Patient children are less likely to share with unknown children

Human prosociality is a valuable but also deeply puzzling trait. While several studies suggest that prosociality is an instinctive and impulsive behavior, others argue that patience and self-control are necessary to develop prosocial behaviors. Yet, prosociality and patience in children have rarely been studied jointly. Here, we measured patience (i.e. delay-of-gratication) and prosociality (i.e. giving in a dictator game to a known or unknown partner) in 250 4- to 6-year-old French schoolchildren. We found that sharing with an unknown partner was negatively linked to patience in children but observed no relationship between patience and sharing with a familiar partner. Taken together, our results support the hypothesis that children are intuitively prosocial independent of strategic concerns and that patience is therefore not necessary to act prosocially during early childhood. Future studies investigating whether and why prosociality show a non-linear developmental trajectory across the lifespan are warranted.


Introduction
From biological to social sciences, the origins of prosociality such as sharing, cooperating and helping (1, 2) fascinated researchers due to its key role in the functioning of advanced human societies (3). While several developmental studies suggest that patience (4) and social norms are necessary to exert prosocial behaviors (5) others argue that prosociality is an instinctive and impulsive behavior (6, 7). We might thus wonder how prosociality can be the result of both impulsivity and rationality.
Patience (i.e. self-control) and age are positively correlated with prosocial actions. Children who perform better on inhibitory control tasks are more prosocial in cooperative play tasks with peers (8) and give more in sharing games (9). And more generally younger children (3-to 4-year-old), that usually have lower self-control, display more sel sh behaviors than older children (7-to 8-year-old) in sharing games (10,11).
In addition, children better in inhibition as evaluated by parents, are more able to follow a norm with respect to altruistic giving in a dictator game (5) and display more in-group favoritism and egalitarian tendencies (10). Thus, prosociality shaped by norms and social structure seems to increase with age as with cognitive development including greater patience (12).
Importantly however, young children engage spontaneously in prosocial actions such as spontaneous helping (13-15) before the development of patience or the internalization of social norms. Furthermore, both children (16) and adults (6) are more prosocial in social dilemmas when given less time to respond and prompted to act "intuitively" suggesting that intuition supports prosocial impulse but that rational thought can hinder them. The development of certain cognitive abilities related to strategic thinking and reduced temporal discounting might thus have a negative impact on such prosocial impulses.
Consequently, while engaging more and more in rational cognitive processes, prosociality could be reduced (17).
Studying the relationship between prosociality and patience in children will allow us to better understand the impulsive and rational components of prosociality. One important rational competence, developing between ages 4 to 6, is patience (18)(19)(20). Patience in children is traditionally evaluated with delay of grati cation tasks, i.e. a trade-off between a small immediate reward vs. a larger future reward (21)(22)(23).
While the investigation of the development of patience and prosociality is not new, there is a lack of studies that link the development of delay of grati cation to the development of prosocial sharing. Only a limited number of studies looked at how delay of grati cation is related to children's prosocial sharing. A positive relationship between delay of grati cation and reciprocal sharing was reported by Moore & Macgillivray (24) and Koomen et al. (25) in a modi ed delay of grati cation task in which children could wait to provide bene ts to a familiar third party and by Sebastián-Enesco & Warneken (26) where resources obtained in a delay of grati cation task could be shared with a puppet. To our knowledge, no study has yet measured delay of grati cation and prosocial sharing using two independent tasks where decisions regarding delay are independent from sharing decisions, and involving "real" same-age children as partners testing in a realistic environment (i.e. school).
Here, we evaluate whether a relationship can be observed between patience and prosocial sharing in 4-to 6-year-old children. Speci cally, we tested 250 4-to 6-year-old children in a quiet room of their school in three different tasks: 1) a delay of grati cation task involving a choice between one candy now vs. two candies at the end of the testing session, 2) a dictator game with an unknown child from another school, and 3) a dictator game with their self-reported "best-friend". The two dictator games allow us to disentangle whether prosociality is in uenced by strategic considerations.
Giving to unknown individuals can be interpreted as "pure" prosociality and has been extensively evaluated using dictator games (10,27,28). The fact that partners are unknown and the giving anonymous distinguishes these situations from situations involving reciprocal prosociality, where strategic concerns might in uence behavior. Indeed, what might seem to be a prosocial behavior might be motivated by sel sh motives if future bene ts can be expected from the partner. In practice reciprocal prosociality might happen when interacting with known, familiar partners for the sake of future bene ts (11). Our results provide evidence that patience in 4-to 6-year-old children is negatively linked to pure prosociality, however we observed no relationship between patience and reciprocal prosociality.

Results
Overall, 72% of children (180 out of 250) chose to wait for the larger reward in the delay of grati cation task. In line with previous studies, older children delayed more often than younger ones (χ2 = 11.8; P < 0.005) with 59 % of the 4-year-old (47 out of 79), 73% of the 5-year-old (80 out of 109) and 85% of the 6year-old (53 out of 62) delaying grati cation. Since recent papers suggest that delay of grati cation performance can be affected by the motivation for the reward (29,30), motivation for the larger reward was explicitly measured (30) in asking children whether they preferred a small or a large reward when no delay was applied. Motivated children delayed more often than non-motivated children (χ2 = 21.3; P < 0.0001); 85% of children (212 out of 248) selected the larger reward in the motivation task and among the For individuals who chose the larger reward in the motivation task (i.e. motivated children: N = 212), we found that the number of stickers given was signi cantly impacted by performance in delay grati cation task (patient vs. impatient) and partner status (unfamiliar vs. familiar partner) while controlling for parents' income, number of siblings, age in months, sex and protocol order (ordered logistic regression analysis, see Methods and Table 1). Speci cally, patient children gave signi cantly less stickers to unknown partners than impatient children did (Fig. 1).

Discussion
In the present study, we examined the relationship between delay of grati cation and prosociality in 4-to-6-year-old children. Our ndings provide experimental evidence of a negative relationship between patience and the number of stickers given in a dictator game with an unknown partner (i.e. pure prosociality). These ndings corroborate the hypothesis that rather than requiring patience and social norms, pure prosociality is an impulsive response (6) and t with previous ndings reporting that infants are naturally altruistic (31), prosocial responses are more rapid than sel sh ones in adults (6, 25) and prosocial responses do not activate brain regions associated with cognitive control (32).
Importantly however, the negative relationship between patience and prosociality was only found in situations in which children donate stickers to an unknown partner (i.e., pure prosociality) and not with respect to reciprocal prosociality (i.e., giving to a familiar partner). This suggests that the negative relationship between giving to an unknown partner and patience might be linked to strategic and rational considerations. A strategic agent might indeed both decide to wait for the larger reward to maximize the payoff (i.e. be patient) and might also decide to donate some of his/her stickers to a familiar partner (i.e. reciprocal prosociality) but not to an unknown partner (i.e. pure prosociality) to elicit reciprocity from the recipient (10,33).
Indeed, the ability to delay grati cation has been shown to be a rational behavior (18, 19). Rational thinking starts to develop at 5-to 6-year-old and is consistently observed in children over 7 (10,17) who are able to internalize social norms (5) including parochialism (10). Likewise, a recent study reports that preschoolers (4-year-old) were more egalitarian than older children (9-year-old) because older children tend to be more selective depending on the recipient (17). The authors conclude that "with cognitive maturation […] children become less generous as they age" (17) or in other terms, they become more rational and select the partners according to their probability to reciprocate. Our ndings go beyond such results and suggest that children who are developing rational thinking, exert patience and favor reciprocal prosociality (give more to familiar and potentially reciprocal partners than unknown ones) at the expense of pure prosociality (Fig. 1).
Taking the above together suggests an answer to the question of how prosociality can be the result of both impulsivity and rationality. While some cognitive abilities follow a linear developmental trajectory (34), others show non-linear developmental changes (35). These transitions might be driven in the case of prosociality by the development of other abilities, for example patience, rational thinking and social norm internalization.
In conclusion, our ndings provide experimental evidence of a negative link between pure prosociality and patience in children. Developing patience is therefore not necessary to act prosocially but is somehow hindering pure prosociality which might allow parochialism to develop.
This study provides an important and crucial input towards future studies focusing on whether, how and why shifts in prosociality develops across the lifespan and is surely important to consider with respect to educational policies aimed at improving patience (Alan & Ertac, 2018) and social skills in children All parents signed an informed consent form for their children and only children who gave their verbal assent were included. Among the 250 questionnaires sent to parents, 238 were returned allowing us to obtain information about parents' income, number of siblings and native language. Forty-ve percent of children were from low and middle-income backgrounds (0 to 30,000 euros/year). Participants had 1.1 siblings on average: 23% were an only child, 55% had only one sibling, 15% had two siblings and 7% had more than 2 siblings. All children (except one) were native French, and all children were French speaking. Two trained female experimenters tested children individually in a single video-recorded session in an available room at their schools.

Experimental session
The experimenter told children that they would play some games together. The child and the experimenter then entered the testing room (i.e., an available room at their school). Participants completed a series of tasks including a delay of grati cation task and two dictator games in one single 15-min long session.
Participants performed the delay of grati cation task with candies. They had to choose between one candy immediately and two candies available at the end of the testing session. After this, participants performed two dictator games with stickers, one with respect to an unknown child and one with respect to a familiar child. The order of the two dictator games was counterbalanced across participants. After this, children who chose to delay in the delay of grati cation task received their reward. Finally, children were given a single choice between receiving the small versus the large candy reward "now" to evaluate their motivation for the larger versus the smaller reward. In the following we will discuss each task in detail.

Delay of grati cation
For the delay of grati cation task, a cup containing one candy and another one containing two candies, were presented to the child. The experimenter said: "Do you want to have one candy now or do you want to wait until the end of the games to get two candies?" If the child chose the one candy option, he/she received one candy immediately. If the child chose to wait, the experimenter put the two cups away and continued the testing session.
Dictator games The children engaged in two dictator games. The order of the two conditions was counterbalanced across children. Before the dictator games, the experimenter asked the child to name the child he/she liked the most to play with. The named child was the "familiar recipient" in the dictator game with a familiar partner.
In both games, four very attractive stickers were presented to the child. The experimenter asked the child whether he/she liked the stickers, if not the stickers were changed until the child said that he/she liked all the stickers. After this, the experimenter explained to the child that he/she could give some of the stickers to another child. The experimenter explained to the child that this decision would not be observed by the experimenter and demonstrated this by closing her eyes while the child allocated the stickers to two different envelopes.
In the game with an unknown partner, the experimenter explained that the recipient child would be a child from another school, randomly selected and would receive the stickers from another experimenter. It was made explicit that the child would not know who would receive the stickers.
In the game with a familiar partner, the experimenter explained to the child that the recipient would be his/her best friend and that the participating child would hand over the stickers in person. It was made explicit that the recipient was the best friend and that the recipient would know who donated the stickers. Again however, the experimenter explained that she would not know about the decision and demonstrated this by closing her eyes while the child allocated the stickers to the two different envelopes.

Motivation task
Motivation for the larger reward was explicitly measured. For the motivation task, the experimenter said: "If you could have one candy now or two candies now, what would you choose?"

Statistical analyses
All analyses were performed in the R environment for statistical computing version 3.3.6 (Coreteam, 2018).
The variable of interest from the delay of grati cation task was binary, namely either "wait" or "don't wait". Similarly, motivation was evaluated through a single choice between the larger and smaller reward used in the delay of grati cation task. In the dictator games, we observe donations between 0 and 4.
For individuals who chose the larger reward in the motivation task (N = 212), we investigate the relationship between delay of grati cation and prosociality using an ordered linear model (LM; 'polr' function in R). We asked whether the number of stickers given was related to performance (patient vs. impatient) and partner status (familiar vs. unknown) in the delay grati cation task while controlling for parents' income, number of siblings, age in months, sex, and protocol order.
For each xed effect, statistical signi cance was evaluated by likelihood ratio tests of the full model against the same model without the tested xed effect. We report likelihood ratio t-values and P-values.

Declarations
Author Contributions: GBJ and AC designed research, GBJ conducted research and experiments, GJB wrote the rst draft and GBJ and AC wrote the paper.
Competing Interest Statement: We declare no competing interests.
Classi cation: Social Sciences/Economic sciences; Biological Sciences/Psychological and Cognitive Sciences