Problem Set 3
Problem 1,6 Pts
A) 3 points Replicate Table 3, Model 4 in Fish (2002) “Islam and Authoritariansism” World Politics 55(1): 4-37. Interpret (Explain what the coefficient estimates mean, pay attention to the possible values of the independent variables, i.e. are they binary, continuous, etc). Use the FishData.dta file, you can find the paper Fish (2002) in the module for week 11.
B) 3 points Now add an interaction term between Islamic religious tradition and economic development. Create a table of the results. Interpret by graphing the Democracy predicted values by economic development for Muslim and Non-Muslim countries, holding OPEC membership constant at its mode. You may want to refer to Lab 6 for code on how to generate this plot.
Problem 2, 6 Pts
This problem uses variables from the 2010 subset of the Quality of Government (QoG) dataset called QoG 2010.dta (see table below for the variables in QoG), as well as World Bank data from 2010 on life expectancy at birth.
Variable Description
cname Country name
fh_cl Freedom House measure of civil liberties
fh_pr Freedom House measure of political rights
fh_status Freedom measure (categorical)
undp_gdp Gross domestic product (GDP) per capita
undp_gini Gini coefficient (economic inequality)
wdi_gris Gender ratio in schools (girls/boys)×100
2 points Read the “QoG 2010.dta” and “lifeexp.csv” data files into R. Merge these two datasets together (hint: use the merge command in R; note that your final merged dataset should have 121 observations. You should be able to merge the datasets using merge(dataframe1,dataframe2)). Plot the relationship between life expectancy and GDP per capita, and include the bivariate OLS regression line in the same plot. (Y = life expectancy, X = GDP).
B) 2 points Is the effect of GDP on life expectancy non-linear? Assess this question by estimating a model with a quadratic GDP term. Present the results of that model in a presentable table, and interpret the coefficients on the GDP variables.
The effect of GDP on life expectancy is non-linear.
C) 2 points Create a variable equal to the average of the civil liberties (fh cl) and political rights (fh pr) scores. Then create two dummy variables: one called ”free” that is equal to one when the average freedom score is less than 3, and zero otherwise; and another called ”unfree” equal to one when the average freedom score is greater than or equal to 3, and zero otherwise. Regress life expectancy on GDP, GDP 2, Gini, and Free. Then estimate the same regression using the Unfree variable instead of the Free variable. Compare the two models and explain the results. What happens when you include both Free and Unfree in the same model?