This is a poll asking what is your preferred stats program for your data analyses. I've seen plenty of researchers using R, but is it because it's free?
I have recently started to use R and have been learning how to use R via YouTube videos and blogs. It takes a bit of time to get used to the syntax, but is very feasible to learn and to be successful with in a short time. Although point and click programs are indeed faster for simpler analyses, R is more powerful and you can control more aspects of how things are modeled and analyzed. Thus, R is better for more complex questions and 'wonky' datasets. And websites like StackExchange have lots of forums where you can search for code or ask questions to fellow R users. I still use JMP or SPSS for quick tabulations and normally distributed models, but as I am getting more used to R, I'm transitioning to just using R for everything. With respect to creating figures, the visual display options for your data seem just as customizable in R as in SigmaPlot, but SigmaPlot is expensive. The graphic displays in JMP are horrendous, and I can't speak to those that can be created using SPSS, as I have not made any in this program. Two final thoughts: 1) R has a community of diverse users. As you collaborate with more people outside your focal discipline, it is important to use a program that everyone can access, and the combined data can be synthesized into a meaningful product. So, overall, R is the way to go. 2) The only word of caution and concern that I have with R is that there are so many packages, and I don't know they are vetted for accuracy and credibility. Granted, those that are cited in publications are likely highly credible and accurate, so just do your homework when choosing packages.
I prefer to use R, and its popularity is growing (see http://blog.revolutionanalytics.com/2017/01/three-reasons-to-learn-r-today.html, though these posts are a bit biased, I would say).
It's not just because it's free - we have licenses for SPSS or STATA here as well. R, and new developments relating to R's syntax and "grammar" (see http://www.tidyverse.org), is the most efficient tool to do data preparation, data exploration (so called "data wrangling") and data analysis (Python or Julia might be comparable in this issue).
And of course, you have much possibilities to create great looking figures. So, R covers every step from data import to communicating results, and each of these steps can solved in a quite intuitive and very efficient manner.
I grew up on SAS, and I do have access to SAS at work.
I do think the free aspect of R is a large factor. It's very convenient to have on any computer I want without worrying about licenses. And as an instructor, I'm glad I am able to train students on software they'll have access to any time.
R is not as easy to use for beginners, but once you learn to do the things you want to do, it's as easy as anything else. I think it is still a challenge to find clear and complete examples of analyses online, but it's getting better.
I also think that R does have top-notch packages for data analysis. It is _easy_ to do ordinal regression. It is _easy_ to do beta regression. For the most part it is possible to get statistics easily like a p-value for the model, a pseudo r-squared value, a table for effects based on analysis of variance or analysis of deviance.
R is the best one for me. I believe, its popularity is growing much faster compare to other software. Not just because it is FREE. It is very powerful, it is useful in almost every field.
I have recently started to use R and have been learning how to use R via YouTube videos and blogs. It takes a bit of time to get used to the syntax, but is very feasible to learn and to be successful with in a short time. Although point and click programs are indeed faster for simpler analyses, R is more powerful and you can control more aspects of how things are modeled and analyzed. Thus, R is better for more complex questions and 'wonky' datasets. And websites like StackExchange have lots of forums where you can search for code or ask questions to fellow R users. I still use JMP or SPSS for quick tabulations and normally distributed models, but as I am getting more used to R, I'm transitioning to just using R for everything. With respect to creating figures, the visual display options for your data seem just as customizable in R as in SigmaPlot, but SigmaPlot is expensive. The graphic displays in JMP are horrendous, and I can't speak to those that can be created using SPSS, as I have not made any in this program. Two final thoughts: 1) R has a community of diverse users. As you collaborate with more people outside your focal discipline, it is important to use a program that everyone can access, and the combined data can be synthesized into a meaningful product. So, overall, R is the way to go. 2) The only word of caution and concern that I have with R is that there are so many packages, and I don't know they are vetted for accuracy and credibility. Granted, those that are cited in publications are likely highly credible and accurate, so just do your homework when choosing packages.
admittedly, R is free and once you get the hang of the syntax and the various packages available, it will probably suit all your needs. Therein, however, lies the problem. Compared to SPSS, R requieres an awful lot of learning and if it is sufficient enough for you to carry out calculations and read the output without knowing the intricacies of the calculations themselves, SPSS would be the item of choice for me. Personally, I use both SPSS and MPlus for any latent modelling.
For many purposes R is what people want, but it will not suite everyone (that is why languages like Julia exist). The metaphor I like comparing R with packages like SPSS is that R is like driving between two places. You have to know where you are going and how to drive, but it can get you where you want. SPSS (at least the GUI) is more like a bus. Easier to get on, and it will take you to some popular places, but it might not take you exactly where you want. My advice is to find out what your colleagues used and what packages/languages you will be able to find lots of in-person help with where you are when learning any package.
I will add to R fans. R is free and that it is important in many context specially where money is a concern like in every developing country (or worst) as mine. I am not sure that it is difficult only because are more easy options out there any scientist could use it, the problem in this vein could be that there are many ways using it and the outcomes are actually different when you try some of the possibilities that the language gives you.
I also think that R is part of a giant community developing, sharing and helping others that could be slower than a help desk but are very deeper that any help desk. I also has to add that many of the statistical guru's are themselves or by means of their students developing packages about their sustantive area of knowledge.
But in line with Daniel words, it will not happen that only one software will prevail because all humans are different and all of us experiment different needs then any of us will be confronted with decisions about which one is better in some specific appliacation. But R as statistical package seems to me as Excel as spreadsheet, is the best first option to deal with any problem.
I like Minitab because it does what I need to do much faster than R. I can enter data, perform an analysis, and get the output in the same time that people are trying to figure out what subroutine to use in R, much less entering and debugging the code before performing the analysis. I am also concerned about the open source nature of R. Who is running the quality control on these what are essentially third party programs? What standards are they following for accepting open source code? If an error is found in the code, who is responsible? One would hope it would be the coder(s) but that is not a given. All of these things have already been addressed in Minitab. Minitab is much more expensive but as in most things in life and especially science, you get what you pay for.