PDA

View Full Version : SPSS for Linux comming soon



sonny
June 23rd, 2007, 12:05 AM
Hi, I was looking at the development of the SPSS replacements under linux, I've used several apps but none as good as SPSS (in my own opinion), the only one close enough is the R-project but you have to learn the language. So I e-mailed the guys at SPSS and here is the response I got from them:


Thank you for your note and your interest in SPSS. When our SPSS 16 version is released, we will have a client version for Linux available. SPSS 16 is expected to be release in late September 2007.

Best Regards,
Brian
SPSS Inc,So, we won't have to wait too long... we'll have SPSS soon. That's good news, at least for me.

Footissimo
June 23rd, 2007, 12:23 AM
I'd love to be able to afford SPSS =( but it's great news that it is being released for linux. For those who don't know - SPSS is a application (well a set of applications) that are used for statistical analysis - as far as I know, it is the best stats package out there.

emrextreme
June 23rd, 2007, 12:31 AM
Thank for the disclosure. It would be such great for these professional programs to be releasing for Linux. That's the proof of upswing of Linux.

Taquoriaan
June 23rd, 2007, 12:49 AM
As a psychology student I'm familiar with SPSS. As an ubuntu user I feel forced to maintain a dual boot system in order for SPSS to function properly. My university has a volume license for SPSS, so I was able to buy it for €12,50. It's a pity indeed, it wasn't available for linux, it would save me so much hard disk space. SPSS is the only reason for me to run windows on another partition. I also have a student licensed copy of Windows XP, so it was affordable, I paid about €25 for a copy of Windows XP Professional. It expires when I leave university, but I don't need it afterwards, so that's not a big deal for me.

At the other hand: it there was a linux version available for linux, my university would probably not have a volume licence due to the lack of students using linux. So I don't know if a linux version would solve any problems.

There is another statistical package besides SPSS: JMP/SAS. AFAIK the package has the same functionality, but the big difference is that it's available for Windows, Linux and OSx. When you can prove that you're a student, you can get a student's license which costs about $85. It's more expensive than SPSS is, but if you consider the costs of an OEM-licence needed to run Windows, it's actually cheap.

An ordinary licence (non-student) for both SPSS and JMP/SAS costs about $1100, so for most of us that's no option whatsoever.

More information on using JMP/SAS as a student go over to www.jmp.com (http://www.jmp.com/academic/student.shtml)

5-HT
June 23rd, 2007, 12:54 AM
:p
I was using an much earlier version (6 I believe) for Solaris, then got into R. Glad to have a native Linux version, but I really hope they don't just port over the exact same thing as the windows version-- I don't like what they've done in recent versions so much...

Now if only NCSS was available...;)

Taquoriaan
June 23rd, 2007, 01:02 AM
:p
I was using an much earlier version (6 I believe) for Solaris, then got into R. Glad to have a native Linux version, but I really hope they don't just port over the exact same thing as the windows version-- I don't like what they've done in recent versions so much...

Now if only NCSS was available...;)

Yeah, I do recognize it. I'm finishing my bachelor's degree and have to do multivariate analysis. In order to do it correctly I'm a heavy user of SPSS's syntax-editor and basically run all analysis from the syntax editor. The menu options and buttons don't offer the things I need. It's fine for basic and moderate statistical analysis, but when you reach an advanced level, it's not sufficient anymore.

HotShotDJ
June 23rd, 2007, 01:07 AM
More information on using JMP/SAS as a student go over to www.jmp.com (http://www.jmp.com/academic/student.shtml)I tested JMP/SAS and found it inadequate. I used some test data and was only able to run some basic univariate and multivariate analysis. For undergraduate research methods and statistics, I can see it being useful. But in real research settings, it simply lacks the functionality that SPSS provides.

I look forward to being able to run SPSS on my Linux box! I'm quite tired of running it from my university's server using Citrix.

Taquoriaan
June 23rd, 2007, 01:18 AM
I tested JMP/SAS and found it inadequate. I used some test data and was only able to run some basic univariate and multivariate analysis. For undergraduate research methods and statistics, I can see it being useful. But in real research settings, it simply lacks the functionality that SPSS provides.

I look forward to being able to run SPSS on my Linux box! I'm quite tired of running it from my university's server using Citrix.

I never tried JMP/SAS, for the record. I only saw some demonstrations of it. Since SPSS means Statistical Package for Social Sciences and I'm a psychology student, SPSS is the logical option. You need also more advanced statistics research in psychology, than for instance in marketing. So for some people JMP/SAS will do, for others it won't. For me even SPSS isn't sufficient. Sometimes I need Matlab in addition to SPSS. I wonder if they have a Linux version, too.

I just wanted to say there's a possible alternative to SPSS, I did not want to claim that it would be exactly the same package.

HotShotDJ
June 23rd, 2007, 01:37 AM
. For me even SPSS isn't sufficient. Sometimes I need Matlab in addition to SPSS. I wonder if they have a Linux version, tooI believe Matlab has a Linux version. However, I've found that I can usually get what I need in SPSS, even if not by any "point-and-click" method. For example, while there is no GUI to calculate Mahalanobis distance to detect multivariate outliers, it can be done pretty easily by using the syntax editor.

UbuWu
June 23rd, 2007, 01:42 PM
Hi, I was looking at the development of the SPSS replacements under linux, I've used several apps but none as good as SPSS (in my own opinion), the only one close enough is the R-project but you have to learn the language.

You don't have to learn R to use it. There are several GUI's available. The one that comes closest to SPSS is RKWard (http://rkward.sourceforge.net/).

euler_fan
June 23rd, 2007, 02:39 PM
The part about learning R is true, but that's where its real power lies--being able to write scripts that will automate your statistical analysis. Combine that with the wealth of packages and there is little not recommending it. But I do agree, it does have a pretty long learning process because of the complexity.

HotShotDJ
June 23rd, 2007, 06:00 PM
The one that comes closest to SPSS is RKWard (http://rkward.sourceforge.net/).Thank you for sharing that information! I would certainly prefer a F/OSS solution to a proprietary solution. I know that R is a very capable statistical package, but I've never had the time to deal with the learning curve -- I've got enough things to do as a middle-aged graduate student. :) RKWard may be just the tool I need to get me using R -- and it is in the Ubuntu repositories!

orthopod
March 2nd, 2008, 10:33 PM
I wound up using JMP for stats, and I like it quite a bit. Fairly easy to use, point and click, does a fair amount ,ANOVA, multivariate, curve fitting, survival curves, etc.

I use it for my medical research