Impulsive Selsh and Contemplated Prosocial Behavior in Children

Cooperation is required for human beings to survive and thrive. In the past decade, to deepen the understanding of human cooperation, more attention has been paid to default prosocial behavior and calculated selsh behavior in the adult population. Whether prosocial behavior is due to an intrinsically altruistic nature or to internalized social norms remains controversial. We approached this question by examining the relationship between cognitive control and decision time, and prosocial behavior in children. We analyzed the data obtained from 226 children aged from 8 to 11 years. The results indicated that greater cognitive control and longer decision time were independently associated with promoted prosocial behavior. That is, the intuitive cooperation model of prosocial behavior established in the adult population may not be supported in children. Although cognitive control was positively associated with prosocial behavior, even children with poor cognitive control promoted prosocial behavior when they were given sucient decision time. Our ndings support the view that prosocial behaviors are automated as habits by the internalization of social norms through lived experiences, and that human beings have an intrinsically selsh nature. the adult population, the intuitive cooperation model of prosocial behavior may be not supported in children. The present study suggests that human beings are naturally disposed to selshness, and a prosocial disposition is acquired through life in society. Also, even children with poor cognitive control can control the natural selshness when they are given sucient decision time.


Introduction
The question of whether human beings are naturally sel sh or prosocial has occupied social scientists for a long time. Over the last decade, social scientists have approached this question by examining the relationship between cognitive control (also referred to as executive function), decision time in social decision-making, and prosocial behavior 1,2 . Previous literature illustrated the nature of intuitive prosocial 3 and heuristic cooperation 4 in humans as evidenced by the shorter prosocial decision time, relative to sel sh decision time 5,6 . Additionally, the prosocial behavior is promoted under cognitive load [7][8][9] and time pressure 5,6 , suggesting that humans have a default prosocial nature but behave sel shly when using cognitive control and su cient decision time. However, since all the above previous studies focused on the adult population, whether the suggested human prosocial nature is true nature, or formed by the internalization of social norms through lived experiences, remains unknown. To address this question, we tested the impact of cognitive control and decision time on prosocial behavior among children.

Results
A total of 226 children completed this study. Participants' demographic data and performance in the anker and social mindfulness tasks are presented in Table 1. The results of multiple regression analyses are summarized in Table 2. The interference score on the anker task re ects the ability of cognitive control (lower score indicates better cognitive control). A socially mindful choice, in the social mindfulness task, re ects on the magnitude of prosocial behavior (higher score indicates greater prosocial behavior). The mean time taken to respond in the social mindfulness task was seen as a measure of decision time. The socially mindful choice was signi cantly associated with interference score and decision time, after controlling for confounders (Table 2). These results indicated that better cognitive control and a longer decision time were associated with a higher proportion of prosocial behavior, and these associations were independent of each other. In the intra-individual level analysis, the decision time of socially mindful choices was signi cantly longer than that of the socially hostile choices (t 286 = 2.01, p = .046, Cohen's d = .15; Table 1). This result indicated that the positive association between the decision time and the probability of prosocial behavior was also observed at the intra-individual level. The interference score was signi cantly positively and negatively associated with decision time when participants made socially mindful choices and socially hostile choices, respectively (Table 2), after controlling for confounders. These results indicated that children with poorer cognitive control need a longer decision times to switch between sel sh and prosocial behavior.  Note: SE = standard error; multicollinearity was inconsequential (variance in ation factor values < 1.6).

Discussion
The present study was designed to understand the relationship between cognitive control and decision time, and prosocial behavior among children. The main ndings showed that better cognitive control and a longer decision time were positively associated with promoted prosocial behavior, and found that the relationship was inverted from results seen in adults [5][6][7][8][9] . Further, the mean proportion of a socially mindful choice (50%) in the present study is smaller than that found in the adult population (60-70%) as examined in previous studies 10,11 . These ndings, coupled with the fact that prosocial behavior increases with advancing age 12 , suggest that the true nature of human beings is sel sh, and previously reported default prosocial nature in an adult population [5][6][7][8][9] is formed through the internalization of social norms through life in society.
The present study also extends previous ndings by exhibiting the independent and interactive relationship between cognitive control and decision time, and prosocial behavior. Previous literature focused either on the impact of cognitive control or that of decision time on prosocial behavior [5][6][7][8][9] . This study provided evidence that better cognitive control and a longer decision time were independently associated with promoted prosocial behavior. In the intra-individual level analysis, prosocial behavior took a longer decision time relative to sel sh behavior, indicating that children could switch their default sel sh behavior to prosocial behavior when taking longer decision time. Additionally, the difference in decision time between prosocial and sel sh behavior was moderated by cognitive control, indicating that children with poorer cognitive control needed a longer decision time to switch their default sel sh behavior to prosocial behavior. These results have the important implication that even children who have poor cognitive control can promote prosocial behavior by taking su cient decision time.
Given the cross-sectional nature of this study which does not manipulate decision time, the present results should be interpreted with caution. It is noteworthy that self-paced decision time can be affected by both the levels of deliberation and decision con ict. Several adult studies manipulated decision time by imposing time pressure 5,6 to minimize the effects of decision con ict, and showed that time pressure promoted prosocial behavior. 13 Based on these ndings of manipulated decision time, they suggested that human adults have a default prosocial nature.
Since the present study only examined the relationship between self-paced decision time and prosocial behavior, we could not deny the possibility that the observed association between prosocial behavior and decision time would have been largely affected by decision con ict. However, given that general decision time (decision time on control trials in the social mindfulness task) was controlled in the regression models, which should re ect the amount of decision con ict, the in uence of decision con ict should have been minimized in the present study. Although the current approach, focusing on children, provides novel insight into the prosocial behavior in human beings, further well-designed studies, such as using manipulation of decision time, are needed to conclusively determine whether human beings have an intrinsically sel sh nature.

Conclusions
Collectively, contrary to previous ndings in the adult population, the intuitive cooperation model of prosocial behavior may be not supported in children. The present study suggests that human beings are naturally disposed to sel shness, and a prosocial disposition is acquired through life in society. Also, even children with poor cognitive control can control the natural sel shness when they are given su cient decision time.

Cognitive control
Cognitive control was evaluated using the modi ed anker task, which has been widely used for this purpose 14 .
The distance between the participants' eyes and the display was approximately 50 cm. In the anker task, the participants were instructed to press either the right or left button as quickly and accurately as possible corresponding to the direction in which the middle sh was facing. This task consisted of two conditions: the target was anked by a non-target stimuli, which corresponded either to the same directional response as the target (congruent condition) or to the opposite direction (incongruent condition). All stimuli were presented in yellow against a blue background on the monitor for 200 ms each, and the stimulus-onset asynchrony was set at 1400- Prosocial behavior and decision time Prosocial behavior and its decision time were evaluated using the social mindfulness task 15 . In this task, the participants were instructed to choose one among four objects in a series of different categories, for example, pens, baseball caps, owers, wrapped gifts, or watches. Children were instructed to keep in mind that they were playing this task together with one other person in a dyadic interaction and to imagine that they both would get to take home one of the four objects. This task consisted of two conditions: two sets of the objects were identical (e.g., two green watches and two blue ones; control trials) or three of the objects were identical, and the fourth only differed in a single aspect (e.g., one green watch and three blue ones; experimental trials). The participants were instructed to

Socioeconomic status
To assess their socioeconomic status, we asked the participants' parents or guardians about their household income (four-point questionnaire ranging from "< 3 million yen" to "> 9 million yen"), and maternal and paternal educational attainments ( ve-point questionnaire ranging from "completed junior high school" to "entered a graduate school"). Socioeconomic status measures were collected from only 217 participants for household income, 222 participants for maternal education, and 221 participants for paternal education, as there were some non-respondents.

Statistical analysis
Three participants were excluded from the data set because their task accuracy was lower than chance (50%) on the anker tasks. To evaluate the relationship between cognitive control and decision time, and prosocial behavior, multiple regression analyses predicting the number of socially mindful choices were conducted. The interference score on the reaction time in the anker task and the decision time on the experimental trials in the social mindfulness task were entered as independent variables. The school, sex, grade, household income, maternal and paternal education, reaction time on the congruent trials, and decision time on the control trials in the social mindfulness task were entered as the confounders. To evaluate the difference in the length of decision time between prosocial and sel sh behaviors at the intra-individual level, a paired t-test was performed to compare the decision time in socially mindful and hostile choices. Then, to test the relationship between cognitive control and decision time in prosocial and sel sh behaviors, multiple regression analyses were carried out to predict the decision times in socially mindful and hostile choices. The interference score on the reaction time in the anker task was entered as an independent variable. The sex, grade, household income, maternal and paternal education, and reaction time on the congruent trials in the anker task, and decision time on the control and experimental trials in the social mindfulness task were entered as the confounders. Six participants were excluded from this intra-individual level analyses because they had no socially hostile choices. Multiple regression analyses were performed with fullinformation maximum likelihood estimation using the lavaan package in the R Studio software, version 1.1.463. All statistical analyses were conducted with α = 0.05.