After checking with describe, my data are all numeric though -time- is in
STATA date format which may or may not be a problem with aucdist. (Please
see my response to Thomas Steichen's message)
Message: 16
Date: Tue, 24 Sep 2002 19:20:08 +0100
From: "Nick Cox" <[email protected]>
Subject: st: RE: Area under the curve trouble with aucdist
[email protected]
>
> I need to calculate the area under the curve (viral load
> over time) for
> about 1000 study participants using the trapezoidal rule.
> I have used
> aucdist but I am not getting a result and am not sure if
> I'm getting an
> error message.
>
> id = patient id
> time = time of viral load test
> conc = viral load value
>
> aucdist id time conc, t generate(auc)
>
> (t specifies that the integral be computed using the simple
> trapezoidal
> rule, and generate(auc) specifies to create a new variable
> with the AUC (C
> max, T max, log conc 0 -inf.))
>
> but all I'm getting is (2000) in return which is supposed
> to mean that I
> have no observations on which to perform the statistical operation.
>
> I have 9500 observations with multiple observations per
> patient, there are
> no missing data. Any suggestions as to what I may be doing wrong?
Probably something is really a string variable
which shouldn't be, itself possibly a legacy from some
previous incarnation of the data within Microsoft software,
or just that your -id- contains names.
Take a look with -describe- and -destring- or -encode-
any offender.
r(2000) strictly means "no numeric values
on which to do that".
I have previously suggested to Stata Corp that
"no observations" is not the best wording for
this error.
Of course, I approve wholeheartedly of their
reluctance to change longstanding syntax,
which is a hangover from Stata 1.0, I guess,
when there were no string variables.
Nick
[email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/