For most statistical techniques, all of the programs will perform well. SAS has the advantages of a long history, commons usage by biostatisticians and a wide range of statistical procedures. It's data step is very powerful for manipulating data, but does have its limitation. Output quality from stat procedures is good. I find the graphics only fair and confusing. Current versions of SAS are syntax driven, so you have to write a syntax command file, which requires some familiarity with what you want to do.

At this point R is the most complete statistics system with libraries available for nearly everything you might want. The advantage of R is that you can do anything you want as its a programming language that mixes statistical analysis with standard programming capabilities (interpreted, interactive). Unfortunately, this is also its main disadvantage as it leads to a high learning curve. Another strong advantage of R is that it is free. There are some IDEs to ake it a little easier, but in the end R is a programming language you would need to learn, not hard if you have any programming experience.

SPSS is the most user friendly system and produces nice output. Most work can be done via fill in the blank dialog boxes, yet it has most of the capabilities of SAS. This is definitely the easiest .

There are many free programs (such as openstat) that do basic analyses. It's a good idea to verify that they work correctly before production use.

Hi,

It is depending what you want to do and process. I’m using SAS, R and SPAD, and I teach of all these. If you want free software with a lot macro available, you could use R. It is good with not so strong errors. If you want to develop some heavy calculus and routines on large databases, so SAS is the best. For the rest, SPSS is going well.

Best, Sébastien

For most statistical techniques, all of the programs will perform well. SAS has the advantages of a long history, commons usage by biostatisticians and a wide range of statistical procedures. It's data step is very powerful for manipulating data, but does have its limitation. Output quality from stat procedures is good. I find the graphics only fair and confusing. Current versions of SAS are syntax driven, so you have to write a syntax command file, which requires some familiarity with what you want to do.

At this point R is the most complete statistics system with libraries available for nearly everything you might want. The advantage of R is that you can do anything you want as its a programming language that mixes statistical analysis with standard programming capabilities (interpreted, interactive). Unfortunately, this is also its main disadvantage as it leads to a high learning curve. Another strong advantage of R is that it is free. There are some IDEs to ake it a little easier, but in the end R is a programming language you would need to learn, not hard if you have any programming experience.

SPSS is the most user friendly system and produces nice output. Most work can be done via fill in the blank dialog boxes, yet it has most of the capabilities of SAS. This is definitely the easiest .

There are many free programs (such as openstat) that do basic analyses. It's a good idea to verify that they work correctly before production use.

If you deal with classical analyses (e.g., ANOVAs, t-tests, GLM, GLMM, MANOVAs, regressions), SAS is a very powerful package. The students in my grad courses find it a lot more user-friendly than R. However, if you are planning to work with unusual tests (e.g., MRT), R is a better package as you might simply not find these tests in SAS. Another major problem with SAS is cost and access.

If SAS were a nice, comfortable Audi, R would e a Lamborghini you can drive for free

I'd favor SAS over SPSS due to the extensibility of the product. SAS can handle immense datasets and has plenty of documentation of how to setup your appropriate analyses. You may also want to check out JMP, which is a SAS product. It has SAS hooks and has a friendly, low learning curve interface.

So how do you go about acquiring the certification of SAS. Because I make use of SAS software for my data analysis and I find it very useful and user friendly, especially when the Analyst version is used. I feel SPSS is too straight jacketed and is not as flexible as SAS. So please Sami DIAF, can you enlighten me on how I can link up with SAS institute for their certification. Thank you.

With SAS recoding your variables is very flexible and easy. R is good too, but am not sure it can handle huge datasets. As others have said SPSS has the best looking output and the most user freindly. R has the best graphs. SAS graphs are not great and seem clumsy at times but they're improving them all the time.

Different packages have different strengths and weaknesses. For example, for analysing designed experiments GenStat (which is one of my favourites) is probably the best (in fact the only one that seems to use John Nelder's theory of general balance). GenStat also has excellent menu and command mode interfaces and very powerful spreadsheet capabilities. On the other hand GenStat is not the package of choice for survival analysis. SAS has many good procedures but getting them to talk to each other is difficult and you probably need to learn about five different languages to use SAS effectively: Base SAS, Proc IML, macro, ODS and SQL.. This sort of programming is much easier in R or GenStat. R is very powerful, flexible and free! It has excellent graphics. One disadvantage is that it is rather unstable. Programs may not work if you switch machine.

Answer depends of

your experience in programming

your experience in statistic

how much time you can invest on it

do you need to analyze of data

To my mind, none of SAS, R and graphpad are perfect, but they are nicely complementary.

I use R because the open source format is providing nicer exchange (forum or coffee machine) and because most of companies do not figure out correctly troubleshooting help.

Best

Just to add a somewhat different thought: If you are interested in working in the pharmaceutical industry and programming or doing statistical analyses for regulatory agencies (e.g. FDA, CHMP, SDA, etc) then you will find that SAS experience is essential.

Application of statistical tools are different in each case of research.. SPSS, MInitab, Sigma Plot, or MS excel with Solver.. etc.. i use all the above ...