Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Question on graph twoway and graph combine

From   Alberto R Osella <>
Subject   Re: st: Question on graph twoway and graph combine
Date   Wed, 07 Dec 2011 10:39:23 +0100

Dear Nick,
I'm using -seqlogit-. Here a portion of code:
xi: seqlogit steato_grade i.StatCiv Etarecl glicemia i.scalacat GOT GPT ///
         if Sesso == 0, or                 ///
         tree(0:1 2 3 , 1:2 3 , 2 : 3)     ///
         ofinterest(BMI) over(Etarecl)          ///
         levels(0=0, 1=1.5, 2=4, 3=5.1)


keep if e(sample)

replace StatCiv=2

sum BMI, meanonly
replace BMI = r(mean)
replace _BMI_X_Etarecl = BMI*Etarecl
bys Etarecl : keep if _n == 1

predict wM* , trweight
_estimates hold model1

mkspline  edad 8= Etarecl if Sesso==0
regress wM1 edad1-edad8
predict SwM1
regress wM2 edad1-edad8
predict SwM2
regress wM3 edad1-edad8
predict SwM3
drop edad1-edad8

_estimates unhold model1

forvalues i = 1/3 {
gen contr`i' = SwM`i'*([#`i']_b[BMI]+ [#`i']_b[ _BMI_X_Etarecl]*Etarecl)

gen contr0 = 0
replace contr2 = contr1 + contr2
replace contr3 = contr2 + contr3

twoway rarea contr0 contr1 Etarecl  || ///
       rarea contr1 contr2 Etarecl  || ///
       rarea contr2 contr3 Etarecl,  ///
       xlab(2(2)8)                ///
       legend(order( 1 "Light or more vs. no" ///
                     2 "Moderate or more vs. light or less" ///
                     3 "Severe vs Moderate or less" ) pos(6))
graph save uomini, replace

I've said "I can't use". It may be: " I don't know how to use the by() option".

Alberto R. Osella, MD, PhD
Laboratorio di Epidemiologia e Biostatistica
IRCCS Saverio de Bellis
Via Turi, 27
70013 Castellana Grotte (BA)
Tel:    +39 0804994655
Fax:    +39 0804994650

Il 07/12/2011 10:16, Nick Cox ha scritto:
That's good advice, but I don't understand why -by()- couldn't be used
in the first place. What does any estimation process have to do with


On Wed, Dec 7, 2011 at 9:09 AM, Phil Clayton
<>  wrote:

You could try -findit grc1leg-

On 07/12/2011, at 8:09 PM, Alberto R Osella wrote:

I'm using Stata 12.0 on W7. I've drawn two graphs (one for each sex) by using -graph twoway- and then I've combined them by using -graph combine-. I've drawn graphics separately because I can't using the by() option as a result of the estimation process. The resulting combine graph has two legends (the same) and I want a graph with only one legend.
*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index