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

st: RE: Issue with -mspredict- after -movestay-

From   "Nick Cox" <[email protected]>
To   <[email protected]>
Subject   st: RE: Issue with -mspredict- after -movestay-
Date   Sun, 14 Nov 2004 16:54:02 -0000

In general, I think it is usually best in 
such circumstances to contact authors of 
user-written programs directly. I am copying this to them 
just in case they are not currently members 
of Statalist. 

Emailing Statalist directly would, 
conversely, be a sensible thing to 
do when it appears that the authors 
have passed away or disappeared; 
or if you attempt to contact the authors 
and get no response. 

In this particular case, as the original 
program appeared in the SJ and I think 
Renzo has identified a bug, I shall 
approach the authors directly myself
to discuss publication of a fix as a 
Software Update in a future issue of 
the Journal. 

[email protected] 

Renzo Comolli
> After installing -movestay- you can use -mspredict- to form predicted
> values.
> There is an issue with -mspredict- though when the list of 
> regressors is
> very long.
> I believe the issue is with the following line
> . local param = "`e(ivar_prob)' 1"
> I think the equal sign is what gives the problem. Possibly 
> taking away the equal sign solves all the problems (but I wait for 
> confirmation on this).
> If you want to see the issue do the following
> . use
, clear
. drawnorm variable1 variable2 variable3 variable4 variable5 variable6 //
the only reason for this line is to make a dataset that is big enough go
beyond the maximum character of the local
. movestay lmo_wage age age2 edu13 edu4 edu5 reg2 reg3 reg4, select(private
=m_s1 job_hold  variable1 variable2 variable3 variable4 variable5 variable6)
. mspredict newvar, xb1

You should get the error message "variabl ambiguous abbreviation".

In a previous posting on -movestay-, Nick Cox suggested that the authors of
-movestay- may not be Statalist members. In general what is the best course
of action? Email Statalist or email the authors directly? Or both?

*   For searches and help try:

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