I am trying to create an index using principal component analysis as i lot of variable for investment. so i hope these are the following command to generate an index...
pca (variables)
rotate
predict (new index name)
do i need to standard these variables is my question? then i run a tobit..with the index as my dependent variable and also specify the ul( the highest values in the index).here are my command
tobit sav_index for_remitt_hh DRatio i.sex_hh i.age_group_hh i.edu_hh i.eco_activity_hh i.married_hh i.hhsize i.religion_hh i.District_Code i.land_acresnew i.type_of_house_hh i.type_of_fuel i.bank_acc_hh, ul(22)
margins,dydx(*)atmeans predict (e(0,.))
margins,dydx(*)atmeans predict(ystar(0,.))
margins, dydx(*) predict(p(0,.))

i understand that we use margins to see the change.. but why do we run three margins.. and which margins do we report ? How do we decide.