Simulations in SAS and R

spunky

Doesn't actually exist
#4
of course they are, because for many, many years SAS or S+ were the only reasonable alternatives to statistical analysis.

but the problem is... what is SAS gonna do if generation after generation of data analysts stop being trained in SAS. they rely on undergrads buying and using their products to remain relevant and if professors are not using them anymore then they'll start losing a big share of their market.

that's why they get all cranky whenever R comes around because they know that, although they dominate the industry now, they could lose a big chunk of their market unless people keep on being trained on it.
 

noetsi

Fortran must die
#5
They are going to rely on the fact that the companies that pay them money, and this means the executives, will insist their employees report results from the software the company buys :p

I doubt SAS loses much sleep over R. They make fortunes and much of their bussiness these days is in data mining and operations research not stats anyway.
 

spunky

Doesn't actually exist
#6
it's just like SPSS. they don't really lose too much sleep over R because business analytics is where their clients are. they couldn't care less about factor analysis or SEM. like i just had the displeasure of dealing with AMOS the other day. oh god... the horror! but it makes sense... they don't care about AMOS because not enough people buy it to make it profitable.

they do care about their text analytics add-in modules though because it lets people mine facebook statues and tweets and stuff.

i guess the only people who are losing sleep are academics who are proficient in SAS, because the paper you attached complains towards the end about academics abandoning SAS
 

hlsmith

Not a robit
#7
I guarantee that they lose sleep, why do you think they came up with the free university package. I use SAS and I get your points noetsi, since I primarily worked in the private sector - I am sure they will lose lots of money.
 

noetsi

Fortran must die
#8
I dont know I run into academics using SAS everday as I study new methods. It is impossible to know what percent of academics use software. I imagine that in statistics R users dominate. In the social sciences I suspect this is not the case simply because schools (as my program) continue to stress one of the well known softwares.

SAS and SPSS appear at times to follow very different routes in statistics. SAS is constantly improving its statistical capacity (which can be anoying since something used in the last release has been replaced by something new making articles more than 5 years old pretty useless for SAS).SPSS has adapted R in its R studio. It might be remembered that the person that owns SAS (it is privately held not publically traded) created it to run statistics. I don't know if he was a statistician, but SAS did mainly statistics at first.

I found this comment interesting in that light.

"In SAS 9.1 a function called RAND() was added to SAS. This random number generator is based on the complicated Mersenne-Twister algorithm, which is similar to the default random number generator that many use in R (WikiBooks 2011). Both software packages have many random number generation methods available, however using the RAND() function in SAS would be equivalent to using the default in R."
 

noetsi

Fortran must die
#9
I guarantee that they lose sleep, why do you think they came up with the free university package. I use SAS and I get your points noetsi, since I primarily worked in the private sector - I am sure they will lose lots of money.
My guess is that little of SAS's profit comes from statistics these days. And I doubt that companies, especially large ones most likely to do statistics, will quickly shift to open source code from a known product. But we will see.
 

spunky

Doesn't actually exist
#10
well, but SAS must feel some sort of pressure from R because, as hlsmith said, not only did they make available their free university package after R started gaining a mass following, they also have a lot of plug-ins and interfaces to have SAS talk to R.

if they were not concerned about R, why are they taking so many preventive steps?
 

noetsi

Fortran must die
#11
It might be preventive. Or it might be that they remain comitted to improving their statistics (since they are owned by someone who spent a lot of his life around that) and thus when they see something R does well they copy it. The owner of SAS is quite rich and has made it clear in the past that more money really is not his driving interest anymore. If it ever was.
 

noetsi

Fortran must die
#15
True. Of course someday it will turn out AIM spread its nanoviruses through the open source coding in R so its all cool. The trick to world domination is to own all the sources and pretend they are fighting each other. The true genius behind my order.
 

hlsmith

Not a robit
#16
Spunky,

SAS logo: Power is Knowledge

I agree with a virus issue in the future - probably not nano, just good ol' code. But biologically nano all the way.
 

spunky

Doesn't actually exist
#17
well, the good thing for me is that here over social-science land, SPSS is still the king. and the one thing i've learnt over the years about it is that even a properly-trained chimpanzee is capable of conducting analyses on SPSS.

R kind of gives me this warm, fuzzy feeling of seclusion. like a scholastic monk working away in on some hidden task in a magical language
 

noetsi

Fortran must die
#18
I am not sure, since I lack the expertise, if it is easier to spread viruses/steal information etc through R or SAS. But if a virus is introduced then R has two problems SAS does not have. First, SAS as a company has control over its coders in a way that R (a community of volunteers that is designed to be open) can never have. And executives, in both the private and public sector are inherently going to trust a corporation that can be sued more than a community that can not be effectively. Actually they would probably always trust a corporation over a group of academics (which is what R really is).

I had to appeal over my IT group to a bureau chief (one step from the top of the agency) to get R added specifically because of trust issues. They know SAS, they did not know R.

When the first virus comes out I guess we will know.

R kind of gives me this warm, fuzzy feeling of seclusion. like a scholastic monk working away in on some hidden task in a magical language
That pretty much describes academics in general. And why, one reason anyhow, why R is so popular. It sets academics above the low life who can run SPSS and SAS :p Us chimps....
 

hlsmith

Not a robit
#19
We are all chimps on different branchs of elevation. Something will come along and be harder than R, then there will be seperation again. Hey, I code only in binary, R chimp boy.