heteroskedasticity simulation

[Karen Arulsamy]

Hi there,
I would like to run the following code:

clear all
program define mcsimul1, rclass
version 10.0
syntax [, c(real 1)]
tempvar e1 e2
gen double ‘e1’=invnorm(uniform())*‘c’*zmu
gen double ‘e2’=invnorm(uniform())*‘c’*z_factor
replace y1 = true_y + ‘e1’
replace y2 = true_y + ‘e2’
summ y1
return scalar mu1 = r(mean)
return scalar se_mu1 = r(sd)/sqrt(r(N))
summ y2
return scalar mu2 = r(mean)
return scalar se_mu2 = r(sd)/sqrt(r(N))
return scalar c = ‘c’
end

local reps 1000

forv i=1/10 {
qui webuse census2, clear
gen true_y = age
gen z_factor = region
sum z_factor, meanonly
scalar zmu = r(mean)
qui {
gen y1 = .
gen y2 = .
local c = ‘i’*10
simulate c=r(c) mu1=r(mu1) se_mu1=r(se_mu1) ///
mu2=r(mu2) se_mu2=r(se_mu2), ///
saving(cc‘i’,replace) nodots reps(‘reps’): ///
mcsimul1, c(‘c’)
}
}

use cc1
forv i=2/10 {
append using cc‘i’
}
gen het_infl = se_mu2 / se_mu1
save cc_1_10,replace

tabstat mu1 se_mu1 mu2 se_mu2 het_infl, ///
stat(mean) by(c)

as described in slide 9-10 here http://fmwww.bc.edu/GStat/docs/StataSimul.pdf

I have tried adjusting the ` ' signs on the code but there seems to be other issues. Could I get some help in running this?

Many thanks
Karen

[Maarten Buis]

The code runs without problem on my computer. Did you copy that code in your do-file editor, and than run the entire code and not line by line?

[Karen Arulsamy]

Thanks Maarten. I tried running it all at once and also by sections. That's odd!

[Andrew Musau]

local c=`i' +10

Pay attention to the left and right single quotes. These are a grave accent and an apostrophe, respectively. Perhaps you are copying the code to a word editor before copying to the do-file and this changes the characters.