2.1 - What are Potential Outcomes?

In this part of the Introduction to Causal Inference course, we outline week 2's lecture and walk through what potential outcomes are. Please post questions in the YouTube comments section.

[maryammoradbeigi]

I just finished the introduction and they were incredible. Thank you so much. I am so looking forward to learning more

[AT-igtd]

Thank you so much. your course is excellent. i found your website a few days ago and i was so happy because this also the topic of my interest. i found it quite confusing with Yi(T=1) and Y|T=1 when i read some papers because of a lack of basic knowledge in the causal inference. but after reading first two chapters in your textbook, everything is now clear. Thanks again.

[niharamzan]

I have a confusion here, would be grateful if you could please clarify it. If I don't take the pill and still my headache goes away, how does it ensure that taking the pill does not have a causal effect on the going away of the headache? Causal Relations don't imply that only the cause being considered is supposed to affect the outcome of interest, there can be many other causes as well that could have a causal effect on the outcome of interest, but we might not be interested in those causes, in the underlying experiment. In my opinion, we can only say, that taking a pill does not have any causal effect on going away of the headache if we encounter some cases in which even taking the pill, doesn't let our headache go away. Will be waiting for your reply. Thank you.

[emanalsuradi]

Why do we call potential outcomes "potential"?

[tetelim]

Thank you so much, it's incredabliy clear

[chancychan]

I think I better ask the question back here hh. So if my treatment value can possibly take 0, 1, 2, ...n(take n pills) and my outcome value can take 0, 1, 2, ..., m(m+1 health levels). Under this situation what is the definition of potential outcome? And when my treatment and outcome become continuous, say T belongs [0, 2] and Y belongs [-1, 1], what's the definition of potential outcome?

[santoshpalaskar]

What are the methods to draw casual graphs