For more information.
- If you hold a valid NYU ID, you may attend any of our talks -- no need to request approval!
- If you do not have a valid NYU ID, you must request approval to attend via this form: goo.gl/forms/lNMfPHNnntqCYtSt1
- We can only grant approval to close collaborators of NYU Data Science researchers, researchers at peer academic institutions, and NYU graduate alumni.
- Requests must be submitted by 10AM on Wednesday, 4/17, to be considered in time for this talk.
A key determinant of whether social movements achieve their policy goal is how many people protest. How many people protest is in turn partially determined by protester and state violence, the movement's ability to ally with individuals across social cleavages, and dampening effects from free riding. Technological capabilities and data cost have forced scholars to rely on coarse approximations of size, violence, cleavages, and free riding and analyze their effects in separate models. This paper generates continuous estimates of protest size and the three mechanisms using new techniques in computer vision (convolutional neural networks) applied to geolocated social media images. Twenty three protests across six countries since 2014 are analyzed. Consistent effects for violence and free riding are found: protester violence, high levels of state violence, and the sharing of many photos of large groups decrease protest size, while low levels of state violence and sharing some photos of large groups correlates with larger subsequent protest. Findings for cleavages are less consistent, with age diversity often leading to larger protests and racial diversity sometimes not. The paper ends with a discussion of ethical concerns and improving data collection in order to apply the analysis to poorer or less populous countries.
Zachary C. Steinert-Threlkeld is an assistant professor of public policy at the University of California, Los Angeles’ Luskin School of Public Affairs. He uses computational methods to study protest dynamics, with a particular interest in how social networks affect individuals’ decision to protest. Text analysis has studied mobilization during the Arab Spring; information warfare in Ukraine; and activists’ online strategies. Work with images measures how violence, social cleavages, and free riding affect protest dynamics. Other work includes simulations of protest diffusion and studying how governments attempt to influence individuals’ online behavior.