As a matter of curiosity, I checked out the websites for SUDAAN and

WesVar. It's patently clear that SUDAAN can perform many more

statistical analyses than WesVar.

Unfortunately this comes with a price. It costs $495 for a permanent

commercial or government license of WesVar. Academic users only need to

pay 350 dollars for a permanent license.

On the other hand, to license Sudaan v. 9 the government license costs

$1015 dollars; the commercial license costs 1,140 dollars, and an

academic license costs costs 875 dollars. Obviously, SUDAAN is much

more expensive than WesVar. Also WesVar offers student versions for a

mere 29 dollars.

Each of these survey research methods software programs is advertised

as complementing both SAS and SPSS.

So that's the scoop.

As far as using simulations to compute sample size, you must still need

to prespecify your type I and type II errors; otherwise you cannot

interpret the results of a given simulation. Yet, once you've stated

your alpha and power requirements, you need only use a simple

mathematics formula to determine sample size requirements. Thus,

running simulations is hardly necessary, since you still need to state

your alpha and power requirements.

joe teitelman

Upon inspection, it's clearly easy to see that SUD

>>> Joseph Teitelman temp2 7/12/2006 4:27 PM >>>

SUDAAN was developed by the Research Triangle Institute, which is

surrounded by Duke, UNC, and NCSU. As such, SUDAAN is quite popular

around that area of NC. WesVar directly competes with SUDAAN.

I've used SUDAAN but not WesVar.

I'd suspect that the only differences between these two packages which

specialize in the analysis of complex and sometime multistage sampling

survey data.

In the 90s, neither SAS nor SPSS provided facilities for analyziing

survey data. Now, both software packages perform analyses of survey

data. Yet, they were never developed to specialize in the analysis of

sample survey data. Whereas, SUDAAN and WesVar were.

Since you probably don't have 1000 dollars to purchase SUDAAN or

WesVar, I'd suggest that you use the SPSS add-on developed for that very

purpose. Personally, I'm a SAS user, but have been forced to learn SPSS

as part of my job.

I hope this information proves helpful

Joe Teitelman

>>> "Will Bailey [Statman]" <

[hidden email]> 7/11/2006 8:27 AM

>>>

There also is Westat's WesVar and SUDAAN

One point, IMHO, none is better than the other only a personal

preference.

Also, each handles the sample development slightly different.

The SPSS module is fully integrated into SPSS and a great benefit if

one's

primary stat pack is SPSS. WesVar & SUDAAN are pretty much standalone.

I

have used WesVar for several years but know very little about the

SUDAAN.

W

-----Original Message-----

From: SPSSX(r) Discussion [mailto:

[hidden email]] On Behalf

Of

Kristina

Sent: Monday, July 10, 2006 7:53 PM

To:

[hidden email]
Subject: Complex Samples - SPSS vs Stat

I am using a national dataset that employs a complex sampling design.

I

understand the reasons for using software that accounts for the

weighting

and complex sampling design. However, I see that there are a few

statistical

packages available to assist with complex samples. I would like to use

either SPSS Complex Samples or Stata. Is one package better, easier,

etc.

than the other?

Thanks,

Kristina