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

Thanks (was st: Float-type variables)


From   "Hiroshi Maeda" <hmaeda1@uic.edu>
To   "Statalist" <statalist@hsphsun2.harvard.edu>
Subject   Thanks (was st: Float-type variables)
Date   Thu, 26 Feb 2004 10:21:12 -0600

Dear Drs. Cox and Sorensen:

Thanks for your reply to my inquiry on float-type variables (especially Dr.
Cox, who must have spent an hour or so writing up such a lengthy detailed
response). As Dr. Cox suggests of other users, I, too, mostly deal with
categorical variables and felt a little annoyed by the problem I encountered
with float-type variables, although I had known about it (I am also aware of
the "set type double" command Dr. Sorensen suggested). I just find rather
counter-intuitive the way Stata "demotes" computationally generated "double"
variables to "float" ones (as documented in the FAQ page entitled "How many
significant digits are there in a float?"), producing seemingly inaccurate
results which are indeed accurate. But it's merely one person's
opinion/feeling. Anyway, I will make a mental note to watch for a value
larger than 16,777,215, the threshold beyond which Stata becomes seemingly
inaccurate but actually not. Thanks. ---Hiroshi

P.S. To those who don't see the context here, what I've written above is a
thank-you note to Dr. Cox's lengthy response to my even more lengthy initial
inquiry, both of which I've chosen not to reproduce here to save space.


Hiroshi Maeda
University of Illinois at Chicago
hmaeda1@uic.edu


*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index