Agree with Warren... it depends on your needs. I think that R has the steepest learning curve and SPSS the easiest. Also, I think that R, Stata, and SAS are generally more capable in the arena of more complex models. I think that SAS and R might be considered to be industry standards so, from a career perspective, they might be the two best to invest your time in.(hopefull your company makes SAS available).

Mark

Having experience with SAS, R, and SPSS, I usually prefer R but I'm glad that I am familiar with the others. If I were to rank the three for my needs it would be R, SAS, then SPSS. I do have to say that the only reason I use SPSS is because my university and colleagues strongly encourage me to my statistics courses using it.

I would also recommend R.

Taking note of the previous comments into account, I would say that it do depends on the need but it also depends on other factors such as your computational skill, knowledge, familiarity of the platform and environment etc.

I also agree that R is the best but in terms of its wide range of packages available for almost all types of statistical analyses but it may not be best for a person who is comfortable with GUI based software and is not familiar or have difficulty in comprehending the script, syntax, and commands of the R.

Further, each statistical analysis involve computation of a number of parameters, and estimates and most of the software provide a handful sample of such parameters and estimates. Thus, it is likely that some software may provide those parameters, estimates and tests that one need for a particular analysis but not for other analysis. For instance, the SPSS ROC analysis gives the ROC curve and a table of sensitivity and specifity for a range of cutoff scores for a given indicator along with the value of AUC. But if one wants to compare two AUC and examine the significance of the difference between two AUC it is not available in it. For this purpose, one may prefer the SigmaPlot. which provide all the output available in the SPSS plus many other facilities including comparison of AUC. However, if one is interested in testing the diagnostic performance of an indicator in relation to a given cutoff of the other indicator, both SPSS and SigmaPlot are not of much use. In this situation one may need the MedCalc.

Thus, for the same analysis, which is avialble in several packages, one software may prove to be the best in one situation but not in the other.

I like SPSS because of its ease of use compared to SAS. Some people prefer SAS specifically because of its obtusiveness, which makes them feel more like an expert in using it. SPSS also publishes its algorithms so I know what it is actually doing. I also prefer the matrix programming language in SPSS.

I did not compare with R, because this is not what you asked.

Prepare to get an on-slaught of differing opinions, because asking what is the better statistical package is liking asking what is the better religion.

SAS and SPSS do nearly the same things. They both have drop down menus and both can use code. They both have most of the same analysis, so neither is "better". It depends on which you like. But as some mention, they both cost a lot, while R is free. There are plenty of tutorials for R and also some graphical interfaces to make it easier. I list most of them here

http://gsociology.icaap.org/methods/soft.html

It depends upon what you want to do and what expertise you have in stats. There are several other programs, which can do every thing even spreadsheet program like MS Excel.

SPSS is good for non-statisticans as it has a user-friendly interface and drop down menus. However, if you have more than 10,000 observations (sometimes less, even 7,000) it will crash. This is where SAS is superior as it reads the data off a network and will never crash. It does however require some programming skills.

I haven't experienced SPSS crashing with large sample sizes but it does have limitations compared to SAS. In my mind, SPSS language is a bit less robust than SAS's and there definately are procedures that SAS can do that SPSS cannot. If you have limited needs then SPSS is fine. Otherwise, I'd commit to learning either SAS (especially, if it's paid for) or R (I've never used it but others in my group think very highly of it)

To give a recent concrete example regarding something SAS can do that SPSS can't (without an expensive add on program). I'm planning a new statistics course for my department's masters students and I need software that can run Structural Equations Models. SAS can do this but SPSS can't without purchasing an expensive extra piece of software called AMOS that my university does not have a site license to.

Hi dear colleagues

I prefer SPSS because older and more used by many researcher.

In addition, the help of SPSS software is too good for whom that start using SPSS.

But in SPSS, diagrams are not good and you must use other software such as excel and graph pad.

Sincerely

I don't think there is a best statistical package (i use SPSS and had a look on SAS few years ago). R should be encouraged because it is a free software and comprehensive but i personnaly did not have enough time/motivation to learn to use it and am still using SPSS not because I think it is best but because i know it and it fits my (simple) needs.

Difference between SAS and SPSS are not noticeable for most of simple current statistics, the problem is availability (both are expensive) and ergonomy for basic users.

If I was new in the field, I would start with R.

I don't agree with Sinead O'Neil on size limitation of SPSS : I usually handle files with 200 000 to more 1 million observations with SPSS 17.0 without problems (of course treatment is slow but crash are extremely rare)

Sincerly

Consider Stata--much easier coding than SAS, but much of the same functionality (if not more). R is also a good recommendation since it is free.

Thanks for all. I may conclude that there is no consensus on the best statistical program, as each program has its own characteristics.

I used to analyze my data using a software called "Statistix", the main advantage of it is the simplicity in inputting data and in selecting the proper analysis. In other words, if you are not familiar with statistics, so, this program is your primary choice.

The software has a 30 days full features trial (with limited variables and cases numbers), find it free at http://www.statistix.com/free-trial/

Note:  I do not have any conflict of interest with Statistix software and Analytical Software company.

 How can I get free R software for data analysis?

SPSS is the most user friendly, by far. Though I have not attempted using SPSS for time-series papers, a cursory exam makes me suspect that SPSS is severely lacking in this area. But for everything else, it is wonderful.  And incredibly expensive. Including taxes an individual license costs more than $6,000. STATA costs a few hundred.  

It depends on your research field, SPSS is originally designed and developed for social science and have pretty good ability for nonparametric designs. it is user friendly software. However SPSS is not good software for experimental designs in biology and engineering and you should use SAS or R software’s instead. For researchers that work on curve fitting and regression, the Sigmaplot is good software.

However, all of mentioned software’s are able to perform common statistical procedures if we know how to use them.

After 10 years work with mentioned statistical software’s, I suggest SAS especially for agriculture, biology and engineering researchers. Please consider these abilities too;

1-       Solution®analysis®analyst for spread sheet format of SAS that is similar to SPSS window.

2-       Html format of SAS is pretty nice. tools®option®preference®results®create HTML

3-      Import and save procedure shortcuts in file shortcut part

I've found good comparison of some softwares here http://www.statistical-software.net/ Hope it'll help. :)

one more website about spss http://www.ibm-spss-online.com/ :)

Dear Khan

1. There are many similarities between Statistical Software Packages (SPSS, SAS, R, Stata, JMP, …) in the logic and wording they use even if the interface is different.

2. Many schools offer only a site license for only one package, and it may not be the one you’re used to.

3. Statisticians, social scientists, … should generally learn SPSS as their main package, mainly because that is what their colleagues are using.

4. SPSS is used by market researchers, health researchers, survey companies, government entities, education researchers, marketing organizations, data miners, and many more for the processing and analyzing of survey data.

5. IBM SPSS Statistics software can help you find new relationships in the data and predict what will likely happen next.

Regards,

Zuhair