اثرات مقایسه اجتماعی بر احساسات و رفتار اجتماعی در دوران کودکی:رشد شناسی حسادت و احساس خوشحالى از بدبیارى دیگران، تغییرات رشدی در تصمیم گیری های مربوط حقوق صاحبان را پیش بینی می کند

Abstract Social comparison can elicit emotions such as envy, which can affect social interactions. The emergence and development of such social emotions through ontogeny, and their influence on social interaction, are unknown. We tested 182 children from 7 to 13 years of age with a novel monetary reward-and-punishment task measuring envy and Schadenfreude (i.e., gloating or taking delight in someone else’s misfortune). Children were either rewarded or punished in a trial-by-trial evaluation of their performance on a speeded reaction time task. In a social condition, feedback of their own and a competitor’s performance was given for each trial. Afterward, children rated how they felt about the outcome. The ratings suggest that when children won, they felt better if the competitor lost instead of winning (i.e., Schadenfreude). Conversely, when children lost, they felt worse if the competitor won instead of losing (i.e., envy). Crucially, levels of envy and Schadenfreude decreased with age. We also studied how these emotions relate to social decisions made separately during three resource allocation paradigms. In each, children chose between two options that differed in the distribution of valuable tokens between themselves and an anonymous other. The combination of choices allowed the measurement of inequity aversion (i.e., equality for all) and spite (i.e., self-profit to maximal disadvantage of the other). We found an age-related increase in inequity aversion and decrease in spite. Crucially, age-related changes in both envy and Schadenfreude predicted the developmental change in equity-related decisions. These findings shed light on the development of social emotions and demonstrate their importance in the development of prosocial behavior in children.

Introduction Comparing oneself with others happens universally across human cultures (Gibbons and Buunk, 1999 and White and Lehman, 2005). From personal attributes such as beauty and intelligence to possessions such as cars and homes, social comparison processes can significantly influence how good one feels about oneself (Festinger, 1954). Think about how men compare the size of their vehicles; to improve their sense of self-worth, they compare themselves with those with smaller or cheaper cars or perhaps perform actions that sabotage those with larger or more expensive ones (see Taylor & Lobel, 1989, and Zizzo, 2003, for similar behavior in other domains). Particularly in competitive environments, it has been shown that social comparison leads to envy as well as Schadenfreude (i.e., gloating or taking delight in someone else’s misfortune) (Smith & Kim, 2007). It has been argued that envy and Schadenfreude are strongly linked emotional states (Smith and Kim, 2007 and Smith et al., 1996), as indicated by an alleviation of envy when misfortune befalls the envied other, resulting in Schadenfreude (Smith et al., 1996 and Takahashi et al., 2009). Furthermore, both envy and Schadenfreude have been linked to grossly antisocial behavior (Hein et al., 2010 and Zizzo, 2003). Given these potentially detrimental effects of emotions resulting from social comparison, understanding their emergence and development during childhood is of particular interest. However, to date, little is known about the ontogeny of envy and Schadenfreude and how they relate to the development of prosocial behavior. The literature on the development of social comparisons has focused predominantly on the age at which children begin to compare themselves with their peers and how these comparisons affect subsequent self-appraisal, motivation, and performance judgments during competitive tasks (Butler, 1989, Pomerantz et al., 1995, Rhodes and Brickman, 2008 and Ruble et al., 1994). Between 5 and 10 years of age, relative failures typically lead to negative self-appraisals and decreased motivation (Boggiano and Ruble, 1979, Butler, 1989, Pomerantz et al., 1995, Ruble et al., 1980, Ruble et al., 1994, Ruble, Feldman, et al., 1976 and Ruble, Parsons, et al., 1976). Given that the effect of social comparison on emotional states in adults is considerable (Smith & Kim, 2007), it is all the more surprising that only a few studies have investigated a direct link between social comparison and affective states in children (Carlson and Masters, 1986 and Masters et al., 1985). Importantly, to our knowledge, only one study has addressed the direct effects of both upward and downward social comparison (i.e., comparison with someone currently better and worse off, respectively) on affective states in children (LoBue, Nishida, Chiong, DeLoache, & Haidt, 2011), and none has studied how these effects change with age. Previous research has reported an increase in the extent to which social comparison occurs during middle childhood (Butler, 1989, Pomerantz et al., 1995 and Ruble et al., 1994). Therefore, one hypothesis is that these increases should be accompanied by a rise in the associated emotions such as envy and Schadenfreude. At the same time, however, it has been shown that with age, children become increasingly aware of the negative aspects of social comparison (Pomerantz et al., 1995). Thus, an alternative hypothesis is that increases in social comparison lead to greater regulation of negative emotions, resulting in less envy and Schadenfreude during childhood. To directly address these hypotheses, we devised a novel experimental paradigm and studied whether and how emotions that arise out of social comparison change through development in children from 7 to 13 years of age. Even though previous studies have focused predominantly on younger ages, we were interested in further developmental changes right throughout middle and late childhood. In addition, piloting suggested that the paradigm was easily followed by 7-year-olds but less so by younger children. Pragmatically, this represented the lower bound of our age range. Because our focus was on the development of social emotions during childhood, the onset of adolescence, typically estimated to occur between 12 and 14 years of age, represented the upper bound. Using such an age range permits us to chart the development of emotions arising out of social comparison and whether an age-related increase in social comparison occurs with a concomitant increase or decrease in associated emotions. Effects of envy are also interesting in terms of their putative influence on human social decisions. This is, for example, evidenced by the commonly displayed preference for dividing a resource into equal parts, also known as inequity aversion ( Fehr & Schmidt, 1999). Numerous studies on the development of this preference provide evidence for age-related changes in behavior reflecting inequity aversion ( Almas et al., 2010, Blake and McAuliffe, 2011, Fehr et al., 2008, Harbaugh et al., 2003, Moore, 2009, Shaw and Olson, 2012, Steinbeis et al., 2012 and Thompson et al., 1997). It has been shown that with age, children increasingly reject unfair offers in the Ultimatum Game ( Harbaugh et al., 2003 and Steinbeis et al., 2012), which can be interpreted as resulting from increasing inequity aversion. Blake and McAuliffe (2011) further reported that whereas 4- to 7-year-olds reject offers that put themselves at a disadvantage, children of approximately 8 years of age also reject offers that put another at a disadvantage, a finding that was recently replicated ( Shaw & Olson, 2012). In addition, inequity aversion develops through adolescence, as indicated by developmental changes in the explicit justification of inequality ( Almas et al., 2010). Using a set of three economic games, Fehr et al. (2008) showed that between 3 and 8 years of age, children increasingly choose sharing options that ensure equal pay-offs for all parties involved. In each game, a child needed to decide how a number of sweets were to be divided between himself/herself and an anonymous other and was faced with two options. One option divided the sweets equally, and an alternative option favored one of the two players. For the child making the decision, the alternative options could be either advantageous (i.e., one sweet for the child and none for the other) or disadvantageous (i.e., one sweet for the child and two for the other). These games provide a systematic test for how strongly children of a certain age prefer equality depending on the costs incurred to create it. The study found an age-related increase in strong inequity aversion, as indicated by a greater willingness to incur costs to one’s own pay-off to obtain equality. Similarly, there was an age-related decrease in spiteful decisions, namely those that always left the other disadvantaged, even when this did not incur any additional benefits to the child making the choice. In the current study, we employed the same measures to test for inequity aversion and spite. Given our tested age range, we modified the incentive structure and used age-appropriate toys instead of sweets as rewards. It has been shown that the type and value of the experimentally implemented currency can have a considerable impact on equity-related decisions in children (Blake & Rand, 2010). Therefore, we expected to replicate the pattern of an age-related increase in inequity aversion and decrease in spite, as reported by Fehr et al. (2008), given the use of salient and age-appropriate attractive rewards. Importantly, we sought to test whether emotions associated with social comparison play a role in bringing about equity-related decisions. In spite of the evident wealth of studies on the development of inequity aversion, the mechanisms underlying age-related changes in this preference remain elusive and emotional development may constitute a plausible candidate. Whereas it has been shown that changes in explicit norm-related reasoning may play a role (Blake & McAuliffe, 2011), nothing is known of the possible involvement and development of emotional processes. So far, only few studies have addressed the role of emotions in social decision making (Gummerum et al., 2010, Kogut, 2012 and LoBue et al., 2011) without providing a direct link between the emotions and decisions. Therefore, we were interested in whether age-related differences in social emotions (i.e., emotions that occur exclusively in real or imagined social situations), such as envy and Schadenfreude, could predict developmental changes in inequity aversion as measured by equity-related decisions. To address this, we linked our measures of envy and Schadenfreude with those of equity-related decisions derived from children’s choices in the economic games. It has been shown that the emotional consequences of decisions can be anticipated, which in turn influences the decisions themselves (Coricelli et al., 2005 and Loewenstein et al., 2001). Thus, it is likely that anticipated emotions affect decisions regarding the distribution of resources. For instance, individual differences in the experience of an emotion, such as regret, predict the extent to which that same emotion is anticipated to occur as a result of a decision (Coricelli et al., 2005). Therefore, we hypothesized that our measures of envy and Schadenfreude should constitute predictors of equity-related decisions. Specifically, we hypothesized that anticipating feelings of envy (i.e., feeling worse when the other has more than oneself compared with when both have equally little) should lead to decisions that aim to increase the other’s disadvantage (i.e., spiteful decisions). It has been argued that Schadenfreude and envy are tightly coupled emotional states (Smith and Kim, 2007 and Smith et al., 1996). This appears to be due to the fact that the experienced envy is alleviated when seeing misfortune befall the envied other, something to result in Schadenfreude (Smith et al., 1996 and Takahashi et al., 2009). Therefore, we predicted that anticipating feelings of Schadenfreude (i.e., feeling better when one has more than the other compared with when both have equally as much) should also lead to decisions that maximize the disadvantage of the other (i.e., spiteful decisions). Thus, we hypothesized that individual differences in our measures of envy and Schadenfreude would be predictive of spiteful decisions, as measured by the choices made in the context of economic games. Furthermore, we hypothesized that age-related changes in feelings of envy and Schadenfreude would mediate the observed age-related changes in equity-related decisions

Results For the sake of comparability with previous studies using some of the same paradigms, we analyzed the data by age groupings used previously (Fehr et al., 2008). Envy and Schadenfreude task Single condition Emotion ratings in the envy and Schadenfreude task were subjected to a repeated-measures mixed analysis of variance (ANOVA). Effects of gains and losses in the absence of the other’s pay-off were looked at first (single condition) with the factors gain (gain/loss) and age (7–8 years/9–10 years/11–13 years). There was a main effect of gain, F(1, 171) = 675.886, p < .01, View the MathML sourceηp2 = .798, indicating significantly different ratings for gains (M = 7.29, SD = 0.19) than for losses (M = −3.71, SD = 0.33). There was also a significant interaction between the factors gain and age, F(2, 171) = 8.111, p < .01, View the MathML sourceηp2 = .087, indicating that the strength of ratings in the single condition decreased with age. Separate analyses of ratings for gains and losses revealed that the age decrease in ratings was found only for gains, F(2, 173) = 15.216, p < .01, View the MathML sourceηp2 = .15, and not for losses (p > .10). These data show that the current paradigm was successful in inducing separable effects of positive and negative emotions in response to winning or losing in a competitive speeded reaction time task. Double condition To test for an effect of the other’s pay-off on ratings of wins and losses (double condition), a repeated-measures mixed ANOVA with the factors self (self win/self loss), other (other win/other loss), and age was conducted. There was a significant effect of the factor self, F(1, 171) = 891.239, p < .01, View the MathML sourceηp2 = .84, as well as an interaction between the factors self and age, F(1, 171) = 9.095, p < .01, View the MathML sourceηp2 = .097, suggesting that ratings for one’s own wins and losses also decreased with age in the double condition. In addition, there was a significant effect of the factor other, F(1, 171) = 32.193, p < .01, View the MathML sourceηp2 = .159, indicating that the competitor’s wins and losses affected one’s own emotion ratings, which also decreased with age, as indicated by a significant interaction between the factors other and age, F(1, 171) = 7.215, p < .01, View the MathML sourceηp2 = .078. Furthermore, there was a significant interaction between the factors self and other, F(1, 171) = 12.700, p < .01, View the MathML sourceηp2 = .07, indicating that effects of the other’s wins and losses were stronger on participants’ ratings when the children themselves had lost compared with when they had won. This implies that the effects of envy were stronger than the effects of Schadenfreude, a factor that did not interact further with the factor age. Given this interaction, we carried out separate analyses for the two social emotions. Effects of envy and Schadenfreude For gains, analyses were conducted using the factor Schadenfreude (self win/other win or self win/other loss) and age. For losses, analyses were conducted using the factor envy (self loss/other loss or self loss/other win) and age. There was a significant effect of Schadenfreude, F(1, 171) = 9.166, p < .01, View the MathML sourceηp2 = .051, as well as envy, F(1, 171) = 38.872, p < .01, View the MathML sourceηp2 = .186. Importantly, there was an interaction with the factor age for both Schadenfreude, F(2, 171) = 4.240, p < .05, View the MathML sourceηp2 = .047, and envy, F(2, 171) = 5.640, p < .01, View the MathML sourceηp2 = .062, revealing a decrease in both with age (see Fig. 1C). When controlling for individual differences in emotional response to winning and losing in the single condition, effects of envy remained significant (main effect of envy: F(1, 166) = 24.795, p < .01, View the MathML sourceηp2 = .13; interaction between envy and age: F(2, 166) = 7.060, p < .01, View the MathML sourceηp2 = .078), whereas those of Schadenfreude did not or only marginally did so (main effect of Schadenfreude: p > .10; interaction between Schadenfreude and age: F(2, 166) = 2.846, p < .10, View the MathML sourceηp2 = .033). In addition, we found a significant correlation between envy and Schadenfreude scores (r = −.424, p < .01), suggesting that the same children who had a tendency to feel strong feelings of envy were also those who experienced more Schadenfreude (see Fig. 1D, upper left quadrant, for effects of envy and Schadenfreude). This correlation was equally strong for all age groups, as indicated by a nonsignificant interaction with the strength of the correlation between envy and Schadenfreude and age, F(3, 172) = 2.744, p < .10. Equity games To look at age-related changes in the context of the three games testing for equity-related decision making, choices made in each game were subjected to logit regressions. There was a significant increase in inequity-averse choices in both the prosocial game (β = −0.916, SE = 0.238, Wald = 14.841, p < .01) and the sharing game (β = −0.687, SE = 0.275, Wald = 6.245, p < .05), but not in the disadvantageous inequality game (see Fig. 2A). Using the previously established decision-making profiles, we could show that there was an increase in strong inequity aversion with age (β = −1.283, SE = 0.503, Wald = 6.508, p < .05) as well as a decrease in spite with age (β = 0.880, SE = 0.244, Wald = 12.971, p < .01) (see Fig. 2B). Behavior in the decision-making task. (A) Percentage of children who made ... Fig. 2. Behavior in the decision-making task. (A) Percentage of children who made inequity-averse choices for each age group and for each of the three games. There was an increase in inequity-averse choices in both the prosocial and sharing games. (B) Decision-making profiles constructed on the combination of choices made in the three games for each of the three age groups. There was an increase in inequity-averse decisions with age as well as a decrease in spiteful decisions with age. Figure options Envy and Schadenfreude and equity games To investigate the predictive relationship between social emotions as measured by the envy and Schadenfreude task in classifying subjects into their decision-making profile, we performed regression analyses. Given that we predicted a specific link between envy and Schadenfreude and spite, we tested the hypotheses of whether greater Schadenfreude and envy are linked to a greater likelihood to behave spitefully. A logit regression revealed that individual differences in envy could successfully predict individual differences in spiteful choices, whereby more envious individuals in the envy and Schadenfreude task were more likely to also behave spitefully in the resource allocation paradigms (β = 0.445, SE = 0.147, Wald = 9.131, p < .01) (see Fig. 3A), which persisted when controlling for age (β = 0.389, SE = 0.153, Wald = 6.49, p < .05). We could also show that individual differences in Schadenfreude were predictive of individual differences in spiteful choices, whereby individuals with greater Schadenfreude were more likely to also behave spitefully in the resource allocation paradigms (β = −0.356, SE = 0.141, Wald = 6.352, p < .05) (see Fig. 3A), which also persisted when controlling for age (β = −0.288, SE = 0.144, Wald = 4.003, p < .05). We also tested for explanatory power of individual differences in envy and Schadenfreude for the other decision-making profiles (strong and weak inequity aversion, strong and weak generosity). None of these relationships was significant. However, given the low incidence of inequity aversion in the current population, we pooled strong and weak inequity-averse participants to test for an association between our measures of social emotions and inequity aversion per se. A logit regression revealed that individual differences in envy could successfully predict individual differences in inequity-averse choices, implying that less envious individuals were also more likely to opt for inequity-averse choices (β = −0.276, SE = 0.137, Wald = 4.046, p < .05) (see Fig. 3B), which remained marginally significant when controlling for age (β = −0.229, SE = 0.137, Wald = 2.79, p < .10). This link was not found for Schadenfreude and inequity aversion. Envy and Schadenfreude scores for children classified as either spiteful or ... Fig. 3. Envy and Schadenfreude scores for children classified as either spiteful or inequity averse. (A) Those classified as spiteful were significantly more envious and showed more Schadenfreude than those classified as not spiteful. (B) Those classified as inequity averse were significantly less envious than those classified as not inequity averse. Figure options We had also hypothesized a possible relationship between age-related changes in social emotions and equity-related decisions, so that we tested for whether the observed age effects on envy and Schadenfreude might explain the age-related changes observed in spiteful and inequity-averse decisions. A significant mediation would imply that the observed age-related decrease in envy and Schadenfreude could account for the age-related decrease in spiteful decisions and possibly also the increase in inequity aversion. To do so, we conducted mediation analyses where age was the predictor, envy and Schadenfreude were the mediators, and spiteful or inequity-averse choices were the outcome variables. Analyses were conducted using bootstrapping procedures recommended for smaller samples and dichotomous outcome variables (Preacher & Hayes, 2004) and operationalized in an SPSS Macro (Preacher & Hayes, 2008). We used 10,000 bootstrap resamples of the data with replacement. Statistical significance with alpha at .05 is indicated by the 95% confidence intervals not crossing zero. As in previous analyses, we treated envy and Schadenfreude separately, assuming that convergent results would be found. We found a significant mediation effect of envy with respect to the relationship between age and spiteful choices (indirect effect = .18, SE = .11, 95% confidence intervals = .037, .47) (see Fig. 4A) and found no such effect for inequity-averse choices. We also found a significant mediation effect of Schadenfreude with respect to the relationship between age and spiteful choices (indirect effect = .13, SE = .08, 95% confidence intervals = .015, .37) (see Fig. 4B) and again found no such effect for inequity-averse choices. In each case, mediation was partial, meaning that the direct effect of age alone could still predict significant portions of the variance observed in spiteful decisions. This suggests that age-related differences in social emotions may be one important underlying mechanism but not the only one driving observed age-related decreases in spite. Mediation models for the effects of age on spiteful decisions via social ... Fig. 4. Mediation models for the effects of age on spiteful decisions via social emotions. Values are standardized regression coefficients, and asterisks indicate significant coefficients (∗p < .05). (A) Feelings of envy significantly mediated age effects on spite. (B) Feelings of Schadenfreude significantly mediated age effects on spite.