Politics Thesis

Thu Apr 23 2020

tags: public ppe_work pdf

tl;dr summary here Thesis on GitHub here

Gibbs Prize for best thesis.

Very proud of this work. I spent almost an entire year, and hundreds of hours on this. It couldn't have been done without the time and effort Bassel, Filip, Martin, Tak Huen, Sergi, Andy, Jonathan Rodden, Nick Eubank, and Prof. Stephanopoulos lavished on me.

Comments from markers here and here.

Comments from Stephanopoulos (via email):

Hi Zhenghong,

Just a quick note that I read, and was very impressed by, your thesis. I think it's a remarkable piece of work: one of the first pieces to explore the tradeoffs between different redistricting goals using simulations. If I can assist in any way down the road, please let me know.


On Thu, May 7, 2020 at 3:56 PM Nicholas Stephanopoulos nicholas.stephanopoulos@gmail.com wrote:

Hi Zhenghong,

Nice to hear from you again! I'm delighted that my spatial diversity work was useful, and I'm looking forward to reading your thesis. I love the idea of looking at how different redistricting criteria are related to one another in hypothetical maps. Hope you're managing okay in this crazy time.


On Thu, May 7, 2020 at 10:41 AM Lieu Zheng Hong lieuzhenghong@gmail.com wrote:

Dear Prof Stephanopoulos,

I hope you and your loved ones are well during these unprecedented times.

I have recently submitted my undergraduate thesis. In it I examine whether or not there is a trade-off between compactness and communities of interest when drawing districts, both of which are mandated by legislatures in the U.S. I operationalise communities of interest with your measure of spatial diversity. I draw hundreds of thousands of counterfactual plans, meant to represent the set of plans an unbiased districting commission would conceivably draw.

I find the following results:

1. Districts' spatial diversity varies a lot between states (your finding), but not so much within states. Each state occupies only a narrow band in the range of possible spatial diversity scores, no matter how we draw the districts. 2. Small urban districts are usually more spatially diverse. 3. Whether or not there is a trade-off between compactness and spatial diversity depends on the compactness measure we choose (Polsby-Popper, Convex Hull, Reock, etc..). Some compactness measures trade-off spatial diversity with compactness, while others synergise well with compactness.

I've attached it here in case you wanted to take a look. Comments are most welcomed, of course.

Without exaggeration, I couldn't have done it without your measure and your data. The spatial diversity measure forms the bedrock of all analyses---without the measure, there would be no thesis. So please accept my sincere gratitude!

Thanks very much and take care, Zhenghong Lieu