PDA

View Full Version : Summary: Statistical Software



Theresa Foti
11-04-1998, 08:18 AM
Below is a summary of responses and all comments I received for the
following question. Thanks to all that replied!

>Hello All,
>I am looking into options for statistical software. Does anyone >know of
any options comparable to SPSS for Windows, SYSTAT, and >SAS JMP that I
could check into? Preferably ones that you use >and like.
>
>Thanks,
>
>Theresa Foti
>Motion Analysis Laboratory
>Shriners Hospital for Children
>Greenville, SC 29605
>tfoti@infoave.net

The follwing statistical packages were recommended (in addition to SPSS,
SYSTAT, and SASJMP):

-Minitab
-Statistica
-Statview
-Sigma Stat
-Prism 2
-SWF "Statistics With Finesse"
-S Plus
-Statistix
-NCSS & PASS (Number Cruncher Statistical Software & Power Analysis
Statistical Software)
----------------------------------------------------------------
Hi Theresa, i always liked MINITAB, i found it very user
friendly and very comprehensive. I am not sure where info. can be found
regarding this package, although Liverpool John Moores University do have
it networked.
----------------------------------------------------------------
I have used and like Minitab for Windows.
--------------------------------------------------------------
This e-mail is actually in response to one
that was posted by somebody else in the listserv.
I have heard from a reliable source that the
minitab software is not very accurate as tested
using various test data sets, but I guess he could
be wrong. Systat, Statview, and some of the other
packages you mentioned are supposed to be among the
better statistical software out there. Anyway,
just wanted to keep you informed.
--------------------------------------------------------------
Try Minitab for windows. It is quite user friendly, and the new
version 12 has a lot of biomechanically useful tools in it (e.g.
GLM for unequal randomized and nested designs). Go to:

httP://www.minitab.com/
--------------------------------------------------------------
I would agree with this recommendation. (see below) I have not personally
used Minitab but several of my colleagues have. It does not have the
versatility of SAS, SPSS, or SYSTAT but it has proved to be functional for
our applications.

For PC-based statistical package you may wish to try: minitab - a demo of
which is available at their web site: www.minitab.com
----------------------------------------------------------------
I've used lots of stats programs, but the best program you've not listed in
your options. have a look in Statistics from Statsoft Inc.. They are
probably presenting the version 6.0. There are many other advantages in a
friendly environment. There are some features useful tools for
biomechanicists, e.g. filters (no many).
----------------------------------------------------------------
I really like Statistica. I've used SYSTAT as well, but like Statistica
better - it's more intuitive to figure out what you need to do and the
output pages are well formatted to provide the information I need. It has
separate "modules" for the basic statistics (means, t-tests, etc) and more
complex models (multivariate, non-parametric) which makes it easy to
isolate the exact test you want. It imports/loads Excel spreadsheets
easily, as well as ascii files. It also allows for easier and more
efficient spreadsheet/data manipulation than SYSTAT. The graphing options
are good, especially if you want to print directly from Statistica.
However, I find that I usually want to export a graph into another program
(such as PowerPoint or Illustrator) in order to add to the image, and that
can be difficult with Statistica.

For product information, contact StatSoft at:

email: info@statsoft.com
web: www.statsoft.com
phone: (918) 749-1119 (in Tulsa, OK)
fax: (918) 749-2217
----------------------------------------------------------------
We are very please and impressed with the power and ease of use of
Statistica. Give it a try.
----------------------------------------------------------------
You should check into Statistica from Statsoft, Inc. I use it and find it
to be very good.
----------------------------------------------------------------
I personally enjoy using the Statistica package from StatSoft. Newer
versions of the package probably feature more bells and whistles, but the
4.5 version I currently use works fine for my analysis of variance,
post-hoc tests, etc. Its graphics capabilities are also a major asset in
figure production.
----------------------------------------------------------------
I use and like STATISTICA - it has a very comprehensive range of
statistical procedures and excellent graphical options. I use it mostly
for analysis of variance, for which it is really excellent. It's available
from StatSoft, 2300 E.14th Street, Tulsa, OK 74104
email: info@statsoft.com
Web: http://www.statsoft.com
---------------------------------------------------------------
I have been using SYSTAT and SPSS for a while but then got a chance to
switch to STATISTICA - I found it easier to use (module based, e.g. you
open a module for parametric stats, another one for non-linear stuff and so
on; and it has a great data management option. In addition, I found its
graphic capabilities very powerful. Its definitely recommendable.
Unfortunately I do not know about web demo versions.
----------------------------------------------------------------
I myself use SPSS for windows however many of my colleagues use STATISTICA.
I am in the minority here but everyone has indicated that it is easy to
work with. It is somewhat similar to SPSS althought the layout may and
protocol may take some time in getting used to.
----------------------------------------------------------------
Hi Theresa - I like Statistica and another person here really likes
StatView. If you can't find contact info for these, let me know and I can
probably dig up phone numbers or web addresses.
----------------------------------------------------------------
Try Statview (now owned by SAS). Version 5.0 is the latest and is operable
on Windows and Macs. Works for me.
----------------------------------------------------------------
I have found Statview to be extremely versatile and
very easy to use. It is reputed to be one of the
more accurate statistical software programs out there.
The user-friendly interface is definitely a plus.
----------------------------------------------------------------
STAT VIEW for MacIntosh is nice and easy to work with. All statistics
available, in specific good graphics.
---------------------------------------------------------------
We have been using the PC version of StatView for a
few years and it is pretty slick, easy and moderately
powerful. The company that developed the software (Abacus
Concepts) has recently sold it to SAS and the 5.0 version
they have released is great in some areas but has a few
quirks, like being unable to do all possible comparisons on
some posthoc tests.
We are now looking at Statistica and it looks very
good. I had all but given up finding a package with an
easy to learn interface, but this software appears
excellent, and has won awards from ergonomists for its
design and interface. It is also extraordinarly complete
and feature-packed. It was raised and nurtured in an
diverse academic environment so it can handle a vast array
of different needs. See their web page at
http://www.statsoft.com
----------------------------------------------------------------
Take a look at a package called SIGMA STAT. It is part of the SIGMA PLOT
family made by Jandel Scientific in San Rafael CA. I think they also add
on packages for image analysis too.

I found SIGMA STAT quite useful and easy to use. The SIGMA PLOT package
lets you do lots of variations on multiple x-y plots with flexible error
bars etc.
---------------------------------------------------------------
We use Sigma Stat by Jandel Scientific. It's pretty straightforward and
easy to use.
---------------------------------------------------------------
I'm quite fond of SigmaStat (though it's overpriced), in that it does a
nice job of reporting in declarative English sentences, and does power and
sample size analyses. A good alternative to SAS or BMDP, which are
basically the final measuring sticks.
---------------------------------------------------------------
Try Prism 2 it is wonderful. We have been using it for years. The only
problem is that you can't enter multiple X and multiple Y.
http://www.graphpad.com/www/welcome.html
----------------------------------------------------------------
SAS - versatile, comprehensive, a new programming language for you to learn

StatView - spreadsheet format, menu driven, can not do repeated measures
ANCOVA for example
----------------------------------------------------------------
SWF "Statistics With Finesse" is not very sophisticated, but very friendly
and semi public domain. Written by James Bolding from U of Arkansas,
available from W. C. Brown Publishers/Brown and Benchmark, Dubuque, Iowa.
----------------------------------------------------------------
I suggest S-PLUS.
http://www.mathsoft.com/splus/
----------------------------------------------------------------
As for statistical software, I'm partial to S-Plus. It's strengths are
linking data with some flexible object-oriented functions and graphics
which enables you to visualize patterns in your data that you might not see
with other packages. The graphics are easily imported to and easily
manipulated in MS Word for Windows.

They've improved the GUI so that it's easier for the beginning user. The
true power comes when you start creating your own functions and learn the
"S" language. I must caution you on my bias towards the flexibility of a
well-scripted command line although I must admit that I'm attracted to
using the mouse to format my graphics. In the command line window the
responses are highly interactive, akin to the feedback in the Matlab
command line responses. This makes writing functions fairly easy.

For statistical work, I've used S-Plus mainly for exploratory analysis,
repeated measure ANOVAs, regression, and discriminant analysis. I actually
miss SAS's volumes of documentation and their mean separation procedure
with repeated measure ANOVAs although the most recent S-Plus version (I'm
still using version 4.0) might better handle the latter problem.
----------------------------------------------------------------
Not sure what you are looking for? Something on a Mac platform? Some
specific statistical procedure? These are all very powerful statistical
packages that no longer require mainframe computing power and are therefore
convenient, and accessible. I use both JMP and Systat and like both of
them. JMP is my preferred program, although I use Systat for some of the
more complex RM-ANOVA and regression applications. Frankly, I see no reason
to look any further.
----------------------------------------------------------------
We are using SPSS Win 7.5 and are very pleased about it
----------------------------------------------------------------
I have found Number Cruncher Statistical Systems (NCSS)to be quite
versatile and user friendly. You can easily import Excel or Quattro Pro
Spreadsheet data. You can find it at: http://www.ncss.com
----------------------------------------------------------------
I like the statistical software called Statistix, it is window based and
very easy to use.
----------------------------------------------------------------

************************************************** **********************
Theresa Foti, Ph.D.
Shriners Hospitals for Children
950 W. Faris Road
Greenville, SC 29605
(864) 240-6289
(864) 271-4471 FAX
tfoti@infoave.net
************************************************** **********************

-------------------------------------------------------------------
To unsubscribe send UNSUBSCRIBE BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://www.lri.ccf.org/isb/biomch-l
-------------------------------------------------------------------