When do socially accepted people feel ostracized? Physical pain triggers social pain

Abstract

The present research investigated whether physical suffering would cause people to feel ostracized even when they are accepted by their social interaction partners. Participants were instructed to place their hands either into a circulated cold water bath (to induce physical pain) or into a water bath at room temperature while they were either included or ostracized during an online ball tossing game—Cyberball. We found that physical pain led people to experience social pain, while they are socially accepted during a social interaction. Our findings suggest that people with physical suffering may need extra attention in social interactions to satisfy their threatened social needs.

Keywords::

• physical pain
• social pain
• ostracism
• belonging

People often use words related to physical injuries (e.g., hurt, wounded) to describe emotional suffering caused by social exclusion. These metaphors are consistent with the Pain Overlap Theory (Eisenberger & Lieberman, 2004; MacDonald & Leary, 2005), which proposes that social pain and physical pain share neurological underpinnings (but see also Cacioppo et al., 2013; Iannetti, Salomons, Moayedi, Mouraux, & Davis, 2013 for recent debates about the theory). Recent empirical investigations have provided evidence to support the validity and applicability of the theory. In particular, research has shown that methods or variables that reduce one type of pain can reduce the other type of pain (pain regulation effect; Eisenberger & Lieberman, 2004). For example, drug studies conducted by DeWall et al. (2010) showed that participants who took an over-the-counter physical painkiller, Tylenol (acetaminophen), for three weeks demonstrated a significant reduction in daily hurt feelings and less activity in pain-related brain regions such as the dorsal anterior cingulate cortex (dACC), when they were socially excluded. In addition, social support that mitigates the distress of social pain (Teng & Chen, 2012) has been shown to reduce people's sensitivity to physical pain in a cold-pressor task (Brown, Sheffield, Leary, & Robinson, 2003) and heat-stimulation task (Master et al., 2009). Even the mere presence of a friend (Brown et al., 2003) or a partner's picture (Master et al., 2009) can reduce people's perceived physical pain.

Social pain and physical pain are similar not only in how people cope with them, but also in how they are detected and experienced. Recent research has suggested that social pain can lead to the experience of physical pain, even when people do not have physical injuries. For example, social exclusion activates the dACC and anterior insula, which are brain areas specifically involved in the processing of pain affect (Eisenberger, Lieberman, & Williams, 2003). Social exclusion can also activate brain regions that are involved in the sensory experience of physical pain (secondary somatosensory cortex and dorsal posterior insula; Kross, Berman, Mischel, Smith, & Wager, 2011). Would physical pain lead to social pain or hurtful feelings when people are not objectively ignored or excluded by others in social interaction? In the following, we review evidence suggesting such a possiblity.

Physical sufferings trigger feelings of exclusion

Past correlational studies that examine the interconnection between physical pain and social suffering suggest that physical pain may lead to social pain. For instance, persistent physical pain is associated with greater proneness to hurt feelings (MacDonald & Leary, 2005). Individuals sensitive to physical pain are also sensitive to social pain (Eisenberger, Jarcho, Lieberman, & Naliboff, 2006). Anxious attachment mediated the positive relationship between physical pain and depression and anxiety (MacDonald & Kingsbury, 2006). Recent experimental research (Riva, Wirth, & Williams, 2011) found that participants who relived a physically painful experience were more inclined to aggressive behavior (Study 1). More relevant to the current investigation, Riva et al. (2011) found that participants who experienced physical pain by placing their hands in cold water reported being ignored and excluded (Study 2). As thwarted social needs and aggression are typical responses after socially painful experiences (see Williams, 2007), these findings suggest that physical pain can induce hurt feelings in the absence of social interaction.

Whereas Riva et al. (2011) demonstrated that physical pain can trigger social pain in the absence of social interaction, what remained unclear was whether physical suffering influences hurtful feelings during potentially positive social acceptance. The current research sought to extend this previous research by examining the role of physical pain in triggering feelings of hurt within the context of a social interaction.

The need to gain social acceptance is a fundamental human motivation (Baumeister & Leary, 1995). During social interactions, people have a normative expectation to engage in reciprocal social interaction by receiving and offering social acceptance and inclusion. When this inclusionary norm is unexpectedly violated, people experience aversive emotional and behavioral reactions, which are immune to the influence to situational and dispositional factors (for a review, see Williams, 2009). In addition, although social exclusion would immediately induce physical pain (e.g., Eisenberger et al., 2003; Kross et al., 2011), it can also numb people to physical pain following the initial painful reaction (Bernstein & Claypool, 2012; Borsook & MacDonald, 2010; Chen, Williams, Fitness, & Newton, 2008; DeWall & Baumeister, 2006). Thus, we predicted that physical pain would not influence the hurt feelings of people who are excluded. More importantly, we examined whether physical pain may influence people's hurt feelings when they engage in normative social interactions, that is, when they experience social inclusion and acceptance by their interaction partners. We predicted that people who suffer from physical pain would feel socially hurt by reporting lower satisfactions in belonging, self-esteem, control, and meaningful existence when they are in fact (or objectively) accepted by others.

In addition, because people suffering from physical pain are more readily to perceive cues of social rejection (e.g., MacDonald & Leary, 2005), it is possible that people suffering from physical pain would underestimate the amount of acceptance they receive from others. We also explored whether this biased perception of social acceptance might mediate the effect of physical pain on people's feelings of social exclusion in a socially accepted interaction.

The present study

The present study tested the hypothesis that socially accepted people will feel excluded when they experience physical pain. During the experiment, participants experienced either physical pain or no pain. At the same time, they played an online ball-tossing game—Cyberball (Williams, Cheung, & Choi, 2000), in which they were either included or excluded. We tested whether physical pain condition would interact with social interaction to influence people's hurt feelings, with a focus on whether physical pain would lead participants who were included during the game to experience the feeling of exclusion. Also, we explored whether physical pain would bias the perceived number of tosses received during the game for participants in the inclusion condition, which would further account for the expected effect of physical pain on feelings of exclusion. Next, we reported all data exlcusions (if any), all manipulations, and all measurements.

Method

Participants and design

One hundred and thirty-eight undergraduate students1 (51 men; mean age = 19.61 years, SD = 1.88) from the University of Hong Kong participated in this study in exchange for a course credit or a small monetary reward. Participants were randomly assigned to a 2 (Cyberball: social inclusion vs. exclusion) × 2 (cold-pressor task: physical pain vs. no pain control) between-subject design.

Procedure and measures

Participants were tested individually, and the room temperature was set at 25°C. After providing the informed consent, participants put their non-dominant hand into a circulating cold water bath (3 ± 0.05°C) to measure their baseline physical pain threshold (i.e., the time they started to feel pain) and tolerance (i.e., the time they could not tolerate the pain). The baseline measure of physical pain threshold and pain tolerance were standardized and averaged to create a covariate in our data analyses reported below.2 Following the baseline measure, participants put their hands into a water bath at room temperature until they no longer felt any physical pain.

Participants proceeded to play a ball-tossing game—Cyberball (Williams et al., 2000). They were led to believe that they were playing the game via the Internet with two other players, who were in fact computer programmed. During the game, participants either received two tosses only at the beginning (exclusion condition) or received one-third of the total 30 tosses (inclusion condition). While playing the game, participants put their non-dominant hand into a water bath. By random assignment, the water was either cold (3 ± 0.05°C; physical pain condition) or at room temperature (no pain control condition). Participants in the physical pain condition could pull their hand out when the pain became unbearable and put it back into the water bath when the pain returned to an acceptable level. On average, participants spent around 3 min in the Cyberball game.

Next, participants rated two items (“I was ignored” and “I was rejected”) on a five-point scale (1, not at all; 5, extremely). The scores correlated strongly (r = .93) and were averaged to check the manipulation of social exclusion. They also estimated the percentage of tosses they received during the game. In addition, participants indicated the level of physical pain they experienced on a visual analog scale (0, no pain; 10, intense pain), to check the manipulation of physical pain.

Following the manipulation check questions, participants were asked to answer a series of questions in a randomized order to assess their feelings of belongingness, self-esteem, control, and meaningful existence during the ball tossing game, on a five-point scale (1, not at all; 5, extremely; e.g., Carter-Sowell, Chen, & Williams, 2008; see Williams, 2009 for the full scale). Furthermore, participants filled out the positive affect and negative affect schedule (PANAS; Watson, Clark, & Tellegen, 1988) on the same five-point scale. Finally, participants were thanked and debriefed.

Results

Manipulation checks

Cyberball experience

A 2 by 2 ANCOVA on perceived rejection/ignorance revealed only a significant main effect of the Cyberball experience, such that excluded participants reported being more rejected and ignored (M = 4.23, SD = 0.98) than included participants (M = 2.78, SD = 1.07), F(1, 133) = 68.51, p <  .001,  = .34. The main effect of physical pain manipulation, F(1, 133) = 0.81, p = .37,  = .01, and the interaction effect were not statistically significant, F(1, 133) = 2.53, p = .11,  = .02.

Another ANCOVA on perceived percentage of ball tosses received revealed only a significant main effect of the Cyberball experience, such that excluded participants (M = 8.94, SD = 10.30) reported that they received a lower percentage of ball tosses than included participants (M = 28.30, SD = 12.47), F(1, 133) = 109.24, p < .001,  = .45. The main effect of physical pain manipulation, F(1, 133) = 2.95, p = .09,  = .02, and the interaction effect were not statistically significant, F(1, 133) = 0.45, p = .50,  = .003. Therefore, the social exclusion manipulation was effective.

Cold-pressor experience

A 2 by 2 ANCOVA on post-manipulation physical pain level revealed only a significant main effect of the cold-pressor experience, such that participants in the physical pain condition (M = 7.45, SD = 2.32) experienced higher levels of physical pain than participants in the control condition (M = 0.77, SD = 1.35), F(1, 133) = 444.78, p < .001,  = .77. The main effect of social exclusion manipulation, F(1, 133) = 0.20, p = .66,  = .001, and the interaction effect were not statistically significant, F(1, 133) = 3.00, p = .09,  = .02. Therefore, the physical pain manipulation was effective.

Four fundamental needs

Did physical pain cause socially accepted participants to experience distress of ostracism? A 2 by 2 ANCOVA on perceived fundamental needs revealed a significant interaction effect, F(1, 133) = 10.35, p = .002,  = .07 (see Figure 1).3 The main effect of the Cyberball experience was also significant, F(1, 133) = 86.67, p < .001,  = .40. However, no main effect of the cold-pressor task experience was found, F(1, 133) = 1.05, p = .31,  = .01.

Figure 1 Participants' four fundamental needs as a function of physical pain condition and exclusion condition. Error bars represent ± 1 standard error.

When participants experienced social acceptance, physical pain (M = 2.83, SD = 0.39) caused them to report lower feelings of fundamental needs than participants who did not experience physical pain (M = 3.25, SD = 0.67), F(1, 133) = 9.86, p < .01. However, when they were excluded, participants in the physical pain condition (M = 2.25,  SD = 0.60) and the physical control condition (M = 2.04, SD = 0.59) did not differ in their feelings of fundamental needs, F(1, 133) = 2.47, p = .12.

When participants experienced physical pain, social exclusion caused them to report lower feelings of fundamental needs than participants who experienced social acceptance, F(1, 133) = 78.55, p < .001. Similarly, when participants did not experience physical pain, social exclusion also caused them to report lower feelings of fundamental needs than participants who experienced social acceptance, F(1, 133) = 18.80, p < .001. Thus, physical pain increased feelings of social exclusion among participants who experienced objective social acceptance (as indicated by their satisfactions in the four fundamental social needs), whereas physical pain did not increase feelings of social exclusion among participants who had experienced actual social exclusion.

Similar analyses were performed on each of the four fundamental needs, which were commonly used in Cyberball research (cf. Williams, 2009; see Table 1 for the descriptive and inferential statistics). The interaction effect of social exclusion and physical pain manipulation was significant on all the four needs (i.e., belonging, self-esteem, control, and existence), F's(1, 133)> = 3.90, p's < = .05, s> = .03. The main effects of exclusion versus inclusion were significant on all the four needs, F's(1, 133)>13.70, p's < .001, s>.09. The main effects of physical pain manipulation on the four needs were not significant, F's(1, 133) < = 3.51, p's>.06, s < .03.

Table 1 The descriptive and inferential statistics for the dependent measures.

	Exclusion		Inclusion
Interaction	Physical pain	No pain	Physical pain	No pain
Pain threshold (in seconds)F(1, 134) = 0.42, p = .52	10.76 (6.74)	11.27 (10.33)	8.08 (3.97)	10.35 (9.08)
Pain tolerance (in seconds)F(1, 134) = 0.07, p = .93	33.94 (29.68)	34.79 (39.20)	29.54 (28.59)	31.34 (35.11)
Manipulation check F(1, 134) = 2.59, p = .11	4.17 (0.99)	4.29 (0.99)	3.00 (0.87)	2.56 (1.21)
Percentage of ball tosses F(1, 134) = 0.56, p = .46	11.11 (13.55)	6.71 (4.42)	29.06 (14.21)	27.56 (10.68)
Belonging (α = .85) F(1, 134) = 11.54, p = .001	2.05 (0.75)	1.88 (0.58)	2.96 (0.60)	3.59 (0.81)
Self-esteem (α = .69)F(1, 134) = 11.84, p = .001	2.60 (0.55)	2.28 (0.75)	2.61 (0.56)	3.06 (0.72)
Control (α = .77)F(1, 134) = 4.22, p = .04	1.90 (0.64)	1.76 (0.63)	2.64 (0.64)	2.97 (0.76)
Existence (α = .79)F(1, 134) = 4.04, p < .05	2.43 (0.83)	2.22 (0.74)	3.08 (0.51)	3.38 (0.83)
Four needs (α = .92)F(1, 134) = 10.60, p = .001	2.25 (0.60)	2.04 (0.59)	2.83 (0.39)	3.25 (0.67)
Positive affect (α = .89)F(1, 134) = .012, p = .91	2.32 (0.67)	2.16 (0.94)	2.30 (0.74)	2.11 (0.66)
Negative affect (α = .90)F(1, 134) = 3.02, p = .09	1.56 (0.51)	1.81 (0.72)	1.75 (0.78)	1.58 (0.76)

Note: All the interactions, means, and standard deviations (in parentheses) reported in the table are without covariate and unadjusted.

It should be also noted that no main effects nor interaction effect on the PANAS was found, F's(1, 133) < 2.83, p's>.09, s < .02.4

Discussion

Social pain can trigger physical pain (e.g., Eisenberger et al., 2003; Kross et al., 2011). However, whether physical pain causes social pain in a socially accepted situation has not been explored. Shared neurological overlap between social and physical pain systems led us to predict that physical suffering may induce feelings of exclusion even when people experience objective social acceptance (Eisenberger & Lieberman, 2004; MacDonald & Leary, 2005). In addition, we explored a mechanism to this effect. Because physical suffering demands attention and makes people perceive signals of rejection more readily (e.g., MacDonald & Leary, 2005), we further predicted that physical suffering may make people underestimate the amount of acceptance they received, which may account for the effect of physical pain on feeling of exclusion.

The findings of the present study supported part of our predictions. During the experiment, participants were either included or excluded during an online ball-tossing game with two confederates while they experienced either physical pain or no pain. We found that participants who experienced physical pain during the interaction, relative to those not experiencing physical pain, felt more socially excluded when they were included during the game because they reported lower satisfactions in belonging, self-esteem, control, and meaningful existence. However, we did not find that socially included participants in the physical pain condition underestimated the percentage of ball tosses they received during the interaction game. Therefore, there was no statistical evidence that biased perception of social inclusionary signals could account for the observed effect.

Whereas Riva et al. (2011) found that physical pain can induce the feelings of social exclusion in the absence of social interaction, the current investigation extend this finding by showing that physical suffering can lead socially accepted people to feel excluded. We found that when people with physical suffering are accepted by others in social interaction, their need for belonging, self-esteem, control, and meaningful existence is still threatened. These findings suggest people who suffer from physical pain would still experience less satisfaction in belonging, self-esteem, control, and meaningful existence even when they experience mild social acceptance. Therefore, people with physical suffering may need extra care and attention from others to feel socially connected, which may help them cope with the pain. Future research may also examine whether prolonged social acceptance from others can offset the negative impacts of physical pain on one's four fundamental needs.

These findings provide novel evidence in support of Pain Overlap Theory (Eisenberger & Lieberman, 2004; MacDonald & Leary, 2005) by showing that the experience of physical pain can thwart people's satisfactions in basic social needs, which are typical responses following social pain experience. However, it should be noted that recent work begins to question the validity and applicability of the theory. For example, a recent study with a larger sample size has failed to support the hypothesis that social pain activates the neural matrix identified in studies of physical pain (Cacioppo et al., 2013). Moreover, Chen et al. (2008) found that social pain could be re-experienced more easily and readily than physical pain, which further highlighted the differences between the two types of pain. Future research may examine when the two types of pain overlap and when they differ.

Limitations and future directions

This investigation in general supported our hypotheses about the effect of physical suffering on feelings of exclusion, as indicated by participants' satisfactions on the four fundamental social needs. Nevertheless, there were some limitations, which may inform future research.

First, we showed that physical pain triggers feeling of exclusion when participants were included during the ball-tossing game. However, when people are excluded during the game, participants in the physical pain condition and the no pain control condition did not differ on their feelings of exclusion. This finding on the exclusion condition could be due to a ceiling effect, such that participants in the exclusion condition already experienced high levels of exclusion, thus physical suffering could not further threaten their satisfactions of these social needs. This finding could also be attributed to the powerful impact of social exclusion. Evolutionarily speaking, social exclusion means “death” for social animals as it deprives of resources vital for survival (e.g., Gruter & Masters, 1986; MacDonald & Leary, 2005). Hence, social species should be alarmed to actual or potential threat of social exclusion. Previous studies consistently showed that the immediate distress caused by social exclusion or ostracism is immune to moderating effects of many individual differences and situational factors (see Williams, 2007, 2009 for reviews). Further research may compare the effect and mechanism of different levels of social exclusion. For example, experiencing partial social exclusion may reduce the possible ceiling effect of social exclusion observed in the current research and examine whether physical suffering would influence people's feelings of exclusion.

Second, we explored a mechanism to explain why socially accepted participants would feel rejected when they experienced physical pain. We predicted that participants who experienced physical pain would underestimate the social acceptance they received, which accounts for their thwarted social needs during social acceptance. As the present findings did not support this mediation model, there are likely other mechanisms through which physical pain and social exclusion affect feelings of social acceptance. Two possibilities, which are supported by prior work, are that socially included participants who experience physical pain may have a hostile cognitive bias (DeWall, Twenge, Gitter, & Baumeister, 2009) or increased entitlement (Poon, Chen, & DeWall, 2013) that may increase their propensity to feel lower levels of social needs satisfaction. Future research may further explore these possibilities.

Third, we used the cold-presser paradigm to induce feelings of physical pain. Although this paradigm is popular and well validated (see Mitchell, MacDonald, & Brodie, 2004), the feelings of physical pain induced were confounded with the feelings of physical coldness. Given that social exclusion can be linked with the feelings of coldness (Zhong & Leonardelli, 2008), future studies may try to use different paradigms to induce the feelings of physical pain to rule out method artifact.

Finally, we did not include a control condition in this study. Although previous studies have shown that people usually react similarly following social acceptance and other neutral experiences (e.g., Nordgren, Banas, & MacDonald, 2011; Poon et al., 2013), future studies may employ a neutral no-interaction condition to address this limitation.

Conclusion

Social exclusion hurts, but no prior work has explored whether physical pain makes socially accepted people feel excluded during a social interaction. Overall, the present findings offer novel support to the hypothesis that social and physical pain share neurological overlap (Eisenberger & Lieberman, 2004; MacDonald & Leary, 2005) by showing that physical pain would make socially accepted people to experience social pain.

Additional information

Funding

This research was supported by the Hong Kong Research Grants Council's GRF (grant number HKU742411H, to Chen and DeWall).

Notes

^1. The data were collected at two time points (phase 1: 53 participants; phase 2: 85 participants).

^2. The covariate only significantly influenced negative affect, F(1, 133) = 4.99, p = .03. The covariate did not influence other dependent measures, F's(1, 133) < 1.94, p's>.16. Moreover, removing the covariate in the analyses did not substantially alter the results. Because previous studies examining the potential interaction between physical and social pain tend to control for individuals' preexisting physical pain sensitivity (e.g., DeWall & Baumeister, 2006), we follow this practice by reporting our findings by controlling participants' preexisting pain sensitivity.

^3. As we collected the data in the separate phases, we conducted a 2 × 2 × 2 ANOVA to examine whether phase would interact and affect our results reported in the main text. No three-way interactions in all of the measures were statistically significant, F's(1, 130) < 1.94, p's>.16. Therefore, we reported the combined data in the main text for the sake of charity. Nevertheless, for the data of phase 1, we got a significant interaction effect on belonging, but the interaction effects on the other three needs were not significant. Moreover, we got a significant interaction effect on perceived percentage of ball tosses received. Perceived percentage of ball tosses received also mediated the interactive effect of physical and social pain manipulation on belonging. However, the interaction effect on percentage of ball tosses received and the mediation effect was no longer significant when we combined the data from two phases.

^4. The interaction effect on the overall four needs was still significant when both positive affect and negative affect were controlled.