We will practice rapidly reading, understanding and summarizing academic arguments. You should complete the Paper Summary Template for each of the following papers:
a. Maoz and Russett 1993
b. Avelino et al 2005
c. Darden and Gryzmala-Busse 2006
d. Titiunik 2011
For the next five days we will use the study by Titiunik 2011 to help understand what is happening in the data when we make a causal argument/critique.
The research question is “Does incumbency in Brazil improve your performance at the next election?”. The research methodology is a regression discontinuity comparing close winners and losers. Use the Mayoral Election Data from 2000 and Mayoral Election Data from 2004 and need to prepare them for the analysis by following the questions below (but do not do the analysis itself).
What is the data sample used to answer the research question? What alternative data samples might have been used?
Filter both datasets to exlude the data from second-round elections. How does this alter the data sample?
What is the unit of analysis in the dataset? What alternative units of analysis could have been used?
With the 2004 data, calculate the outcome variable, “performance at the next election”. What different measures are available?
With the 2000 data, create a measure of how “close” were the winners and losers to the threshold (of a tied election). What alternative measures could have been used?
With the 2000 data, create a measure of ‘incumbency’ for the period 2000-2004. What are the alternative measures available?
Merge the 2000 and 2004 data into a single dataset. What ‘keys’ do you need to use? What type of join did you perform?
Save your processed dataset! We will use it tomorrow.