, 2005). Thus, the data show that the motives that were considered in this study were those that the respondents recalled. However, some motives might be only weakly accessible to awareness, some motives might even fail to be recalled, and some motives might be completely inaccessible to the mind. We also should acknowledge that some motives may purposely not be reported either because of social desirability concerns or simply because there are motives that are inaccessible to the mind.
Positively and negatively valenced episodes were then compared for the occurrence of the various categories of motives, and the results of this comparison are displayed in Table 4. Positive episodes were more frequently shared for purposes of “rehearsing,” “arousing empathy/attention,” or “informing and/or warning,” whereas negative episodes were more frequently shared for purposes of “venting,” “assistance, support, and comfort/consolation,” “clarification and meaning,” or “advice and solutions.” These findings are consistent with the literature on emotional regulation, which indicates that regulation needs differ according to whether an emotion is positive or negative (Gross & Thompson, 2007; Koole, 2009). After a positive emotion, people predominantly want to amplify the pleasantness that is felt (upregulation), whereas after a negative emotion they are in need of cognitive and emotional assistance to gain control over this emotion (downregulation). No difference occurred between the two types of episodes in the frequency of the bonding motive.
Comparison of positive and negative events for frequency of occurrence of the various categories of motives in respondents’ answers.
*p < .001.
A final version of the item list was obtained by eliminating items in each of the categories of motives that were redundant or lacked clarity and by keeping only the most representative items in each category. In each category, a total of 9 items that were representative of the category to which they belonged were preserved. The 72 resulting items were organized in random order and thus constituted the Social Sharing Motives Scale (SSMS), which was then tested in Study 2. Whereas in the first study, participants referred to either a negative or a positive emotional experience; but in the second study, it was decided to collect the data in reference to emotion categories rather than to emotional valence. We adopted the four emotion categories that were common to classic research on emotional expression (see Ekman, Friesen, & Ellsworth, 1972) and emotional experience (Scherer, Wallbott, & Summerfield, 1986; Shaver, Schwartz, Kirson, & O’Connor, 1987), that is, joy, anger, sadness, and fear.
Study 2
Method
Participants
In Study 2, a total of 770 participants (245 males and 525 females) were invited to participate in a study about the emotions. They were asked to recall a recent emotional event that they personally experienced, before completing a questionnaire. The participants were randomly distributed in one of the four conditions. A quarter of the participants was randomly assigned to recall an event of joy, another quarter had to recall an event of sadness, and the two other quarters were to recall an event of fear or an event of anger. Overall, the participants had shared their emotional events in more than 93.4% (n = 719 of 770) of cases. The data from those 719 participants were then taken into account in the analyses reported hereafter. There were 193 participants in the joy condition (M age = 19.55, SD = 2.97; 62.7% females), 166 in the sadness condition (M age = 18.8, SD = 2.78; 70.5% females), 167 in the fear condition (M age = 19.18; SD = 2.83; 73.6% females), and 193 participants in the anger condition (M age = 18.5, SD = 2.95; 69.9% females). The majority of participants were college (n = 443, 57.53%) or university students (n = 327, 42.47%).
Measurements
Manipulation check
The participants first rated the emotional valence of the event that they had experienced on a 7-point scale (1 = not negative at all to 7 = very negative) and then rated the intensity of their subjective emotion on a 10-point scale from 1 = not upset at all to 10 = very upset. The respondents then evaluated the primary emotions that they had felt in this situation by rating each of the four primary feelings (anger, joy, fear, and sadness) on a 7-point scale (not at all/very strong).
Social sharing
The participants were asked if they had spoken to other person(s) about the episode. Answers were collected for 5 successive items: (a) yes or no; (b) if yes, how long after the emotional event did you discuss it for the first time? (the same day/the same week/more than a week later); (c) with whom did you discuss the event? (partner/friend/family/relative); (d) how often did you discuss it? (once or twice/three to four times/five or more times); and (e) with how many people did you discuss it? (1 to 2/3 to 4/5 or more).
Social Sharing Motive Scale
The SSMS began with the instruction “We would like you to report the reasons that you shared this episode with other people. To this end, please rate the following propositions by indicating how much you agree or disagree with each of them (1 = not at all; 7 = very much).” The participants then rated the 72 items of social sharing motives on the SSMS resulting from Study 1.
Results and discussion
Manipulation check
As expected, the participants who recalled a negative social sharing situation (anger: M = 5.75, SD = 1.50; sadness: M = 5.70, SD = 1.71; and fear: M = 5.59, SD =1.88) reported that these events were more negative than those in the positive condition (joy: M = 1.48, SD = 1.18; F(3, 715) = 335.82, p < .001). Experienced emotions were congruent with the assigned condition, with a high level of anger in the anger condition (M = 6.19, SD = 1.16), of sadness in the sadness condition (M = 5.37, SD = 1.69), of fear in the fear condition (M = 4.98, SD =1.94), and of joy in the joy condition (M = 6.22, SD =1.10). Moreover, the intensity of the recalled emotional episodes was generally high, with M = 8.37 (
Parameters of social sharing
The participants initiated this sharing on the day that the episodes occurred in 70.74% of cases, during the same week in 25.24% of instances, and more than a week later in 4.02% of cases. The latency of sharing initiation varied across conditions, F(3, 715) = 3.25, p < .05). Joyful events were shared more rapidly (M = 1.25, SD = .47) than sad events (M = 1.40, SD = .62) or angry events (M = 1.40, SD = .58). Overall, social sharing occurred 1 time or 2 times in 26.33% of cases, 3 to 4 times in 26.62% of instances, and 5 or more times in 47.05% of instances. The respondents in the joy condition discussed their episodes more recurrently (M = 2.47, SD = .70) than did the respondents in the anger condition (M = 2.05, SD = .82) or fear condition (M = 2.04, SD = .86; F(3, 715) = 12.35, p < .001). The number of social sharing partners amounted to one or two in 2.77% of cases, three to four in 26.77% of cases, and five or more in 70.46% of instances. The participants in the joy condition shared with more persons (M = 2.88, SD = .33) compared with those in the other three conditions: anger (M = 2.54, SD = .57), fear (M = 2.64, SD = .56), or sadness (M = 2.64, SD = .55; F(3, 715) = 15.55, p < .001). The social sharing partners were an intimate in 94.04% of cases (companion: 22.75%, family member: 31.90%, and friend: 39.39%) or an acquaintance in 5.96% of cases. This pattern was independent of event valence, F(3, 715) = .50, p > 1). These results were consistent with those of previous studies and existing literature about social sharing (e.g., Rimé et al., 1998, 2009).
Social sharing motive scale
The analysis of the ratings of the respondents on the SSMS was completed in two steps. First, the correlation matrix for the 72 items was inspected, and redundant items were eliminated to avoid generating spurious factors. Item distributions were inspected to eliminate skewed items that were likely to bias the factor analysis. Finally, a preliminary factor analysis was conducted to identify items failing to load onto any factor. In a second factor analysis, the remaining items were reanalyzed and refined until a satisfactory factor structure was obtained.
Item analysis
The 72 × 72 correlation matrix was inspected to detect potential redundancies (r > .60). A semantic analysis showed an absence of redundancy in the majority of cases. However, 5 pairs of items were determined to be similar; thus, 1 item was removed from each pair. Second, the item distributions were examined (Gorsuch, 1997) for skewness and kurtosis. The item distributions were examined following Kendall and Stuart (1958). In the purpose of factorial analysis, Kline (1998) states that nonnormality is not problematic unless skewness >3 and kurtosis >10. In this study, items that had response distributions with high skewness (approaching 3.0) and high kurtosis (greater than 7.0) were eliminated, which is slightly more severe than suggested by Kline. In exploratory analyses, factor loadings are generally considered to be meaningful when they exceed .40 (Floyd & Widaman, 1995). If an item or items fail to have any substantially high loadings on any factor, these items may be deleted from the analysis and the factor analysis may be recomputed on the remaining subset of items (Floyd & Widaman, 1995). Simple structure is achieved when each factor is represented by several items that each load strongly on that factor only (Pett, Lackey, & Sullivan, 2003). Tabachnick and Fidell (2001) suggest that the secondary loading (or cross-loading) should be no greater than .32. Beavers et al. (2013) states the requirement of having a difference between highest loading and other loadings greater than .3, but this rule is very strong, mostly available for cognitive constructs, and we take .2 as final criterion. Two items had high skewness (greater than 3.0) and kurtosis (greater than 7.0) and were thus eliminated, as highly skewed items can significantly bias the results of factor analyses (Lyne & Roger, 2000). In total, this item analysis resulted in a loss of 7 items.
Factorial refinement
During the second phase, the questionnaire was reanalyzed. In order to investigate its factorial structure, the data were submitted to an exploratory factor analysis (EFA) with Statistical Package for Social Sciences Version 18 and then to a confirmatory factor analysis (CFA) with LISREL 8.8 software (Jöreskog & Sörbom, 2001). The EFA was conducted on a random sample of 382 participants and the CFA on a sample of 387 participants.
Exploratory factor analysis
The responses to the remaining 65 items were subjected to a common factor analysis using the principal axis factoring method of extraction and oblimin oblique rotation to allow for correlations among factors. To obtain an understandable, parsimonious, and stable structure, we defined the selection criteria to determine which items should be included in each factor. To be included, an item must have satisfied the following requirements: the highest loading needed to be above .40; the loadings on the other factors needed to be below .30; and the difference between the highest loading and the other loadings needed to be greater than .20. After the analysis of the factor loadings and iterative eliminations of items that did not fulfill the selection criteria, 39 items remained and were organized into seven factors that explain 66.7% of the total variance (Table 5). The translation of these items is presented in Appendix 1.
Loadings of the 39 items in the seven-factor solution.
Note. N = 382, extraction: principal axis factoring method, rotation: oblimin method. Loadings above .30 are shown in boldface. For the sake of readability, loadings below .20 are not shown.
Confirmatory factor analysis
The results of this analysis generally suggest a good model fit, χ2(387) = 1,789.95; p < .001; root mean square error of approximation = .0661; goodness-of-fit index = .81; normed fit index = .94. The remaining 65 items were subjected to a principal component analysis using pairwise missing data deletion, which yielded 720 valid cases. Three methods were employed to estimate the optimal number of components to be retained: the scree test (Cattell, 1966), Kaiser–Guttman’s criterion (Kaiser, 1961), and component representativeness. The Kaiser criterion generally leads to overestimation of the number of dimensions (Tzeng, 1992), and the scree test is a rather subjective evaluation that can also slightly overestimate the number of factors (Zwick & Velicer, 1986). The representativeness of each component after rotation gives the number of nonnegligible dimensions. Eleven factors had an eigenvalue above 1 (Kaiser criterion), and the shape of the eigenvalue curve suggested that seven components should be retained (scree test). Moreover, examinations of the eight- and seven-component solutions consistently indicated that only seven components had at least three loadings above .40. Thus, a seven-dimensional structure was retained.
Factor 1 “clarification and meaning” (6 items) comprises the various strategies that are intended to achieve an understanding of an emotional experience and to assign meaning to it. Factor 2 “rehearsing” (5 items) refers to a person’s willingness to reexperience an emotion, memorize it, recall it, and even amplify it through rehearsal. Factor 3 “venting” (6 items) reflects a desire to reduce the emotional weight that is associated with an experience. Factor 4 “arousing empathy/attention” (7 items) covers the various strategies that are developed to describe emotions to listeners and elicit their sympathy. Factor 5 “informing and/or warning” (5 items) represents the intention to lend one’s own experience to others for their benefit. Factor 6 “assistance/support and comfort/consolation” (5 items) reflects the willingness to obtain some form of emotional support. Finally, Factor 7 “advice and solutions” (5 items) represents the pursuit of intellectual and/or practical support.
Scores for the 7 subscales were calculated by averaging the ratings of the respondents for the various items involved. Cronbach’s α coefficients were then used to assess the internal consistency of the subscales. As shown in Table 6, the internal consistency of the 6 subscales ranged from .75 to .92, and the correlations among the subscales were satisfactory.
Properties of the seven subscales of the SSMS and their intercorrelations.
Note. Bravais–Pearson’s r (N = 720).
p < .001: correlations in boldface.
The influence of event valence on social sharing motives
The motives for the social sharing of emotion varied across emotional conditions (see Table 7). Overall, the motives for sharing a positive event (joy) differed significantly from those for sharing a negative event (anger, fear, or sadness). Compared with negative events, positive experiences were shared more frequently for the purposes of reexperiencing the event and arousing empathy/attention and less frequently for the purposes of venting and seeking understanding, support, and advice. There were also numerous differences in alleged motives between the various types of negative episodes (implying anger, fear, or sadness); thus, the discriminative power of the SSMS was confirmed.
Motives of social sharing by emotional event condition (means and SD).
*p < .01; **p < .001; a,b,c p < .05.
Impact of gender on the alleged motives for social sharing
As shown in Table 8, all the alleged motives for social sharing were influenced by gender. Women reported more than men talking about their emotional experiences for making sense of their emotional experience, venting, seeking help/support, and obtaining advices/solutions. Men confided themselves more in order to relive the event, to elicit empathy/attention, and to inform/warn others.
Motives of social sharing by gender (mean and SD).
*p < .05; **p < .01; ***p < .0001.
Links between the social sharing parameters and the alleged motives for social sharing
The links between the emotional intensity of the event, the characteristics of the social sharing (delay, frequency, and number of sharing partners), and the seven categories of alleged motives for social sharing are shown in Table 9. This correlation matrix clearly demonstrates the discriminative power of the SSMS-39, as each of the four parameters of social sharing appears to be linked to a specific pattern of alleged motives. Regarding the relationship between the type of partner the episode was confided to for the first time and the reasons why social sharing was undertaken, it appears that seeking help/support and venting are not undertaken to the same extent depending on the type of sharing partner (Table 10).
Correlations between the intensity of the emotional episode, the parameters of social sharing (delay, frequency, and number of persons), and the alleged motives for social sharing (n = 719).
*p < .05; **p < .01.
Motives of social sharing by type of sharing partner (mean and SD).
*p < .05; a,b,c p < .05.
General discussion
The purpose of the current studies was to construct a reliable questionnaire to assess the major classes of motives for the social sharing of emotion. In the first study, we collected a large number of motives that the respondents could identify for having shared a personal emotional experience with others. Subsequently, the collected motives were organized into categories and converted into items to create the intended questionnaire of alleged motives for sharing an emotional experience. In the second study, this questionnaire was tested on a larger sample of respondents. We will first examine how far our findings were consistent with the results of the literature on the social sharing of emotion. We will then discuss the comparability of our findings regarding social sharing motives with those of the three existing studies that investigated this question. Next, we will comment on the motives that prevailed in our respondents’ answers and on the instrument that was created in the present study to assess such motives in a systematic way in the future. Finally, we will discuss the findings resulting from our first empirical use of this instrument, in the comparison of motives according to the type of emotions on the one hand and according to respondents’ gender on the other hand.
Did the data reported by our respondents in relation to the social sharing of emotional episodes confirm previous observations regarding the social sharing of emotion? The results of our two studies in this respect were largely consistent with those of previously published studies. In particular, emotional episodes were predominantly found to be socially shared within a short delay. In total, these episodes had been shared on the same day that they occurred 63% and 73% of the time in Studies 1 and 2, respectively, each of which is close to the value of 60% that has typically been reported in previous studies (for a review, see Rimé, 2009). The major traits of the social sharing of emotion were also confirmed by our findings, which indicate that the reported episodes had been modally shared several times with several persons and that these persons were nearly always individuals with whom the respondents reported having close relationships, such as friends, companions, spouses, or family members (e.g., Rimé et al., 1998, 2009; Skowronski, Gibbons, Vogl, & Walker, 2004). The fact that in Study 1 less participants (54.4%) had talked about their emotional experience to a large number of persons than was the case in Study 2 (70.46%) likely results from the difference in the intensity of the emotional episodes collected in the two studies (M = 2.10 for moderate intensity events in Study 1 vs. M = 8.00 for more intense events in Study 2); it is indeed well established that the more emotionally intense the experienced event was, the more people talk about it (Luminet et al., 2000; Rimé et al., 1998).
How consistent were our findings on motives for the social sharing of emotion with regard to preexisting studies? Of the 12 classes of motives, 8 proposed by Rimé (2007) were confirmed by the open answers that were provided by our participants. The motives that did not emerge from these answers were “legitimization” (i.e., receiving approval, being legitimized, and being understood) and “entertaining” (i.e., amusing another person). However, these two classes of motives partially overlapped with the motives of “clarification and meaning” and “gaining attention”, respectively. In addition, the motives of “emotional social support” and “gaining attention and empathy” became broader categories, with “assistance, support, and comfort/consolidation” in one category and “arousing empathy/gaining attention” in the other category. The results of the present study are also similar to those of the three preceding studies on alleged motives (see Table 11). The results of this study thus confirm the data from the three previous studies (Delfosse et al., 2004; Finkenauer & Rimé, 1996; Nils et al., 2005, cited by Rimé, 2007) with a sample whose average age is higher (M = 30, SD = 12.08; Study 1). In addition, in our sample there was as many women as men and as many students as salaried employees (Study 1). Now, insofar as it is well established that the emotion regulation needs evolve according to age and gender, having a heterogeneous sample in terms of age and gender constitutes an undeniable contribution to the existing data.
Comparison between the results of the present study and those of the three previous studies about alleged motives for social sharing.
In our data, the most popular motives were those that involved the notion of venting (i.e., expressing one’s emotion and relieving an emotional load), which was immediately followed by motives that were oriented toward listeners: “informing and/or warning,” “arousing empathy/attention,” “assistance, support, and comfort/consolation,” and “advice and solutions.” In our discussion of Study 1, we emphasized that “venting” corresponded to the common belief that emotional expression reduces the intensity of emotions or resolves them completely (for a critique, see Nils & Rimé, 2012). The next most popular motives that were mentioned above involve the expectation of an active contribution of the social sharing target to the emotion regulation process of the narrator. These expected contributions can take either a cognitive form, such as assistance in search for meaning, or a socioemotional form, such as emotional social support and manifestations of empathy (Rimé, 2007). The importance of interpersonal relationships in emotional sharing motives is also evidenced by the presence of the motive of “help/support and comfort/consolation.” This evidence indicates that the social sharing of emotion is not only primarily addressed to close persons with important demands for assistance in the emotion regulation process; this sharing also occurs with the open purpose of strengthening preexisting social ties. As was already stated earlier, we insist that the motives considered in the reported studies were restricted to those that respondents recalled. Obviously, some motives may be inaccessible to the mind. Despite this caveat, collecting self-report data about motives for social sharing of emotions means identifying the reasons why, according to them, people talk about their emotional states. Consequently, it provides relevant information on the needs people experience after an emotion and that they feel as central in their interpersonal emotional verbalization.
The SSMS-39 that was developed in the present studies was both consistent and reliable. Future research should aim to verify the psychometric proprieties of this scale among larger and different populations. Most of the participants in Study 2 were women, and the respondents were quite young. This characteristic of the sample could limit the generalization of our results. Future studies will thus have to confirm the psychometric properties of this scale with a more representative sample and to investigate the divergent and predictive validity of this tool. Also, due to the fact the SSMS-39 should differ from the existing measures of disclosure (among others because not focusing on negative emotions and not considering the reasons for confiding as a fixed characteristic but rather as determined by the characteristics of the emotional experience (e.g., valence and type of emotional episode) and of the individuals (e.g., gender and age), its divergent validity with such existing scales will be investigated in future studies. The scale developed in the current study offers a new tool for research that is intended to provide insight into the social dimensions of emotional regulation. In particular, the SMSS-39 will facilitate investigation of the relationships between motives and expectations that are manifested by a narrator, the manner in which listeners respond, and the consequences of such interaction on the various types of benefits that emotional sharing can provide (e.g., Badr & Taylor, 2008; Banthia et al., 2003). Future research in this direction is likely to provide new insight into the manner in which regulation processes and the social regulation of emotions complement one another. Since emotion regulation skills are evolving with age (Carstensen, Fung, & Charles, 2003; Charles & Carstensen, 2007), it is likely that the alleged motives also change with age. Although this could not be shown in our study due to the homogeneity of age in our sample, future studies would have to investigate the impact of this variable. Having a scale which evaluates the alleged motives for social sharing then allows to better understand the emotional regulation needs the individuals search to fulfill when they talk about their emotions with their close relatives and more generally the interpersonal side of emotion regulation.
The data collected with the SSMS-39 led to two comparisons offering an opportunity to test the capacity of this new tool to reveal differences in respondents’ alleged motives for social sharing. A first comparison regarded the type of emotion (joy, fear, anger, or sadness) respondents had to refer to. Paying attention to these four emotions (joy, fear, anger, or sadness) in particular was justified by the fact that they are easily identifiable, what ensured us that the participants in each condition would report an event related to the emotion cited. This variable was found to determine markedly the reasons for sharing (Study 2). Thus, compared with the sharing of negative episodes, the social sharing of episodes of joy is more motivated by a desire to revive such an event by discussing it and to elicit empathy/attention. These findings are coherent with those from the capitalization studies. Gable et al. (2004) have indeed shown that sharing positive experiences within the couple results not only in intraindividual benefits (increased well-being and positive affects by the pleasurable sensations associated with the evocation of the episode) but also in interindividual benefits. In particular, these studies show that socially sharing positive events even more strengthens the relationship when the listener responds in an empathical way (Gable et al., 2004). As evidenced by Gable and Reis (2010), talking about one’s positive emotions to an empathic listener increases the appreciation and confidence that we have in this person (interpersonal benefits). In the light of these studies on “capitalization,” it thus seems logical that individuals look for partners’ empathy after they experienced positive emotions, since this type of reaction will draw both intrapersonal and interpersonal benefits. In the same way, the findings pertaining to the negative emotions are not surprising. Negative emotional episodes are shared more frequently for the purpose of finding meaning and seeking distress-buffering attitudes and actions from others. However, differences emerged according to whether the emotional episodes involved sadness, anger, or fear. Emotional episodes of sadness or anger are more frequently shared to obtain counsel from others than are episodes of fear. The data that allowed the construction of the questionnaire were collected on respondents’ emotions distinguished by their valence in the first study (positive and negative emotions) or by their category in the second one (joy, anger, fear, or sadness). The question then arises whether the scope of the questionnaire constructed in this manner is limited to these particular emotions or if it extends to the whole realm of emotions. In fact, we have stressed that the different social sharing motives evidenced in our work largely overlapped those that emerged from Rimé’s (2007) review of the sharing motives recorded in the three previous studies on alleged motives for social sharing. In these studies, participants were instructed to refer to “an emotion that they had experienced recently and of which they had spoken.” They thus referred to an emotional experience they themselves had selected, a selection procedure that allowed collecting a large variety of emotional experiences. The overlap between the reasons highlighted in these studies and those that emerge from the present work constitutes an important argument in favor of the generality of the scope of our questionnaire. Future studies will refine these preliminary data, by investigating the impact of other types of emotions on the alleged motives.
The second comparison conducted with the new SSMS-39 demonstrated that male and female respondents differ significantly—and sometimes at very high levels of significance—in their level of endorsement of all seven assessed motives of social sharing. In comparison with men, women reported sharing emotions much more often for the purpose of venting their feelings, for receiving help and comfort, for obtaining advice and solutions, and for understanding their emotions. By contrast, men express their emotions much more in order to reexperience the episode, to arouse empathy or attention, and to inform people around them. In sum, after an emotional episode, women experience needs that are different from those experienced by men. It has already been argued that emotional regulation strategies depend on gender (e.g., Haga, Kraft, & Corby, 2009; Zlomke & Hahn, 2010). In particular, Haggard, Robert, and Rose (2011) argued that compared to women, men are more inclined to see problem talk as a puzzle to be solved. This is suggesting that males and females have complementary ways of discussing emotional experiences. Indeed, social sharing theory proposed that two different social sharing modes can be adopted after an emotion and that each mode is having specific consequences for the regulation of the emotion (Nils & Rimé, 2012; Rimé, 2007, 2009). The cognitive mode takes place when the social sharing involves cognitive work, with distancing, perspective taking, reframing, and reappraisal of the episode. This mode favoring the processing of the emotional experience is proper to bring emotional recovery or a significant reduction in the impact of the episode’s memory. The socioaffective mode provides the narrator with social responses involving help, support, comfort, consolation, legitimization, attention, bonding, and empathy perspective. In contrast to the previous one, this mode does not bring emotional recovery but well a strong but temporary alleviation and feeling of relief. The socioaffective mode may usefully pave the way to the more demanding cognitive mode. The underlying rationale of this two-mode view is that as long as the cognitive appraisal of a past emotional episode remains unchanged, the memory of this episode necessarily triggers the same emotional state as the one experienced initially (Rimé, 2007). Experimental findings by Lepore, Fernandez-Berrocal, Ragan, and Ramos (2004) and by Nils and Rimé (2012) provided support for the two-mode view. In the framework of such concepts, the gender differences evidenced here are suggesting that the social sharing motives of males and females are complementary. It should be remembered that at adult age, in both genders, the predominant sharing partners are the spouses (e.g., Rimé et al., 1998). These findings definitely call for future work on the way males and females, on the one hand, and members of couples on the other hand do mutually regulate their emotions in the social sharing process. In any case, the strong and consistent differences evidenced in social sharing motives of males and females prove that the new scale developed in the reported studies is capable of evidencing new and heuristic findings.
Footnotes
Motives for Social Sharing Scale (SSMS-39)
Pouvez-vous nous dire à présent pour
“We would like you to report the reasons that you shared this episode with other people. To this end, please rate the following propositions by indicating how much you agree or disagree
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.