--- On Thu, 16/7/09, Nils Braakmann wrote:
> thanks. However, I would like the plotted interval to
> reflect the precision of the underlying parameter
> estimates.
OK, that can make sense too: the coefficients are in this
case the marginal effects, i.e. the first derivative of
the curve I gave in the previous post.
*-------------------- begin example -------------------
use http://www.stata-press.com/data/r10/nlswork, clear
xtset idcode year
mkspline agespl 20 = age, dis
matrix knots = r(knots)
sum age, meanonly
local min = r(min)
local max = r(max)
xtreg ln_wage agespl*, fe cluster(idcode)
preserve
drop _all
set obs 21
gen x = `min' in 1
gen b = _b[agespl1] in 1
gen se = _se[agespl1] in 1
forvalues i = 2/20 {
replace x = el(knots,1,`=`i'-1') in `i'
replace b = _b[agespl`i'] in `i'
replace se = _se[agespl`i'] in `i'
}
replace x = `max' in 21
replace se = _se[agespl20] in 21
expand 2
sort x
by x : gen first = _n == 1
replace b = b[_n-1] if first
replace se = se[_n-1] if first
gen lb = b - 1.96*se
gen ub = b + 1.96*se
twoway rarea lb ub x || ///
line b x, yline(0)
restore
*-------------- end example ------------------
Hope this helps,
Maarten
-----------------------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
