Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: "twoway function" (& closure on "test 2")

From   Nick Cox <>
Subject   Re: st: "twoway function" (& closure on "test 2")
Date   Wed, 11 May 2011 11:06:01 +0100

Allan seems to want StataCorp to be responsible for documenting what
happens when you don't use a command as intended but decide to branch
out on your own.

That is, none of his examples make references to x.

Interesting principle. I think StataCorp have better things to do in this case.


On Wed, May 11, 2011 at 10:57 AM, Allan Reese (Cefas)
<> wrote:
> As a health & safety champion, I have to demur from Nick Cox's attitude
> to IT pratfalls.
> From: Nick Cox Subject: Re: st: plotting a regression function with
> time-dummies indicating structural breaks
> - -twoway function- has been around since Stata 8 (2003). Its use of x
> regardless of whatever x might be in the data is documented. Its
> capacity for small surprises when you want to use some or all of your
> data as well is less surprising on brief examination. I agree that
> there is a case for considering a tweak of the documentation here, but
> I see no evidence that -twoway function- does not work as intended.
> --------
> In the first place, I'm not sure I understand how, or if, y=function(x
> z) was intended to operate.  It is documented that y and x are
> interpreted as dummy variables and not affected by "y" or "x" being in
> the current dataset.  The documentation does not say how z will be used
> when it is in the dataset, and it appears to be controlled by the number
> of calculated points (default 300) and number of observations in the
> dataset.  Eg, using the auto data,
>     twoway function y=mpg  /* plots over 74/300 of the "x" range */
>     twoway function y1=mpg, n(37) range(0 .5) || function y2=mpg, n(74)
> /* doesn't quite match */
>     twoway function y1=mpg, n(37) range(0 `=36.5/74') clcol(blue) ||
> function y2=mpg, n(74) clpattern(_) /* kludge! */
> The H&S approach to hazards is hierarchical: (1) can the hazard be
> removed? (2) can it be circumvented or mitigated? (3) protect and warn
> people who might meet it.
> Computers are deterministic machines, and I don't appreciate small
> surprises.  Over to the Corp...

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index