You are planning to estimate whether voter-protection laws increase voter turnout. You note that, in 2015, a lot of new voter-protection laws were enacted in some provinces but not in others. Conveniently, no new laws were enacted in 2012, 2014, or 2016, so you decide to use 2012 and 2014 as your “before” periods and 2016 as “after”. a) Which of the following best describes what you’d want to regress state-and-year level “voter turnout” measures on? i. An indicator for whether the state is treated, and an indicator for whether the year is 2016. ii. A set of fixed effects for state, and a set of fixed effects for year. iii. An indicator for whether the state is treated, a set of fixed effects for year, and an indicator for whether the state is currently treated. iv. A set of fixed effects for state, and for year, and an interaction between “is 2016” and “is a treated state”. v. This design should not be estimated using a regression. b) Unless you chose the final option in the previous question, specify which coefficient in that regression would give you the DID estimate.
You are planning to estimate whether voter-protection laws increase voter turnout. You note that, in 2015, a lot
of new voter-protection laws were enacted in some provinces but not in others. Conveniently, no new laws
were enacted in 2012, 2014, or 2016, so you decide to use 2012 and 2014 as your “before” periods and 2016 as
“after”.
a) Which of the following best describes what you’d want to regress state-and-year level “voter turnout”
measures on?
i. An indicator for whether the state is treated, and an indicator for whether the year is 2016.
ii. A set of fixed effects for state, and a set of fixed effects for year.
iii. An indicator for whether the state is treated, a set of fixed effects for year, and an indicator for
whether the state is currently treated.
iv. A set of fixed effects for state, and for year, and an interaction between “is 2016” and “is a
treated state”.
v. This design should not be estimated using a regression.
b) Unless you chose the final option in the previous question, specify which coefficient in that regression
would give you the DID estimate.
Step by step
Solved in 2 steps