ORIGINAL ARTICLE
Physical pain increases interpersonal trust in females C. Wang1, J. Gao1, Y. Ma2, C. Zhu1, X-W. Dong1,3 1 Key laboratory of Brain Functional Genomics, MOE & STCSM, Institute of Cognitive Neuroscience, School of Psychology and Cognitive Science, East China Normal University, Shanghai, China 2 State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing, China 3 NYU-ECNU Institute of Brain and Cognitive Science at NYU Shanghai and Collaborative Innovation Center for Brain Science, Shanghai, China
Correspondence Xiao-Wei Dong E-mail: [email protected] Funding sources This study was supported by grants from the National Natural Science Foundation of China (31600890) and China Postdoctoral Science Foundation (2015M571516). Conflicts of interest None declared.
Accepted for publication 13 August 2017 doi:10.1002/ejp.1111
Abstract Background: People behave and interact with others differently when experiencing physical pain. Pain has dramatic effects on one’s emotional responses, cognitive functions and social interaction. However, little has been known about whether and how physical pain influences interpersonal trust in social interaction. In the present study, we examined the influence of physical pain on trusting behaviour. Methods: A total of 112 healthy participants were recruited and assigned to physical pain condition (induced by Capsaicin) and control condition (with hand cream), respectively. Thirty minutes after pain induction, three decision-making tasks were conducted to measure behaviours in social interaction, including trust and trustworthiness (trust game), nonsocial risk-taking (risk game) and altruism (dictator game). Results: Results showed that physical pain increased interpersonal trust among females, but not among males. Pain did not influence non-social risk-taking, altruism or trustworthiness, as evaluated by monetary transfers in those tasks. Moreover, the effect of physical pain on interpersonal trust was fully mediated by expectation of monetary profit. Conclusions: These findings demonstrate an effect of pain on interpersonal trust and suggest a reciprocity mechanism that the effect may be driven by self-interest rather than altruistic motivation. The pain effect on trust was evident only in females, implying distinct pain coping strategies used by both genders. Significance: The present work highlights the social component of pain and extends our understanding of mutual interactions between pain and social cognition.
1. Introduction Pain, as an unpleasant sensory and emotional experience, considerably modulates one’s emotional responses and cognitive functions. It also exerts significant impact on social cognition and behaviours (Williams and Craig, 2016). In a biopsychosocial model of pain (Gatchel et al., 2007), pain and social cognition are suggested to mutually interact with each other. In support of this model, previous studies have shown that pain perception is modulated by © 2017 European Pain Federation - EFICâ
social factors, such as social support (L opez-Martınez et al., 2008) and cultural mindsets (Wang et al., 2014). Conversely, physical pain can also affect social interactions. Physical pain has been found to promote prosocial behaviour. Prosocial behaviour is a social behaviour that benefits other people or society as a whole, such as helping, sharing, donating, cooperating, etc. (Batson and Powell, 2003). A recent study reported that shared pain experience promoted social bonding and enhanced cooperation with strangers (Bastian et al., Eur J Pain
(2017)
–
1
Pain increases trust in females
2014). The enhancement of group salience was thought to be a main factor attributed to these observed changes in social interaction. However, interpersonal trust has also long been recognized as another important element in constituting the foundation of cooperative behaviours (McAllister, 1995; Jones and George, 1998). Thus, physical pain might promote cooperation by increasing trust in others. Meanwhile, studying the influence of physical pain on interpersonal trust may facilitate our understanding of the underlying mechanism for pain mediated effects on social cognition. Different theories have been proposed to explain the motives of human prosocial behaviour, including the reciprocity theory and empathy-altruism theory. The key difference of these two theories is whether the prosocial behaviour involves a component of the motive of self-benefit (Batson and Powell, 2003; Hein et al., 2016). These two theories hypothesize different mechanisms underlying the effects of physical pain on trusting behaviours. From a reciprocity perspective, it could be hypothesized that individuals in physical pain choose to trust others, with the consideration that the cooperative behaviour will eventually benefit themselves. Indeed, it was found that pain would shift one’s attention and drive individuals to be more self-focused (Pincus and Morley, 2001). Alternatively, from an empathy-altruism perspective, it could be hypothesized that individuals in physical pain tend to trust others purely for altruistic reasons, that is, to benefit others regardless of what they will gain. Gender differences in pain perception and trusting behaviour have been well-documented. First, females were more sensitive to pain and they often exhibited a higher degree of anxiety than males (Berkley, 1997; Armstrong and Khawaja, 2002). In addition, it was reported that men were rated as less willing to report pain than women (Robinson et al., 2001). Thus, the motivation to deal with noxious stimuli and pain coping strategies could differ between them. As a result, women were more likely than men to use emotion-focused techniques and social strategies including social support (Fillingim, 2000; Bartley and Fillingim, 2013). Second, females were found to display higher scores of personality trust than males (Feingold, 1994). Females and males also manifested different preference in trusting behaviour. Thus, males would trust others based on shared group memberships; whereas females would trust those who shared direct or indirect relationship connections (Maddux and Brewer, 2005). Therefore, we hypothesize a gender difference that females are 2 Eur J Pain (2017) –
C. Wang et al.
more prone than males to show interpersonal trust when under the influence of physical pain. The current study aims to test whether physical pain promotes interpersonal trust and to further explore the underlying mechanism by examining the empathy-altruism hypothesis and the reciprocity hypothesis. In addition, the gender difference of the pain effect on trust is to be examined. In our experiment, physical pain was induced in participants by applying Capsaicin on one of the forearms. We were particularly interested in the trusting behaviour in the person in pain, but not in the observer witnessing others in pain. Three decision-making tasks were employed, including trust game, dictator game and risk game, to measure trust (as well as trustworthiness), altruism and risk-taking tendency, respectively. The present study complements the biopsychosocial model of pain (Gatchel et al., 2007) and may provide important information for physical pain intervention.
2. Methods 2.1 Participants In the experiment, 112 university students (28 groups of 4 people of the same gender) in Shanghai participated in the study as paid volunteers. Participants were recruited through advertisements placed on the internet. They were informed to be paid 25 Yuan plus the gains from the monetary decisionmaking tasks during the experiment. Usually, they would get 35–50 Yuan (equivalent to US $5–$7.20). A total of 105 participants were used for formal data analyses (53 females, 52 males, mean age = 21.5 years, SD = 2.3), with seven participants being excluded due to failure of pain manipulation. The exclusion criterion was that subjective pain rating was lower than 3 out of 10 in painful condition (five participants) or higher than 3 of 10 in control condition (two participants). The current sample size was pre-determined using a statistic estimation (VanVoorhis and Morgan, 2007) based on an assumed medium effect size of 0.4 and a power of 0.8. Ethnicity and age were controlled for in the four sub-groups separated by gender and treatment in the study. All participants self-reported no psychiatric illness or chronic pain disorders. This study was approved by the University Committee on Human Research Protection of East China Normal University and was carried out in accordance with the approved guidelines. Written informed consent was obtained prior to participation. © 2017 European Pain Federation - EFICâ
Pain increases trust in females
C. Wang et al.
2.2 Pain induction and assessment The heat/capsaicin sensitization model (Modir and Wallace, 2010) was adopted in our study to generate stable, long-lasting pain sensation. It was a safe and noninvasive paradigm. In the painful treatment, Capzasin-HP cream (Capsaicin 0.1%) was brushed to a 2 9 2 cm area on the volar side of the dominant forearm and then was covered with plastic film to insure skin contact and prevent evaporation. In the non-painful treatment, hand cream was administrated to the same area. Both Capsaicin and hand cream are white and odorless, which cannot be distinguished by participants at the very beginning. Capsaicin or hand cream was applied for 60 min. Pain sensation was assessed using a subjective numerical pain rating with visual analog scale (VAS; scale of 0–10, with 0 corresponding to ‘no pain at all’ and 10 corresponding to the ‘worst imaginable pain’; Carlsson, 1983; Huang et al., 2013). At the end of the experiment, the skin was cleaned with soapy water after Capsaicin cream was wiped off with tissue paper.
2.3 Experimental paradigm Four participants of the same gender who did not know each other came to the laboratory at a time. After filling out a few questionnaires, they were divided into two competing groups by drawing lots (with card labelled White Team-player 1, White Team-player 2, Black Team-player 1 and Black Teamplayer 2). Either Capsaicin or hand cream was then applied to each individual. Importantly, we made sure that participants did not know the treatments of each other as so to avoid the effects of shared painful experience or empathic responses to others’ pain on social interaction. In reality, the player-1 of each team was always assigned to the painful condition and the player 2 was assigned to the control condition. After pain induction, participants interacted with each other for 25 min by playing a popular poker game. The game allowed players to cooperate with partners and compete with rivals. This session was referred as group formation, which aimed to provide real social interaction and to build up intergroup relationship. The perceived relationship was measured by subjective rating of closeness which was illustrated by different levels of overlap between two circles (1 = ‘not close at all’, 6 = ‘highest closeness’). Successful group manipulation was confirmed by that individuals showed higher closeness to in-group members than to out-group members (3.49 1.04 vs. 2.80 1.23, t (104) = 6.69, p < 0.001).
© 2017 European Pain Federation - EFICâ
Thirty minutes after pain induction by when the Capsaicin would produce stable moderate painful feelings, monetary decision-making tasks were conducted. This session lasted for 30 min and consisted of three typical economic games (Kosfeld et al., 2005; Cameron et al., 2013) successively: the dictator game (DG), the trust game (TG), and the risk game (RG). In this sequence, the tasks became more difficult and the interference from the previous task on the latter ones would be reduced to a minimum. To prevent any verbal communication or eye contact during decision making, participants were asked to sit back to back. Before the experiment, participants were told that all the decisions they made would be taken into account for the final payment.
2.4 Decision-making tasks 2.4.1 Dictator game The dictator game is designed to measure the degree of altruism of the participants. The game has two roles, a dictator and a receiver. The dictator is first endowed with 20 Yuan (the basic unit of the Chinese currency), and then must decide his/her endowment (0 ≤ endowment ≤ 20) given to the receiver. The receiver just takes all the money with no need to pay any amount of money back to the dictator. Altruism, is indexed as the money endowment by the dictator. Each participant played two rounds, always being assigned as a dictator, while the receiver would either be an in-group member or an out-group member. 2.4.2 Trust game The trust game measures the extent to which participants trust others and the degree of their trustworthiness. The game has two roles, an investor and a trustee. The investor is first endowed with 20 yuan, and then must decide the endowment (0 ≤ endowment ≤ 20) given to the trustee. The endowment will be tripled by the experimenter and the trustee has the opportunity to return some amount of money (0 ≤ money return ≤ 3*endowment) back to the investor. If the investor trusts and believes that the trustee will return considerable amount of money back, he/she will endow as much money as possible. Trust, is indexed as the money endowment by the investor and trustworthiness, is indexed as the ratio of money return by the trustee to what they have received. After endowment, the investor is asked to estimate how much money he/she will get
Eur J Pain
(2017)
–
3
Pain increases trust in females
at the end of a round. Expectation (of money profit), is indexed as the estimation of the total of monetary gain. All the participants played four rounds: the first two rounds as an investor and the last two rounds as a trustee. Each of the two rounds, the partner would either be an in-group member or an outgroup member. 2.4.3 Risk game The risk game is treated as a controlled non-social trust game which measures the extent to which participants take risk on investment. The game has two roles, an investor and a box. The rule is similar with the trust game, except that a box is used instead of the trustee. The box contains 11 balls labelled 0%, 10%, 20% to 100%, denoting the percentage of money back to the investor. After tripling the endowment, participant randomly picks up a ball in the box, and then is returned with the certain amount of money (0 ≤ money return ≤ 3*endowment). Risk-taking, is indexed as the money endowment by the investor. Each participant played the game once as an investor.
C. Wang et al.
dictator game, and risk-taking in the risk game. With a hypothesis-driven data analysis strategy, these dependent variables were tested by several univariate or repeated measures ANOVAs. In the ANOVAs, between-subject variables Treatment (painful, nonpainful), Sex (female, male) and within-subject variable Relationship (in-group, out-group) were included. Paired t-tests were also conducted to directly compare the differences in dependent variables between the painful condition and the control condition. 2.6.2 Effect size estimation Effect size (Cumming, 2014) was calculated to estimate the power of pain effect on interpersonal trust. Original effect size was defined as the difference of interpersonal trust between the painful and control group. Standardized effect size was indexed by Mcontrol)/Scontrol; where Mpain Cohen’s d: d = (Mpain and Mcontrol were the means of interpersonal trust in the painful condition and the control condition, and Scontrol was the standard deviation of interpersonal trust in the control condition.
2.5 Questionnaires
2.6.3 Mediation analysis
Upon arrival, participants were asked to fill in a few questionnaires, including the Self-esteem Scale (SES; Rosenberg, 1965) and the Life Orientation Test Revised (LOT-R; Scheier et al., 1994). The Selfesteem Scale assessed 10 items with 4-point scale (1 = strongly disagree, 4 = strongly agree). The LOTR, which was modified into a Chinese version (a = 0.8) and assessed 11 items with 7-point scale (0 = strongly disagree, 6 = strongly agree), evaluated tendency of optimistic or pessimistic value. These questionnaires were selected based on the consideration that both self-esteem and optimism/pessimism may influence individuals’ behaviour in social interaction. In addition, right before decision-making session (25–30 min after pain treatment), participants completed the positive and negative affect schedule (PANAS; 5 point; Watson et al., 1988), which contained 10 items of positive affect and 10 items of negative affect with 5-point scale (1 = very slightly or not at all, 5 = extremely).
Mediation analyses were conducted to explore the underlying mechanisms of pain effects on trust. Mediation effect was tested by linear regression combined with bootstrap analysis, which was implemented by a SPSS INDIRECT Macro Syntax (Preacher and Hayes, 2008). In these analyses, expectation (of money profit), anxiety-related irritable feelings and the combination of the two would be entered into the regression model as mediators, respectively.
2.6 Data analysis 2.6.1 Analyses of variance (ANOVA) Four dependent variables were defined, i.e., trust and trustworthiness in the trust game, altruism in the 4 Eur J Pain (2017) –
3. Results 3.1 Manipulation check: induction of physical pain Physical pain was elicited in the participants by topical application of Capsaicin cream to their dominant forearms. Subjective pain ratings were assessed using visual analog scale (Table 1). To check whether Capsaicin induced considerable pain intensity in the experiment, an analysis of variance (ANOVA) of Treatment (pain, control) and Time (0, 25, and 60 min) was conducted. This resulted in significant main effects of Treatment (F (1, 103) = 1050, p < 0.001) and Time (F (2, 102) = 200, p < 0.001), as well as a significant Time * Treatment interaction © 2017 European Pain Federation - EFICâ
Pain increases trust in females
C. Wang et al.
Table 1 Subjective pain assessments in the control and painful conditions.
Control condition (n = 54) Painful condition (n = 51)
0 min
25 min
60 min
0.35 (0.68) 0.35 (1.11)
0.46 (0.57) 6.14 (1.95)
0.46 (0.82) 5.75 (1.78)
(F (2, 102) = 185, p < 0.001). Post hoc analyses showed that Capsaicin, but not hand cream, induced significant painful sensation 25 min after application compared to the onset of the treatment (Capsaicin: t (50) = 19.5, p < 0.001; hand cream: t (53) = 1.29, p = 0.200). The pain remained for approximately 60 min (Capsaicin: t (50) = 17.5, p < 0.001), which ensured a sustained painful state of the participants in the painful condition throughout the entire decision-making session. It was important that participants were blinded to the Capsaicin or hand cream treatment of others. To verify this, participants were asked to make guesses of the types of the treatment applied to other participants after the experiment. Each guess would be either true or false. A total of 97 subjects answered the questions and the mean accuracy was 53%, which was not significantly different from a 50% random chance (t (96) = 0.295, p > 0.25). The participants were all blinded to the treatment of both in-group members and out-group members (ps > 0.1).
3.2 Effects of physical pain on interpersonal trust To investigate the effect of experiencing physical pain on the interpersonal trust, trust was assessed by the amount of money transferred to the trustee in the trust game (Table 2). To avoid possible influence of closeness within in-group member on the pain effect, the trusting behaviour towards out-group member was first assessed by an ANOVA of Treatment (pain, control), Sex (female, male). It revealed significant main effect of Sex (F (1, 100) = 4.14, p = 0.044), suggesting that males had higher trust level than females. Although, the main effect of Treatment was not significant (F (1, 100) = 1.38, p = 0.243), the interaction of Treatment 9 Sex was found significant (F (1, 100) = 4.89, p = 0.029; Fig. 1A) when the individual differences in risk-taking tendency was controlled. It indicated a differential effect of pain on trust between males and females. Post hoc analyses confirmed that female subjects experiencing physical pain showed more trust in others than those in the control condition © 2017 European Pain Federation - EFICâ
Table 2 Money endowment in trust game and risk game. Trust game Condition Female Control (n = 27) Painful (n = 26) Male Control (n = 27) Painful (n = 25)
Out-group trust
In-group trust
Risk game Risk-taking
7.85 (3.7) 10.9 (6.2)
11.5 (3.6) 14.0 (5.1)
11.9 (4.5) 11.6 (3.8)
13.2 (6.0) 12.1 (5.1)
15.0 (5.3) 14.9 (4.5)
14.2 (5.3) 15.1 (4.9)
(t (51) = 2.17, p = 0.035; Effect Size, d = 0.82), whereas no effect of pain on trust was observed in males (p > 0.5). Furthermore, questionnaire measurements further clarified that the effect was not due to sample differences in personal traits, such as optimism or self-esteem (independent sample t-tests, ps > 0.25). A similar effect of physical pain on trust was also observed when the trusting behaviour towards ingroup member was assessed. Females in pain condition displayed higher level of trust towards in-group members compared to those in the control condition (t (51) = 2.11, p = 0.040; Effect Size, d = 0.69; Fig. 1A), a phenomenon which was not observed in males (p > 0.25). In addition, to evaluate the ingroup/out-group difference of pain effect, a threeway ANOVA of Treatment, Sex and Relationship (out-group, in-group) was conducted. The triple interaction was not statistically significant (F (1, 101) < 1, p > 0.25), indicating that the pain effect on trust was not differentiated between in-group and outgroup members. The main effect of Relationship was significant in both females (F (1, 51) = 43.7, p < 0.001) and males (F (1, 50) = 31.3, p < 0.001), suggesting that generally individuals showed more trust in in-group members than on out-group members.
3.3 Role of social interaction in the influence of pain on trust In the trust game, the investor took a risk of the uncertainty of the trustee’s behaviour. Thus, the observed increase in the interpersonal trust under pain condition could be due to the enhancement of risk-taking tendency. To test this possibility, risk-taking tendency, measured by the amount of money that participants invested in the risk game, was entered into ANOVA of Treatment and Sex (Table 2). It revealed significant main effect of Sex (F (1, 101) = 10.4, p = 0.002), suggesting that males had higher risk-taking tendency than females. Eur J Pain
(2017)
–
5
Pain increases trust in females
C. Wang et al.
The results confirmed a gender difference of pain effect on trusting behaviour and also implied that females in physical pain may tend to trust others rather than rely on random fortune.
3.4 Mechanism underlying the effects of physical pain on interpersonal trust
Figure 1 Effects of physical pain on interpersonal trust and risk-taking tendency. (A) Bar charts represent the degrees of trust in either an out-group member or an in-group member in the Trust Game (TG) indexed by money endowment (unit: RMB, range: 0–20 Yuan). (B) Bar charts represent the degrees of trust in individuals in Trust Game (TG) or the degrees of trust in randomness (risk-taking behaviour) in Risk Game (RG) indexed by money endowment (unit: RMB, range: 0–20 Yuan). Participants are categorized by gender and treatment (pain, control). Error bars denote standard error. *denotes p < 0.05.
However, main effect of Treatment was not significant (F (1, 101) < 1, p > 0.5), nor did the interaction of Treatment 9 Sex (F (1, 101) < 1, p > 0.5). The results indicated that physical pain did not modify one’s risk-taking behaviour in a non-social situation. The amounts of money transferred in the risk game and the trust game were also compared to examine the contribution of social interaction in the observed pain effect on trust. The monetary transfers in these two games were subjected to an ANOVA with Treatment (pain, control), Sex (female, male) and Target (trust in randomness, trust in person). A triple interaction of Treatment 9 Sex 9 Target was significant (F (1, 101) = 4.68, p = 0.033; Fig. 1B). Separate analyses confirmed that, the interaction of Treatment 9 Target was significant in females (F (1, 51) = 6.99, p = 0.011), but not in males (F (1, 50) = 0.718, p > 0.250). Post-hoc analysis showed that, for females, physical pain increased behaviours of trusting on person (t (51) = 2.308, p = 0.024) but not on randomness (t (51) < 1, p > 0.5). For males, neither the trusting behaviour on person nor that on randomness was modulated by physical pain (ps > 0.5). 6 Eur J Pain (2017) –
We further tested two potential mechanisms underlying the effects of physical pain on interpersonal trust in the female sample, i.e., the reciprocity hypothesis and the empathy-altruism hypothesis. To test the empathy-altruism hypothesis, altruism was measured by the monetary transfers in the dictator game. Results showed that the amounts of money endowment were unaffected by pain condition in either the out-group condition (3.70 vs. 4.81; t (51) = 1.11, p = 0.273; Fig. 2A) or the in-group condition (7.52 vs. 8.69; t (51) = 1.10, p = 0.277). Meanwhile, we also measured the trustworthiness in the trust game. While physical pain increased interpersonal trust, it in opposite tended to decrease the trustworthiness towards both out-group member (0.50 vs. 0.43; t (51) = 1.42, p = 0.161; Fig. 2B) and in-group member (0.54 vs. 0.48; t (51) = 1.72, p = 0.091). Participants experiencing pain tended to return less money to the investors and kept more for themselves. These results indicated that the paininduced increase in interpersonal trust may not due to general increase in prosocial inclinations. To test the reciprocity hypothesis, we examined the role of expectation (of money profit) in the pain effect on interpersonal trust by measuring the amount of money one would expect to gain as an investor in the trust game. Considering that the pain effect on trust was not differentiated by relationship, the variable Relationship was collapsed and interpersonal trust was then evaluated by averaging the amounts of money transferred to in-group and outgroup members. First, it was found that individuals who expected higher profit tendered to trust more on others by transferring more money (r = 0.826, p < 0.001; Fig. 3A). Second, mediation analysis was then performed to test whether expectation of profit mediated the pain effect on trust. It revealed that the pain treatment was a significant predictor of expectation of profit (a path, B = 2.38, p = 0.047), as well as a significant predictor of interpersonal trust (c path, B = 2.79, p = 0.024). Importantly, the inclusion of expectation into the regression model predicting interpersonal trust from pain treatment resulted in an absence of a significant treatment effect (c’ path, B = 0.810, p = 0.279; see Table 3 for © 2017 European Pain Federation - EFICâ
Pain increases trust in females
C. Wang et al.
Figure 2 Effects of physical pain on altruism and trustworthiness. (A) Bar charts represent altruistic behaviour to in-group/out-group members in Dictator Game (DG), indexed by money endowment (unit: RMB, range: 0–20 Yuan). (B) Bar charts represent trustworthiness to in-group/out-group members in Trust Game (TG), indexed by the ratio of money return to the amount of money they have received (range: 0–1). Participants are females and are categorized by treatment (pain, control). Error bars denote standard error.
statistic details; Fig. 3B). Third, a bootstrap resampling analysis (N = 5000) of the effect size showed that this mediation effect was different from zero with 95% confidence (R2 = 0.69; Effect Size = 1.98; CI (Confidence Intervals) = [0.04–4.03]). Furthermore, we also conducted mediation analyses with in-group/out-group trust being taken into account; however, this resulted in marginally significant effects. We speculated that this might be due to a large variation among participants and an averaged in-group/out-group effect may help to some extent overcome this problem. These results indicated that the expectation of profit fully mediated the effect of pain on trust, which supported a reciprocity mechanism that individuals exhibited more interpersonal trust for self-benefit when experiencing physical pain. In addition, physical pain significantly increased subjective rating of individuals’ anxiety-related irritable feeling (control: 1.59, pain: 2.69; t (51) = 3.37, p = 0.001) and moreover, such irritable feeling was positively correlated with interpersonal trust (r = 0.355, p = 0.009). The results implied that irritable feeling might play a role in the pain effect on trust. This was confirmed by mediation analysis that anxiety (irritable feeling) was a significant mediator in the pain effect (Fig. 3C; R2 = 0.17; Effect Size = 1.15, CI = [0.019–3.00]). Furthermore, when putting both expectation and anxiety (i.e., irritable feeling) into a regression model as two mediators, a robust indirect effect was observed (Fig. 3D; R2 = 0.69; Effect Size = 2.22, CI = [0.22–4.27]). However, when these paths were inspected individually, only the expectation path was significant (CI = [0.021–4.00], p < 0.001). These results suggested that the modulation of pain on one’s cognitive processes (i.e., expectation) and affective mental © 2017 European Pain Federation - EFICâ
states (i.e., irritable feeling) contributed to the paininduced increase of trust.
4. Discussion The present study sought to explore the influence of physical pain on prosocial behaviours in social interaction by examining the effects of experiencing physical pain on interpersonal trust. Our experiment revealed that, first, physical pain induced a differential effect on interpersonal trust between genders, in which increased trust was only observed among females but not males; second, the effect of physical pain on trust was specifically associated with social interaction, whereas it was not driven by risk-taking tendency in non-social situation; third, the changes in interpersonal trust among females did not result from the alteration of altruism or trustworthiness by pain; and fourth, the enhancement of monetary expectation by physical pain fully mediated the pain effect on trust. Our findings provide important information for the understanding of the mechanism underlying how physical pain increase interpersonal trust. Motivation has been found to play an important role in prosocial behaviour such as trust. The current study examined the relative involvements of the two motives, i.e., the reciprocity and empathy-based altruism (Batson and Powell, 2003), in pain-induced facilitation of interpersonal trust. If the increase in trust was due to the pain effect on altruism towards other, we would expect more money transfer in the dictator game in the painful relative to control condition. However, we did not see such a difference in the dictator game. On the other hand, we found that individuals in physical pain transferred more money to others with an increased expectation of monetary
Eur J Pain
(2017)
–
7
Pain increases trust in females
C. Wang et al.
Figure 3 The pain effects on trust were mediated by expectation and anxiety. (A) The correlation between expectation of money profit and trust level (money endowment in Trust Game). (B) The mediation effect of expectation of money profit between treatment (painful condition, control condition) and trust level (money endowment). (C) The mediation effect of anxiety (irritable feeling) between treatment and trust level. (D) The mediation effect of expectation of money profit and anxiety (irritable feeling) between treatment and trust level. The red arrows denote the direct effect and blue arrows denote the total effect. *,**,*** denote p < 0.05, p < 0.01, p < 0.001, respectively.
Table 3 Testing expectation as a mediator on the effect of pain on trust. Coeff. B
SE of B
t
P
c path: Total Effect of IV (Treatment) on DV (Trust) 2.79 1.20 2.33 0.024* a path: IV (Treatment) to Mediator (Expectation) 2.38 1.17 2.04 0.047* b path: Direct Effect of Mediator (Expectation) on DV (Trust) 0.833 0.085 9.77
Physical pain increases interpersonal trust in females C. Wang1, J. Gao1, Y. Ma2, C. Zhu1, X-W. Dong1,3 1 Key laboratory of Brain Functional Genomics, MOE & STCSM, Institute of Cognitive Neuroscience, School of Psychology and Cognitive Science, East China Normal University, Shanghai, China 2 State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing, China 3 NYU-ECNU Institute of Brain and Cognitive Science at NYU Shanghai and Collaborative Innovation Center for Brain Science, Shanghai, China
Correspondence Xiao-Wei Dong E-mail: [email protected] Funding sources This study was supported by grants from the National Natural Science Foundation of China (31600890) and China Postdoctoral Science Foundation (2015M571516). Conflicts of interest None declared.
Accepted for publication 13 August 2017 doi:10.1002/ejp.1111
Abstract Background: People behave and interact with others differently when experiencing physical pain. Pain has dramatic effects on one’s emotional responses, cognitive functions and social interaction. However, little has been known about whether and how physical pain influences interpersonal trust in social interaction. In the present study, we examined the influence of physical pain on trusting behaviour. Methods: A total of 112 healthy participants were recruited and assigned to physical pain condition (induced by Capsaicin) and control condition (with hand cream), respectively. Thirty minutes after pain induction, three decision-making tasks were conducted to measure behaviours in social interaction, including trust and trustworthiness (trust game), nonsocial risk-taking (risk game) and altruism (dictator game). Results: Results showed that physical pain increased interpersonal trust among females, but not among males. Pain did not influence non-social risk-taking, altruism or trustworthiness, as evaluated by monetary transfers in those tasks. Moreover, the effect of physical pain on interpersonal trust was fully mediated by expectation of monetary profit. Conclusions: These findings demonstrate an effect of pain on interpersonal trust and suggest a reciprocity mechanism that the effect may be driven by self-interest rather than altruistic motivation. The pain effect on trust was evident only in females, implying distinct pain coping strategies used by both genders. Significance: The present work highlights the social component of pain and extends our understanding of mutual interactions between pain and social cognition.
1. Introduction Pain, as an unpleasant sensory and emotional experience, considerably modulates one’s emotional responses and cognitive functions. It also exerts significant impact on social cognition and behaviours (Williams and Craig, 2016). In a biopsychosocial model of pain (Gatchel et al., 2007), pain and social cognition are suggested to mutually interact with each other. In support of this model, previous studies have shown that pain perception is modulated by © 2017 European Pain Federation - EFICâ
social factors, such as social support (L opez-Martınez et al., 2008) and cultural mindsets (Wang et al., 2014). Conversely, physical pain can also affect social interactions. Physical pain has been found to promote prosocial behaviour. Prosocial behaviour is a social behaviour that benefits other people or society as a whole, such as helping, sharing, donating, cooperating, etc. (Batson and Powell, 2003). A recent study reported that shared pain experience promoted social bonding and enhanced cooperation with strangers (Bastian et al., Eur J Pain
(2017)
–
1
Pain increases trust in females
2014). The enhancement of group salience was thought to be a main factor attributed to these observed changes in social interaction. However, interpersonal trust has also long been recognized as another important element in constituting the foundation of cooperative behaviours (McAllister, 1995; Jones and George, 1998). Thus, physical pain might promote cooperation by increasing trust in others. Meanwhile, studying the influence of physical pain on interpersonal trust may facilitate our understanding of the underlying mechanism for pain mediated effects on social cognition. Different theories have been proposed to explain the motives of human prosocial behaviour, including the reciprocity theory and empathy-altruism theory. The key difference of these two theories is whether the prosocial behaviour involves a component of the motive of self-benefit (Batson and Powell, 2003; Hein et al., 2016). These two theories hypothesize different mechanisms underlying the effects of physical pain on trusting behaviours. From a reciprocity perspective, it could be hypothesized that individuals in physical pain choose to trust others, with the consideration that the cooperative behaviour will eventually benefit themselves. Indeed, it was found that pain would shift one’s attention and drive individuals to be more self-focused (Pincus and Morley, 2001). Alternatively, from an empathy-altruism perspective, it could be hypothesized that individuals in physical pain tend to trust others purely for altruistic reasons, that is, to benefit others regardless of what they will gain. Gender differences in pain perception and trusting behaviour have been well-documented. First, females were more sensitive to pain and they often exhibited a higher degree of anxiety than males (Berkley, 1997; Armstrong and Khawaja, 2002). In addition, it was reported that men were rated as less willing to report pain than women (Robinson et al., 2001). Thus, the motivation to deal with noxious stimuli and pain coping strategies could differ between them. As a result, women were more likely than men to use emotion-focused techniques and social strategies including social support (Fillingim, 2000; Bartley and Fillingim, 2013). Second, females were found to display higher scores of personality trust than males (Feingold, 1994). Females and males also manifested different preference in trusting behaviour. Thus, males would trust others based on shared group memberships; whereas females would trust those who shared direct or indirect relationship connections (Maddux and Brewer, 2005). Therefore, we hypothesize a gender difference that females are 2 Eur J Pain (2017) –
C. Wang et al.
more prone than males to show interpersonal trust when under the influence of physical pain. The current study aims to test whether physical pain promotes interpersonal trust and to further explore the underlying mechanism by examining the empathy-altruism hypothesis and the reciprocity hypothesis. In addition, the gender difference of the pain effect on trust is to be examined. In our experiment, physical pain was induced in participants by applying Capsaicin on one of the forearms. We were particularly interested in the trusting behaviour in the person in pain, but not in the observer witnessing others in pain. Three decision-making tasks were employed, including trust game, dictator game and risk game, to measure trust (as well as trustworthiness), altruism and risk-taking tendency, respectively. The present study complements the biopsychosocial model of pain (Gatchel et al., 2007) and may provide important information for physical pain intervention.
2. Methods 2.1 Participants In the experiment, 112 university students (28 groups of 4 people of the same gender) in Shanghai participated in the study as paid volunteers. Participants were recruited through advertisements placed on the internet. They were informed to be paid 25 Yuan plus the gains from the monetary decisionmaking tasks during the experiment. Usually, they would get 35–50 Yuan (equivalent to US $5–$7.20). A total of 105 participants were used for formal data analyses (53 females, 52 males, mean age = 21.5 years, SD = 2.3), with seven participants being excluded due to failure of pain manipulation. The exclusion criterion was that subjective pain rating was lower than 3 out of 10 in painful condition (five participants) or higher than 3 of 10 in control condition (two participants). The current sample size was pre-determined using a statistic estimation (VanVoorhis and Morgan, 2007) based on an assumed medium effect size of 0.4 and a power of 0.8. Ethnicity and age were controlled for in the four sub-groups separated by gender and treatment in the study. All participants self-reported no psychiatric illness or chronic pain disorders. This study was approved by the University Committee on Human Research Protection of East China Normal University and was carried out in accordance with the approved guidelines. Written informed consent was obtained prior to participation. © 2017 European Pain Federation - EFICâ
Pain increases trust in females
C. Wang et al.
2.2 Pain induction and assessment The heat/capsaicin sensitization model (Modir and Wallace, 2010) was adopted in our study to generate stable, long-lasting pain sensation. It was a safe and noninvasive paradigm. In the painful treatment, Capzasin-HP cream (Capsaicin 0.1%) was brushed to a 2 9 2 cm area on the volar side of the dominant forearm and then was covered with plastic film to insure skin contact and prevent evaporation. In the non-painful treatment, hand cream was administrated to the same area. Both Capsaicin and hand cream are white and odorless, which cannot be distinguished by participants at the very beginning. Capsaicin or hand cream was applied for 60 min. Pain sensation was assessed using a subjective numerical pain rating with visual analog scale (VAS; scale of 0–10, with 0 corresponding to ‘no pain at all’ and 10 corresponding to the ‘worst imaginable pain’; Carlsson, 1983; Huang et al., 2013). At the end of the experiment, the skin was cleaned with soapy water after Capsaicin cream was wiped off with tissue paper.
2.3 Experimental paradigm Four participants of the same gender who did not know each other came to the laboratory at a time. After filling out a few questionnaires, they were divided into two competing groups by drawing lots (with card labelled White Team-player 1, White Team-player 2, Black Team-player 1 and Black Teamplayer 2). Either Capsaicin or hand cream was then applied to each individual. Importantly, we made sure that participants did not know the treatments of each other as so to avoid the effects of shared painful experience or empathic responses to others’ pain on social interaction. In reality, the player-1 of each team was always assigned to the painful condition and the player 2 was assigned to the control condition. After pain induction, participants interacted with each other for 25 min by playing a popular poker game. The game allowed players to cooperate with partners and compete with rivals. This session was referred as group formation, which aimed to provide real social interaction and to build up intergroup relationship. The perceived relationship was measured by subjective rating of closeness which was illustrated by different levels of overlap between two circles (1 = ‘not close at all’, 6 = ‘highest closeness’). Successful group manipulation was confirmed by that individuals showed higher closeness to in-group members than to out-group members (3.49 1.04 vs. 2.80 1.23, t (104) = 6.69, p < 0.001).
© 2017 European Pain Federation - EFICâ
Thirty minutes after pain induction by when the Capsaicin would produce stable moderate painful feelings, monetary decision-making tasks were conducted. This session lasted for 30 min and consisted of three typical economic games (Kosfeld et al., 2005; Cameron et al., 2013) successively: the dictator game (DG), the trust game (TG), and the risk game (RG). In this sequence, the tasks became more difficult and the interference from the previous task on the latter ones would be reduced to a minimum. To prevent any verbal communication or eye contact during decision making, participants were asked to sit back to back. Before the experiment, participants were told that all the decisions they made would be taken into account for the final payment.
2.4 Decision-making tasks 2.4.1 Dictator game The dictator game is designed to measure the degree of altruism of the participants. The game has two roles, a dictator and a receiver. The dictator is first endowed with 20 Yuan (the basic unit of the Chinese currency), and then must decide his/her endowment (0 ≤ endowment ≤ 20) given to the receiver. The receiver just takes all the money with no need to pay any amount of money back to the dictator. Altruism, is indexed as the money endowment by the dictator. Each participant played two rounds, always being assigned as a dictator, while the receiver would either be an in-group member or an out-group member. 2.4.2 Trust game The trust game measures the extent to which participants trust others and the degree of their trustworthiness. The game has two roles, an investor and a trustee. The investor is first endowed with 20 yuan, and then must decide the endowment (0 ≤ endowment ≤ 20) given to the trustee. The endowment will be tripled by the experimenter and the trustee has the opportunity to return some amount of money (0 ≤ money return ≤ 3*endowment) back to the investor. If the investor trusts and believes that the trustee will return considerable amount of money back, he/she will endow as much money as possible. Trust, is indexed as the money endowment by the investor and trustworthiness, is indexed as the ratio of money return by the trustee to what they have received. After endowment, the investor is asked to estimate how much money he/she will get
Eur J Pain
(2017)
–
3
Pain increases trust in females
at the end of a round. Expectation (of money profit), is indexed as the estimation of the total of monetary gain. All the participants played four rounds: the first two rounds as an investor and the last two rounds as a trustee. Each of the two rounds, the partner would either be an in-group member or an outgroup member. 2.4.3 Risk game The risk game is treated as a controlled non-social trust game which measures the extent to which participants take risk on investment. The game has two roles, an investor and a box. The rule is similar with the trust game, except that a box is used instead of the trustee. The box contains 11 balls labelled 0%, 10%, 20% to 100%, denoting the percentage of money back to the investor. After tripling the endowment, participant randomly picks up a ball in the box, and then is returned with the certain amount of money (0 ≤ money return ≤ 3*endowment). Risk-taking, is indexed as the money endowment by the investor. Each participant played the game once as an investor.
C. Wang et al.
dictator game, and risk-taking in the risk game. With a hypothesis-driven data analysis strategy, these dependent variables were tested by several univariate or repeated measures ANOVAs. In the ANOVAs, between-subject variables Treatment (painful, nonpainful), Sex (female, male) and within-subject variable Relationship (in-group, out-group) were included. Paired t-tests were also conducted to directly compare the differences in dependent variables between the painful condition and the control condition. 2.6.2 Effect size estimation Effect size (Cumming, 2014) was calculated to estimate the power of pain effect on interpersonal trust. Original effect size was defined as the difference of interpersonal trust between the painful and control group. Standardized effect size was indexed by Mcontrol)/Scontrol; where Mpain Cohen’s d: d = (Mpain and Mcontrol were the means of interpersonal trust in the painful condition and the control condition, and Scontrol was the standard deviation of interpersonal trust in the control condition.
2.5 Questionnaires
2.6.3 Mediation analysis
Upon arrival, participants were asked to fill in a few questionnaires, including the Self-esteem Scale (SES; Rosenberg, 1965) and the Life Orientation Test Revised (LOT-R; Scheier et al., 1994). The Selfesteem Scale assessed 10 items with 4-point scale (1 = strongly disagree, 4 = strongly agree). The LOTR, which was modified into a Chinese version (a = 0.8) and assessed 11 items with 7-point scale (0 = strongly disagree, 6 = strongly agree), evaluated tendency of optimistic or pessimistic value. These questionnaires were selected based on the consideration that both self-esteem and optimism/pessimism may influence individuals’ behaviour in social interaction. In addition, right before decision-making session (25–30 min after pain treatment), participants completed the positive and negative affect schedule (PANAS; 5 point; Watson et al., 1988), which contained 10 items of positive affect and 10 items of negative affect with 5-point scale (1 = very slightly or not at all, 5 = extremely).
Mediation analyses were conducted to explore the underlying mechanisms of pain effects on trust. Mediation effect was tested by linear regression combined with bootstrap analysis, which was implemented by a SPSS INDIRECT Macro Syntax (Preacher and Hayes, 2008). In these analyses, expectation (of money profit), anxiety-related irritable feelings and the combination of the two would be entered into the regression model as mediators, respectively.
2.6 Data analysis 2.6.1 Analyses of variance (ANOVA) Four dependent variables were defined, i.e., trust and trustworthiness in the trust game, altruism in the 4 Eur J Pain (2017) –
3. Results 3.1 Manipulation check: induction of physical pain Physical pain was elicited in the participants by topical application of Capsaicin cream to their dominant forearms. Subjective pain ratings were assessed using visual analog scale (Table 1). To check whether Capsaicin induced considerable pain intensity in the experiment, an analysis of variance (ANOVA) of Treatment (pain, control) and Time (0, 25, and 60 min) was conducted. This resulted in significant main effects of Treatment (F (1, 103) = 1050, p < 0.001) and Time (F (2, 102) = 200, p < 0.001), as well as a significant Time * Treatment interaction © 2017 European Pain Federation - EFICâ
Pain increases trust in females
C. Wang et al.
Table 1 Subjective pain assessments in the control and painful conditions.
Control condition (n = 54) Painful condition (n = 51)
0 min
25 min
60 min
0.35 (0.68) 0.35 (1.11)
0.46 (0.57) 6.14 (1.95)
0.46 (0.82) 5.75 (1.78)
(F (2, 102) = 185, p < 0.001). Post hoc analyses showed that Capsaicin, but not hand cream, induced significant painful sensation 25 min after application compared to the onset of the treatment (Capsaicin: t (50) = 19.5, p < 0.001; hand cream: t (53) = 1.29, p = 0.200). The pain remained for approximately 60 min (Capsaicin: t (50) = 17.5, p < 0.001), which ensured a sustained painful state of the participants in the painful condition throughout the entire decision-making session. It was important that participants were blinded to the Capsaicin or hand cream treatment of others. To verify this, participants were asked to make guesses of the types of the treatment applied to other participants after the experiment. Each guess would be either true or false. A total of 97 subjects answered the questions and the mean accuracy was 53%, which was not significantly different from a 50% random chance (t (96) = 0.295, p > 0.25). The participants were all blinded to the treatment of both in-group members and out-group members (ps > 0.1).
3.2 Effects of physical pain on interpersonal trust To investigate the effect of experiencing physical pain on the interpersonal trust, trust was assessed by the amount of money transferred to the trustee in the trust game (Table 2). To avoid possible influence of closeness within in-group member on the pain effect, the trusting behaviour towards out-group member was first assessed by an ANOVA of Treatment (pain, control), Sex (female, male). It revealed significant main effect of Sex (F (1, 100) = 4.14, p = 0.044), suggesting that males had higher trust level than females. Although, the main effect of Treatment was not significant (F (1, 100) = 1.38, p = 0.243), the interaction of Treatment 9 Sex was found significant (F (1, 100) = 4.89, p = 0.029; Fig. 1A) when the individual differences in risk-taking tendency was controlled. It indicated a differential effect of pain on trust between males and females. Post hoc analyses confirmed that female subjects experiencing physical pain showed more trust in others than those in the control condition © 2017 European Pain Federation - EFICâ
Table 2 Money endowment in trust game and risk game. Trust game Condition Female Control (n = 27) Painful (n = 26) Male Control (n = 27) Painful (n = 25)
Out-group trust
In-group trust
Risk game Risk-taking
7.85 (3.7) 10.9 (6.2)
11.5 (3.6) 14.0 (5.1)
11.9 (4.5) 11.6 (3.8)
13.2 (6.0) 12.1 (5.1)
15.0 (5.3) 14.9 (4.5)
14.2 (5.3) 15.1 (4.9)
(t (51) = 2.17, p = 0.035; Effect Size, d = 0.82), whereas no effect of pain on trust was observed in males (p > 0.5). Furthermore, questionnaire measurements further clarified that the effect was not due to sample differences in personal traits, such as optimism or self-esteem (independent sample t-tests, ps > 0.25). A similar effect of physical pain on trust was also observed when the trusting behaviour towards ingroup member was assessed. Females in pain condition displayed higher level of trust towards in-group members compared to those in the control condition (t (51) = 2.11, p = 0.040; Effect Size, d = 0.69; Fig. 1A), a phenomenon which was not observed in males (p > 0.25). In addition, to evaluate the ingroup/out-group difference of pain effect, a threeway ANOVA of Treatment, Sex and Relationship (out-group, in-group) was conducted. The triple interaction was not statistically significant (F (1, 101) < 1, p > 0.25), indicating that the pain effect on trust was not differentiated between in-group and outgroup members. The main effect of Relationship was significant in both females (F (1, 51) = 43.7, p < 0.001) and males (F (1, 50) = 31.3, p < 0.001), suggesting that generally individuals showed more trust in in-group members than on out-group members.
3.3 Role of social interaction in the influence of pain on trust In the trust game, the investor took a risk of the uncertainty of the trustee’s behaviour. Thus, the observed increase in the interpersonal trust under pain condition could be due to the enhancement of risk-taking tendency. To test this possibility, risk-taking tendency, measured by the amount of money that participants invested in the risk game, was entered into ANOVA of Treatment and Sex (Table 2). It revealed significant main effect of Sex (F (1, 101) = 10.4, p = 0.002), suggesting that males had higher risk-taking tendency than females. Eur J Pain
(2017)
–
5
Pain increases trust in females
C. Wang et al.
The results confirmed a gender difference of pain effect on trusting behaviour and also implied that females in physical pain may tend to trust others rather than rely on random fortune.
3.4 Mechanism underlying the effects of physical pain on interpersonal trust
Figure 1 Effects of physical pain on interpersonal trust and risk-taking tendency. (A) Bar charts represent the degrees of trust in either an out-group member or an in-group member in the Trust Game (TG) indexed by money endowment (unit: RMB, range: 0–20 Yuan). (B) Bar charts represent the degrees of trust in individuals in Trust Game (TG) or the degrees of trust in randomness (risk-taking behaviour) in Risk Game (RG) indexed by money endowment (unit: RMB, range: 0–20 Yuan). Participants are categorized by gender and treatment (pain, control). Error bars denote standard error. *denotes p < 0.05.
However, main effect of Treatment was not significant (F (1, 101) < 1, p > 0.5), nor did the interaction of Treatment 9 Sex (F (1, 101) < 1, p > 0.5). The results indicated that physical pain did not modify one’s risk-taking behaviour in a non-social situation. The amounts of money transferred in the risk game and the trust game were also compared to examine the contribution of social interaction in the observed pain effect on trust. The monetary transfers in these two games were subjected to an ANOVA with Treatment (pain, control), Sex (female, male) and Target (trust in randomness, trust in person). A triple interaction of Treatment 9 Sex 9 Target was significant (F (1, 101) = 4.68, p = 0.033; Fig. 1B). Separate analyses confirmed that, the interaction of Treatment 9 Target was significant in females (F (1, 51) = 6.99, p = 0.011), but not in males (F (1, 50) = 0.718, p > 0.250). Post-hoc analysis showed that, for females, physical pain increased behaviours of trusting on person (t (51) = 2.308, p = 0.024) but not on randomness (t (51) < 1, p > 0.5). For males, neither the trusting behaviour on person nor that on randomness was modulated by physical pain (ps > 0.5). 6 Eur J Pain (2017) –
We further tested two potential mechanisms underlying the effects of physical pain on interpersonal trust in the female sample, i.e., the reciprocity hypothesis and the empathy-altruism hypothesis. To test the empathy-altruism hypothesis, altruism was measured by the monetary transfers in the dictator game. Results showed that the amounts of money endowment were unaffected by pain condition in either the out-group condition (3.70 vs. 4.81; t (51) = 1.11, p = 0.273; Fig. 2A) or the in-group condition (7.52 vs. 8.69; t (51) = 1.10, p = 0.277). Meanwhile, we also measured the trustworthiness in the trust game. While physical pain increased interpersonal trust, it in opposite tended to decrease the trustworthiness towards both out-group member (0.50 vs. 0.43; t (51) = 1.42, p = 0.161; Fig. 2B) and in-group member (0.54 vs. 0.48; t (51) = 1.72, p = 0.091). Participants experiencing pain tended to return less money to the investors and kept more for themselves. These results indicated that the paininduced increase in interpersonal trust may not due to general increase in prosocial inclinations. To test the reciprocity hypothesis, we examined the role of expectation (of money profit) in the pain effect on interpersonal trust by measuring the amount of money one would expect to gain as an investor in the trust game. Considering that the pain effect on trust was not differentiated by relationship, the variable Relationship was collapsed and interpersonal trust was then evaluated by averaging the amounts of money transferred to in-group and outgroup members. First, it was found that individuals who expected higher profit tendered to trust more on others by transferring more money (r = 0.826, p < 0.001; Fig. 3A). Second, mediation analysis was then performed to test whether expectation of profit mediated the pain effect on trust. It revealed that the pain treatment was a significant predictor of expectation of profit (a path, B = 2.38, p = 0.047), as well as a significant predictor of interpersonal trust (c path, B = 2.79, p = 0.024). Importantly, the inclusion of expectation into the regression model predicting interpersonal trust from pain treatment resulted in an absence of a significant treatment effect (c’ path, B = 0.810, p = 0.279; see Table 3 for © 2017 European Pain Federation - EFICâ
Pain increases trust in females
C. Wang et al.
Figure 2 Effects of physical pain on altruism and trustworthiness. (A) Bar charts represent altruistic behaviour to in-group/out-group members in Dictator Game (DG), indexed by money endowment (unit: RMB, range: 0–20 Yuan). (B) Bar charts represent trustworthiness to in-group/out-group members in Trust Game (TG), indexed by the ratio of money return to the amount of money they have received (range: 0–1). Participants are females and are categorized by treatment (pain, control). Error bars denote standard error.
statistic details; Fig. 3B). Third, a bootstrap resampling analysis (N = 5000) of the effect size showed that this mediation effect was different from zero with 95% confidence (R2 = 0.69; Effect Size = 1.98; CI (Confidence Intervals) = [0.04–4.03]). Furthermore, we also conducted mediation analyses with in-group/out-group trust being taken into account; however, this resulted in marginally significant effects. We speculated that this might be due to a large variation among participants and an averaged in-group/out-group effect may help to some extent overcome this problem. These results indicated that the expectation of profit fully mediated the effect of pain on trust, which supported a reciprocity mechanism that individuals exhibited more interpersonal trust for self-benefit when experiencing physical pain. In addition, physical pain significantly increased subjective rating of individuals’ anxiety-related irritable feeling (control: 1.59, pain: 2.69; t (51) = 3.37, p = 0.001) and moreover, such irritable feeling was positively correlated with interpersonal trust (r = 0.355, p = 0.009). The results implied that irritable feeling might play a role in the pain effect on trust. This was confirmed by mediation analysis that anxiety (irritable feeling) was a significant mediator in the pain effect (Fig. 3C; R2 = 0.17; Effect Size = 1.15, CI = [0.019–3.00]). Furthermore, when putting both expectation and anxiety (i.e., irritable feeling) into a regression model as two mediators, a robust indirect effect was observed (Fig. 3D; R2 = 0.69; Effect Size = 2.22, CI = [0.22–4.27]). However, when these paths were inspected individually, only the expectation path was significant (CI = [0.021–4.00], p < 0.001). These results suggested that the modulation of pain on one’s cognitive processes (i.e., expectation) and affective mental © 2017 European Pain Federation - EFICâ
states (i.e., irritable feeling) contributed to the paininduced increase of trust.
4. Discussion The present study sought to explore the influence of physical pain on prosocial behaviours in social interaction by examining the effects of experiencing physical pain on interpersonal trust. Our experiment revealed that, first, physical pain induced a differential effect on interpersonal trust between genders, in which increased trust was only observed among females but not males; second, the effect of physical pain on trust was specifically associated with social interaction, whereas it was not driven by risk-taking tendency in non-social situation; third, the changes in interpersonal trust among females did not result from the alteration of altruism or trustworthiness by pain; and fourth, the enhancement of monetary expectation by physical pain fully mediated the pain effect on trust. Our findings provide important information for the understanding of the mechanism underlying how physical pain increase interpersonal trust. Motivation has been found to play an important role in prosocial behaviour such as trust. The current study examined the relative involvements of the two motives, i.e., the reciprocity and empathy-based altruism (Batson and Powell, 2003), in pain-induced facilitation of interpersonal trust. If the increase in trust was due to the pain effect on altruism towards other, we would expect more money transfer in the dictator game in the painful relative to control condition. However, we did not see such a difference in the dictator game. On the other hand, we found that individuals in physical pain transferred more money to others with an increased expectation of monetary
Eur J Pain
(2017)
–
7
Pain increases trust in females
C. Wang et al.
Figure 3 The pain effects on trust were mediated by expectation and anxiety. (A) The correlation between expectation of money profit and trust level (money endowment in Trust Game). (B) The mediation effect of expectation of money profit between treatment (painful condition, control condition) and trust level (money endowment). (C) The mediation effect of anxiety (irritable feeling) between treatment and trust level. (D) The mediation effect of expectation of money profit and anxiety (irritable feeling) between treatment and trust level. The red arrows denote the direct effect and blue arrows denote the total effect. *,**,*** denote p < 0.05, p < 0.01, p < 0.001, respectively.
Table 3 Testing expectation as a mediator on the effect of pain on trust. Coeff. B
SE of B
t
P
c path: Total Effect of IV (Treatment) on DV (Trust) 2.79 1.20 2.33 0.024* a path: IV (Treatment) to Mediator (Expectation) 2.38 1.17 2.04 0.047* b path: Direct Effect of Mediator (Expectation) on DV (Trust) 0.833 0.085 9.77
