People are social, living life as part of a rich social network.  Friends and family can be sources of pressure- sometimes subtle, sometimes less so- to think and do certain things.  Our study was designed to understand how this social pressure shapes feelings towards members of other groups (so-called "out-groups") and to see if it can be harnessed to improve relations. 

We conducted a field experiment in rural Uganda, a setting in which whole communities' social networks could be traced, that attempted to warm attitudes towards refugees that had arrived from Democratic Republic of Congo and South Sudan.  Uganda hosts one of the top five largest refugee populations in the world, and the largest in Sub-Saharan Africa. As for all countries that host large migrant populations, in Uganda the refugee influx has led to some tension, and even occasional instances of violence.  Research has shown that "perspective-taking" exercises in which a person puts themselves in the shoes of someone from the out-group can warm attitudes towards them.  We used similar exercises with new features which, combined with our measures of word-of-mouth social networks in these communities, revealed important findings.

First, our results show that a perspective-taking intervention shown to reduce prejudice towards outgroups in the US does also reduce prejudice against refugees in Uganda. Average effects are sizeable and endure for at least two weeks.  This is important not only for learning that this kind of intervention can work in Uganda, but also for learning that it can work in areas that are rural and that have scarcer resources. 

Second, our results show that perspective-taking exercises work better when they are conducted in groups than when they are one-on-one.  In fact, the more people who are present during the perspective-taking discussion, the more impactful it is at warming the attitudes of the participants.  It appears that this is because community-level treatments are especially impactful at changing beliefs about what attitudes others in the community hold.

Finally, our results confirm that social networks matter for how well interventions such as these work in the longer-term.  While attitudes warm in the short-term in predictable ways, in the long term, spillovers through networks make things more complicated.  These interventions kick off a social process where people vet, deliberate, and update their views based on others in the village social network.  The people whose attitudes warm the most from our intervention in the long-term are those whose contacts in the social networks most support the viewpoint of the intervention.  Our study also warmed the attutdes of others in the community who were not themselves part of the intervention because they wound up beign included in the social processing that took place afterwards.

In short, our findings demonstrate that social processing is an important part of what we typically observe as individual-level attitude changes.  While individuals react on their own right away, their ultimate attitude after our study depends on an active round of social processing with members of their community to land on an updated attitude. 

Our study has resulted in two journal publications, three articles in preparation, and a book proposal.  It has also generated a large dataset that can be used by other scholars which contains rich information about village life, social networks, interactions with out-groups, and feelings towards refugees.  This dataset and the tools built around it are also useful for instruction in statistics, research design, and the newer field of social network analysis.

Last Modified: 12/19/2024
Modified by: Jennifer Larson