Your question implies one of these two is the BEST. Why do you make this assumption? Also, as previous people have noted, "it depends" which of these is likely to be better on both your circumstances and what you are specifically working on. I use both, but one about 95% of the time and one about 1% of the time (and 4% on other ones), but I don't know what you are trying to do.

Both are excellent software. As a statistician, I prefer R over SPSS. SPSS is easy to use, but you have better control (in terms of coding, analysis, graphs, etc.) in R. Also, R is an open-source tool, and you have expansive free options compared to SPPS. Check what suits you.

https://1lib.us/ireader/5825760 I prefer R I have attached a couple of examples of R's power and flexibility. See the above link for a very powerful introduction. Good luck, D. Booth

If you are new to data analytics then SPSS is a better choice because of its user-friendly interface to perform statistical analysis with ease from SPSS you can create basic visualization this problem can be overcome by R, R has a wide range of visualizations.

R is best for (EDA) exploratory data analysis. R and SPSS both are slow when it comes to handling large data to solve this problem you have to go for another tool.

Dear David Eugene Booth

As Abas M. Sharif said, if one is new to data analytics then SPSS might be a good start. Also if one is just working on an individual project, in a way SPSS can help. However, for heavy duty workload like the case of Bio-statisticians (I am one), SPSS is by far less useful when compared to R. With the point and click in SPSS and of course though some programing, a bio-statistician workload can not be handled by SPSS since there is a range of projects to deal with within a short space of time.

Remember SPSS was put together by some individuals who thought of some ideas on what are the most likely needs for data analysis and the updates come at some point each year. Real life data problems are so dynamic and require one to reprogram in order to come up with the best solution. With that SPSS is far outweighed by R.

I visit SPSS on my way to AMOS. Otherwise, I am with R about 99.999% of the time. R can do almost anything from numbers, text and images. Most importantly, data transformation where SPSS has no match at all. R macros so called fucntions are extremely amazing than SPSS. In short SPSS is suitable for beginners in data analytics.

Your question implies one of these two is the BEST. Why do you make this assumption? Also, as previous people have noted, "it depends" which of these is likely to be better on both your circumstances and what you are specifically working on. I use both, but one about 95% of the time and one about 1% of the time (and 4% on other ones), but I don't know what you are trying to do.

Interesting insights, all. Thanks a lot for sharing your take about both in terms of practicality as perceived from your own experience.

As mentioned before, there is no best, but you have to look at what suits you. As mentioned before, SPSS is expensive outside the academic world and R is free. With R, you can install a menu like structure using Rcmdr. This will give some connection to SPSS and might help you to learn both. Otherwise, buy the book by Muechen's 'R for SPSS and SAS users'. It shows you how to do most of what you have learned in SPSS to do in R.

Another option would be to use JASP. This is also open source and in effect is a menu on top of R (https://jasp-stats.org/). It is still in development, but I have used it with factoranalysis because parallel analysis can be included as an option.

I am a poor SPSS user, it is much maligned but for some aspects of work it is vital - it is very impressive with large and complex surreys with millions of records. Indeed some of the data I work on comes as a SPSS saved file. I then manipulate it in SPSS before transferring selected data to software like MLwiN and Mplus that has specialist models and 'fast' estimation. You need to be aware of the capability of different software and you need to argue the case for access to multiple facilities. I still do a lot of my work in Minitab as I can write executable macros which speeds up work, and there is good editable graphics. - I have been using it since the late 1970s. It also has a very effective interface for modelling discrete outcomes. It was purchased for all UK universities after evaluation - at the time it was the judged 'best' interactive statistical package. So for me it is horses for courses. as per Daniel Wright

I have been using SPSS. It is a very helpful and useful statistical tool.

SPSS is the one I am use to, I have not been able to use R. Therefore, I will like to subscribe to SPSS for now.

Kelvyn Jones indeed executable macros are a way to go if one is involved in a mass production of statistical results. Hassan Izzeddin Sarsak also, SPSS is a useful tool. Daniel Wright true, it depends on what you are trying to do.

To help someone who is fairly new to data analytics, I will put it this way:

There are many statistical methods out there and each may be using some underlying algorithms of which there are usually several options for one to choose from. Software like SPSS is designed to group together these different statistical methods and under each group, subgroups are also given for further selection. Hence, by just exploring the menus in SPSS say or similar point and click software, one can learn more of the available statistical methods out there. Bearing in mind that the grouping of these statistical methods is a way of guiding the user in navigating the field of statistics without much theory involved as there are just headings and subheadings or topics. This becomes handy for a new person. However, with the grouping in place, you are kind of having blinkers as you are relying on what the authors prefer in terms of the analysis procedure. You have the limitation of not having full control of the data and the output. Taking for example SPSS, you have to thoroughly clean your data before uploading. Of course there are some few functions for data transformations and depends on what the authors have built in. So, in general, SPSS and similar software are good for understanding statistics from the basics and working on an individual project where you have all the time to run the analysis and manually tweak the graphics e.g editing the titles if say your variable names were having underscores. Further, with SPSS, you will need to re-write the tables into most acceptable publication format. Not the whole lot of SPSS results tables may be useful for publication material and you will have to take what you want. All the results presentation is based on the authors' vision.

Now that one knows what is happening in software like SPSS, that is where you can think of software like R. First of all, R requires you to know the statistical method that is applicable to what you are working on. Having that on the table, you only need to activate the specific library for that statistical method and the depends comes along. The method comes with all the options and you have to know what you are doing - the theory is provided in R studio and gives you the opportunity to read and decide on what you exactly want. Remember this is just on one statistical method and this was assuming that your data was already cleaned up and if not there are tons of other libraries to make use of prior to the actual data analysis.

To then better understand the capabilities of R, you need to think of a production line in a factory. Raw materials come in and processed in a series of stages until the final product. To avoid doing each step manually, machines and robots automate the work since one programmed everything to be repeated the same way when a new raw material batch comes in. This is where R comes in when you program and do most of the things on the fly. Unlike SPSS where you follow the software author's vision, with R you own the world and do what you want and be able to execute the same way for a new dataset. Once programmed today, tomorrow requires you to press Ctrl+A to highlight all and click run. The results become available with no time and that same code can be used for many datasets. That includes repeating all the graph editing and table formatting to the selection of what you really need as before. Take note that with R you can program to send your results to Ms Word and Excel with graphs and tables already formatted in some ways that are even difficult to do in Ms Word. After all, if graphs need the titles to have superscripts, subscripts, rotated, paneled, and even bring several graphs into a single graph, R is built for production and giving the full control of everything you can think of. With graphics, you can have your plots occupying just a small space with all the details readable. With SPSS giving long list of tables for each variable and also graphs elsewhere, R allows you to annotate the statistical tests on the graphs themselves - paves the way for a better story telling.

Hence, R is basically designed to give you the platform to redesign any other statistical software into doing what works best for you and mostly for efficiency where you can add new rules to streamline your work procedures. Just to add, a mix of dialog boxes and R code is the best. You code for R to decide and where it is necessary to stop for user input, that is pops out a input dialog box (see R widgets). The work horse for all the R capabilities is in the "defmacro" function in 'gtools" library. From my experience, imagine the output and google an R function for the task and the next minute you will be amazed. You can also write any R function that has not been imagined by SPSS authors or similar software because real data requires real solutions. The R community shares everything online and there are latest solutions than any other software. Remember, modelling and getting the p-values is the destination. The journey to that is what makes all the difference between R and any other software. With thousands of rows and thousands of columns, you need to sift the right data and do some checks either by plotting or tabulating with R agility.

I hope that gives someone new to data analytics some idea of what is really R and why SPSS should be the starting point.

Best wishes!

To add, most statistical programs have their own programming languages, including SPSS (Sax Basics comparable to VB. SPSS can be called from Office programs).

One of the main differences: R uses objects, SPSS does not as such. In R it is therefore quite easy to ask for the differences in fit for two or more nested models. SPSS does not have that option. R resembles the old GLIM4 (early 90's).

But if you want to have the best of both worlds (large data manipulation in SPSS and fast estimations and some tests of models), integrate SPSS with R and Python. The only limitation for SPSS so far is disk space, but I have created a 650GB file on my home pc. After the integrations of Python and R, you obtain a lot of free additional menu items in all SPSS menu's.

It really depends on 1) how much time and effort you want to put into learning new software and 2) what purpose you want to use the sotfware for. R will take a lot longer to learn properly, but it will have possibilities/options for doing any kind of analysis common to see in the social sciences. The opposite is the case for SPSS.

Interesting ideas! Thank you, all.

In my opinion, If you want to do a project, worksheet base software such as SPSS is very good. But if you are going to write a paper the R software is the best.

I personally think it is a matter of preference and familiarity. I would pick SPSS. I tried R, but it just wasn't working for me. I found it too complicated.

R is program-based and very very very stronger. SPSS is more user-friendly.

As most people said, it depends on what you need. The first software I learned was SPSS. SPSS is user friendly and easy to learn and use. But R is more flexible and can conduct complex analyses that could not be performed in SPSS. I especially like to use R for data visualization. Yeah, sometimes typing long syntax in R is a little bit annoying. When I don’t want to type long syntax, I use Stata for quick analysis and Mplus for the SEM model. This two software are also very good.

Superb!

Thanks to all of you for sharing experiential insights.

Cheers,

For me, SPSS the best because I am familiar with this software compare R

SPSS is much better than R for decision trees because it does not give numerous algorithms. The SPSS interface is considered to be understandable and user-friendly. But the only drawback is, it is not freely available for the users.

Very interesting discussion.

I completely agree with Daniel Wright and Mehmet Mehmetoglu .

None of these two softwares is really "the best".

I think that if you master one of them (it seems to be SPSS in your case), it's generally sufficient. So, stick on it. And if you can learn and master R also, it will be on your advantage. In case you need some specific tests that SPSS cannot perform, then you can still use R for them.

But you don't necessarily need to choose one.

Interesting ideas!

Thank you, all, for sharing.

SPSS is too much user friendly but limitating. I love R, sometimes i waste a lot of time to fix codes but it give me the possibility to fit in the best way the statistical application to all sort of research works

Thank you so much Matteo Velardo and I repeat:

"I love R, sometimes i waste a lot of time to fix codes but it give me the possibility to fit in the best way the statistical application to all sort of research works ".

I appreciate what SPSS and similar software have done to arrange the topics in the GUI but in practice, there are many limitations to what these GUIs can offer in most statistical applications and with efficiency. I have attached some of the work that I do with the R capabilities and would love to hear your views.

Thanks

Thanks, Matteo Velardo & Partson Tinarwo!

Cheers,

For SPSS some extra features are only available if you use Syntaxes. Most people do not have a real grasp of the possibilities of SPSS using only the GUI and menus. One advantage of SPSS: it can work with very large datasets.

I have encountered the limits of R frequently: unable to load a vector of 2.8Gb into memory. You can have more memory, but I can only be addressed if it is at one core.

It is one of the reasons I have SPSS integrated with R and Python. Sometimes I have to go to R as it does not always work as it should work in the combination.

If you want to use a GUI with R, use Rcmdr or one of the extensions.

Thank you Peter Moorer for highlighting the issue of large datasets with R. Indeed integrating the software at times can be the way to go. Although all the three: SPSS, R and Python can use syntaxes, SPSS may not be compared to the other two for syntax capabilities and quality of outputs, especially publication ready material.

As for the Rcmndr, that is far limited to what one can need for flexibility. In short, there is no GUI from any software that can replace the efficiency of syntax. That explains the versions of these software we see each year or from time to time. Daily problems need much flexibility which at times the GUI developers might not have thought of. If the GUIs have not thought of your current problem then you are stuck until the next time someone ask them for that feature.

In my opinion, for a once off project and not troublesome dataset, the GUIs can make it but if handling different projects at the same time and for a long period, the syntax based is the best since this allows you to program and piece together a solution keeping track of every activity and later make changes easily.

Best wishes!

Partson Tinarwo , maybe for R you should try R Markdown. In the last month I had to learn it to create additional material for a publication and R Markdown can create output of high quality. IF you include R Bookdown, you can add a bibliography as well.

I have to explore R Bookdown for its connection to Reference software. I work with RefWorks to include references in WORD. R Bookdown -as far as I am aware- does not connect to RefWorks, but it should connect to Zotero.

IBM SPSS has an add-on option for calling R function. It is called "R Essentials for SPSS". Once you install this add-on, you can use many commonly used R packages in familiar SPSS interface. There is also a plug-in for Python.