Janet I. Lewis


My research interests center on political violence, ethnic conflict, and subnational state formation, especially in Sub-Saharan Africa. Recent and ongoing projects include:

How Insurgency Begins

This book project -- along with related articles, including one in Comparative Political Studies and another in International Organization -- examines rebel groups’ formative initial stages, when some nascent groups are defeated while others go on to become viable challengers to central governments. The former was honored with the annual award for Best Article published in CPS. The latter was honored with the annual Alexander L. George Award for best article or chapter developing or applying qualitative methods from the Qualitative and Mixed Methods Research section of APSA.

The central argument of the book manuscript is that rebel groups can form almost anywhere in weak states – and do so more frequently than most research recognizes – and that kinship networks among civilians where rebels launch importantly influence whether nascent rebel groups become viable fighting forces. This is the case because of how kinship networks shape the spread of trusted rumors about nascent rebels.

The project focuses primarily on the 16 rebel groups that operated in Uganda between 1986 and 2006. In addition to making use of census, geographic, and household survey data, it relies on evidence I collected in over 200 interviews with former rebel leaders, government military and intelligence officials operating in urban and rural areas throughout Uganda, as well as local officials and civilians who lived near initial rebel bases. This project is based on my dissertation, which won the Senator Charles Sumner prize from Harvard`s Department of Government as well as the discipline`s top African politics dissertation honor, the Best Dissertation Award from the African Politics Conference Group and Lynne Rienner.

Ethnic Networks

Active research on a wide range of political contexts centers on ethnicity`s role in collective action. Many theories posit that information flows more easily in ethnically homogeneous areas, facilitating collective action, because social networks among co- ethnics are denser. Although this characterization is ubiquitous, little empirical work assesses it. Through a novel field experiment with co-author Jenn Larson in a matched pair of villages in rural Uganda, this project directly examines word-of-mouth information spread.

In a paper in the American Journal of Political Science, we find that (as expected) information spread much more widely in the homogeneous village. However, unexpectedly, the more diverse village`s network is significantly denser. Using unusually detailed network data, we offer an explanation for why network density may hamper information dissemination in heterogeneous areas, showing why even slight hesitation to share information with people from other groups can have large aggregate effects. The paper won Best Article published in AJPS in 2017.

This project provides a rare, direct look at the process of information transmission in rural societies, and why news and behavioral change may spread more easily in less diverse areas. In a working paper (with Jenn and Pedro Rodriguez), we examine why only some people acted on the information they received. In a paper in Political Science Research and Methods, Jenn and I discuss how to collect network data in the field. In the future, we plan to broaden the study to a larger sample of villages.

Building the Subnational State

In a growing number of developing, democratizing countries, the global shift away from centralized political and economic regimes coincided with a rarely noted phenomenon: the proliferation of subnational administrative units. Why has this occurred, and what does this teach us about center-periphery relations and decentralization efforts in African states?

In a 2014 American Political Science Review paper with Guy Grossman, we posit that administrative unit proliferation occurs where and when there is a confluence of interests between the national executive and local citizens and elites from areas that are politically, economically, and ethnically marginalized. We argue further that although the proliferation of administrative units often accompanies or follows far-reaching decentralization reforms, it likely results in a recentralization of power. We find support for these arguments using original data from Uganda. In a paper in Regional & Federal Studies, I elaborate this logic of why administrative unit proliferation can lead previously-decentralizing states to recentralize: by reducing the intergovernmental bargaining power and administrative capacity of each subnational unit, and by substantially expanding both the reach of the national executive`s patronage network and its ability to monitor emergent security threats on its periphery.