Part 1: Descriptive Statistics
1. Complete the following table of summary statistics:
Table 1: Summary Statistics
|
Variable
|
Mean
|
Standard Deviation
|
Minimum Value
|
Maximum Value
|
n
|
|
ftmpop
|
|
|
|
|
|
|
gdppc
|
|
|
|
|
|
|
lackpfhigh
|
|
|
|
|
|
|
ethnic
|
|
|
|
|
|
|
religion
|
|
|
|
|
|
2. Based on your background knowledge, do you expect there to be a relationship between a country's level of political freedom and amount of terrorism? If so, why?
3. What is the average value of ftmpop among countries with a low level of political freedoms?
4. What is the average value of ftmpop among countries with a high level of political freedoms?
5. Do these numbers allow you to conclude that countries that there is a causal relationship between a lack of political freedoms and the number of terrorist fatalities that occur in a country? Explain.
Part 2: What is the effect of political freedom on the incidence of terrorism?
6. Complete the regression table (Table 2) below. A dash indicates that the independent variable is not included in the model.In each cell, you should include the coefficient as well as the robust standard error. The robust standard error should be placed inside the parentheses.
7. Using Model 1, test the hypothesis that the coefficient on lackpfhigh is zero against the alternative that it is non-zero at the 5% significance level.State the null and alternative hypotheses as well as the result of the test.
8. Using Model 1, interpret the coefficienton lackpfhighin real-world terms.
9. Does the coefficient on lackpfhigh change between Model 1 and Model 2? What does this imply about the presence of omitted variable bias in Model 1? Explain.
10. What is the sign of the omitted variable bias in Model 1? Explain how you arrived at your conclusion.
11. Using Model 3, test the hypothesis that the coefficient on lackpfhigh is zero against the alternative that it is non-zero at the 5% significance level. Explain why the result of this hypothesis test differs from that in question 7.
12. Using Model 3, calculate the predicted number of terrorist attacks for a country with the following characteristics:
Population: 600 million
Limited political freedoms
Ethnic heterogeneity: .8
Religious heterogeneity: .7
GDP per capita (in thousands of dollars): 4
Region: Asia
13. Using Model 4, explain why Stata drops one of the regional dummy variables when estimating the regression.
14. Using Model 4, interpret the coefficient on lackpfhigh in real-world terms.
15. Interpret the value of the adjusted R2 in Model 4.
16. Which model is best? Why?
17. Using the model you selected in question 16, can we draw a firm conclusion about the causal effect of the lack of political freedom on the incidence of terrorism? Why or why not?
Table 2: The Effect of Political Freedom on Terrorist Fatalities
|
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
|
Dependent
Variable:
|
Terrorist Fatalities
|
Terrorist Fatalities
|
Terrorist Fatalities
|
Terrorist Fatalities
|
|
Lack of Political Freedoms
|
( )
|
( )
|
( )
|
( )
|
|
Per Capita GDP
|
-
|
( )
|
( )
|
( )
|
|
Religious Heterogeneity
|
-
|
-
|
( )
|
( )
|
|
Ethnic Heterogeneity
|
-
|
-
|
( )
|
( )
|
|
Middle East Dummy
|
-
|
-
|
( )
|
( )
|
|
Other Regional Dummies Included (latinam, westeurope, easteurope , eastasia , africa)
|
No
|
No
|
No
|
Yes
|
|
Constant
|
( )
|
( )
|
( )
|
( )
|
|
Regression Summary Statistics:
|
|
R2
|
|
|
|
|
|
Adjusted R2
|
|
|
|
|
|
n
|
|
|
|
|
Notes: Terrorist fatalities are measured using the variable ftmpop. Robust standard errors are given in parentheses under the estimated coefficients. * p< .1, ** p< .05, *** p< .01.
Attachment:- Problem Set.rar