variate_checks

*** Comparing means on your key variable among various demographic groups

*** the ID Scales:
* id_race_scale
* id_gender_scale
* id_PID_scale
* id_religion_scale and/or id_religimp_scale
* id_geo_scale
* id_orientation_scale


*** We need to check to see if demographics influence Strength of ID response:

*** race | Race
tab race
oneway id_race_scale race, t

*** Q8 | gender * 1=women 0=men
tab Q8
generate women = .
replace women = 1 if (Q8 == 2)
replace women = 1 if (Q8 == 4)
replace women = 0 if (Q8 == 1)
replace women = 0 if (Q8 == 3)

tab women
ttest id_gender_scale, by (women) 

*** pid7 | Party ID * 1=Dems 2=Reps 3=Ind+Not Sure
tab pid7
generate pidsimple = .
replace pidsimple = 1 if (pid7==1)|(pid7==2)|(pid7==3)
replace pidsimple = 2 if (pid7==5)|(pid7==6)|(pid7==7)
replace pidsimple = 3 if (pid7==4)|(pid7==8)

tab pidsimple
oneway id_PID_scale pidsimple, t bon sid sch

*** religpew | Religion
tab religpew

oneway id_religion_scale religpew, t

*** Q7 | Urban density
tab Q7
generate geo = Q7
oneway id_geo_scale geo, t

*** Q9 | Sexual Orientation
tab Q9
generate orientation = Q9
replace orientation = . if (Q9==4)|(Q9==5)|(Q9==8)

tab orientation
oneway id_orientation_scale orientation, t


*** Then we can check to see if the same demographics affect Affective Polarization
oneway AP_th race, tab
oneway AP_th women, tab
oneway AP_th pidsimple, tab
oneway AP_th religpew, tab
oneway AP_th geo, tab
oneway AP_th orientation, tab


*** this is the fine details:
twoway lfitci id_PID_scale women, by(pidsimple)