Contact's true effect size

For 70 years, a simple idea has shaped efforts to reduce prejudice: put people from different groups together under the right conditions, and contact reduces prejudice. Gordon Allport proposed it in 1954. A landmark 2006 meta-analysis of 515 studies seemed to confirm it, reporting an average effect of 0.4 standard deviations on prejudice measures. That paper has been cited more than 14,000 times. The credibility revolution has undermined this evidence, by correcting for publication bias that meant null results were seldom published. 

Matt Lowe of the Vancouver School of Economics has published a new review of 41 pre-registered studies, and he finds the average effect is one-tenth of a standard deviation. Those 41 pre-registered intergroup contact experiments cover nearly 40,000 participants across a wide range of countries, roughly half of them in the Global South. He tells Tim Phillips that the effects are real, consistently positive … but consistently small. 

Contact interventions are a waste of time. Costs can be low, and the alternatives have not yet been held to the same rigorous standard. But the gap between what the old literature promised and what careful experiments deliver is large enough to matter for anyone designing programmes to reduce prejudice between groups.

The research behind this episode:

Lowe, Matt. 2025. "Has Intergroup Contact Delivered?" Annual Review of Economics 17.

To cite this episode:

Phillips, Tim. 2026. "Has Intergroup Contact Delivered?" VoxDev Talk (podcast). 

Assign this as extra listening: the citation above is formatted and ready for a reading list or VLE.

About Matt Lowe

Matt Lowe is an assistant professor at the Vancouver School of Economics at the University of British Columbia, a CIFAR Azrieli Global Scholar, and a J-PAL faculty affiliate whose research spans intergroup relations, development, and political economy. His website is at mattjlowe.github.io. He has previously been published in VoxDev discussing his field experiment on collaborative and adversarial caste integration through cricket leagues in India.

Research cited in this episode

Allport, Gordon W. 1954. The Nature of Prejudice. Addison-Wesley. The founding text of intergroup contact theory, which proposed that contact between groups reduces prejudice when it meets four conditions: equal status, common goals, intergroup cooperation, and support from authorities.

Pettigrew, Thomas F., and Linda R. Tropp. 2006. "A Meta-Analytic Test of Intergroup Contact Theory." Journal of Personality and Social Psychology 90 (5). The 515-study meta-analysis that established the 0.4 standard deviation benchmark for contact effects and became the dominant reference point for the field.

Paluck, Elizabeth Levy, Roni Porat, Chelsey S. Clark, and Donald P. Green. 2021. "Prejudice Reduction: Progress and Challenges." Annual Review of Psychology 72. A review of 418 experiments on prejudice reduction from 2007 to 2019, identifying troubling signs of publication bias and finding that most studies evaluate light-touch, small-scale interventions with uncertain long-term effects.

Scacco, Alexandra, and Shana S. Warren. 2018. "Can Social Contact Reduce Prejudice and Discrimination? Evidence from a Field Experiment in Nigeria." American Political Science Review 112 (3). A randomised field experiment mixing Christian and Muslim young men in a vocational training programme in Kaduna, Nigeria. Contact reduced discriminatory behaviour but did not change attitudes.

Mousa, Salma. 2020. "Building Social Cohesion between Christians and Muslims through Soccer in Post-ISIS Iraq." Science 369 (6505). Randomly assigned Iraqi Christian displaced persons to football teams with Muslim teammates. Effects were positive on behaviours within the intervention but did not generalise to interactions with Muslim strangers outside it.

Chakraborty, Anujit, Arkadev Ghosh, Matt Lowe, and Gareth Nellis. 2024. "Learning About Outgroups: The Impact of Broad Versus Deep Interactions." SSRN Working Paper. A field experiment in India finding that broad contact (meeting many different outgroup members) corrects misperceptions about outgroups, while deep contact (sustained interaction with one person) builds social and economic ties. Neither type generalises fully to the wider outgroup.

Lowe, Matt. 2021. "Types of Contact: A Field Experiment on Collaborative and Adversarial Caste Integration." American Economic Review 111 (6). Randomly assigned Indian men from different castes to cricket teams or control groups, finding that collaborative contact increased cross-caste friendships and efficiency in trade while adversarial contact reduced them.

More VoxDev Talks on this topic

Promoting national integration in Nigeria: Tim Phillips talks to Oyebola Okunogbe about her research on the Nigerian National Youth Service Corps, which posts university graduates to states other than their own to promote national integration through intergroup contact.

Peacemaking, peacebuilding and post-war reconstruction: Salma Mousa and Lisa Hultman discuss what the evidence shows about building peace and social cohesion after conflict, including which interventions hold up and which do not.

Building social cohesion in ethnically mixed schools: an intervention in Turkey: Sule Alan discusses a programme designed to build cohesion between children from different ethnic backgrounds in Turkish schools, with effects on peer violence, reciprocity, and interethnic friendships.

Related reading on VoxDev

How competition between villages helped divided communities in Indonesia: in ethnically diverse or divided settings, shared efforts towards a collective external goal can help bridge internal divides and build a shared identity.

Reducing prejudice towards forced migrants through perspective taking: evidence on how perspective-taking interventions affect attitudes towards refugees and displaced populations.

How a documentary film fostered interethnic harmony in Bangladesh: a media-based approach to reducing intergroup prejudice, examining what content and delivery can shift attitudes at scale.