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

RE: st: RE: re. downloading the STB ado files collin and tabrate

From   "Nick Cox" <[email protected]>
To   <[email protected]>
Subject   RE: st: RE: re. downloading the STB ado files collin and tabrate
Date   Thu, 11 Nov 2004 14:06:26 -0000

I agree with Richard here. In addition, 

1. -collin- and other ado files are transparent 
in the sense that you can open up the files 
and examine the code. In practice, people may 
not have the time, the patience or the skills
to do that, but that's not the programmer's 

2. I guess there's a social niche for some
one to set themselves up as a Stata critic 
and start pronouncing on what's good and what's 
bad, just as people do on cars, films, restaurants, 
whatever. That would make for interesting reading. 
But, more immediately, statements of invalidity 
are sometimes straightforward, as when a bug is 
obvious to any competent person, but statements of 
validity are often difficult and always provisional. 

3. In broad terms, you needn't install anything 
you can't be sure of trusting. But it's pretty clear
that even as official Stata continues to grow, the 
fraction of user-written material that is adopted 
by StataCorp will continue to go down. What StataCorp 
thinks of its quality and importance is not crucial 
here, but rather that a hyper-bloated Stata is in no 
one's long-term (or even short-term) interests, or so 
I assert. So the question raised by Roger, while impossible 
to answer fully, is not going to go away. Users 
are going to have to make their own judgements about what
is worth trying, let alone trusting. 

[email protected] 

Richard Williams
> roger webb

> >Thanks for your help Nick,
> >As 'collin' is not part of 'official Stata', to what degree can I be
> >confident of its statistical validity?
> You can ask the same question about any user written routine! 
>  And even 
> official Stata programs can have bugs.
> When possible, I think it is a good idea to try to 
> cross-validate some 
> results using different programs.  As for -collin-, I've been able to 
> replicate its results using SPSS and also printed results, so 
> as far as I can tell it is fine.

*   For searches and help try:

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