Thanks for the videos, could you explain how the other probabilities are computed for complier and always-taker at 34:39? More specifically, my question is shouldn't 0.3112 (1-Pr(never-taker)) = Pr(always-taker) + Pr(complier), leading to Pr(always-taker) = 0.187 and Pr(complier) = 0.1242?
@@jayagupta9239 Check the relevant slide. There are four equations and four unknowns. From this system of equations you can back out the share of compliers.
This is the best verbal explanation of LATE anywhere after 6 years of study.
Thanks for the videos, could you explain how the other probabilities are computed for complier and always-taker at 34:39? More specifically, my question is shouldn't 0.3112 (1-Pr(never-taker)) = Pr(always-taker) + Pr(complier), leading to Pr(always-taker) = 0.187 and Pr(complier) = 0.1242?
Hi Akilesh, thanks a lot for spotting this. You are right. I'll change that soon. Regards,
Ben
@@ben_elsner hii, can you please explain how to compute the probability of complier
Please check the post of Akilesh further above. His probabilities are correct.
@@jayagupta9239 Check the relevant slide. There are four equations and four unknowns. From this system of equations you can back out the share of compliers.
33:13
GOAT
I love you after watching this set of videos. If I were a woman, I would want to marry you.
I can give you my dad’s number, perhaps you could talk about the dowry.
Perhaps a typo at 17:14 : The second term in the last line should be E[Y_i0 | never-taker].
These are great lectures, thanks a lot!
I believe you're correct. Thanks for pointing this out. Will correct next week.