We talked about Causal graphical models. This is a tool that makes thinking about causality very transparent. Basic statistical theory alone appears insufficient for understanding the causal concepts. Equipped with this theory, we will see how useful this will be later on.

(Note that there are different views on the DAGs. In case you are interested, under this blog from 2014 http://causality.cs.ucla.edu/blog/index.php/2014/10/27/are-economists-smarter-than-epidemiologists-comments-on-imbenss-recent-paper/ there is an exchange between Judea Pearl (DAGs proponent) and Guido Imbens (Potential Outcomes proponent) in the discussion. Much later (in 2020), here is an interview with Pearl https://www.youtube.com/watch?v=hB9xDcumnHY, with some questions from Imbens. This Imbens' recent essay summarises his views: https://arxiv.org/pdf/1907.07271.pdf  This is not required reading/video, watch only if you want to learn more about the ongoing debate.)

Having some firmer grounds on causal models, we are ready to study causal effects under different circumstances. First we will talk about Randomized controlled trials, the benchmark for causality. Then we turned our attention to situations when using observable information can make the treatment of interest "as good as random". There are different statistical tools how to estimate the causal effects. We saw pros and cons of some of them.

We saw many empirical applications along the way, with some demonstrations in R. 

Note that at the end of each lecture notes, there is an extensive list of references. These are of different levels and depth. Some are useful for grasping the main concepts only, while others may bring you closer to the research frontier. Try to organize your time wisely.