Happy Little Benefactor: Prosocial Behaviors Promote Happiness in Young Children From Two Cultures

Song, Yue; Broekhuizen, Martine Louise; Dubas, Judith Semon

doi:10.3389/fpsyg.2020.01398

ORIGINAL RESEARCH article

Front. Psychol., 30 June 2020

Sec. Developmental Psychology

Volume 11 - 2020 | https://doi.org/10.3389/fpsyg.2020.01398

Happy Little Benefactor: Prosocial Behaviors Promote Happiness in Young Children From Two Cultures

$\r\nYue Song*$ Yue Song^1*

Martine Louise Broekhuizen²

Judith Semon Dubas¹

¹Developmental Psychology, Utrecht University, Utrecht, Netherlands
²Child, Family, and Education Studies, Utrecht University, Utrecht, Netherlands

Evidence that young children display more happiness when sharing than receiving treats supports that humans, by nature, are prosocial. However, whether this “warm glow” is also found for other prosocial behaviors (instrumental helping and empathic helping) and/or in different cultures is still unclear. Dutch (studies 1 and 2) and Chinese (study 3) young children participated in a sharing task, followed by instrumental helping and empathic helping tasks in which they were praised (thanked) if they helped. Consistent results were found across three studies, showing that (1) participants displayed more happiness after giving than receiving treats; (2) toddlers displayed more happiness after instrumental helping than initially interacting with the experimenter; and (3) toddlers’ happiness remained the same after positive social feedback (i.e., being thanked). Taken together, these results indicate that independent of culture, both sharing and instrumental helping are emotionally rewarding, supporting an evolutionary origin of these behaviors.

Introduction

Sacrificing one’s own resources to support others, even strangers, is an exceptional human ability that has puzzled researchers for many years. While essential for establishing large-scale social cooperation in early human groups (e.g., Darwin, 1871; Wilson, 1975; Ebstein et al., 2010) and maintaining large organizations in the modern world (Wilson, 1975; Fehr and Fischbacher, 2003), the question is what benefits are there, if any, for the individual who is sacrificing? While theories such as kin selection (Hamilton, 1964) or reciprocal altruism (Trivers, 1971) focus on the potential, long term rewards that sharing or cooperation may bring (Aknin et al., 2015), a growing body of studies shows that the prosocial behavior itself might also be emotionally beneficial for an individual, suggesting additional short-term rewards (e.g., Dunn et al., 2008; for a review, see Aknin et al., 2018). For adults, prosocial behaviors (e.g., spending money on others) have been found to lead to an increase in happiness in both western and non-western cultures and across diverse socio-economic contexts (e.g., Dunn et al., 2008; Aknin et al., 2009, 2013, 2015). Based on these findings, these emotional rewards have been proposed as an evolved psychological mechanism that sustains prosocial behaviors even when it costs individual resources (e.g., Dunn et al., 2008; Aknin et al., 2013).

This “positive feedback loop” (Aknin et al., 2018, p. 55) also has drawn great attention from researchers studying young children. While many studies have focused on how (positive) emotional arousal serves as a precursor to toddlers’ and young preschoolers’ prosocial behaviors (e.g., Hepach, 2017; Cowell et al., 2018; Miller, 2018), another line of research has focused on whether and in which social context prosocial behaviors themselves are emotionally rewarding. A large number of studies have shown that prosocial behaviors can alleviate children’s negative emotions. For instance, even 2 years old negative arousals were alleviated both when they helped a stranger themselves (i.e., helped to retrieve the object that was out of the experimenters’ reach), and when a third-party provided help (Hepach et al., 2016). In fact, children who exhibit more empathic concern for others, helped faster (for a review, see Eisenberg and Miller, 1987). Furthermore, some researchers even propose that some prosocial behaviors (e.g., comforting) are motivated by alleviating one’s own negative feelings (for a review, Paulus, 2014).

In comparison, fewer studies have focused on how prosocial behaviors contribute to an increase of positive feelings, rather than just alleviating negative feelings. Three studies directly explored whether prosocial behavior leads to an increase of happiness in young children, by rating and comparing their happiness before and after sharing. The first study, conducted in Vancouver, Canada, gave toddlers windfalls (i.e., 8 treats), and then asked them to share one of these treats with a toy monkey (i.e., “I do not see any more treats, will you give one to monkey?”), and also share from other resources (i.e., look, I found one more treat, will you give it to monkey?). Results showed that 22 months old were rated as happier when they shared treats rather than when they received treats, and happier under a costly sharing condition (i.e., sharing from their own resources), rather than under a non-costly sharing condition (i.e., sharing from other resources) (Aknin et al., 2012). This study demonstrates that the relation between sharing and happiness already exists when sharing first emerges, and the results suggest this immediate emotional reward of sharing is a proximate mechanism that facilitates prosocial behavior despite its costs (Aknin et al., 2012). Using the same experimental paradigm, these results were replicated in a second study conducted among 2–5 years old children living in a remote, small rural village on Tanna Island (Aknin et al., 2015). More recently the emotional benefits of sharing was found among 3 and 5 years old Chinese children but only when sharing was autonomous rather than obligated (Wu et al., 2017). Nevertheless, before firm conclusions about this mechanism can be drawn, more studies are needed to test this universal claim.

Therefore, the first aim of the current study was to replicate and extend these findings by examining whether sharing behavior itself can lead to an increase of happiness among young children (i.e., toddlers and young preschoolers) in different cultures. For this purpose, we employed the same experimental procedure and rating scales used in the original study (Aknin et al., 2012), and we tested this proposed mechanism in three samples (i.e., Dutch toddlers, Dutch preschoolers, and Chinese preschoolers). More specifically, the first study aimed to replicate previous findings in toddlers. In addition, even though the warm glow is potentially detectable in all cultures, it “may vary in degree of expression according to the cultural context” (Aknin et al., 2013, p. 636). Considering that upon reaching the preschool age, children begin to internalize cultural norms, and orientations (Schuhmacher and Kärtner, 2015), it is plausible that the cultural context could play a role in young preschoolers’ prosocial behaviors, and further affect the happiness shown after prosocial behavior at this age. Thus, we also examined the warm glow in 3 years old among Dutch (Study 2) and Chinese (Study 3) preschoolers, when the warm glow may begin to vary in different cultural contexts. If the emotional reward of generosity is a universal and robust psychological mechanism, both Dutch and Chinese toddlers and younger preschoolers should express more happiness after sharing rather than receiving treats, and also after costly sharing compared with non-costly sharing.

Next to replicating these findings, our second aim was to examine whether the number of resources available when sharing affects the degree of happiness shown. There are reasons to think that resource availability might influence the happiness gain after sharing for toddlers and young preschoolers. In order to share, children need to overcome their own desire for the material resources (Dunfield, 2014). However, young children just engage in, and are sometimes even reluctant to share (Ulber et al., 2015), indicating that it may be difficult to overcome these desires. Accordingly, it is plausible that the increase in happiness after sharing is larger when individuals have a larger compared to a smaller amount of resources. That is, the more resources available, the more likely it is that their own material needs are still fulfilled when they share one resource, leading to a higher increase in happiness compared to children with fewer resources. In addition, cognitively, even toddlers can already discriminate a small set of objects (the numbers “one” to “four”) from a larger set of objects (more than four) (Xu, 2003; Sarnecka and Gelman, 2004; Le Corre and Carey, 2007). Thus, the understanding of number may allow toddlers to recognize the cost involved in sharing based on the number of resources available. Finally, studies on sharing behavior showed that young preschoolers are more likely to share after receiving more resources (Posid et al., 2015). The current study further examines whether children’s happiness after sharing is affected by resource availability. In previous studies on the happiness after sharing, children received a relatively large number of resources (8 treats) (Aknin et al., 2012, 2015). The current study went a step further by comparing the increase in happiness in three resource conditions: 2, 4, and 8 treats. If the increase in happiness after sharing is larger in the large set condition than in small set conditions, then this would support the idea that children’s own material needs need to be met (i.e., having enough resources) as a prerequisite for the happiness after costly sharing at this age.

The third aim of this study was to examine whether the emotional rewards of sharing also apply to other prosocial behaviors. Studies on the emotional rewards of prosocial behaviors mainly focus on sharing (e.g., Aknin et al., 2012, 2015), but there are more types of prosocial behaviors, such as instrumental helping and empathic helping, that already emerge at toddlerhood (e.g., Dunfield et al., 2011; Svetlova et al., 2011; Brownell et al., 2013; Dunfield and Kuhlmeier, 2013). More importantly, these three prosocial behaviors (i.e., sharing, instrumental helping, and empathic helping) are distinct from each other (i.e., requested for by different cognitive abilities, and driven by different motivations, for a review, see Padilla-Walker and Carlo, 2014). Thus, in addition to sharing, examining happiness after other types of prosocial behaviors would help to further investigate whether the emotional reward is a psychological mechanism that sustains several kinds of prosociality. Accordingly, the current study examined happiness after two additional types of prosocial behaviors (instrumental helping and empathic helping).

The link between empathic helping and happiness has not been investigated. As for instrumental helping, recently, using body posture as a measurement of positive emotion, researchers found that 2 years old show positive emotions after they helped an experimenter achieve her goal (Hepach et al., 2017a). However, in this study instrumental helping is a byproduct of play situations rather than an intentional behavior, as children had no chance to decide whether they wanted to help the experimenter; instead, they just happened to find the object that the experimenter needed during the session. Thus, it is still unclear whether and how these emotions change when children actually want to and succeed in helping others. Hence, more studies are needed to directly examine how emotions change when children initially intend and succeed in instrumental helping (Hepach et al., 2017a). In the current study we used facial expression to measure happiness, and we expected that if the emotional rewards of prosocial behaviors are ingrained in human nature, then most kinds of prosocial behavior should lead to an increase of happiness, not only sharing.

An alternative explanation for the increase of happiness is that it is the positive social feedback (such as being thanked or praised) after prosocial behaviors, but not the prosocial behaviors themselves, that promotes happiness. However, praising does not affect toddler’s instrumental helping subsequently (Warneken and Tomasello, 2008). Nevertheless, in order to further test this alternative explanation, more research on children’s happiness after prosocial behaviors and after receiving positive feedback is needed. Therefore, the current study evaluated toddler’s and prechooler’s happiness immediately after they performed instrumental helping or empathic helping behaviors, and after being thanked for this behavior. If happiness is a result of the social rewards/interactions after the prosocial behaviors, then happiness after being thanked would be higher than before they heard “thank you.” It should be noted that because we replicated Aknin’s methods for the sharing task (Aknin et al., 2012, 2015), in which the child was not thanked, we did not test this explanation in sharing.

Current Study

The overarching goal of the current set of studies is to examine the link between prosocial behaviors and happiness for toddlers and young preschoolers. Specifically, we extend prior research by (1) replicating previous studies in two different cultures; (2) exploring whether and how resource availability would play a role in this (potential) relationship; (3) examining whether this positive relationship also exists for other prosocial behaviors (instrumental helping and empathic helping); and (4) examining the role of positive social feedback (i.e., being thanked) as an alternative explanation for the increase in happiness. For these purposes, we used a series of tasks to test Dutch toddlers’ (study 1), Dutch preschoolers’ (study 2) and Chinese preschoolers’ (study 3) prosocial behaviors and observe their happiness accordingly.

Study 1

Materials and Methods

Participants

Children were mainly recruited through daycares in several urban areas across the Netherlands, with nine participants recruited from posters in the university and word of mouth. After the daycare agreed to participate, the researchers sent active consent forms to the parents who had a toddler between 16 and 27 months of age. Parents were given a brief explanation of the tasks and told that the experiment would be conducted at the daycare center by two experimenters. In total, 122 toddlers were initially tested, with the majority being Dutch (95.9%), and coming from middle to upper middle-class families. Thirty-one toddlers did not show any of the targeted behaviors (sharing, instrumental helping, and empathic helping) in the experiment, leaving it impossible to measure their happiness while doing these behaviors. The study was approved by the Ethics committee of the Faculty of Social and Behavioral Sciences, Utrecht University. Informed written consent was obtained from all parents before the experiment. Previous studies had 20 children per sample (Aknin et al., 2012, 2015). In addition, a priori sample size calculation (using G^∗Power 3.1.9.2, Faul et al., 2007) showed a total sample of 52 is needed to detect an effect size of 0.46 (with α at 0.05 level and 1 – β error prob at 0.90 level). Thus, the current sample size should be sufficient in detecting associations between happiness and prosocial behaviors.

General Procedure

The experiment was conducted in the child’s daycare (e.g., separate room or a separate corner of a larger room) by a main experimenter (E) and an assistant experimenter (AE). Both the E and AE arrived at the daycare at least 20 min before the testing and helped the teacher in arranging some classroom activities, so that the child could become familiar with them. Once the child felt comfortable with both the E and AE, he or she was invited to do the experiment. During the experiment, both the child and the E were videotaped. Neither the teachers nor parents were present during the testing, except for three cases where the children were too fussy to leave the teacher. In addition, the nine children that were recruited through posters and word of mouth were tested in either child’s daycare (n = 1), university lab (n = 1), or families’ home (n = 7). Parents were not present in the room during testing except for three cases where parents remained quiet and did not interact with the child. Each child completed a sharing, instrumental helping and empathic helping task, with the sharing task always coming first, following by the instrumental helping and empathic helping tasks in counterbalanced order. The session lasted approximately 10 min.

Sharing task (based on Aknin et al., 2012)

Warm up phase

The child got familiar with both the treats and the receiving/giving actions during the warm up phase. Each child met four stuffed animals (in order: mouse, rabbit, cat and panda) with their bowls in front of them, and were told that all animals “love to eat the treats.” Next, the E gave the child a bowl. Then the E gave each animal a treat, and the animal “ate” the treat by making eating sounds like “Yummm” and pushing the treat through a false bottom. After that, the E gave the child one treat, and asked the child whether he/she would like to eat the treat. In the next step, the E gave the child an extra bowl with five treats, and asked the child to share the treats with the four animals. The animal “ate” the treat when given. If, and only if the child hesitated in sharing, the E prompted him/her step by step (see Aknin et al., 2012 for detailed information). If the child shared with the animals, or if he/she refused to share after two rounds of prompts, the E put the animals (including their bowls) away, took away any remaining treats in the child’s bowl, and then moved to the formal test.

Test phase

The children were divided into three conditions (8, 4, or 2 treats they received in the test phase), balanced according to age and gender. The test included the following phases: (a) Meeting the monkey: The E introduced a new stuffed animal (Monkey) who also loves the treats, and encouraged but not forced the child to interact with the monkey (e.g., petting or touching). The E stressed that “now both you and the Monkey have no treats,” indicating to the child that the treats are a limited resource. (b) Receiving treats: The E “found” 2, 4, or 8 treats (depending on the condition), showed these treats to the child and put them in the child’s bowl while saying “Look, I found 2/4/8 treats, and I am giving them all to you”; (c) The E found one more treat and gave this treat to the Monkey; (d) The E found one more treat and asked the child to give this treat to the Monkey; (e) The E acted as if she could not find treats anymore, then asked the child to give one treat from his/her own bowl to the monkey. In order to prevent the order effects, phases (c)–(e) were performed in a counterbalanced order within each treat condition. For example, some toddlers were asked to share from their own resources first then to share from common pool. Non-costly or costly sharing behavior was identified if the child gave any treats from the common pool in phase d, or from his/her own bowl to the Monkey in phase e.

Instrumental helping and empathic helping (based on Svetlova et al., 2011)

In the instrumental helping task, the E wrapped five blocks one-by-one by using napkins that were placed on the table (one napkin within the reach of the child but not the E, and four within the reach of the E but not the child). After the E was out of the four napkins, she asked the child to hand her the final napkin by using eight sequential prompts (see Svetlova et al., 2011, for detailed information), and each prompt was present for 5–7 s. If the child helped, the E stopped providing prompts and took the napkin. After at least 3 s since the child helped, they said “thank you,” but did not give any other praise or rewards. Instrumental helping was identified if the child handed the napkin to the E before or directly after the last cue.

In the empathic helping task, the E showed the child a blanket, folded it around her shoulders, and stated “it makes me warm.” In this way, the link between the blanket and warmth was made explicit. Then the E put the blanket within the reach of the child while she was finding a toy bear, and let the child play with the bear. After 60 s, the E suddenly felt cold and shuddered, and prompted the child to hand the blanket to the E (similar prompts and time for each prompt as in the helping task). If the child helped the E by giving or pushing the blanket to her, the E stopped providing prompts and took the blanket. After at least 3 s since the child helped, they said “thank you,” but did not give any other praise or rewards. Empathic helping was identified if the child handed the blanket to the E before or directly after the last cue.

Coding Procedure

Two coders who were blind to the experimental hypotheses coded the children’s behavior (how many treats the child shared in phase e, and how many cues the child needed before instrumental/empathic helping). Intra-Class Correlations (ICC) were high, with a mean of 0.91 in coding behaviors (ICC _{number shared} = 0.98, ICC _{steps before instrumental helping} = 0.98, and ICC _{steps before empathic helping} = 0.77). The videos in which the children showed sharing, instrumental helping or empathic helping were given to new coders who only coded happiness using the same coding procedures as Aknin et al. (2012).

Happiness coding for children

We used the same rating scales that already had been used in various cultures (Canadian, Aknin et al., 2012; Vanatu, Aknin et al., 2015; and Chinese, Wu et al., 2017). In total five coders who were blind to the experimental hypotheses and the behavior coding rated the children’s happiness by using a 7-point scale (1 = not happy at all (very unhappy or crying), 2 = big frown, 3 = small frown, 4 = neutral/flat affect, 5 = small smile, 6 = big smile, and 7 = very happy/laughing) (Aknin, personal communication, November 15, 2014).

For sharing, they coded the child’s happiness during phases (a)–(e). For instrumental helping and empathic helping, they coded happiness in the following three phases: (a) when the child watched the E wrapping the blocks (instrumental helping task)/showing the blanket (empathic helping task), prior to when the child was prompted to help; (b) when the child helped the E, but before the E said “thank you” (this phase lasted about 2 or 3 s); (3) after the E said “thank you” (this phase lasted about 2 or 3 s). The ICC for children’s happiness across coders was high, ranging from 0.85 to 1.00 for sharing, from 0.71 to 0.83 for instrumental helping and from 0.80 to 0.90 for empathic helping. Eighty-seven percent of the cases were coded by three coders, and the rest (13%) were coded by two coders. The happiness ratings represent the mean value across the coders.

Happiness coding for the E (and the Monkey they used)

Although all Es were blind to the experimental hypotheses and were trained to remain neutral during the experiment, another two raters coded the E’s/monkey’s “reaction” (combined happiness and enthusiasm) during the tasks (sharing, instrumental helping and empathic helping) in the same phases mentioned above for the children’s happiness rating, using the same 7-point scale (1 = not happy/enthusiastic at all, 4 = neutral, and 7 = very happy/enthusiastic). The two coders’ absolute agreement was high, ranging from 79.5 to 100% across all coding phases with a mean of 90.1%. The ICC could not be calculated, because the means of the coded values ranged from 3.99 to 4.17, and the SD’s from 0 to 0.46 on a 7-point scale. These results indicate that both the E and monkey showed a consistent, neutral emotion during the experiment.

Results

Sharing

In total 73 of 122 participants shared one or more treats with the Monkey. For those who shared, three children’s happiness could not be coded in one or several of the phases as their faces moved out of the camera frame, and was excluded pairwise for the main analyses (including 2 cases in phase a, 1 in phase b, and 1 in phase d). In addition, for those who did not share, six children’s happiness could not be coded in phase a and b.

Preliminary analyses compared children who shared to those who did not on their levels of happiness when meeting the monkey (phase a) and receiving the treats (phase b). The independent samples t-tests showed that, compared with those who shared, children who did not share exhibited less happiness when meeting the monkey, t(111) = 2.11, p = 0.04, Cohen’s d effect size (d) = 0.41, and inclined to exhibit less happiness when receiving the treats, t(112) = 1.95, p = 0.05, d = 0.35.

For the main research questions, we only included children who shared in the experiment. For the first research question (concerning whether toddlers show higher levels of happiness after sharing, compared with receiving treats), paired samples t-tests showed that, compared with receiving treats from E (phase b, Mean_b = 4.24, SD = 0.43, see Figure 1A), the participants expressed more happiness after they shared the treat from their own bowl (phase e, Mean_e = 4.67, SD = 0.53), t(70) = 3.08, p < 0.01, d = 0.83, and after they shared the treat from the common pool (phase d, Mean_d = 4.44, SD = 0.46), t(71) = 6.71, p < 0.01, d = 0.36. As for whether happiness is higher after costly than non-costly sharing, results showed children’s happiness to be higher in phase e than phase d, t(71) = 4.97, p < 0.01, d = 0.50.

FIGURE 1

Figure 1. Observed happiness across the five phases in the sharing task, aggregated across all treat conditions. (A) presents the results for study 1, (B) for study 2, and (C) for study 3, respectively. Error bars display standard error of the mean. ^∗p < 0.05.

For the second research question (whether the number of resources would affect the increase of happiness after costly sharing), we first calculated the difference between phase e and b for each child. Next, a one-way ANOVA test (across the 3 resource conditions) was conducted on the calculated difference. Results indicated no difference in children’s increased happiness (from phase b to e) among the different treat conditions, F(2, 71) = 1.99, p = 0.14¹.

In order to test whether we replicated the previous studies (Aknin et al., 2012, 2015), we analyzed data for the 8-treat condition only (n = 29). Paired samples t-tests showed that, compared with receiving treats from E (phase b, Mean_b = 4.41, SD = 0.53, see Figure 2A), the participants expressed more happiness after they shared the treat from their own bowl (phase e, Mean_e = 4.80, SD = 0.48), t(28) = 3.83, p < 0.01, d = 0.77, but not after they shared the treat from the common pool (phase d, Mean_d = 4.51, SD = 0.52), t(28) = 1.05, p = 0.30, d = 0.19). In addition, the participant expressed more happiness in phase e than in phase d, t(28) = 3.74, p < 0.01, d = 0.60, reflecting a higher level of happiness in a costly, compared with a non-costly sharing condition. Thus, we found the same pattern as previous studies. That is, sharing leads to an increase of happiness, and costly sharing leads to a greater increase in happiness compared with non-costly sharing.

FIGURE 2

Figure 2. Observed happiness across the five phases in the sharing task, 8-treat condition. (A) presents the results for study 1, (B) for study 2, and (C) for study 3, respectively. Error bars display standard error of the mean. ^∗p < 0.05, ⁺p = 0.06.

Findings in the 4-treat condition were consistent with findings in the 8-treat condition. Compared with phase b, the participants expressed more happiness in phase e, t(20) = 4.68, p < 0.01, and in phase d, t(20) = 2.67, p = 0.02. Additionally, the happiness is higher in phase e than phase d, t(20) = 2.86, p = 0.01. Findings in the 2-treat condition revealed some different patterns. Similar to findings in the 8- and 4-treat condition, compared with phase b, participants expressed more happiness in phase e, t(21) = 3.33, p < 0.01. However, no difference was found between phase b and d, t(20) = 1.60, p = 0.13. Moreover, happiness was only (marginally) higher in phase e than phase d, t(20) = 1.98, p = 0.06.

Instrumental Helping and Empathic Helping

Because we counterbalanced the order between the instrumental helping and empathic helping tasks, we conducted an independent sample t-test to examine a possible task order effect. No significant differences were found, p_s > 0.24, indicating that there was no order effect on the happiness in these helping tasks. In total 87 participants engaged in instrumental helping and 66 participants engaged in empathic helping. For those who helped, one children’s happiness could not be coded in two phases (after they helped and being thanked) in the empathic helping task, as the face was blocked by the blanket. In addition, for those who did not help, 11 children’s happiness cannot be coded, including 11 cases in phase a in the instrumental helping task, and 5 cases in phase a in the empathic helping task.

Instrumental helping

Preliminary analyses compared children who helped to those who did not help on their levels of happiness when watching the E wrapping the block (phase a), and found no differences, t(107) = 1.08, p = 0.28, d = 0.26. For the main research questions, we only included children who helped in the experiment. For the third research question (whether instrumental helping leads to an increase of happiness), paired samples t-tests showed that compared with watching E wrapping the block (phase a, Mean_a = 4.15, SD = 0.36, see Figure 3A), the participants expressed more happiness after they helped the E (phase b, Mean_b = 4.32, SD = 0.48), t(86) = 3.01, p = 0.00, d = 0.40, and after being thanked (phase c, Mean_c = 4.27, SD = 0.47), t(86) = 2.24, p = 0.03, d = 0.29. For the fourth research question (whether positive social feedback would lead to an increase of happiness after instrumental helping), paired samples t-test showed happiness did not differ between phase b and c, t(86) = 1.37 p = 0.17, d = 0.10.

FIGURE 3

Figure 3. Observed happiness across three phases in the instrumental helping task. (A) presents the results for study 1, (B) for study 2, and (C) for study 3, respectively. Error bars display standard error of the mean. ^∗p < 0.05.

Empathic helping

Preliminary analyses compared children who helped to those who did not help on their levels of happiness when watching E showing the blanket (phase a), and found no differences, t(112) = 1.41, p = 0.16, d = 0.28. For the main research questions, we only included children who helped in the experiment. For the third research question (whether empathic helping leads to an increase of happiness), paired samples t-tests showed that compared with witnessing E showing the blanket (phase a, Mean_a = 4.34, SD = 0.64, see Figure 4A), the participants showed the same level of happiness after they helped the E (phase b, Mean_b = 4.39, SD = 0.55), t(65) = 0.66, p = 0.52, d = 0.08), and after being thanked (phase c, Mean_c = 4.31, SD = 0.45), t(64) = 0.28, p = 0.77, d = 0.03). For the fourth research question (whether positive social feedback would lead to an increase of happiness after empathic helping), results showed happiness ratings did not differ between phase b and c, t(64) = 1.56, p = 0.12, d = 0.16.

FIGURE 4

Figure 4. Observed happiness across the three phases in the comforting task k. (A) presents the results for study 1, (B) for study 2, and (C) for study 3, respectively. Error bars display standard error of the mean. ^∗p < 0.05, ⁺p = 0.06.

Follow-Up Analyses on E’s Happiness Potentially Influencing Children’s Happiness

We further examined whether E’s happiness in each phase was related to children’s happiness in the same phase. For instance, in phase b for the sharing task, we analyzed the correlation between E’s happiness when giving the treats to the participants, and participants’ happiness when receiving the treats. If a significant correlation between E’s happiness and participants’ happiness was found, we re-ran the main analyses (using repeated measure analyses) involving that phase controlling for E’s happiness.

Results

For sharing, no significant correlations were found between E’s happiness and children’s happiness ratings in any phases, p_s > 0.23. Thus, E’ happiness does not appear to play a role in the findings on sharing. For instrumental helping, we found a correlation between E’s happiness and toddlers’ happiness when watching the E wrapping the block (phase a), r = 0.26, p = 0.02, but not after they helped the E (phase b), p = 0.11, or being thanked (phase c), p = 0.82. After controlling for E’s happiness in phase a, we still found that children exhibited more happiness in phase c than a, p = 0.02. However, the difference between phase a and b diminished to trend levels of significance, p = 0.08, indicating that E’s happiness appears to account for some but not all of the increase in children’s happiness after instrumental helping. For empathic helping, no significant correlations were found between E’s happiness and children’s happiness ratings in any phase. Thus, E’ happiness appears to not play a role in the findings on empathic helping. Overall, E’ happiness appears to not play a role in the main findings in study 1.

Discussion

Thus, similar to previous studies, in study 1 we found sharing leads to a higher level of happiness after sharing, and higher levels of happiness in costly than non-costly sharing. Also, we replicated the main findings in previous studies for 8-treat condition. In addition, the increased of happiness after costly sharing did not differ based on number of treats received, implying the number of resources may not affect the link between sharing and happiness. Moreover, we found instrumental helping but not empathic helping to lead to higher levels of happiness. In order to examine whether this proposed mechanism is robust at different ages, in study 2 we further examined whether this warm-glow of prosocial behaviors remains when children are young preschoolers.