Research Projects

Funding:   Harry Frank Guggenheim Foundation

My book manuscript explores why individuals choose to participate in ethnic riots in contemporary Nigeria. The rich existing literature on ethnic conflict focuses heavily on top-down, elite-centered processes, which leaves us with few answers as to why ordinary people would follow their leaders and voluntarily engage in actions that are often fraught with extreme risk. The answer I give in my book manuscript is that the interaction between poverty and neighborhood-level social networks dramatically increases the likelihood of riot participation. While poverty may increase a person's willingness to riot, it is centrality in certain types of social networks that transforms potential into actual rioters.

To provide evidence for this narrative, I conducted an original survey of over 800 rioters and non-rioters in two cities in northern Nigeria, Kaduna and Jos, both of which have experienced riots in the past. I also completed a survey of all 70 neighborhood chiefs in the two cities; absentee surveys with family members of those randomly sampled individuals who had died or moved away; and 40 in-depth interviews with riot participants and riot organizers, which are used to contextualize and interpret survey findings. This fieldwork was completed in the fall of 2007 and the summer of 2008.
 

Coinvestigators:  
Bernd Beber (WZB/RWI), Macartan Humphreys (WZB/Columbia), and Dean Yang (U Michigan)

Funding:  
WZB Berlin Social Science Center; Innovations for Poverty Action (IPA) Peace and Recovery Full Study Grant

At the end of 2018, I implemented (with Bernd Beber) qualitative interviews and a representative survey in Benin City, the epicenter of irregular migration out of Nigeria, itself the single largest sub-Saharan African source of irregular migrants to Europe. We find in a first paper, which is to be submitted soon, that information campaigns appear to make false assumptions about potential migrants’ state of knowledge.

Last year we then participated in evaluating outreach activities of the UK-funded anti-trafficking Not for Sale program in Nigeria’s Edo state (with Florian Foos), and I am now the lead investigator for a large-scale randomized controlled trial in Edo and Delta states, which includes door-to-door information provision about risks along the journey to Europe. Baseline surveys were conducted in the fall of 2019 and in March of this year. Activities are now paused due to the coronavirus, except for a phone survey centered on corona-related issues.
 

Coinvestigator:   Shana Warren (IPA)

Funding:
United States Institute of Peace (USIP) Grants Competition
NYU University Research Challenge Fund (URCF)

Status:
Core field activities completed in 2014 and 2015. First paper published in the American Political Science Review, second paper using follow-up information under review.

We conducted an education-based field experiment—the Urban Youth Vocational Training (UYVT) course—to test whether sustained contact between members of groups in conflict can reduce prejudice and discrimination, increase cooperation, and limit support for violence. The UYVT program took place in the fall of 2014 in Kaduna in northern Nigeria. Kaduna has experienced deadly Christian-Muslim clashes over the past decade and has recently suffered from several brutal Boko Haram attacks.

Subjects were drawn randomly from the city's most conflict-prone neighborhoods, participation in the UYVT program was offered to a random sample of 450 respondents, and participants were randomly assigned to religiously heterogeneous or homogeneous classrooms and study pairs. We find no meaningful changes in prejudice but a significant and robust reduction in discriminatory behaviors among participants assigned to the social contact treatments. To our knowledge, this is the first time social contact theory has been tested using empirically rigorous field methodology in an ongoing conflict environment, despite the fact that intergroup contact and engagement projects are one of the international community's most popular policy tools in addressing interethnic strife.
 

Coinvestigators:   Bernd Beber (WZB/RWI), Philip Roessler (William & Mary)

Funding:
UNHCR support for research in Sudan and South Sudan
Additional funding from the British Academy

Status:
Fieldwork completed in 2010 to 2012. First paper published in the Journal of Politics. Additional papers complete and soon to be submitted for review.

In Sudan and South Sudan, we conducted a panel survey of about 1,400 individuals and additional qualitative interviews around the time of the partition of the country. We successfully tracked about 80% of our contacts, despite significant partition-induced relocation among peripheral minorities. The first of several papers to have resulted from this project discusses the effects of violence on public opinion. Some of our northern Sudanese respondents were affected by riots that swept through Khartoum in 2005 while other similar respondents were not, and we find and discuss that those exposed to violence are more likely to support concessions in response to separatist demands, in the expectation that doing so could improve their personal security.

A second paper analyzes the migration decisions of South Sudanese and argues that “middle class” households are particularly likely to try to stay in relatively more developed northern Sudan. Compared to others, these households are especially dependent on access to the kind of economic opportunities that South Sudan does not adequately provide. They try to remain in Sudan, despite the risk of hostile actions being taken by either the government in Khartoum or ordinary non-Southern citizens. This also has implications for South Sudan and similarly situated countries, in that they are less likely to be able to build the kind of vibrant “middle class” that is generally associated with improvements in development outcomes.

A third paper investigates partition-related changes in self-identification, in particular among minorities remaining in northern Sudan, and finds that subject-reported identity attributes are surprisingly responsive to economic and security pressures. This affects both the activation and ranking of preexisting identity components (e.g. whether subjects prioritize their tribal or national identities) and the adoption of entirely new characteristics (e.g. Christians self-representing as Muslims). We show that some identity markers (such as language) are more malleable than others (such as religion), but overall the evidence suggests that subjects seek to adapt as necessary in the pursuit of security and well-being.
 

Coinvestigator:   Salma Mousa (Stanford)

Funding:
IPA Peace and Recovery Full Study Grant
Society Needs Science Grant

Status:   Pilot intervention in progress (currently paused)

We are implementing a field experiment in Lebanon that brings together hundreds of Syrian refugees and Lebanese youths for several months of soccer practice and league play in order to gauge the effect of such social contact on intergroup prejudice. We vary the composition of teams, and whether auxiliary programming focuses on reconciliation or health issues. In addition to baseline and endline survey data, we also collect qualitative information to carefully trace potential causal mechanisms. While the contact theory literature is voluminous, very few systematic empirical studies address contact between migrants or refugees and members of host communities, and fewer still can separate the effects of social contact from different types of programming. Our study is groundbreaking in this regard. A pilot intervention and data collection effort have been underway, but are currently paused due to the coronavirus.
 

Coinvestigators:
Manuel Bosancianu (WZB), Hanno Hilbig (WZB/Harvard), Macartan Humphreys (WZB/Columbia), Nils Lieber (WZB/Bayreuth), Sampada KC (WZB)

Status:   Working paper currently in circulation

We (i) identify national level political and social measures that past research suggest can help explain variation in a society’s ability to respond to adverse shocks and (ii) report on the evolving covariance between these measures and the cross national distribution of COVID-19 burdens.