Experiment studies young children's altruism and their parents' role modeling
Posted September 23, 2017
Whether people are innately good, bad, or amoral, is a much debated question, though perhaps too simplistically posed, for most purposes. Some stripped-down versions of economic theory imply that people are innately selfish and that any indications of altruism or caring for others that might seem to be evident in their behaviors are in one way or another explicable by the selfish long run benefits that they are in fact after. These days, though, at least some economists are taking seriously the possibility that genuine empathy, altruistic impulses, or concerns about others may be part of the mix of explanations for seemingly other-regarding actions, including the donation of tens of billions of dollars to charitable organizations each year.
If wishing to help others is significant and real for at least some people, what determines who has how much of this motivation? To what extent is it innate, to what degree attributable to socialization in the family or to the broader social and cultural environment? Should parents try to encourage benevolence in their children, and are children susceptible to such influence, if offered it?
An opportunity to investigate some of these questions arose when scholars at the University of Chicago began participating in the management of a pre-school program in a low-income neighborhood of the city. Several studies of early childhood have by now been undertaken there, with the helpful control for unobserved influences afforded by the fact that children were admitted to the program by lottery, and that ones judged equally qualified but not admitted were also followed for some purposes. The team has included experimental economist John List of the university’s Economics Department (whose work is known to some readers through reference to it in the book SuperFreakonomics, and by his own book [see note]), List’s then post-doctoral collaborator Anya Samek, and teams of other collaborators who proposed study ideas taken up by the Chicago group.* My frequent collaborator Avner Ben-Ner and I teamed up with List and Samek to design one of the experiments that would engage pre-schoolers and their parents as participants. The goal was to see whether adults would make an effort to model altruistic behavior to their child, if offered the chance, and how easily children are influenced by the adult behaviors in question. Parental and child choices independent of role modeling would also be observed as baselines, permitting similarities and differences to be studied.
Child and adult participants were brought to different rooms and given similar instructions, though in age-appropriate manners. Each participant took part in three or four decision interactions in each of which she or he was linked to a different anonymous individual not at the research site. For each decision, an adult participant was given six dollars, a child participant six stickers from a sticker sheet she had selected according to her own fancy. The decision-maker had to decide each time whether to give any of the six dollars or six stickers to the anonymous, linked individual. Any money or stickers kept were then set aside, and a fresh six dollars or six stickers were brought out for the next interaction with a different individual. Similar adult and child decisions had already been studied by others. What we did differently was that after each adult and child made an initial decision without knowing that the other was making a similar choice and without access to information about what that choice was, the adult was told that their next decision would be shown to a child, with most being truthfully told it would be their own child. Then, after the adults made these sharing or non-sharing choices, the experimenter working with the child informed him or her that they would see, via a simple cartoon-like image, what choice their parent (or in one treatment, an adult other than their parent) had made in a similar situation. The child was then shown an image of hands moving some (or no) dollars from the plate for self to the plate for other, then was paired with a new anonymous, non-present child and given six new stickers (from a different sheet) to divide or simply keep.
What decisions were made? First, most participants chose to share their dollars or stickers with their anonymous counterparts. As List has argued, such behaviors in response to the choice offered by a researcher should not be viewed as very strong evidence of altruism. This is because the situation is somewhat contrived, and despite the researcher telling the participants that their decision will not be traceable to themselves as individual subjects, people feel that their values are being tested. They are also prone to make decisions to help them feel good about themselves, when made self-conscious about it in the experimental situation. Researchers accordingly focus more on variation in the amount shared than on the fact of sharing; for example, do people give more to a poorer recipient, to one of their own ethnic group, or when told there is a chance someone will learn their decision? Do people of society X tend to give more than those of society Y? Do X’s give more to other X’s than to Y’s? And conversely? What about middle school versus elementary school versus pre-school aged children? In the event, very young children have been found to share less than older ones, suggesting that fairness norms tend to emerge or strengthen over time. In our experiment, substantially more pre-schoolers than adults kept everything in their initial interaction (about 7% of adults versus 32% of the children).
Our study yielded mostly “null” findings. In particular, a pre-schooler’s initial sharing decision is not especially correlated with that of his or her parent, and evidence that parents shared more when their decision would be shown to their own child was surprisingly weak. The main positive finding is that children who shared little, especially zero, did tend to increase their sharing after being shown the parent’s choice, if it was a larger amount. However, the level of such emulation was not demonstrably greater when the adult decision shown was that of the parent, than when it was that of another adult. The children might even simply have been mimicking a decision shown them; a follow-up study would be needed to see whether emulation would be any different if the displayed decision were said to be that of another child, for instance.
On another note, the weak or absent indication that the adults engaged in role modeling could partly be due to the fact that they were informed that even those of their decisions that would not be shown to their own child could be shown to another child, something that might already have put them “on better behavior.” Also, an exception to the null role modeling finding is that parents of those children who were unusually generous in their own initial choices tended to share more in a decision that would be shown to their child. Possibly the parents had engaged in some role modeling or talking about such things with the child on earlier occasions, with real effect, and the parents wanted to live up to the standard they had set, when the child could “watch.” Alternatively, some of these parents may have known their child to be generously inclined, and did not want their own behavior to disappoint the child. They might even have felt embarrassed into being more generous in front of their generous child.
All of these results are derived from a relatively small set of observations, so I would argue that the study is more important for raising the issue of whether and how parental influence affects the altruistic or fairness inclinations of children, and what examples parents want to set in these respects, than for obtaining definitive answers. Earlier economists, especially the Nobel laureate Gary Becker (also long a Chicago faculty member), had investigated parental efforts to impart altruism in their children almost exclusively from the vantage point of the parent trying to induce their child to care for the parent in old age. But as stated above, more economists are now prepared to accept that altruistic impulses are relatively widespread. That many very young children may have a natural desire to help others, e.g. an adult “confederate” of an experimenter, has been demonstrated rather convincingly by Warneken and Tomasello (2009). But there remains a myriad of questions to be addressed, including what values parents should want to help their children develop, even if only from the standpoint of their children’s long-term well-being as members of society. If dispositions to help are generally beneficial to the functioning of societies, and more importantly if helpful individuals tend in the long run to be rewarded for it by attracting more cooperative interaction partners (including friends), then it seems likely that parents will do their part to nurture such impulses, especially if the seeds of those impulses are already present in the young child’s make-up.
* The pre-school itself is described in a chapter of List’s book with Uri Gneezy, The Why Axis, 2013.
Avner Ben-Ner, John List, Louis Putterman and Anya Samek, “Learned Generosity? An Artefactual Field Experiment with Parents and their Children,” Journal of Economic Behavior and Organization (in press).
Uri Gneezy and John List, The Why Axis (2013).
Felix Warneken and Michale Tomasello, “The Roots of Human Altruism,” Br. J. of Psychology (2009).