Not all social exclusions are created equal: Emotional distress following social exclusion is moderated by exclusion paradigm

Although we may wish it to be otherwise, our lives are wrought with occurrences of social exclusion; we experience hundreds, even thousands, of such episodes over the course of our lifetimes (Williams, Citation2001), and thus it comes as no surprise that psychologists have taken up the mantle to examine their consequences. As it turns out, these consequences are abundant and variable, with some reactions appearing to be adaptive and/or prosocial (e.g., Bernstein, Sacco, Brown, Young, & Claypool, Citation2010; Bernstein, Young, Brown, Sacco, & Claypool, Citation2008; Gardner, Pickett, & Brewer, Citation2000; Lakin, Chartrand, & Arkin, Citation2008; Williams & Sommer, Citation1997) and others appearing maladaptive and/or antisocial (e.g., Baumeister, DeWall, Ciarocco, & Twenge, Citation2005; Baumeister, Twenge, & Nuss, Citation2002; Twenge, Baumeister, Tice, & Stucke, Citation2001).

Although these consequences of social exclusion clearly vary, most people would intuitively agree that social exclusion is emotionally painful, resulting in feelings of social distress and a worsened mood; indeed, when researchers (e.g., DeWall et al., Citation2011) have asked participants to imagine how they would feel after experiencing a type of social exclusion, they readily forecast they would be very upset. Although this assumption that exclusion hurts emotionally is very intuitive and sensible, the research supporting this naïve belief is far murkier.

For example, although many studies find that excluded individuals report decreases in their senses of belonging, control, meaningful existence, and self-esteem as well as worsened mood (e.g., Bernstein et al., Citation2008; Bernstein, Sacco, Brown, et al., Citation2010; Eisenberger, Lieberman, & Williams, Citation2003; Williams, Cheung, & Choi, Citation2000; Zadro, Williams, & Richardson, Citation2004), others have found quite the contrary, with evidence suggesting that social exclusion results in a feeling of emotional numbness (e.g., Baumeister et al., Citation2005; DeWall & Baumeister, Citation2006; Twenge, Catanese, & Baumeister, Citation2003).

The question concerning whether social exclusion results (or does not result) in emotional distress is so compelling that in recent years two meta-analyses were published with the goal of determining the answer. In these meta-analyses distress was operationalized as negative impacts on mood, self-esteem, and other need-based reactions (feelings of belonging, control, and meaningful existence; although these latter needs were only examined in Gerber & Wheeler, Citation2009a). Interestingly these meta-analyses arrived at opposing conclusions. Based on their analysis, Gerber and Wheeler (Citation2009a) concluded that social exclusion does indeed elicit emotional distress, by decreasing mood, lowering self-esteem, and thwarting the needs of belonging and control. In stark contrast Blackhart, Nelson, Knowles, and Baumeister (Citation2009) concluded, based on their own meta-analysis, that although social exclusion does result in lower relative mood compared to accepted and control participants, excluded people do not feel objectively bad or distressed, instead feeling fairly flat affect (fairly neutral). Moreover these same researchers concluded that exclusion does not lower self-esteem compared to those in neutral control conditions.

Thus, despite two recent meta-analytic investigations, it seems the question about whether or not social exclusion causes emotional distress remains unanswered. We suspect this is because the impact of social exclusion on emotional reactions is complex and multiply determined. There may be a number of moderators that dictate when exclusion triggers emotional distress and when it triggers numbness. And thus no individual investigation, not even a meta-analytic one, may fully elucidate the connection between exclusion and emotional distress. Although the present work also may be unable to fully explain this connection, we hope to examine one piece of the puzzle by focusing on one possible moderator, exclusion paradigm, as a potential determinant of exclusion's impact on emotional distress.

More concretely, we speculate that the manner in which social exclusion is achieved may play an important role in shaping emotionally distressing outcomes because recent work in our lab has shown its important role in shaping physically painful reactions. Specifically, Bernstein and Claypool (2012) put some participants through the Cyberball paradigm (e.g., Williams et al., Citation2000) and others through the Future-Life paradigm (e.g., Twenge et al., Citation2001). In Cyberball participants believe they are playing a virtual game of ball toss with two other individuals over a computer. In the exclusion condition the other “players” (actually controlled by a computer program) toss the ball to the participant a few times at the beginning and then never again. In the inclusion condition the participant receives the ball consistently throughout. The Future-Life paradigm works differently. In it participants first complete a personality inventory and then receive feedback about their future lives, allegedly based on their responses. This bogus feedback tells those in the exclusion condition that their futures will be devoid of meaningful relationships. For those in the inclusion condition, however, they learn of a future filled with caring, loving social connections. Bernstein and Claypool (2010) found that those excluded in Cyberball showed increased sensitivity to physical pain (compared to their baseline levels), whereas those excluded in the Future-Life paradigm showed a numbing of pain (again, relative to their baseline).

Based on these findings, we hypothesize that Cyberball exclusion may induce emotional distress (e.g., reduced mood, and lowered self-esteem and other needs), whereas Future-Life exclusion may induce emotional numbing. The basis of this prediction is straightforward: numerous theoretical approaches and empirical findings have elucidated a connection between experiences of physical and social pain. Indeed, both pain overlap theory (e.g., Eisenberger & Lieberman, Citation2005) and social pain theory (MacDonald, Kingsbury, & Shaw, Citation2005; MacDonald & Leary, Citation2005) contend that both types of pain share a neurological basis, and thus variables that impact one type of pain should have a similar effect on the other (e.g., acetaminophen reduces both physical pain as well as the neurological response to the pain of social rejection, see DeWall et al., Citation2010). Thus, if paradigm moderates physical pain, it should also moderate social pain in the same manner (Bernstein & Claypool, 2012).

In addition to showing that Cyberball exclusion triggers physical pain sensitivity and Future-Life exclusion triggers physical pain numbing, Bernstein and Claypool (2012) argued that these effects were driven, at least in part, by differences in the severity of the “social injury” instantiated by them. To explain, these researchers noted that the relationship between physical injury severity and experienced pain is not linear. Physical injuries that result in mild degrees of tissue damage (i.e., low severity ones, such as stubbing one's toe) result in less pain than do those that result in greater amounts of tissue damage (like breaking a finger). Nonetheless, if the severity of a physical injury becomes too great (e.g., car accident, gunshot), numbing occurs rather than pain. When a physical injury is massive, shock occurs during which both physical and emotional responses are dulled (for a review see Kandel, Schwartz, & Jessell, Citation2000). The authors reasoned that because physical and social pain share the same underlying “neural hardware,” less-severe social injuries (e.g., finding out that a first date wasn’t interested in you) should result in pain whereas a very severe social injury (e.g., the end of a life-long marriage) may result in numbing to pain. In other words, social injuries should behave like physical injuries; at low levels of damage (social or tissue) physical pain will ensue; at extreme levels physical numbness will set it.

Pilot data presented by Bernstein and Claypool (2012) confirmed that participants perceive Future-Life exclusion as a more severe social injury than they do Cyberball exclusion. Importantly, when these same researchers manipulated the severity of the social injury within the Future-Life paradigm they found that the highly severe version (wherein participants were told of a future life devoid of almost all relationships) led to numbness, whereas a less-severe version (wherein participants were told of a future life with fewer social relationships than they might desire) led to hypersensitivity. Thus this work was consistent with the hypothesis that Cyberball exclusion may differ from Future-Life exclusion in terms of perceived social-injury severity, and this difference may result in differential physical pain responses. Importantly, although the authors did theorize about parallel effects for social pain, they presented no direct evidence on this point. Thus the broad purpose of this work is to investigate the hypothesis that Cyberball exclusion will trigger social pain sensitivity, whereas Future-Life exclusion will trigger social pain numbing, as assessed via measures of emotional distress.

This hypothesis flows seamlessly from the findings of Bernstein and Claypool (2012) and is consistent with observations and speculations of other theorists in the literature that “future alone” might produce reactions that differ from other paradigms. Indeed, the Blackhart and colleagues (Citation2009) and the Gerber and Wheeler (Citation2009a) meta-analyses examined paradigm as a possible moderator of exclusion's impact on emotional responses. Regarding affective reactions, Blackhart and colleagues found a “larger effect size when participants were left out of a group … than when they anticipated future rejection” (2009, p. 295), suggesting that a paradigm like Cyberball might lead to greater emotional distress than one like the Future-Life paradigm. Gerber and Wheeler (Citation2009a), however, found no significant effect of paradigm on mood in their meta-analysis, but in a follow-up article (Gerber & Wheeler, Citation2009b) speculated that it may be the case that “the life-alone paradigm [does] not produce emotional responses but that other paradigms [do]” (p. 494).

Given the somewhat inconclusive judgment one might draw regarding whether there is or is not a paradigm (e.g., a Future-Life versus a Cyberball) effect on emotional distress following exclusion based on these two meta-analyses, it seems prudent to examine this question directly in an empirical study. That is, randomly assigning participants to one paradigm or the other (and to exclusion and inclusion conditions within them) and assessing the same measures of emotional distress in all conditions will allow us to directly test whether our hypothesis is viable. Thus in these studies we intend to test the hypothesis that Cyberball exclusion will lead to emotional distress and that Future-Life exclusion will not. To maintain some consistency with the aforementioned meta-analyses, we (like them) will assess emotional distress by tapping participants’ mood, self-esteem, and other need-based reactions (of control, meaningful existence, and belongingness).

OVERVIEW

In two studies we compared how social inclusion and exclusion affected various measures of emotional distress and whether exclusion paradigm acted as a moderator. We hypothesized that Cyberball should elicit increased distress (e.g., lower self-esteem, basic needs, and mood) as compared to inclusion conditions, whereas no parallel differences in distress should occur for participants in the Future-Life conditions.

STUDY 1

Results and discussion

We hypothesized that whereas Cyberball exclusion should decrease mood and the fulfillment of basic needs relative to inclusion, the same pattern would not occur in the Future-Life paradigm, for whom we expected participants’ mood and basic needs to be no different between inclusion and exclusion conditions. We began by averaging the eight mood items to form a composite mood score (α = .87). Because the basic need subscales of belonging, self-esteem, control, and meaningful existence were highly correlated (α = .88), we also formed a composite of the basic needs (for similar procedure, see Bernstein, Young, Sacco, Hugenberg, & Cook, Citation2010; van Beest & Williams, Citation2011). We then conducted 2 (Social Experience: Inclusion or Exclusion) × 2 (Paradigm: Cyberball or Future-Life) between-participants ANOVAs on the composite basic needs score as well as the average mood score.

The ANOVA on the composite basic needs score revealed the predicted interaction, F(1, 50) = 11.12, p = .002,  = .18. Simple effect analyses revealed that although future-belonging (M = 4.04, SD = .44) and future-alone conditions (M = 4.11, SD = .31) did not differ (p > .67), Cyberball exclusion (M = 3.28, SD = .68) did result in less satisfaction of these needs than did inclusion (M = 4.05, SD = .34, p < .001). Further, although inclusion conditions did not differ across paradigm, (p > .97), exclusion conditions did (p < .001, see Figure 1).

Figure 1. Study 1: Composite basic needs score as a function of Social Experience and Paradigm.

Display full size

To further examine the effect that Social Experience and Exclusion Paradigm had on the basic needs, we conducted 2 × 2 ANOVAs on each need separately. The results are displayed in Table 1. For each need, the 2 (Social Experience: Inclusion or Exclusion) × 2 (Paradigm: Cyberball or Future-Life) interaction was significant (all ps < .05). Further, in each case Cyberball exclusion led to a significant reduction of the need relative to Cyberball inclusion, whereas future-belonging and future-alone conditions did not differ. Also, and again for every need, inclusion conditions did not differ between Cyberball and Future-Life, whereas Cyberball exclusion led to reduced needs relative to the future-alone condition.

TABLE 1 Study 1

Download CSVDisplay Table

We next examined the mood measure and found the same significant interaction, F(1, 50) = 4.13, p = .047,  = .076 (see Figure 2). Simple effect analyses revealed that future-belonging (M = 4.31, SD = .54) and future-alone conditions (M = 4.19, SD = .55) did not differ (p > .50); however, Cyberball exclusion (M = 3.80, SD = .58) did result in lower mood than did inclusion (M = 4.48, SD = .27, p = .001). Further, although inclusion conditions across paradigms did not differ, (p > .36), Cyberball exclusion led to marginally worsened mood relative to future-alone (p = .057).

Figure 2. Study 1: Mood as a function of Social Experience and Paradigm.

Display full size

The results of Study 1 strongly support our hypothesis. Exclusion in the Future-Life paradigm, contrary to what people may intuitively expect, resulted in no change in basic psychological needs or mood, whereas exclusion via Cyberball did. In fact, not only did Cyberball exclusion feel worse than Cyberball inclusion, it felt worse than Future-Life exclusion too, indicating this difference cannot be explained simply as an increase in needs or mood due to inclusion. These differences were true for both the overall composite measure as well as the individual basic needs. To further generalize and replicate our findings we conducted Study 2, which used a different measure of mood, commonly used in exclusion paradigms and particularly common with the use of the Future-Life manipulation. Our predictions, however, remained the same.

STUDY 2

Results

We again hypothesized that basic needs and mood should be lessened by exclusion via Cyberball as compared to inclusion, but that no differences should exist between future-alone and future-belonging conditions. Because of the results of Study 1, in which each individual basic need showed the same pattern as the composite measure, we formed a composite of the needs again in Study 2 (α = .95) and only performed analyses using this measure. We conducted a 2 (Social Experience: Inclusion or Exclusion) × 2 (Paradigm: Cyberball or Future-Life) between-participants ANOVA on the composite basic needs score, which revealed the predicted interaction, F(1, 47) = 4.81, p = .033,  = .093. Simple effect analyses revealed that, while future-belonging (M = 5.7, SD = .94) and future-alone conditions (M = 5.7, SD = .99) did not differ (p > .99), Cyberball exclusion (M = 3.92, SD = 1.61) did result in less satisfaction of the basic needs compared to inclusion (M = 5.35, SD = .98, p = .005). Further, although inclusion conditions did not differ, (p > .45), exclusion conditions did (p < .001, see Figure 3).

Figure 3. Study 2: Composite basic needs score as a function of Social Experience and Paradigm.

Display full size

We next conducted a 2 (Social Experience: Inclusion or Exclusion) × 2 (Paradigm: Cyberball or Future-Life) between-participants ANOVA on the composite PANAS score (α = .94), which again revealed the predicted interaction, F(1, 47) = 4.04, p = .05,  = .079 (see Figure 4). Simple effect analyses revealed that though future-belonging (M = 5.5, SD = .60) and future-alone conditions (M = 5.61, SD = .90) did not differ (p > .73), Cyberball exclusion (M = 4.06, SD = 1.19) did result in decreased mood compared to inclusion (M = 4.98, SD = .92, p = .02). Further, although inclusion conditions did not differ, (p > .16), exclusion conditions did (p < .001).Footnote1

Figure 4. Study 2: Mood as a function of Social Experience and Paradigm.

Display full size

GENERAL DISCUSSION

Social psychologists have shown a surge of interest in social exclusion research during the past decade. Although many of the consequences of social exclusion appear in conflict as to whether individuals become pro-social or anti-social following such experiences, the perhaps most surprising conflicting consequence is the different effects social exclusion has on emotional distress. The current work shows, in two studies, that the type of social exclusion experience matters a great deal and may, in part, explain why some have concluded that exclusion causes emotional distress, whereas others have concluded that it does not; namely, Cyberball exclusion lowers the satisfaction of basic needs as well as mood. Future-Life exclusion, though, appears to have no impact on any of these constructs.

The Future-Life and Cyberball paradigms differ in a variety of ways, any one of which might explain why exclusion in the former results in emotional numbing whereas exclusion in the latter results in emotional distress. One way in which these exclusion paradigms differ appears to be in the amount of “social damage” they inflict. In our prior work (Bernstein & Claypool, 2012), we showed that highly severe social injuries (ones that cause or imply great degrees of social disconnection and isolation) resulted in physical pain numbing, whereas those that are less severe (ones that cause or imply mild amounts of social disconnection) triggered physical pain hypersensitivity. We also showed in that same work that the future-alone feedback is perceived as a more severe social injury than is Cyberball exclusion. Given the relation between physical and social pain (e.g., Eisenberger & Lieberman, Citation2005; MacDonald et al., Citation2005; MacDonald & Leary, 2005), we suggested in the Bernstein and Claypool (2012) work that highly severe social injuries (as we assert future-alone exclusion to be) should result in social pain numbing, whereas a less-severe social injury (as we assert Cyberball exclusion to be) should result in increased social pain. The current findings are consistent with this perspective.

We should note, however, that while we believe our current data support the hypothesis that Cyberball exclusion elicits social pain whereas Future-Life exclusion does not, what we actually measured in the current work is likely best characterized as emotional distress, rather than social pain per se. According to MacDonald and Leary (Citation2005) social pain is “a specific emotional reaction to the perception that one is being excluded from desired relationships or being devalued by desired relationship partners or groups” (p. 202). Such pain might best and most directly be tapped by assessing “the deep aching of homesickness, grief, abandonment, or longing for a loved one” or feelings of “relational devaluation… known as hurt feelings” (p. 202). Indeed, Chen, Williams, Fitness, and Newton (Citation2008) assessed social pain directly in their work by asking participants to report their responses on a “no pain” to “intense pain” scale (called the “pain slide” measure). Instead we measured mood and satisfaction of the basic needs (of belongingness, control, self-esteem, and meaningful existence). These measures may be correlated with feelings of social pain, but we caution against using these interchangeably with bona fide social pain indices. Some of the basic needs, in particular, may be tapping something related to, but conceptually different from, social pain. For example, being far from a loved one would certainly elicit social pain (as social distance is a trigger for social pain responses; see MacDonald & Leary, Citation2005). Nonetheless, it may not always elicit a decrease in all basic needs, perhaps not impacting, for example, self-esteem or control. Thus factors that elicit, numb, or moderate social pain responses may not do so identically on some of the measures we assessed here, which again are better characterized as emotional distress.

Moreover, it is important to acknowledge that, unlike our previous work, we did not directly manipulate the severity of the social injury presently, as we did in Bernstein and Claypool (2012). Thus, although we believe differences in social injury severity may explain why Cyberball exclusion elicits emotion distress and Future-Life exclusion does not, we must leave room for the possibility that other factors might also play a role. Regardless of the exact mechanism or mechanisms that gave rise to our observed effects, what appears clear is that Cyberball and the Future-Life paradigm operate differently and yield different outcomes. For both theoretical and methodological reasons, it is important to identify such differences across paradigms and better understand why they exist. Such an understanding may help researchers in designing their studies (i.e., for identifying which paradigm may be most effective under various circumstances) and may also reveal a more nuanced understanding of exclusion itself.

Another limitation of this work is that we examined only two paradigms: Cyberball and Future-Life. We choose these methods rather than the litany of other manipulations of social exclusion for a few reasons. First, prior work has shown them to differ on their perceived severity as well as to result in different outcomes for at least one dependent variable (i.e., physical pain sensitivity, see Bernstein & Claypool, 2012). Second, they are two of the most common manipulations of social exclusion in the literature, and some researchers have explicitly speculated that these manipulations (particularly Future-Life) may be creating very different exclusion experiences (e.g., Williams, Citation2007). We thus sought to examine these two paradigms in these studies, but we do wholeheartedly believe that other paradigms should be investigated in future research.

We also did not manipulate the timing of the emotional distress measures following the exclusionary experience. We measured only how people were feeling immediately following the episode. It is possible that more delayed effects following the reflexive response (Williams, Citation2007) to exclusion might have yielded different results. It is plausible, for example, that emotional distress may be high following Cyberall exclusion, which is a relatively low-severity social injury, whereas people's emotional distress may be numbed immediately following Future-Life exclusion, enabling them to deal with the current danger the high-severity exclusion poses (just as the body responds to physical injuries). If this is true, then there remains the interesting possibility that while emotional distress might decrease over time following Cyerball exclusion, the opposite may be true following Future-Life exclusion, with distress becoming more intense as time passes.Footnote2 Future research should examine the immediate and delayed responses to Future-Life exclusion as prior research has done with Cyberball (e.g., van Beest & Williams, Citation2006, Citation2011; van Beest, Williams, & van Dijk, Citation2011).

Another limitation is that this work examines only emotional and need-based responses following social exclusion. It would be very valuable to examine experimentally whether exclusion paradigm moderates outcomes such as aggression or pro-social behaviors. Perhaps it is the case that Cyberball exclusion results in more pro-social acts, whereas Future-Life exclusion might trigger more aggressive or anti-social acts, or vice versa. Future work should examine these behaviors as well as others, paying particular attention to any consequences of social exclusion previously examined that might have resulted because of their confounding with a particular paradigm.

Researchers should also examine possible dispositional moderators that may exacerbate this effects we observed. Research has previously shown that individuals who are intrinsically involved in their faith experience less well-being and behave less pro-socially following reminders of exclusion by God than individuals who were not intrinsically motived by their faith (van Beest & Williams, Citation2011). Presumably this is because, for individuals whom faith is important, exclusion by God is a particularly painful and distressing experience. Other work has shown that other individual differences, such as rejection sensitivity, play a key role in how people respond to social exclusion; individuals high in rejection sensitivity tend to respond most aggressively following social exclusion relative to those low in this individual difference (see, for example, Ayduk, Gyurak, & Luerssen, Citation2008). It is possible, then, than individuals high in such dispositional variables associated with intense reactions to social exclusion might actually elicit stronger reactions to Cyberball exclusion and possibly stronger reactions to Future-Life exclusion.

In conclusion, the current studies add to a literature that has been quick to claim that different manipulations of social exclusion may not have the same effect on individuals, but has not always offered the empirical evidence to support such an assertion. By examining different manipulations within the same study, and randomly assigning participants to paradigms, we can be certain that there is something about the paradigms themselves that causes different emotional reactions. Researchers should keep this in mind when conducting work in the future, and not simply assume that all exclusions behave the same. With this in mind, we may be able to better understand and predict how people respond to social exclusion and find solutions to mitigate its often deleterious consequences.