The Science and Art of Causality (part 2)
There are no test to prove that your effect is causal. Despite this challenge, we can actually do a lot to find evidence, eventually leading to the proof of a causal effect. In this article, I will show you how we could proceed with a fascinating research paper: 'LONDON FOG: A CENTURY OF POLLUTION AND MORTALITY, 1866-1965'^1
Put you in the shoes of a police detective, because that's exactly what we are doing when we want to find a causal effect. We have a suspect. But when we look closer, was it really colonel mustard with the candlestick in the conservatory? Are you sure it was not with a wrench or the crime done by someone else?
In addition to the RDD, I wanted to show you also event study. We will use the data from Hanlon (2018): LONDON FOG: A CENTURY OF POLLUTION AND MORTALITY, 1866-1965. London was a densly habited and heavily polluted area already during the 19th century. Fog reports are consistent through the century and allow to capture day where the air was heavily polluted (fog keep the pollution low).
Data: https://dataverse.harvard.edu/dataset... Paper: https://www.nber.org/papers/w24488
Subscribe and click the bell!
------------------------------------------------------------
Join the discussion:
LinkedIn: / quentin-gallea-phd-74641b56
Instagram: / stats_with_quentin
TikTok: / stats_with_quentin