Part 2

Question 1

ATE, ATET definitions

Answer 1

p9

Question 2

QTE definition

Answer 2

p9 assumpton!

Question 3

DiNardo Lee criteria

Answer 3

behavioural models consist, quality descrip assign mech, testability of implications

Question 4

Sample means ATE randomized experim

Answer 4

p10 ate = atet

Question 5

Regression ATE randomized experiments ZCM

Answer 5

need for beta, for delta only need E |X = E |X, D

Question 6

Bad controls

Answer 6

not pre-determined, affected by treatment

Question 7

Heteo ATE experiments

Answer 7

E Y|X - EY|X, D indep X ate=atet. Stratify sample or OLS formula p10

Question 8

MDE formula

Answer 8

p11

Question 9

Budget MDE

Answer 9

p/1+p = sqrt cost data/cost data+ cost treatment

Question 10

Dinardo random experime

Answer 10

know treat, no form restrict, testable ex X balanced

Question 11

Threats random ex and sutva

Answer 11

Hawthorne, spillover, substitution. SUTVA participating not affect potential outcomes

Question 12

Switching regression potential outcomes vs IV

Answer 12

p12

Question 13

Monotonicity

Answer 13

D_i(1) >= D_i (0)

Question 14

LATE

Answer 14

p13

Question 15

Dinardo LATE

Answer 15

known ass mech, consist beh but not testable excl restri

Question 16

Counterfactual distributions and means

Answer 16

p14

Question 17

Panel DiD pros

Answer 17

treatment continuous, SE, time-varying regressors

Question 18

Ashenfelter dip

Answer 18

select bc doing bad randomly, break str exog

Question 19

Prove str exog DiD

Answer 19

compare pre-interv outcomes conditional on FE, placebo treatments randomly

Question 20

Event study specification

Answer 20

p16, how tr eff changes after intervention, excludes never treated. Still homo tr eff across groups

Question 21

MTE(S)

Answer 21

p17

Question 22

McCrary test

Answer 22

density of S around Sbar

Question 23

Treatment effect regr kink

Answer 23

last

Question 24

Causal effect kink

Answer 24

dealt2-delt1/gamma2-gamma1

Question 25

Issue reg disc

Answer 25

check bunching (less with kink), choice poly, choice bandwidth Cattaneo stata min MSE

Question 26

Dinardo RD

Answer 26

consistent bc few param ass, info on, testable balance cov, continuity ass variabel