People are really quick to sort themselves into categories, or social groups, and to form a preference for their in-group. In-group favoritism starts early, and has been found in children across a wide range of categories, including gender, race or ethnicity, language, nationality, and religion. Intuitively, we may think that in-group favoritism develops because the in-group is meaningful. However, almost 50 years of research on less meaningful groups suggests this is surprisingly not the case.
In 1970, the first "minimal group" study was published. Henri Tajfel and colleagues were surprised to find that people gave more resources to their in-group members even when the groups were based on highly superficial dimensions such as the tendency to overestimate or underestimate dot arrays or an interest in abstract art.
Since then, psychologists have shown over and over again that even under the most minimal conditions, people more positively evaluate their in-group members, allocate more resources to them, and hold stronger implicit favoritism towards them. Minimal in-group bias has been found in young children-- even as young as age three-- highlighting the deeply ingrained nature of this bias among humans.
Still, an important question remains: how does minimal group favoritism compare to biases that arise in real groups? You might think that since minimal group biases lack real-world significance, they would be weaker than real group biases. However, the evidence to date on this question has been mixed. While some studies have found that meaningful groups show greater in-group bias, one recent study on 4- to 6-year-olds found that the effects on generosity were similar in their pattern and magnitude despite fundamental differences between two groups (one group involved shared interests and the other group has minimal group membership). The researchers concluded that their findings "highlight the broad impact of affiliation on young children's sharing behavior."
Until very recently, there has been no direct experimental test of the effect of group meaningfulness and in-group favoritism. However, in a hot-off-the press paper in the Journal of Experimental Child Psychology, Xin Yang and Yarrow Dunham experimentally manipulated meaningfulness in novel social groups among 5- to 8-year-olds and measured any resulting in-group biases.
They manipulated the meaningfulness of the groups by having children place their hand on a machine that the researchers told them would assign them to one of two groups (the "green" and "orange" group). They told different stories about how the two machines worked, however. Half of the children were told that the machine could look deep inside them to reveal their true category membership (meaningful group), whereas the other half of the children were told that the machine assigned them randomly to a group (minimal group). Did they find stronger in-group favoritism in the meaningful group or the group who was told they were assigned at random? What do you think they found?
On the one hand, they did find that children reported higher levels of meaningfulness and essentialism in the meaningful group condition. In the meaningful group condition they were more likely to report that their group members would be more likely to share the same hobby, to remain in the same group even if they changed their colored stickers, and they also reported that when they grow up, they expected to have more friends in their in-group than in the other group. This step was important, because it showed that the researchers were able to successfully create groups that varied in meaningfulness.
Nevertheless, and counter to their original prediction, children in both groups held equally strong in-group favoritism despite their differing reports of meaningfulness. In both conditions, children were more likely to like people in their in-group, were more likely to play with other children in their in-group, were more likely to think the other children in their in-group were similar to them, and were also more likely to share resources with other children from their in-group. This finding suggests that "mere membership" is sufficient to bring out strong in-group favoritism, however random and meaningless the group assignment is.
It is possible, however, that the children didn't really comprehend the randomness of the groups. Even though the researchers emphasized that the machine behaved like flipping a coin, children might have thought they had some control over whether a coin lands on heads or tails, or they may have believed that a coin lands on a certain side for some systematic reason.
To address this concern, the researchers conducted another study in which they really pulled the stops to emphasize the meaninglessness of the minimal group membership. First, they used a real quarter coin to help children understand randomness, and gave the following instructions in the minimal group condition:
“When we flip a coin we sometimes get heads and sometimes get tails. So if we just put people in one group if we get heads [demonstrate with the coin] and put them in the other group if we get tails [demonstrate with the coin], this wouldn’t really tell you very much about what people are like.”
They even switched the groups! After the machine assigned the child to a color group, the experimenter told the child that the lab had just run out of materials of that color and that they would instead be assigned to the other color group. The researchers hoped that by emphasizing that the group assignment was arbitrary, unimportant, and highly unlikely to be based on any deep unchanging aspects of the person, the children would realize the shallowness of the group procedure. Pulling all these stops, were they able to demonstrate a difference in in-group favoritism between the two groups?
Yes and no. Going to such great lengths, they were able to find reduced levels of in-group bias on two of the three measures: similarity and preference. Also, they did see some evidence that children in the minimal condition believed that the groups were less stable and more prone to change than did the children in the meaningful group condition.*
Now, the less encouraging news: The researchers were not able to find a significant difference on the resource allocation measure, in which children were asked to distribute 1, 3, or 5 stickers between the in-group and out-group by putting the stickers into green and orange boxes (they were not allowed to keep any stickers for themselves or leave any stickers on the table). Children in both conditions were equally likely to give more stickers to their in-group.
One possible interpretation of this finding according to the researchers is that even in unimportant and arbitrary groups, cooperating more with in-group members is an adaptive strategy because people expect their in-group members to cooperate with them. As prior research has shown, it's these very expectation of reciprocity from in-group members that is a major source of the in-group favoritism demonstrated even in minimal groups.
Taken together, all of this research shows just how little it takes to elicit strong group attachments-- even in randomly assigned social identities-- and just how much it takes to change people's in-group favoritism. As Yang and Dunham note, their findings have relevance to the broader project of understanding the early emergence of prejudice and discrimination. It really does seem as though mere membership in a group is enough to bring out robust in-group favoritism-- competition and scarce resources are not necessary conditions. A really important line of research is the development of interventions that can help children and adults override this deep bias for prejudice, bias, and exclusion-- not only for more meaningful group categories, but also for the many superficial and random groups that exist in the world today.
* Interestingly, they found that children who made more within-group generalizations and saw group membership as more stable demonstrated stronger in-group biases.