Potential Outcomes Framework

Question 1

Selection bias

Answer 1

The difference in outcomes in absence of treatment

Question 2

Counterfactual

Answer 2

What would have happened to those that received the treatment if they had not received the treatment and to those that did not receive the treatment if they had received the treatment

Question 3

Fundamental Problem of Causal Inference

Answer 3

Only one of the potential outcomes is observed, you can never know with certainty what would have happened in case you made the other decision

Question 4

ATE

Answer 4

Average Treatment Effect: E [Y(1) - Y(0)]

Question 5

ATT

Answer 5

Average Treatment Effect on the Treated = E [Y(1) - Y(0) | T = 1]

Question 6

ATU

Answer 6

Average Treatment Effect on the Untreated = E [Y(1) - Y(0) | Y = 0]

Question 7

With-and-without comparison

Answer 7

ATT + selection bias = E [Y(1) | T = 1] - E [Y(0) | T = 0]

E [Y(1) | T = 1] - E [Y(0) | T = 1] + E [Y(0) | T = 1] - E [Y(0) | T = 0]