Here you can find my paper on measuring media effects with a natiral experiment. Specifically, in a small European country I found that there is a border region where people have a signal of a foreign analog TV. After identifing that the signal is not correlated with all available observables, I treat that TV signal avaiability as a quasi-random variable. Under this assumption, I apply a research design of a natiral experiment. I look if the signal of a foreign analog TV had any effects on electoral outcomes during crucial political events.
In this repository, you can find all necessary replication materials: the paper, appendix, R code, and data.