Any social scientists out there who use R?

I have a question about whether it is really worth using R to do your basic descriptive stats and tests, when at a glance it actually seems much easier/less timely to do it in SPSS... If you are an R user - do you use R for everything in your analysis?

If you know one software then learning another isn't always a benefit although R is free and has a good community for help out there. But to answer I use R for everything to keep the format of data the same.

Thanks kikodog. I'm learning it because I'm doing some multilevel modeling, which apparently R is better for... (I don't know why it is better, but apparently it is)...

I know SPSS, so I am wondering whether to screen my data etc as I normally would in SPSS before conducting the main analyses in R, or whether to just learn how to and do the whole shabang in R.

Hi Tudor,

I am an R user and I absolutely recommend it. I use it for everything, descriptive stats, multilevel modelling, meta analysis, spatial statistics, visualisation, data management, you name it.

It is open source, there is plenty of help and you can write your own packages to do your stuff. it is a steep learning curve in the beginning but it is worth to invest some time to get a basic expertise. I would expect that social scientists need to deal with lots of complex data so this is a skill you should have.

Thanks Dr Jekyll. I agree. I think! I'm just getting familiar with reading csv. files and doing some basic stats on my existing datasets. It sure does feel like a steep learning curve. How did you learn it? Could you recommend any good books?
Thanks
Tudor

No books for R, just "stack overflow" and of course all the CRAN documentation.
For multilevel modelling I recommend online tutorials for MLwiN from university of Bristol. For general statistics I loved open university courses on iTunes from Berkeley.

Awesome thanks!!!

I've come across "The R book" by Crawley, which seems an excellent guide (for those of us who tire of screen reading!)

You should try Field, Miles and Field 'Discovering statistics using R' :)

Thanks pd1598! I have Field's book from the library, but strangely I am not finding it very helpful. It is strange, because I found his SPSS book on statistics SUPER helpful - I don't know how I would have got through my masters without it! But with the R one I'm struggling. For example, in the part where it explains how you import you data by reading an Excel (or csv) file into R, it then doesn't go on to tell you anywhere how to work with the variables that were on that sheet (as far as I could see). It took me HOURS (literally!) to figure it out and seemed odd that such an important thing wasn't in the book (or maybe it was but I just couldn't find it).

I can definitely second MLwiN for multilevel modelling - and the online information course is excellent. Sorry, no help with R, but it does seem like a very steep learning curve to do ANYTHING!

Thanks allyballybee. I am thinking I'd better stick to R, as that is what my supervisor is comfortable with (and he may way to look through my outputs/be able to offer support).

I agree with Dr Jekyll, I used R for just about everything in my biomedical PhD. It was not generally used in our department (they tend to use SPSS), but we had an introductory lecture from a postdoc in the department and I was hooked. I started with a very thin volume (just over 100 pages) called "Getting Started with R: An Introduction for Biologists" by Beckerman & Petchey, which I read from cover to cover, but thereafter my main resource was CRAN and Stack Exchange. I bought "The R Book" by Cawley and "Discovering Statistics Using R" by Field, Miles & Field, but hardly referred to them. I had to consolidate and manipulate data from over 100 spreadsheets, so R was a Godsend. The original data were never touched, thereby avoiding the risks of copy and paste type operations, they were simply loaded, manipulated, then exported to new CSV files. It's certainly a steep learning curve in the early stages, but well worth investing the time.

Cheers ologist - I'll see how I go. Feeling proud that I've done my descriptive stats and reliability checks in R so far. Now for the real stuff...