Which is better, R or SPSS?

Use Graph PAD PRISM V. 5 for any research related bio-statistical work

Thank you Lalit Samant

R is free SPSS not.

I would also consider Stata.

I would go for R if you know that all the statistical tools you need are available in R (through packages). This is because it is free (it doesn't cost a dime), multiplatform (this enables easy collaboration) and open source (it ensures the reproducibility of the analysis).

Thanks guys:)

It depends on your use case: If you need something fast without learn to write coding lines, then you should use SPSS. If you want to dig in into bigger problems, where tweaking and big data sets are important, then you should learn R. (I used many times for the visualization Graphpad Prism as someone suggested before.)

on the long run, R is definitely better, as it is more flexible to your needs.

And R is not a blackbox thing. You know what happens to your data while analysing them. Furthermore, most up-to-date methods are not implemented in SPSS (e.g. Random Forests Classification or Boosted Regression). Finally, most R packages are accompined by a peer reviewed publication, making them more reliable than using an algorithm in SPSS where nowbody knows how it works.

One reason SPSS and SAS are so prevalent is because many older faculty and established research groups have been using it for years. Prior to R, these were clearly better than having to write your own programs. They provided easily repeatable and easy to verify results.

Flash forward to today, we have R which may require some programming skills; however, many packages are available that minimize this need. R is free, new packages are available as quickly as the theory is published, and it is now being accepted by a wider audience as a valid alternative to the commercial software. While SPSS and SAS are likely not going away, but as budget cuts encourage the use of open source and freeware, the younger generation of scientists who learn R will encourage its use in subsequent years.

thank you very much for all opinions guys. Very informative and useful knowledge. Thanks again

"Based on an analysis of Google Scholar data on usage of statistical software, Bob Muenchen makes a forecast: R will overtake SAS and SPSS in 2015". see more at the R cheerleader blog: http://blog.revolutionanalytics.com/2012/05/how-long-before-r-overtakes-sas-and-spss.html

IMO, R takes the cake for its great simulation abilities, which is invaluable to visualize, intuit and explore fundamental assumptions of statistical models

I think R is better.

Their are a lot of topics asking this question already. The most frequent answer are:

SAS is easier to manage data set, more used in company and they are also SAS certification center. It is more powerful in huge data set, and it exist jmp in order to get interface help.

I use R because the open source format is providing nicer exchange (forum or coffee machine). Moreover it is also easier to build complete personalized tool that could used as routine in a team/lab. It exist couple of solution to integrate an interface as RKward or Rcmder

Best

I think that R is better than SPSS

It is obvious from the foregoing posts that R beats SPSS hands down. For someone who seeks to learn a new Statistical package, the writing is on the wall "R" and more "R". That said, SPSS still has enthusiasts and will be with us for sometime to come.

It really depends on your needs SPSS is a brilliant piece of software if you want to use predefined statistical test and you dont want to spend a lot of time preparing code and data sets. And on first glance it is similar to operate.

R on the other hand allows a far wider scope of statistical evaluation, the software allows you to design and code your own test which in essence allows you to create your own tailor made statistical package for your project. R also allows models to be created and linked with other R features so that is a plus.

On first use R does look difficult to use but after you know some key lines of code it is very easy to build upon this crating more advanced codes.

In terms of what is better SPSS or R it depends on your field I believe SPSS is still the top software in the social sciences and is the best tool to teach statistics to all undergraduate students (physical & social science). But I do believe that at doctoral level researchers and above in the physical sciences and engineering should be using R or one of its equivalents.

I think R is more flexible and includes much more statistical models. SPSS is good for beginners.

In my opinion, one of the advantages of using R in statistics is that you know how the test is dealing with your variables. The interface of menus of SPSS or other statistical programs is more like a recipe book, and it is easier to fall into the “garbage in, garbage out”. R might look scary, but once you get familiar with the codes, the benefits outweigh the concerns.

If you work together with colleagues, SPSS is social sciences, SAS is medical, Matlab is engineering, S-plus banking (?), R is open source community. The latter two are implementations of S, where R is line oriented and open source and S-plus is menu oriented and commercial. I happily use R.

BTW line oriented systems requires you to learn command names by heart and have a steeper learning curve.

For the real beginners, the uninitiated, Graphpad instat for analysis and PASS for sample size caculation is a good alternative. it also has a excel like user interface and the learning curve is fairly simple

I agree that someone just starting to learn a statistics package should probably choose R over SPSS; its prevalence and versatility is only going to increase. R does have a steep learning curve for non-programmers – having initially learned SPSS in a UNIX command line environment, I've come full circle with R. Fortunately, there are front ends like R Studio and R Commander (a GUI interface) that help the novice get up to speed quicker.

Depending on the project that you will conduct or the scientific field, one might come handy compared to the other.

SPSS is a packaged software and if you only need basic, pre-defined statistical analysis tools (as in the case of the majority of studies in social sciences), then it would be reasonable to use SPSS. As mentioned in the previous comments, SPSS has an eye-candy, user-friendly interface and you can allocate more time on the research itself, rather than learning the software.

On the other hand, R is actually an environment where you can "code" your own functions and as you learn, you will be surprised how flexible it is. However, the question is whether your project demands such flexibility and advanced statistical analysis. If it does, then it will be worthy to use it in spite of the steep learning curve. There is abundant instructional material for R on the Internet.

Python

Just to be devil's advocate I would say that I have wasted loads and loads of time playing with R (partly because it's fun to learn opensource software/languages). I sometimes wonder if I would be more productive just using the rather boring proprietary tools like SPSS. Same goes for Grass v ArcGIS. I'd say If you enjoy programming go R if you enjoy getting papers out quickly go SPSS.

Its depends on your data and your goals.

some times SPSS and some times R software.

Having used and taught SPSS, SAS, Stata and R (and keeping to the question) I would say R is a MUCH stronger statistics package. I agree with many above that if you analysis is very straight forward, and you don't do too much of it, SPSS may be a better option. However, I have to take exception at the statement that SPSS is "brilliant" (@Craig). Perhaps SPSS's only strength (over other packages) is in the area of Factor Analysis. This is explained by the high prevalence of Factor Analysis in some of the social sciences (rem: SPSS=Statistical Package for the Social Sciences). I would not suggest anyone working in the area of the physical sciences use SPSS (EVER!!!). One of my main problems with SPSS (Agree with @Danny above), is that it so often allows you to do the wrong analyses, switches your outcome event (0->1 and 1->0) or referent groups without asking (more often than you would think random chance would would explain), I don't even want to bring up the problems with its Mixed effect modeling modules (I want to limit my response to <1000 words). My feeling is that for very basic statistics SPSS is probably OK, but it only gives the illusion of being easy for anything beyond classical bivariate methods (all modeling).

For those in the health sciences (clinical sciences / public health / epidemiology) I think Stata is a good compromise (nice easy front end, and very nice succinct language....also MUCH cheaper than SAS or SPSS especially for students). As for SAS, which for many years was considered the gold standard, it's a strong package, but what a COW of a language. Generally pages of code for what can be coded in 10 or so lines in R or Stata.

Finally, I think for anyone who does a decent amount of analysis, or anything that is not 'off-the-shelf', R is a great package. Yes it does involve a little time to get familiar with (as many above point out), but there are so many good resources out there these days (You-tube walk throughs, freely available texts, R dedicated websites etc). I wish these had been available when I first came across R.

Last word....@Adam...Python...for stats...REALLY???

For both SPSS and SAS cost is one of the negative factors but you will want to check to see if your employer has a site license for one or both. That can reduce the cost substantially for either. SAS is probably better than R for really large data sets. R has made great inroads in terms of being known in the various disciplines and there are many books out now about how to use R in different areas. It also links well with other software, e.g.,,MATLAB. Needless to say access to the source codes for the various packages as well as the R software itself is a tremendous advantage.

R is for analyzing data, and helps you do that in original and creative ways. SAS offers a wide repertoire of standard analyses useful in a production context, and I imagine is also focussed on data management. Building and updating a data-base. If you want to both manage and analyze large qantities of data you probably need to combine dedicated data-base software with R. There are excellent free and open-source solutions for the data management side.

@Cameron, Pandas is a nice Python data frame package that is fairly nice to use:

Also it might be worth keeping an eye on the Julia language:

Both packages both R language as SPSS, have weaknesses and strengths. The difference is that R can be made with new routines to suit the new models (R is an ideal language for programming settings of statistical models) unlike SPSS whose calculation routines are pre flexible clip with no exits. However, SPSS has sophisticated models and methods of statistical studies.

R as a freeware is a good platform for statistic analysis. I started to learn about one week. R is really flexiable, if you use spreedsheet programme on your pc. It is better than SPSS.

In addition to stats, I enjoy using R for its nice, flexible plotting and data visualization capabilities. You can also use it for GIS and mapping. There is also a wide network out there of brilliant folks developing packages and users to bounce questions off.

Depends. The statistician in me says R. However, the piratical side of me says if you are looking for a user friendly program that will get you approximately the correct answer than SPSS (though Stata, SAS, and many other programs are way better than SPSS).

I prefer to use R, although it is true, the learning curve for R is very steep, but once you get used to the R environment you will realize of the unbelievable advantages of using it. Besides, currently there are a lot of material to learn it faster and easier than some years ago. So enjoy R and remember... it's free!

Short answer: try both and pick up better for You. It depends on Your needs and skils.

Thanks all :)

R needs programming skills and is more free in the statistic options on the other SPSS is close related program and have limited orders.

Whenever you ask a question such as "which of ****** is better", you need to know or specify the criterion for "better". otherwise you are arguing in circles. If cost is your criterion then R is the only choice since it is open source and freely downloadable whereas SPSS is commercial and you have to pay for it. If you want user manuals, a support service from the provider then you would choose SPSS. Mostly the tools and techniques available in SPSS are those selected by the company so flexibility is limited to what they provide. R is really in two parts, one is the basic software you get by downloading and installing the R software, the other is the ever expanding list of packages/libraries that can be downloaded from within R (from one of the mirror sites). The codes for the various packages/libraries have been written by different people but are required to be standardized in some ways AND most importantly the source code must be available. If you use R Commander you will not need to much programming at all. The source code(s) for SPSS are propietary so you can't see them. An important recent development, Microsoft bought a company that provides support services for R users and Microsoft is now supporting R in various ways.. If you are working with a group that mostly uses SPSS then it may be better to choose SPSS. If you want software that is readily available to students then R is far better.

 

R.

If more options, I prefer Python

I Find that R requires a lot of programming ability ...DOES  not IT?

I first posted on this thread years ago and rather controversially suggesting SPSS was the best option. I think I was going through a corporate phase! Since then, I've spent two years learning R and I've completely changed my mind. Learn to program in R (install R and RStudio), get involved in the huge user community, be amazed at the ever-expanding set of packages available to achieve almost anything, enjoy that R and it's packages are free so everyone benefits - not just IBM and wealthy universities that can afford the site license. You'll never look back. Plus the programming skills are universal so if you ever want to learn another programming language you will have a big head start.

Confirm R. 

FREE $ and develop

SPSS... my opinion

No matter what you are comparing you need to have a clearly stated criterion for deciding. When you look at SPSS you can only consider the features already provided in the program by the company (although SPSS has changed over time and presumably will continue to do so). R is not a uniquely determined set of features/algorithms it is in a constant state of flux and moreover it is really a programming language. R has many packages that are not statistical and hence are not directly comparable with SPSS. You have to think about the intended audience, for non-programmers SPSS certainly is attractive. In terms of cost there is no comparison since R (and all the various packages) are open source (both the source code as well as the binaries). SPSS is proprietary and even with a license one does not have access to the source code, i.e. the details for the various algorithms.are hidden. I suggest it is not useful to ask the question, "which is better R or SPSS?". As a tool for students and teaching open source software will always win out in the end.

R is an open-source powerful tools and its packages are being updated and too many new ones are being added, sometimes daily.

Also, it has nice connection with Latex to document your results directly as PDF.

The amount of programming necessary to use R to accomplish most basic statistical calculations is relatively minimal. General computer programming skills are now being taught in high school, with an increasing number of voices calling for *mandatory* programming courses for all students in high school and even middle school. Since programming skills can be taught to middle school kids, any college or especially graduate level student should be able to learn the basics of working with data using R. Given the lower cost of R (free!), and the clear superiority of R over SPSS for more advanced statistical calculations and analytics (note the comment by Cornelius Senf above), there is no question that the next generation of students should be learning R in their most basic statistics classes. The argument that SPSS is "easier" or "standard" in some way just doesn't hold up any longer. IBM has strong incentive to perpetuate these myths, but they simply are not true.

Thanks Ahmadu Mohammed !

Here is the Datacamp infographic (2014)

Just need to update the number of the current versions for 2018-2019.

It would be very useful if this figure is updated regularly to highlight current differences between the languages.

With R you can really know what you do and push you to understand what is the good parameters.

The documentation is very good and easy to understand.

I work under Linux and Windows and is so easy to work uzing the same files and scripts on dual boot computer.

I think if someone has the knowledge of programming then R is the winner.

SPSS doesn't change anything in past 6-10 years. Compare with huge number of new packages with new functionality: 12000+ !

If You don't like programming use Jamovi instead :-)

Clear understood about licence problem. It is not a problem for scholars of developed countries. But for us no one can believe or can afford by each student or scholar.

R has no alternatives even though SPSS or all other closed software are concerns.

Thanks.

Much appreciation for Paweł Kleka, brilliant methodologist. His answer is applicable - you may use Jamovi instead of SPSS.

I'm used to SPSS & Hays, "R" seems difficult

Which religion is better? This is akin to asking which statistical software package is better? You will get great debates in trying to answer both questions, generating more heat than illumination. The general answer is whatever are you more familiar with or what did you learn first. R is great for programmers, but give me SPSS if I have to teach students in a statistical course.

Sure I could point out limitations in R as well. I don't think you get my argument. R is great for programmers, but not everyone is a programmer (nor needs to learn programming).

BTW, it is possible to extend SPSS with a little programming as well. Check out my program for Hotelling's Profile Analysis in Raynald's repository for SPSS programs.

PSPP is an open source version of SPSS from the Free Software Foundation. Its development was stalled for a very long time, so it had limited functionality and a poor reputation. Suddenly there was a flurry of development for PSPP,, especially for the Windows version (PSPPIRE), and there is also a Mac version. I estimate that PSPP now has about 80% of the functionality of SPSS.

Some people may not know that R is based upon a commercial statistical package S (and Splus), which in turn variants of the C programming language. This may be one case where the popularity open source version (R) vastly exceeds the popularity of the commercial version (S). I am not sure that the latter is even sold anymore.

On the other hand, the earlier versions of source code for SPSS were passed around from computer centre to centre (before the concept of open source came into being), until someone realized that they could make money off of it. There was one extension of SPSS that was very popular, MANOVA, which could analysis several complex ANOVA models (not just multivariate). Unfortunately, SPSS, Inc. could not reach agreement with the original author of MANOVA so it was dropped, and replaced with GANOVA eventually.

I would say that R is great for many a researcher or clinician who are not programmers, but learn to do the necessary code (which is not that complex) and get results they need. Then they are very happy with the flexibility. On a more philosophical note, I think, the digital divide of the future might be between which people do learn a computing language (e.g. R, or some other) and are able to be sufficiently proficient in interacting with machines (grok them, kind of know what to expect from them) and those who continue to shudder at the thought. Maybe not, of course, but I think there is something in this thought.

With (all due) respect, every problem can be solved in Excel. There are a lot of books, Add-Ins, and tutorials for it. I recommend Power BI.

It is depend on the type of research you want to do. If you want to do some social research, than SPSS should be better choice because it is more user friendly and very similar to Excel. But, if you need statistical packet to do some complex calculation with very fast outcomes, then R and RStudio could be better choice. It is correct that almost everything can be calculate in Excel, but often it mean time consuming.

Thanks, Tijana Šoja. Great answer. True. For example, SPSS is great for processing of a survey. On the other hand, R and RStudio are better tools in estimating the probability of bankruptcy of a large number of companies in the sample.

Is not possible to define "Better" in this case because it is too subjective. It has been pointed out by others that depends of your background and the nature of the calculations you want to do with your data. If you are comfortable with programming, R is a great choice due to the fact that is free and has a lot of packages already available.

It depends on your previous skills and financial possibilities, I preffer SPSS because ist the easiest to learn and user friendly, however the cost of licensing is a strong barrier

I use SPSS for Windows but learned it originally using syntax not drop down menus. That gives me the best of both worlds. That said, i know the individual user licence is expense. i dont use R but presume it's not that difficult to learn and imagine that most young researchers (I'm a very old researcher!!!) will gravitate to it.

Here are two excellent and visual overviews of this topic through these well-designed infographics. A crane view of data science in general.

- https://365datascience.com/defining-data-science/

- https://www.datacamp.com/community/blog/data-scientist-vs-data-engineer

Additionally, I find the infographic titled 'Statistical Language Wars' from DataCamp one of, if not 'the' best, in terms of comprehensiveness yet with simplicity, user-friendliness, and self-explanatory as a comparison between SPSS, R, and SAS.

- https://www.datacamp.com/community/tutorials/statistical-language-wars-the-infograph

- https://101.datascience.community/2014/07/07/statistical-programming-languages-infographic/

I am hopeless at statistics but for my PhD I had to include some basic analysis. I had no knowledge of any stats software. I first tried R and spent about a week getting to what I needed. I then tried SPSS and achieved the same results in a couple of hours. If you want to spend more time on your research and analysing data then use SPSS. But if you want to spend time customising results and outputs then R.

Both are useful, indeed.