This week we focus on demonstrating mechanisms of cause and effect, and expressing those mechanisms quantitatively. We want to become more intentional in interrogating data precisely and relentlessly, and to do that we spend some time with Fisher’s suggestion to “make your theories elaborate”.
Read “Elaborate theories,” chapter 7 from Paul Rosenbaum’s beautiful book.1 Rosenbaum (2017) Observation and Experiment, Harvard University Press. (When you have some time, reading the entire book is highly recommended).
Watch the first 20 minutes of the video of Dylan Small’s talk’s in the Online Causal Inference Seminar2 If you visit the webpage of this seminar, you can see videos of prior talks. There is also one by Paul Rosenbaum, so you can put a face to the author of the pages you read.. Watching the entire video is of course also an option—we are focusing on the first 20 minutes only because they beautifully introduce the topic, without getting technical.
Figure out the relationship between Paul Rosenbaum, Dylan Small and the members of the DSSG team.
Now return to the COVID-19 pandemic. Read the document Explaining the pandemic and complete the task there described. Be prepared to share your work on wednesday morning.
[optional] Another spirited and inspiring reading related to this week topic is “The ladder of causation”, chapter 1 of Judea Pearl’s book3 Pearl & Mackenzie (2018) The book of why, Basic Books. My amazon page for the book contains an ample preview. Pearl is definitely not a statistician and he can be quite polemic, but he forcefully argues in favor of the importance of causal reasoning. He has an interesting relation with AI and the role he envisions for imagination and coding (more flashed out in the last chapter of the book) could contribute to spirited discussions with Mike S. in and out of slack…