*This do file requires the cumulative NES 1948-2002 File.

*Extract Variables of Interest
keep VCF0809 VCF0808 VCF0004 VCF0102 VCF0106 VCF0114 VCF0140 VCF0301 VCF0656 VCF0104 VCF0113 

*Rename some of the variables
rename VCF0004 year
rename VCF0102 age
rename VCF0106 race
rename VCF0114 income
rename VCF0140 educ
rename VCF0301 pid
rename VCF0656 tindex
rename VCF0104 gender
rename VCF0113 south

*Call to Merge Aggregate Data with Cross Sectional Data
*Note merge.dta must also be sorted by year
sort year
merge year using e:\merge.dta
tab _merge

*Generate Ideology Proxy Measure
drop if VCF0808==9
drop if VCF0809==9
egen jobs1=std(VCF0808)
egen jobs2=std(VCF0809)

gen jobs3=jobs2
replace jobs3=jobs1 if year==1956
replace jobs3=jobs1 if year==1958
replace jobs3=jobs1 if year==1960
replace jobs3=jobs1 if year==1964
replace jobs3=jobs1 if year==1968
tabstat jobs3,by(year)

drop jobs1 jobs2 

rename south south1
recode south1 (1=1 "South") (2=0 "Non-south"), into(south)
drop south1

rename gender gender1
recode gender1 (1=1 "Male") (2=0 "Female"), into(gender)
drop gender1

rename race race1
recode race1 (1=0 "White") (2=1 "Black") (3=.), into(race)
drop race1


drop if pid==9
*Party Identification Dummies For OLS Models
recode pid (1=1 "Strong Dem") (2/7=0 "Other"), into(str_dem)
recode pid (2=1 "Weak Dem") (else=0 "Other"), into(wk_dem)
recode pid (3=1 "Ind Dem") (else=0 "Other"), into(ind_dem)
recode pid (4=1 "Ind") (else=0 "Other"), into(ind)
recode pid (5=1 "Ind Rep") (else=0 "Other"), into(ind_rep)
recode pid (6=1 "Weak Rep") (else=0 "Other"), into(wk_rep)
recode pid (7=1 "Strong Rep") (else=0 "Other"), into(str_rep)

*Party Identification Dummy For Difference of Means Test
recode pid (1=1/3 "Democrats") (0=5/7 "Republicans") (4=.), into(dems_rep)

*Presidential Party Measure
* 1 equals your party controls that institution.
gen prez=.
replace prez=1 if pid>=4 & year==1958
replace prez=0 if pid<=3 & year==1958

replace prez=0 if pid>=4 & year==1960
replace prez=1 if pid<=3 & year>=1960 & year<=1968

replace prez=1 if pid>=4 & year>=1970 & year<=1976
replace prez=0 if pid<=3 & year>=1970 & year<=1976

replace prez=0 if pid>=4 & year>=1978 & year<=1980
replace prez=1 if pid<=3 & year>=1978 & year<=1980

replace prez=1 if pid>=4 & year>=1982 & year<=1992
replace prez=0 if pid<=3 & year>=1982 & year<=1992

replace prez=0 if pid>=4 & year>=1994 & year<2000
replace prez=1 if pid<=3 & year>=1994 & year<2000

replace prez=1 if pid>=4 & year==2000
replace prez=0 if pid<=3 & year==2000

*Senate
gen senate=.

replace senate=1 if pid<=3 & year>=1958 & year<=1980
replace senate=0 if pid>=4 & year>=1958 & year<=1980

replace senate=0 if pid<=3 & year>=1982 & year<=1984
replace senate=1 if pid>=4 & year>=1982 & year<=1984

replace senate=1 if pid<=3 & year>=1986 & year<=1994
replace senate=0 if pid>=4 & year>=1986 & year<=1994

replace senate=0 if pid<=3 & year>=1996 
replace senate=1 if pid>=4 & year>=1996

*House
gen house=.

replace house=1 if pid<=3 & year>=1958 & year<=1994
replace house=0 if pid>=4 & year>=1958 & year<=1994

replace house=0 if pid<=3 & year>=1996
replace house=1 if pid>=4 & year>=1996


*Difference of Means Tests of Table 1
ttest tindex, by(dems_rep)
*Democratic (Expected Negative Difference)
ttest tindex if year<=1968, by(dems_rep)
*Republican (Expected Positive Difference)
ttest tindex if year>=1970 & year<=1976, by(dems_rep)
*Democratic (Expected Negative Difference)
ttest tindex if year>=1978 & year<=1980, by(dems_rep)
*Republican (Expected Positive Difference)
ttest tindex if year>=1982 & year<=1992, by(dems_rep)
*Democratic (Expected Negative Difference)
ttest tindex if year>=1994, by(dems_rep)

*Models in Table 2
regress tindex prez senate house
regress tindex prez senate house ics prap, ro 
regress tindex prez senate house str_dem wk_dem ind_dem ind_rep wk_rep str_rep jobs3 age educ income race south ics prap, ro