** Change the file path "../MillerData.dta" to match location on your computer or use
use "http://keitheleejr.com/research/SSQ2016/MillerData.dta", clear
set more off

*Replicate Miller Table 2 to set up same(similar) modeling.
*Note: Cooeficients are same but rsquare values different. This is due to different statistical software. 

xtreg logvote logrecord, vce(cluster city1) fe
xtreg logvote logrecord logincvotelag mayorterm teamenter teamexit, vce(cluster city1) fe
xtreg logvote logrecord logincvotelag mayorterm teamenter teamexit unemp, vce(cluster city1) fe
xtreg logvote playoffs_avg logincvotelag mayorterm teamenter teamexit, vce(cluster city1) fe
xtreg logvote playoffs_avg logincvotelag mayorterm teamenter teamexit unemp, vce(cluster city1) fe

*Table 1: Replicates Miller data without clustering/fixed effects due to dummy variable and extends it by adding south and south*logrecord variables.
xtreg logvote logrecord logincvotelag mayorterm teamenter teamexit unemp south slr
xtreg logvote logrecord logincvotelag mayorterm teamenter teamexit south slr
xtreg logvote logincvotelag mayorterm teamenter teamexit unemp south slr playoffs_avg
xtreg logvote logincvotelag mayorterm teamenter teamexit south slr playoffs_avg

*Table 2: Extends Miller data  and the previous table by looking at football only.

xtreg logvote football logincvotelag mayorterm teamenter teamexit unemp south sfb
xtreg logvote football logincvotelag mayorterm teamenter teamexit south sfb
xtreg logvote football logincvotelag mayorterm teamenter teamexit unemp south sfb playoffs_avg
xtreg logvote football logincvotelag mayorterm teamenter teamexit south sfb playoffs_avg

** Use LBL Dataset to Extend Miller and Healy et al. to look at CFB
** Change the file path "../LBL Data.dta" to match location on your computer or use

use "http://keitheleejr.com/research/SSQ2016/LBLData.dta”, clear

*Table 3: Difference of Means Test

ttest GubenatorialVote, by(GovWins2WeeksBefore)
ttest GubenatorialVote, by(GovWinsWeekBefore)
ttest GubenatorialVote, by(GovWinsWeekAfter)
ttest GubenatorialVote, by(GovWins2WeeksAfter)

ttest GubenatorialVote if South==1, by(GovWins2WeeksBefore)
ttest GubenatorialVote if South==1, by(GovWinsWeekBefore)
ttest GubenatorialVote if South==1, by(GovWinsWeekAfter)
ttest GubenatorialVote if South==1, by(GovWins2WeeksAfter)

*Table 4: Predicting Incumbent Party Vote Share in Gubernatorial Elections

reg  GubenatorialVote GovTotalWins GubernatorialPreviousVote GubernatorialUnememploymentRat HighSchoolGrad Black PerCapitaIncome PopDen
reg  GubenatorialVote GovTotalWins GubernatorialPreviousVote GubernatorialUnememploymentRat HighSchoolGrad Black PerCapitaIncome PopDen South SouthGovTotalWins
reg  GubenatorialVote GovTotalWins GubernatorialPreviousVote GubernatorialUnememploymentRat HighSchoolGrad Black PerCapitaIncome PopDen South SouthGovTotalWins ChampionshipTeams ChampTeamsGovTotalWins