Bookmark and Share

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


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

RE: RE: st: A bug in egen and gen?


From   "Liao, Junlin" <junlin-liao@uiowa.edu>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   RE: RE: st: A bug in egen and gen?
Date   Sat, 19 Feb 2011 19:42:48 +0000

I'm not disagreeing that float is sufficient for storing numeric data. I only disagreed where a number stored in double only adds more zeros. Adding zeros change nothing. But double is more accurate than float. I understand why storing raw data in float is sufficient for most of the time (especially of situations you proposed). I'm only preventing situations I had experienced earlier. I do not see bad outcome in storing numbers in double. That's my argument. Comparing to float, double is not noise, on the contrary, float is noisier. That's my point.

Junlin
________________________________________
From: owner-statalist@hsphsun2.harvard.edu [owner-statalist@hsphsun2.harvard.edu] on behalf of Maarten buis [maartenbuis@yahoo.co.uk]
Sent: Saturday, February 19, 2011 11:59 AM
To: statalist@hsphsun2.harvard.edu
Subject: RE: RE: st: A bug in egen and gen?

--- On Sat, 19/2/11, Liao, Junlin wrote:
> For two, case in mind, Stata store 4.1 not as
> 4.1, but as many digits of 9's following 4.0.

What information is stored in the number 4.1.
Since you said 4.1 and not 4.10 or 4.100, we
know that this is some number between 4.05 and
4.15. It obviously makes no sense to try to
store such a number with 16 digits of accuracy.
We only have information on the first two,
storing it as a double is not going to magically
give us information on the remaining 14 digits.
If information just isn't there, and you
will not create it by storing it as a double.

When storing data you need to remember where
data comes from, and data collection is
necesarily a messy process. Two digits of
accuaracy seems to me about right (or even
optimistic) for most measurements. Imagine
a survey: the interviewer needs to ask the
question correctly (they often do not), than
the respondent needs to understant the
question (they often do not), than the
respondents needs to know the answer (they
often do not), than the interviewer needs
to understand the answer, and it needs to fit
it in the answer form (often this is a problem),
than the data needs to be typed in correctly
(ever heard of typos...). It is a miracle we
get any useful information out of a survey
at al! There is no chance that we will get
anything with an accuracy of 16 digits out
of that...

-- Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------




*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


________________________________
Notice: This UI Health Care e-mail (including attachments) is covered by the Electronic Communications Privacy Act, 18 U.S.C. 2510-2521, is confidential and may be legally privileged.  If you are not the intended recipient, you are hereby notified that any retention, dissemination, distribution, or copying of this communication is strictly prohibited.  Please reply to the sender that you have received the message in error, then delete it.  Thank you.
________________________________

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


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