Here we consider the main idea behind instrumental variables analysis again, from a different perspective. We also discuss how this method works in practice.
I am a UNC graduate student and I've been trying to grasp this concept from many different sources. This is the clearest of them all. I'm finally confident that I understand what iv is. Thank you professor!
I think there are two points that I am not agree with. One: randomize Instrument variable only remove the arrow from the confounders to Instrument, does not remove the backward dir. Two: Instrument variable does have some causal effect on the outcome, but not directly impact on the outcome. @author
