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: st: A bug in egen and gen?


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   statalist@hsphsun2.harvard.edu
Subject   RE: st: A bug in egen and gen?
Date   Fri, 18 Feb 2011 11:20:15 +0000 (GMT)

--- On Fri, 18/2/11, Liao, Junlin wrote:
> I further tested Stata in its capability to assign correct
> data type when importing data. I'm really concerned now.
> Apparently the data type setting has far reaching effect
> other than -gen- and -egen- commands. In fact importing
> procedures also depend on it. If you have numeric values
> with many digits, it's very likely you'll get the wrong
> data.

I would say you get the right data, or at least better 
data than you used to have. I am old-fashioned in the 
sense that when I say 4.10 I mean that I checked that it
is not less than 4.095 or more than 4.105, and when I say
4.100 I mean I checked that it is not less than 4.0995 or 
more than 4.1005. So when I look at data 4.1, 4.10, and
4.100 are to me not the same number. I real data I would
consider 4.1 good data, 4.10 exceptional, and I suspect
someone did not follow proper procedure when I see 4.100.
This is what I meant when I said above that the default
of using floats will improve your data: doubles would 
suggest a degree of certainty that is just too preposterous
for words(*). To borrow yet another example from Bill's 
Stata Blog entry:

"Most fundamental constants of the universe are known only 
to a few parts in a million, which is to say, to less than 
float precision, although we do know the speed of light in 
a vacuum to one decimal digit beyond float accuracy; it’s 
299,793.458 kilometers per second. Round that to float and 
you’ll be off by 0.01 km/s."

Can you imagine collecting an entire dataset upto the same
precision as the way the speed of light was measured?

-- Maarten

(*) The one exception to this rule would be identification
numbers, as I mentioned in an earlier post.

--------------------------
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/


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