Re: st: Accessing saved results from -xtmelogit- for postestimation calculations

[Eric Booth <ebooth@ppri.tamu.edu>]

<>

Try running:
 
**
estimates esample:  outcome displacement  rep78, replace
**

before your line:

**
predict re, reffects level(rep78)
**

Your example modified with -estimates example-:

******
cap rm model_1.ster
discard

sysuse auto.dta, clear
*cd "C:\Data"
gen outcome=0
replace outcome=1 in 30/74
xtmelogit outcome displacement || rep78:, variance laplace
**look at example
ereturn li
est stats
di "`r(N)'"
count if e(sample)
estimates esample
di "`r(who)'"
**
estimates save model_1, replace
**check re:
predict re, reffects level(rep78)
su re


///close stata


discard
*cd "C:\Data"
estimates use model_1.ster
sysuse auto.dta, clear
gen outcome=0
replace outcome=1 in 30/74
xtmelogit
**
est stats
estimates esample
estimates esample:  outcome displacement  rep78, replace
estimates esample
**
predict re, reffects level(rep78)
su re
******
- Eric

On Aug 11, 2011, at 5:05 PM, Marshall Garland wrote:

> Hello-
> 
> I'm unable to use saved results from -xtmelogit- for postestimation
> calculations. Results saved using -estimates save- are viewable when I
> -estimates use- and replay them, but when I attempt to calculate
> random effects, I get an "r(2000) no observations" error. I do not
> receive the error when I save and restore estimates from -reg-. I
> received the error both in Stata 12 MP for Windows and Stata 11 SE on
> Unix. I suspected that it might have something to do with properly
> setting the -e(sample)-, but if I include -estimates esample:- before
> -predict-, I receive a "data have changed since estimation r(459);"
> error.
> 
> The syntax below reproduces the error. Has anyone else received this error?
> 
> Thanks.
> 
> -Marshall
> 
> sysuse auto.dta, clear
> cd "C:\Data"
> gen outcome=0
> replace outcome=1 in 30/74
> xtmelogit outcome displacement || rep78:, variance laplace
> estimates save model_1, replace
> ///close stata
> cd "C:\Data"
> estimates use model_1.ster
> sysuse auto.dta, clear
> gen outcome=0
> replace outcome=1 in 30/74
> xtmelogit
> predict re, reffects level(rep78)
> 
> 
> ///now, with regress
> sysuse auto.dta, clear
> cd "C:\Data"
> regress price foreign
> estimates save model_2, replace
> ///close stata
> cd "C:\Data"
> estimates use model_2.ster
> sysuse auto.dta, clear
> regress
> predict b
> *
> *   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/


*
*   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/