[Social-sciences-list] Introduction and stats software inquiry
a.j.godfrey at massey.ac.nz
Mon Oct 29 19:59:44 UTC 2012
Hi Katie et al.,
I use R almost every day at work and often on weekends too if I'm honest.
I've been fighting with statistical software for over 15 years, and
only two software options have left me feeling independent and
useful. They amount to the same thing though as SPlus is built on the
S language that is used in R.
I have in the last two weeks used SAS on a networked machine using
the NVDA screen reader in order to simulate the experiences of a
blind student wanting to work in a lab setting. It worked just fine.
I have used Stata briefly last year and found it could be used but I
really don't like having to export output all the time to get
something I can read. (A text file was recommended in that instance)
SAS output can be read in the same way we read webpages so I found it
quite attractive in an accessibility sense, but I can't get past the
professional dislike I have that it gives me a page of output
(formatted nicely) when all I wanted was a few numbers to tell me
what I need to do next.
I agree that working with the same software as your
classmates/colleagues is a good idea, but sometimes the fight isn't
worth it. I've drawn the line to say that SPSS is not worth the fight
even if it can be made accessible. I'd rank it above Minitab,
Genstat, Statistica and JMP though.
In terms of learning to use software, I have found the resources for
SAS, R, and Stata to be easy to find and read. SAS resource manuals
are arranged as a textbook so make for good reading. R documents are
everywhere and many are awful. Stata documents are often just too
high level for what is required at the outset.
Anyone wanting to get into R could look at the resources I have made
for my own students at:
The document for beginners is called LURN (Let's Use R Now) as I have
found my students unable to find the right R command for the job they
want done. The chapters of LURN are arranged in terms of the tasks
people might want to do, all R input and output is given and I can
guarantee the syntax has no errors. Best of all, the material is all
readable by blindies as I've developed the document in html pages as
well as the pdf favoured by many of my students.
I think anyone who sees themselves doing more than a dozen hours of
statistical analyses each year should consider using R. If you're at
the level of doing a job that needs doing once (ever) then use other
software like SPSS. The real benefits of R and SAS are that we keep
our programs so that we can alter them to suit the next analysis that
is similar. In fact, I have a huge number of programs that are re-run
as data is updated. The re-run takes me less than 10 seconds of my
time to find the right program, and then I walk away while the PC
does all the work and I can drink coffee.
At 07:48 a.m. 30/10/2012, you wrote:
>Welcome to the list! I'm a 4th-year Social Psychology Ph.D. student at
>Yale University, so the accessibility of SPSS is an issue I have been
>dealing with for years. I'm currently running SPSS 19 on a PC (I am
>planning to install SPSS 20 next month), and while the interface is
>not ideal for screen reader access, I find it adequate and use it to
>complete all of my statistical analyses. Like you, I export my output
>to Word or Excel for reading; to minimize the annoyance caused by the
>sluggish JAWS response, I usually do as much of the data clean-up and
>preparation (e.g. arranging the variables in the correct order,
>deleting subjects with extraneous responses) as possible in Excel
>before importing the files into SPSS. I also use a combination of the
>menu commands and syntax, which can speed things up a bit.
>While I have little personal experience with other statistical
>packages, I do know from previous discussions on this list that some
>blind researchers opt to use SAS and R, which appear to require less
>tinkering in the accessibility department than SPSS. I chose to stick
>with SPSS for the same reasons you mentioned-- I was taught how to use
>it in my stat courses and can easily get help from my colleagues if I
>run into any problems. That said, one of my sighted labmates uses R
>for all her analyses and gets her questions answered through web
>searches, so it's perfectly feasible to use a different stat package
>from the mainstream researchers in your field. In my opinion, it's
>mostly a matter of choosing between dealing with the annoyances of
>SPSS accessibility and taking the extra initiative to master a new
>software with relatively little formal support.
>Hope this helps!
>On 10/29/12, Robert Hooper <hooper.90 at buckeyemail.osu.edu> wrote:
> > Hello everyone:
> > My name is Robert Hooper, and I am a third year undergraduate
> student at The
> > Ohio State University. I am majoring in psychology and neuroscience, with a
> > minor in Spanish and biology. I don't remember from where or whom I
> > originally discovered the existence of this list, but here I am.
> > So, I might as well start off with a question, as I have not yet had a
> > community of people with visual disabilities to answer it.
> > My question is: statistics. I enjoy working with them, which is
> > have always loathed working with other types of mathematics. Anyway, I find
> > that with some tinkering, SPSS is halfway accessible. I have installed SPSS
> > 20, along with the Java accessibility bridge provided by IBM. I find the
> > interface a bit sluggish, and the output window completely inaccessible. I
> > can, however, access the output if I export it to a Word
> document. For those
> > of you in need of a statistical package, what do you use? If SPSS, what do
> > you do to make it work? If not, what do you use and how does it compare to
> > SPSS? I have heard of and tried R, but I have no experience with command
> > lines, and using it seemed very tedious-I gave it maybe half an hour before
> > installing SPSS and trying to make that accessible. Microsoft
> Excel can only
> > be used for so much-when you get into regression and multiple regression,
> > etc. it simply falls short (which is to be expected-Excel was
> never marketed
> > as a comprehensive statistical package).
> > If I must use another program, I face the problem of being
> alone-that is, it
> > seems as though most people use SPSS. SPSS is often used in my
> data analysis
> > class, both in the textbook and the classroom. All of those with whom I
> > spoke about research say that they use SPSS. So, if I use another
> package, I
> > fear that I wouldn't be able to get as much support from my
> peers, who would
> > be unfamiliar with whatever software I would be using. Because SPSS is so
> > widely used, there are vast support networks for it. If I had a
> problem with
> > a statistical test or running a statistical test, a professor, peer,
> > colleague, etc. could look over my shoulder and give advice-not so if I
> > forego SPSS.
> > There you have it-I'm sure that this has come up before, and I apologize if
> > I am stirring up an already well-stirred pot. I hope that many
> benefits come
> > from being a member of this community. I'm sure I will be in touch again-I,
> > like many others with a passion for a particular field or subject, will
> > happily discuss, learn, and debate until my fingers fall off. Thanks in
> > advance for any help.
> > One final note-I have a 2012 MacBook Pro with Mountain Lion installed. I
> > also installed Windows 7 via bootcamp on a separate partition. I have SPSS
> > for both Windows and Mac OS, however the installation of SPSS on the Mac
> > side is inaccessible. Therefore, I haven't been able to determine SPSS's
> > level of accessibility on Mac OS.
> > Sincerely,
> > Robert Hooper
> > Hooper.90 at buckeyemail.osu.edu
> > The Ohio State University, Department of Psychology; Department of
> > Neuroscience
> > 572 Stinchcomb Drive #3
> > Columbus, Ohio 43202
> > (740) 856-8195
>Social-sciences-list mailing list
>Social-sciences-list at nfbnet.org
>To unsubscribe, change your list options or get your account info
Dr A. Jonathan R. Godfrey
Lecturer in Statistics
Institute of Fundamental Sciences
Office: Science Tower B Room 3.15
Phone: +64-6-356 9099 ext 7705
Home Address: 52 Linton St, Palm. Nth.
Home Phone: +64-6-353 2224 (Just think FLEABAG)
More information about the Social-Sciences-List