Choosing Units of Analysis in Subnational Research: The Modifiable Areal Unit Problem and the Study of Local Violence During Civil War

Hillel Soifer, Temple University
Monday, October 2, 2017 -
12:00pm to 1:30pm
A71 Louis A. Simpson International Building
Monday, October 2, 2017 -
12:00pm to 1:30pm

DemDev/CP Lecture Series

As political scientists, our research often involves characteristics of subnational units, whether as context or as objects of interest in their own right.  In pursuing this type of research, however, scholars are confronted with a thorny research dilemna: they must identify the appropriate unit of analysis. This issue has received strikingly little attention in political science methodology, despite the fact that when one's research involves aggregates, the appropriate unit of analysis is often unclear. This paper suggests that political scientists might benefit from confronting what geographers call the modifiable areal unit problem and the consequences it poses for the validity of much subnational research. The paper sketches the logic underlying this problem, shows its implications for research on local violence using data from Peru's internal conflict, and provides some implications for the design of research that uses spatial units.

Sponsored by
Project on Democracy and Development/Comparative Politics