Re: st: suest after xtlogit

[Jorge Eduardo Pérez Pérez <perez.jorge@ur.edu.co>]

You do not need -suest- to test for differences between two periods in
this setting. Estimate a joint model with interactions of the dummy
that separates the periods and the explanatory variables, then do a
joint test of the significance of the interactions.

webuse nlswork, clear
xtset idcode year
gen before=year<=80
xtlogit union i.before##(c.collgrad c.ln_wage i.south)
testparm before#c.collgrad before#c.ln_wage before#south



_______________________
Jorge Eduardo Pérez Pérez



On Tue, Apr 17, 2012 at 4:32 PM, Taras Murzenkov <murzenkovtn@gmail.com> wrote:
> Hello!
>
> I am estimating default probabilites using xtlogit with random effect.
> I'd like to test hypothesis whether there is any structural changes in
> my sample. For this reason I'v devided my sample into 2 non
> overlapping sub sample: before specific date and after specific date.
> Then, I've estimated model on this two subsamples and stored results.
> Basicly, what I've done:
>
> xtlogit perfom SK_CA_lag1 SK_CA2_lag1 Z LA_CA_lag1 PZS_CA_lag1
> lnCA_lag1 if !(crisis)//perfom binary variable, 0 for defaulted firm
> est store after_crisis
>
> xtlogit perfom SK_CA_lag1 SK_CA2_lag1 Z LA_CA_lag1 PZS_CA_lag1
> lnCA_lag1 if crisis
> est store before_crisis
>
> suest before_crisis after_crysis //I've recieved the error message
> "unable to generate scores for model before_crisis. suest requires
> that predict allow the score option"
>
> Would you be so kind to tell me how to solve this problem? Thank you in advance!
> *
> *   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/