Stata's strength is that you get a comprehensive stats package, including
survey facilities, in one powerful, reasonably-priced, sanely-licensed
package written by a highly responsive, user-oriented company. It is a
general purpose package, not a single purpose, dedicated survey package.
Stata's survey facilities will get you through almost any survey analyses,
but it would be somewhat unreasonable to expect Stata to do everything that
the best single purpose survey packages can do. If you want a comparison of
everything Stata can do versus everything each of the single-purpose survey
packages can do, the respective lists are long, so you'll have to check out
each website individually (try Westat and SUDAAN). Stata's capabilities are
listed here: http://www.stata.com/info/capabilities/svy/
Stata compares well to other general stats packages:
1. SPSS: Until very recently, SPSS had no survey facilities at all and was
very frustrating for complex survey analysis, especially with regard to
weights. As of this month, Johnny-come-lately SPSS has released an add-on
package for the new SPSS 12.0 called Complex Samples. It is brand-spanking
new, so I haven't tried it. It had better be bloody good though, because
according to their website, the cost to license this add-on package alone is
US$699. Of course, if SPSS is not your primary package, then you'd have no
reason to buy it.
2. SAS: I don't use SAS, but my understanding is that if you want to analyze
complex surveys in SAS, you buy the SAS add-on version of SUDAAN.
3. R: Thomas Lumley has written a survey package for R. It just came out in
version 2 this past August. It favours a GLM approach to survey analysis,
and implements replications and jackknife estimates. It's features are here:
http://probability.ca/cran/src/contrib/Descriptions/survey.INDEX . R is nice
because it is powerful and free, but it is different enough to render
comparison with Stata difficult. To quote Thomas Lumley: "Stata is very good
at the things it is good at (which are in many ways orthogonal to the things
R is good at)"
4. EpiInfo: Free; includes some ability to specify complex designs, but only
rather limited analyses are possible.
Regards,
Lee
Lee Sieswerda, Epidemiologist
Thunder Bay District Health Unit
[email protected]
> -----Original Message-----
> From: wildscop [mailto:[email protected]]
> Sent: Thursday, October 30, 2003 1:13 PM
> To: [email protected]
> Subject: st: why STATA
>
>
> Dear all,
>
>
> Can any one give me an idea why Stata should be
> preferred for analyzing all sort of survey data?
> What are its superiority to other available packages?
>
> Any response / comments will be appreciated.
> Thank you.
>
>
> _______________________
>
> Mohammad Ehsanul Karim <[email protected]>
> Institute of Statistical Research and Training
> University of Dhaka, Dhaka- 1000, Bangladesh _______________________
>
>
> *
> * 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/
>
*
* 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/
