Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: prvalue and the invisible prediction

From   "Nick Cox" <[email protected]>
To   <[email protected]>
Subject   st: RE: prvalue and the invisible prediction
Date   Wed, 20 Oct 2004 10:56:55 +0100

Stata [sic] is innocent here. It is doing what -prvalue- 
tells it to. -prvalue- is, as you know but others may 
not, a program from the -spost- package from Scott Long and Jeremy 
Freese. I am not sure if either is a member of Statalist, 
so if you get no good answers to this, the next step may be 
to contact the program authors directly. 

Can -adjust- help with what you want? 

[email protected] 

Robert Bozick
> I want STATA to estimate predicted probabilities for my 
> binary outcome for 
> different levels of my independent variables.  I am 
> estimating a logit 
> model.  STATA estimates the model without any error messages. 
>  I then use 
> the prvalue commands to obtain the predicted probabilities.  
> To be specific, 
> the code I use is:
> prvalue, x(hourw10=1 hourw20=0 hourw30=0 hourw40=0 hourw40p=0 
> offcampus=1 
> apt=0 fam=0 misres=0 off20=0 off30=0 off40=0 apt10=0 apt20=0 
> apt30=0 apt40=0 
> apt40p=0 fam10=0 fam20=0 fam30=0 fam40=0 fam40p=0 misr30=0 misr40=0 
> misr40p=0)
> Again, I receive no error messages, but STATA does not 
> produce the predicted 
> probabilities (or they are invisible).  The output looks like this:
> logit: Predictions for persist
>   Pr(y=continuo|x):        .   95% ci: (     .,     .)
>   Pr(y=no_persi|x):        .   95% ci: (     .,     .)
> The predictions are not there!  Help!  Does anyone know what 
> I am doing 
> wrong?  Why can I not get STATA to produce the predicted values?

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index