Question

1) As discussed in Economics Video 1.2, the "fundamental problem of causal inference" is that

a) random assignment into treatment and control groups cannot always be done.

b) the counterfactual for the treated cannot be observed.

c) results from controlled experiments cannot be generalized.

d) correlation is not causation.

2) Suppose that reducing class size does not cause student performance to improve. Which of the following are possible reasons why an observed negative correlation between class size and student test performance may nonetheless be observed?
1. Schools that have more resources also have smaller classes.
2. Experienced teachers choose to take smaller classes.
3. Schools may assign low performing students to small classes.

a) 2 and 3 only.

b) 1, 2 and 3.

c) 1 only.

d) 3 only.

e) 1 and 2 only.

3)

	Before Campaign	After Campaign
The Deck (Advertising Campaign)	25%	54%
Techno Edge (No Advertising Campaign)	28%	42%

The table above shows the percentage of canteen customers purchasing healthier choice meals. There was an advertising campaign held in the surroundings of The Deck, while no advertising was done near Techno Edge. Assuming that the common trend assumption holds, what is the best estimate for the effect of the advertising campaign on the percentage of customers purchasing healthier choice meals?

a) +14%

b) +29%

c) +15%

d) +12%

4) Which of the following is true of the Tennessee Project STAR study described in Wheelan (2013)? (Required Reading: Naked Statistics - Chapter 13 Program Evaluation, by Charles Wheelan)

a) The study is an example of regression discontinuity.

b) The identification assumption was affected by intergroup contamination and dropouts, etc., but research findings were still convincing due to a large sample size.

c) The study is a natural experiment.

d) The study found that black students in small classes performed 0.15 standard deviations better than all students in regular-size classes.

5) In Wheelan (2013), what is the unobserved counterfactual to the treatment group in the Klick and Tabarrok study that investigates the effect of increasing police presence on street crime? (Required Reading: Naked Statistics - Chapter 13 Program Evaluation, by Charles Wheelan)

a) The crime rate on Yellow alert days if the alert was Orange instead.

b) The crime rate on Orange alert days.

c) The crime rate on Orange alert days if the alert was Yellow instead.

d) The crime rate on Yellow alert days.

6) In Wheelan (2013), what is the control group in Hjalmarsson's study on the effect of using jail sentences on juvenile criminal offenders to deter future criminal convictions? (Required Reading: Naked Statistics - Chapter 13 Program Evaluation, by Charles Wheelan)

a) Juvenile offenders who were sent to jail.

b) Juvenile offenders who fall just above the threshold for a jail sentence, and were therefore sent to jail.

c) Juvenile offenders who were given probation or fined.

d) Juvenile offenders who fall just below the threshold for a jail sentence, and were therefore given probation or fined.

7) John and Nidhi play a game in which each person chooses to show one side of a coin, i.e. either Heads or Tails. The payoff matrix is as follows:

		Nidhi
		Heads	Tails
John	Heads	4 , 4	2 , 5
	Tails	5 , 2	1 , 1

Reminder: the first number in a cell is John's payoff; the second number is Nidhi's payoff.
Which of the following statements is correct?

a) If both choose Tails, neither player will change his/her choice, given the choice of the other player.

b) (Tails, Heads) is a Nash Equilibrium.

c) Nidhi would prefer to choose Heads if she knows John will play Heads.

d) There is no Nash Equilibrium.

8) Models are necessarily

a) unscientific, because they can never be used to make predictions.

b) simplified, because the real world has too much detail for humans to handle.

c) mathematical, so that they can be called scientific.

d) written down in words, so that others can access them.

9) In the reading "The Evolution of Trust", if the Game of Trust is played once (Required Reading: Nicky Case’s Evolution of Trust, https://ncase.me/trust/)

a) Both Cheat and Cooperate are best moves, no matter what the other player does.

b) There are no moves that are the best, no matter what the other player does.

c) Cheat is the best move for each player, no matter what the other player does.

d) Cooperate is the best move for each player, no matter what the other player does.

10) In the reading "The Evolution of Trust', the payoff matrix for the Game of Trust is (Required Reading: Nicky Case’s Evolution of Trust​​​​​​​, https://ncase.me/trust/)

a)

		Other Player
		Cooperate	Cheat
You	Cooperate	2 , 2	3 , -1
	Cheat	-1 , 3	0 , 0

b)

		Other Player
		Cooperate	Cheat
You	Cooperate	3 , 3	-1 , 3
	Cheat	3 , -1	1 , 1

c)

		Other Player
		Cooperate	Cheat
You	Cooperate	2 , 2	2 , 4
	Cheat	4 , 2	6 , 6

d)

		Other Player
		Cooperate	Cheat
You	Cooperate	2 , 2	-1 , 3
	Cheat	3 , -1	0 , 0

11) Which of the following statements is true in all cases of the tournament in “The Evolution of Trust”? (Required Reading: Nicky Case’s Evolution of Trust​​​​​​​, https://ncase.me/trust/)

a) “Always cheat” is a superior strategy to “Copycat”

b) “Copycat” is a superior strategy to “Always cheat”

c) None of the other options is correct

d) “Copykitten” is a superior strategy to “Copycat”

12) Refer to the reading "The Evolution of Trust". Consider the Repeated Tournament in Section 5: the evolution of distrust. The tournament begins with 23 Always Cooperates, 1 Copycat and 1 Always Cheat, with 10 rounds per match, and zero chance a player makes a mistake each round. Suppose now if the two players both cooperate, each gets only one coin rather than the original two coins. What happens to the result of the Repeated Tournament?  (Required Reading: Nicky Case’s Evolution of Trust​​​​​​​, https://ncase.me/trust/)

a) Instead of Copycats replacing everyone else, now Copycats coexist with Always Cooperates.

b) Instead of Copycats replacing everyone else, now Always Cheats replace everyone else.

c) Instead of Always Cheat replacing everyone else, now Always Cheats coexist with Copycats.

d) Always Cheat replaced everyone previously, and still replaces everyone now.

*PLEASE ANSWER ALL THE ABOVE QUESTIONS & TYPE YOUR ANSWER (NO SCREENSHOTS OR IMAGES) IN OPTION (A), (B), (C), (D) & (E), WITH SHORT EXPLANATION. THANK YOU IN ADVANCE

Answer 1

1 b) the counterfactual for the treated cannot be observed

(The fundamental problem of casual inference is we can only observe potential outcome of one possible treatment,the other remains missing)

2 d) 3 only.

(If low performing students are assigned to small classes,negative correlation between class size and student test performance may nonetheless be observed.They can be monitored throughly)

3 b) +29%

(Consumer behaviour depends upon the marketing strategy)

4 c) The study is a natural experiment.

(natural observation which happened eventually)

5 a) The crime rate on Yellow alert days if the alert was Orange instead

(As per observation the crime rate was pretty high which happened in mere presence of police)

6 b) Juvenile offenders who fall just above the threshold for a jail sentence, and were therefore sent to jail.

(the offeneders were observed relation between jail and educational attainment)

7 c) Nidhi would prefer to choose Heads if she knows John will play Heads.

(asthe matrix value is 4)

8 b) simplified, because the real world has too much detail for humans to handle

(simplified version of reality that can be easily observed)

9 d) Cooperate is the best move for each player, no matter what the other player does.

(as per the game version)

10

c)

		Other Player
		Cooperate	Cheat
You	Cooperate	2 , 2	2 , 4
Cheat	4 , 2	6 , 6

11 c) None of the other options is correct

(As per practical expweience)

12 d) Always Cheat replaced everyone previously, and still replaces everyone now.