Monday, November 13, 2017

Impact Assessment of Gerrymandering: A Discontinuity-based Causal Inference Approach

Suggested by: Bernd Resch and Luke Miratrix description: Gerrymandering is a well-discussed method to artificially manipulate the boundaries of voting districts in order to achieve political advantage. The gerrymandering process has a spatial component that this master thesis aims to investigate. Concretely, the impact of spatial changes in the voting districts over time on different elections' outcomes shall be researched. Therefore, a method shall be developed based on geospatial regression discontinuity design. The output will quantify the causal effects through analysing treatment effects and their temporal development, compared to, for instance, random or regular spatial patterns.
The master thesis will be carried out together with Harvard University’s Department of Statistics and the Center for Geographic Analysis (CGA).


Johnston, R. (2002) Manipulating Maps and Winning Elections: Measuring the Impact of Malapportionment and Gerrymandering. Political Geography, 21(1), 1-31.

Ding, P., & Miratrix, L. W. (2017) Model-free Causal Inference of Binary Experimental Data. arXiv preprint arXiv:1705.08526.

Start: ASAP

Prerequisites/qualifications: experience with geospatial analysis, geostatistical knowledge may be helpful

No comments: