Problem Set. Week 2

2e. What is the p-value for , as estimated in Question 2b above? Please enter your answer as a value between 0 and 1, rounded to the nearest four digits after the decimal place (for example: 0.1111)

0.0056

3d. What is the p-value for under the null hypothesis = 0 when we control for weapon fixed effects? Please enter your answer as a value between 0 and 1, rounded to the nearest four digits after the decimal place (for example: 0.1111)

0.0071

3c. Test the null hypothesis that = 0.1 at the 5% level under this specification. Do you reject or fail to reject?

Fail to reject

1c. True or False: The authors attempt to distinguish whether the effects are driven by successful assassinations, failed assassinations, or both in their main results.

False

1b. The most common weapons used in assassination attempts are:

Guns

2b. Regress whether the institution changed from one year before the attempt to one year after the attempt [absnpolity2dummy11] on whether or not the attempt was successful[ [success]. As in the original paper, make sure this is limited to there being a "serious attempt" [seriousattempt]. Fill in the blanks. The interpretation of is that a(n) _________________________ leads to a(n) ______ percentage point increase in the probability of the institution changing from one year before the attempt to one year after the attempt.

successful assassination attempt, 11

1a. The paper identifies assassination attempts on national leaders worldwide from:

1875-2004

2a. Compute the following summary statistics: Hint: Only consider countries that have defined data in 1900. Please enter each response as an integer between 0 and 100

38, 63, 9, 3

1d. According to Table 5, and the absolute change between democracies and autocracies, Jones and Olken (2009) find that transitions between autocracy and democracy are ______ percentage points more likely following the successful assassination of a leader than following a failed attempt.

9

2d. Test the null hypothesis that = 0 at the 5% level. Do you reject or fail to reject the null?

Reject

2c. What assumptions do we need to interpret this as the expected difference in outcomes between when the attempt succeeds and when it fails?

The error term ε must be uncorrelated with the success dummy.

3a. Why would you want to include these controls?

The type of weapon might be correlated with both the success dummy and the outcome variable.

3b. True or False: The interpretation of changes when we control for weapon fixed effects.

True