Define the scalar variable when categorical variable is used in regression

[Remmy Saimon]

Hi everyone,

Kindly help me on how I can define the content of the scalar variable i.e. "sca b_ `x' " after estimating a regression. I have provided the codes that I am running using the "sysuse auto" data set to illustrate what I am trying to do.

sysuse auto
global X " mpg weight i.rep78"
reg price $X

foreach x of global X {
sca b_`x' = _b[`x']
}

When I run the codes, stata(15) gives me this error " b_i.rep78 invalid name" . If also tried " sca b_ `x' = _b[`x' , 2.`x' .. 5. `x' ] still does not work.

[FernandoRios]

You are close
You just need to call the betas the correct way:

sysuse auto
global X " mpg weight i.rep78"
fvexpand $X
global X2 "`r(varlist)'"
reg price $X

foreach x of global X2 {
local cnt=`cnt'+1
sca b_`cnt' = _b[`x']
}
scalar list

[Remmy Saimon]

Thank you FernandoRios.

I have tried the suggested codes, and they works but they limit me to progress with further arguments in my loop. When I run the codes i get this error " 1 invalid name"

global X " mpg weight i.rep78"
fvexpand $X
global X2 "`r(varlist)'"
reg price $X
sum price 
sca mean_`y' =r(mean)

foreach x of global X2 {
local cnt=`cnt'+1
sca b_`cnt' = _b[`x']
corr length `cnt' ,c
sca cov_`cnt' = r(cov_12)
sum `cnt'  
sca elas_`cnt' = (b_`cnt'*r(mean))/mean_`y'
}