Download this publication


We report results from a randomized field experiment conducted in the Tigray region of Ethiopia, which tests the impact of interaction with migrants on host community members’ attitudes towards migrants. In three treatment groups, host community members were randomly paired with a migrant from a nearby refugee camp to play an incentivized guessing game. In the first of these treatments, the game was neutral in content, in the second it introduced subtle cues to economic matters, and in the third subtle cues to ethnic identity. In a fourth treatment, host community members were paired with other host community members to play the neutral game, and in the control condition host community members did not interact with anyone. The results show that, compared to the control group, interaction with a migrant significantly improved attitudes towards them. Subtle cues to economic matters or identity did not diminish this effect. However, we see similar effects on attitudes to migrants in the treatment group where hosts interacted with other hosts, which suggests that the effects are driven by human interaction in general, rather than by interacting specifically with a migrant. The effects of interaction are not much affected by the characteristics of the paired hosts and migrants, though host respondents in low skill occupations appear to respond more favourably to the treatments. Interestingly, however, we find no effects of the treatments on how migrants believe they are perceived by host community members.

Ivar Kolstad

Associated Research Professor

Sosina Bezu

Associated Senior Researcher

Päivi Lujala

Affiliated Senior Researcher