Discussion: View Thread

Hi,

I use Minitab 16, SAS/JMP 9, Design Expert, NLREG, GPower, and Gnumeric (fixes Excel's wrong algorithms).

I can't afford SAS.

-------------------------------------------
Patrick Spagon
-------------------------------------------

Hi Everyone, What about Stata? I believe the individual license is around $3,000. I haven't used SAS yet, but I think Stata can do a few things that SAS can not. ------------------------------------------- Margaret Linan -------------------------------------------
Original Message: Sent: 08-22-2011 14:57 From: Patrick Spagon Subject: Software Used for Consulting Hi, I use Minitab 16, SAS/JMP 9, Design Expert, NLREG, GPower, and Gnumeric (fixes Excel's wrong algorithms). I can't afford SAS. ------------------------------------------- Patrick Spagon -------------------------------------------

To give some perspective on cost, STATA/SE is $1695, Systat is $1495, Statistix is $495 and R is free.  The cost indicated are for perpetual licenses; that is the cost is one time.  I believe base SAS is $10,000 per year but I believe modules like SAS Graph are extra.  JMP does not sound so bad at $1295 but this also has an annual license fee of about $600.  Moreover, if you want all the bells and whistles you want JMP pro which is $8050 for an initial license and $4000 per year thereafter.  Obviously if you need SAS or JMP you should pay the money, but you can buy a very powerful set of software tools for much less than an annual SAS or even JMP Pro annual license fee.  Also, if you are working with something like NHANES, STATA has a really nice set of tools for dealing with complex survey data and for the analysis of epidemiologic data and is probably, for many, preferable to SAS.

-------------------------------------------
Michael Ginevan
M.E. Ginevan & Associates
-------------------------------------------






Original Message:
Sent: 08-22-2011 15:04
From: Margaret Linan
Subject: Software Used for Consulting

Hi Everyone,

What about Stata? I believe the individual license is around $3,000.
I haven't used SAS yet, but I think Stata can do a few things that SAS can not.

-------------------------------------------
Margaret Linan
-------------------------------------------

You can an individual perpetual JMP license for $1895.  No annual fees.

http://www.jmp.com/software/license.shtml

-------------------------------------------
James Loughlin
J2Logic/Loughlin Consulting
-------------------------------------------






Original Message:
Sent: 08-22-2011 15:54
From: Michael Ginevan
Subject: Software Used for Consulting

To give some perspective on cost, STATA/SE is $1695, Systat is $1495, Statistix is $495 and R is free.  The cost indicated are for perpetual licenses; that is the cost is one time.  I believe base SAS is $10,000 per year but I believe modules like SAS Graph are extra.  JMP does not sound so bad at $1295 but this also has an annual license fee of about $600.  Moreover, if you want all the bells and whistles you want JMP pro which is $8050 for an initial license and $4000 per year thereafter.  Obviously if you need SAS or JMP you should pay the money, but you can buy a very powerful set of software tools for much less than an annual SAS or even JMP Pro annual license fee.  Also, if you are working with something like NHANES, STATA has a really nice set of tools for dealing with complex survey data and for the analysis of epidemiologic data and is probably, for many, preferable to SAS.

-------------------------------------------
Michael Ginevan
M.E. Ginevan & Associates
-------------------------------------------






Original Message:
Sent: 08-22-2011 15:04
From: Margaret Linan
Subject: Software Used for Consulting

Hi Everyone,

What about Stata? I believe the individual license is around $3,000.
I haven't used SAS yet, but I think Stata can do a few things that SAS can not.

-------------------------------------------
Margaret Linan
-------------------------------------------

I have JMP, and possibly I checked the wrong box at the time I ordered.
I know for certain from tech support that I can install my JMP on three separate computers. AFter that I have to buy a new copy.



-------------------------------------------
Chris Barker, Ph.D.
President - San Francisco Bay Area Chapter of the American Statistical Association
www,barkerstats.com
-------------------------------------------






Original Message:
Sent: 08-22-2011 16:15
From: James Loughlin
Subject: Software Used for Consulting

You can an individual perpetual JMP license for $1895.  No annual fees.

http://www.jmp.com/software/license.shtml

-------------------------------------------
James Loughlin
J2Logic/Loughlin Consulting
-------------------------------------------

Though not a biostatistician I use biostatistical methods quite a bit and have used several statistical packages.  SAS was my main tool for many years (though not recently).  I've used Stata for several years now and in terms of value for money I also feel it deserves high marks.  My impression is that it is widely-used in medical research.  Its user group is very sophisticated and energetic and if you have a look at the Stata website it counts among its users some prominent statisticians. 

-------------------------------------------
Kevin Gray
Cannon Gray LLC
-------------------------------------------






Original Message:
Sent: 08-22-2011 15:54
From: Michael Ginevan
Subject: Software Used for Consulting

To give some perspective on cost, STATA/SE is $1695, Systat is $1495, Statistix is $495 and R is free.  The cost indicated are for perpetual licenses; that is the cost is one time.  I believe base SAS is $10,000 per year but I believe modules like SAS Graph are extra.  JMP does not sound so bad at $1295 but this also has an annual license fee of about $600.  Moreover, if you want all the bells and whistles you want JMP pro which is $8050 for an initial license and $4000 per year thereafter.  Obviously if you need SAS or JMP you should pay the money, but you can buy a very powerful set of software tools for much less than an annual SAS or even JMP Pro annual license fee.  Also, if you are working with something like NHANES, STATA has a really nice set of tools for dealing with complex survey data and for the analysis of epidemiologic data and is probably, for many, preferable to SAS.

-------------------------------------------
Michael Ginevan
M.E. Ginevan & Associates
-------------------------------------------






Original Message:
Sent: 08-22-2011 15:04
From: Margaret Linan
Subject: Software Used for Consulting

Hi Everyone,

What about Stata? I believe the individual license is around $3,000.
I haven't used SAS yet, but I think Stata can do a few things that SAS can not.

-------------------------------------------
Margaret Linan
-------------------------------------------

Joe Hilbe is a big time user of STATA.  In the pharmaceutical industry however, SAS is a must.  FDA preferred.
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------






Original Message:
Sent: 08-22-2011 17:21
From: Kevin Gray
Subject: Software Used for Consulting

Though not a biostatistician I use biostatistical methods quite a bit and have used several statistical packages.  SAS was my main tool for many years (though not recently).  I've used Stata for several years now and in terms of value for money I also feel it deserves high marks.  My impression is that it is widely-used in medical research.  Its user group is very sophisticated and energetic and if you have a look at the Stata website it counts among its users some prominent statisticians. 

-------------------------------------------
Kevin Gray
Cannon Gray LLC
-------------------------------------------






Original Message:
Sent: 08-22-2011 15:54
From: Michael Ginevan
Subject: Software Used for Consulting

To give some perspective on cost, STATA/SE is $1695, Systat is $1495, Statistix is $495 and R is free.  The cost indicated are for perpetual licenses; that is the cost is one time.  I believe base SAS is $10,000 per year but I believe modules like SAS Graph are extra.  JMP does not sound so bad at $1295 but this also has an annual license fee of about $600.  Moreover, if you want all the bells and whistles you want JMP pro which is $8050 for an initial license and $4000 per year thereafter.  Obviously if you need SAS or JMP you should pay the money, but you can buy a very powerful set of software tools for much less than an annual SAS or even JMP Pro annual license fee.  Also, if you are working with something like NHANES, STATA has a really nice set of tools for dealing with complex survey data and for the analysis of epidemiologic data and is probably, for many, preferable to SAS.

-------------------------------------------
Michael Ginevan
M.E. Ginevan & Associates
-------------------------------------------






Original Message:
Sent: 08-22-2011 15:04
From: Margaret Linan
Subject: Software Used for Consulting

Hi Everyone,

What about Stata? I believe the individual license is around $3,000.
I haven't used SAS yet, but I think Stata can do a few things that SAS can not.

-------------------------------------------
Margaret Linan
-------------------------------------------

It is sad that a regulatory agency "prefers" one commercial software over others.
I am glad that FDA does not mandate it.

-------------------------------------------
Winson Taam
The Boeing Company
-------------------------------------------






Original Message:
Sent: 08-22-2011 17:28
From: Michael Chernick
Subject: Software Used for Consulting


Joe Hilbe is a big time user of STATA.  In the pharmaceutical industry however, SAS is a must.  FDA preferred.
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------

No it is not sad that the FDA prefers SAS.  They have good reasons to prefer it.  It is very good at handling large data sets. It provides a good audit trail for calculations and it provides just about all the analysis capabilities that one needs for clinical trials. Aside from the price there is no reason to be down about SAS.  The pharmaceutical and medical device companies can afford it and usually have multiple users licenses.

For the small consulting companies it is not affordable.  But as can be seen from the responses on this thread there are many good and less expensive alternatives including Minitab, STATA and JMP.

Also the FDA preference is basically interpreted as a mandate.  The FDA must check the data and the analyses of the companies submitting the NDAs and the PMAs and having a common software package facilitates that.
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:10
From: Winson Taam
Subject: Software Used for Consulting

It is sad that a regulatory agency "prefers" one commercial software over others.
I am glad that FDA does not mandate it.

-------------------------------------------
Winson Taam
The Boeing Company
-------------------------------------------

I don't work with the FDA at all, so I didn't follow the discussion in detail, but this has been discussed in different R fora, and people have said that they use R for FDA data. 

Again, I can't give details, it's not my area. But I do recall that.
-------------------------------------------
Peter Flom
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:23
From: Michael Chernick
Subject: Software Used for Consulting


No it is not sad that the FDA prefers SAS.  They have good reasons to prefer it.  It is very good at handling large data sets. It provides a good audit trail for calculations and it provides just about all the analysis capabilities that one needs for clinical trials. Aside from the price there is no reason to be down about SAS.  The pharmaceutical and medical device companies can afford it and usually have multiple users licenses.

For the small consulting companies it is not affordable.  But as can be seen from the responses on this thread there are many good and less expensive alternatives including Minitab, STATA and JMP.

Also the FDA preference is basically interpreted as a mandate.  The FDA must check the data and the analyses of the companies submitting the NDAs and the PMAs and having a common software package facilitates that.
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------

My understanding is that if you use SAS for an FDA application that to validate SAS you only need to run the SAS validation code.  This is much easier than having to validate all the procedures that you use in SAS.  I'm not sure how one would validate R or the procedures in R. 

SAS spends a lot of effort validating it's code.  So I can see the FDA's concern if you use another package.  It's not to say that R is incorrect it's just how do you validate it and demostrate that it is giving the correct  results for your specific application.

I have always felt that the FDA has gone out of their way in not endorsing SAS or any other statistical software package but they also requre a validation of the software.  So SAS has become the gold standard in the clinical trials area.  And the gold does not just stand for quality $$$$.

-------------------------------------------
Rocco Brunelle
Senior Statistician
Bowsher Brunelle Smith LLC
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:28
From: Peter Flom
Subject: Software Used for Consulting


I don't work with the FDA at all, so I didn't follow the discussion in detail, but this has been discussed in different R fora, and people have said that they use R for FDA data. 

Again, I can't give details, it's not my area. But I do recall that.
-------------------------------------------
Peter Flom
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:23
From: Michael Chernick
Subject: Software Used for Consulting


No it is not sad that the FDA prefers SAS.  They have good reasons to prefer it.  It is very good at handling large data sets. It provides a good audit trail for calculations and it provides just about all the analysis capabilities that one needs for clinical trials. Aside from the price there is no reason to be down about SAS.  The pharmaceutical and medical device companies can afford it and usually have multiple users licenses.

For the small consulting companies it is not affordable.  But as can be seen from the responses on this thread there are many good and less expensive alternatives including Minitab, STATA and JMP.

Also the FDA preference is basically interpreted as a mandate.  The FDA must check the data and the analyses of the companies submitting the NDAs and the PMAs and having a common software package facilitates that.
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------

I have been a SAS user for over 30 years, and
currently provide statistical programming, specializing
in analysis of large discharge level medical data bases,
as well as claim level data. I have been an independent
consultant for the last 11 years, and I consider SAS to
be a minor expense compared to the value it brings to
my organization. I use the BASE product and the STAT
product, single user license. Much of the work I do is
under contract to AHRQ, and SAS is required for the
tasks.

-------------------------------------------
Timothy Kenney
MA Math Stat
MA Educational Psychology
Kenney IS Consulting Inc
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:23
From: Michael Chernick
Subject: Software Used for Consulting


No it is not sad that the FDA prefers SAS.  They have good reasons to prefer it.  It is very good at handling large data sets. It provides a good audit trail for calculations and it provides just about all the analysis capabilities that one needs for clinical trials. Aside from the price there is no reason to be down about SAS.  The pharmaceutical and medical device companies can afford it and usually have multiple users licenses.

For the small consulting companies it is not affordable.  But as can be seen from the responses on this thread there are many good and less expensive alternatives including Minitab, STATA and JMP.

Also the FDA preference is basically interpreted as a mandate.  The FDA must check the data and the analyses of the companies submitting the NDAs and the PMAs and having a common software package facilitates that.
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------

I am also one of the oldies having used SAS for over 30 years.  When I became an independent consultant 5 years ago I also only got the BASE product and the STAT product, single user license.  Within a month I also got the SAS graph, since I am very visual and impatient.  I only use SAS graph for visualization and a quick show of what is going on.  I am also hampered by being a Mac person and having a bootcamp partition to incorporate a PC (with only the SAS program) within the Mac.  One has to restart every time you go from one type to the other.

-------------------------------------------
Susanne Aref
Aref Consulting Group LLC
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:39
From: Timothy Kenney
Subject: Software Used for Consulting

I have been a SAS user for over 30 years, and
currently provide statistical programming, specializing
in analysis of large discharge level medical data bases,
as well as claim level data. I have been an independent
consultant for the last 11 years, and I consider SAS to
be a minor expense compared to the value it brings to
my organization. I use the BASE product and the STAT
product, single user license. Much of the work I do is
under contract to AHRQ, and SAS is required for the
tasks.

-------------------------------------------
Timothy Kenney
MA Math Stat
MA Educational Psychology
Kenney IS Consulting Inc
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:23
From: Michael Chernick
Subject: Software Used for Consulting


No it is not sad that the FDA prefers SAS.  They have good reasons to prefer it.  It is very good at handling large data sets. It provides a good audit trail for calculations and it provides just about all the analysis capabilities that one needs for clinical trials. Aside from the price there is no reason to be down about SAS.  The pharmaceutical and medical device companies can afford it and usually have multiple users licenses.

For the small consulting companies it is not affordable.  But as can be seen from the responses on this thread there are many good and less expensive alternatives including Minitab, STATA and JMP.

Also the FDA preference is basically interpreted as a mandate.  The FDA must check the data and the analyses of the companies submitting the NDAs and the PMAs and having a common software package facilitates that.
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------

I'm not an expert on FDA issues, but if you believe that FDA has a mandate or even a preference for SAS, you should look at a couple of UserR! conference presentations.

Harrell FE. R for Clinical Trial Reporting: Reproducible Research, Quality and Validation. [Accessed October 21, 2009]. Description: The title says it all. This is a PowerPoint presentation from the 2007 Use R! conference. Available at: user2007.org/program/presentations/harrell.pdf

Soukup M. Using R: Perspectives of a FDA Statistical Reviewer. Description: The author provides perspective from the government's side about validation of statistical software in general and validation of R in particular. This is a PowerPoint presentation from the 2007 Use R! conference. [Accessed October 21, 2009]. user2007.org/program/presentations/soukup.pdf

The issue is complicated, and I can see why some pharmaceutical companies might want to mandate the use of SAS for any work that they contract out. But R is a strong competitor and should not be disregarded just because it is "not SAS".

I was attending a talk at a conference where the speaker was looking at various multiple comparison techniques and running Monte Carlo simulations of them. It annoyed me that the only techniques that the speaker considered were techniques available in SAS, even though I knew there were some pretty good techniques that hadn't made it into SAS yet. So I coined the term "SASism" to represent the belief that if it isn't in SAS, it isn't worth talking about.
-------------------------------------------
Stephen Simon
Independent Statistical Consultant
P. Mean Consulting
-------------------------------------------

 I wouldn't say FDA mandates or prefers SAS.  You can use any software you want. I've sent SPLUS and R output to FDA.

SAS (as with other packages) is "21 CFR Part II" compliant". 
And back in the day, the choices were SAS, SPSS, BMDP , eventually SPLUS (in the U.S.) and other less well known statistics packages. I used all of them.

Its entirely my own opinion, I think SAS succeeded in Pharma because of "productivity" tools - they had the best report writer ("data _null_"), and a batch script submission utility. SAS had a "data step" which was sort-of FORTRAN like, handy if you first learned to program in FORTRAN. They had a lot of other "features", say, a dozen or more types of data you could read in (binary, hex, etc.)

Before a report writer,  there was a time when you wrote/typed values into a table, by hand, sent it to the  Word processing group - and when it returned checked that the numbers were correctly typed. Made corrections as needed, including correcting your corrections,  sent it back to word processing and repeated the loop. Eventually you got your table. And if you changed the analysis you updated the table, sent it back to word processing  etc. With the 'data _null_" you could generate a "publication ready" table.

That noted, SPSS is thriving. IBM paid about $1.2 Billion to buy SPSS.
 http://www-03.ibm.com/press/us/en/pressrelease/27936.wss
I was recentlly in Fiji, and at the medical center in Suva Fiji they used SPSS. They commented that SAS was too expensive.









-------------------------------------------
Chris Barker, Ph.D.
President - San Francisco Bay Area Chapter of the American Statistical Association
www,barkerstats.com
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:10
From: Winson Taam
Subject: Software Used for Consulting

It is sad that a regulatory agency "prefers" one commercial software over others.
I am glad that FDA does not mandate it.

-------------------------------------------
Winson Taam
The Boeing Company
-------------------------------------------

I still maintain that FDA prefers SAS.  I have had no experience submitting with other packages but I don't doubt that others have.  I believe all the FDA statisticians have SAS and prefer submissions in SAS.  SAS dominates the pharmaceutical industry.  There is no doubt about it.  I think the audit trail it provides is a major reason.  I am not saying that there aren't other good software packages.  BMD was the first and best back in its day before SAS took over.

-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:43
From: Christopher Barker
Subject: Software Used for Consulting



 I wouldn't say FDA mandates or prefers SAS.  You can use any software you want. I've sent SPLUS and R output to FDA.

SAS (as with other packages) is "21 CFR Part II" compliant". 
And back in the day, the choices were SAS, SPSS, BMDP , eventually SPLUS (in the U.S.) and other less well known statistics packages. I used all of them.

Its entirely my own opinion, I think SAS succeeded in Pharma because of "productivity" tools - they had the best report writer ("data _null_"), and a batch script submission utility. SAS had a "data step" which was sort-of FORTRAN like, handy if you first learned to program in FORTRAN. They had a lot of other "features", say, a dozen or more types of data you could read in (binary, hex, etc.)

Before a report writer,  there was a time when you wrote/typed values into a table, by hand, sent it to the  Word processing group - and when it returned checked that the numbers were correctly typed. Made corrections as needed, including correcting your corrections,  sent it back to word processing and repeated the loop. Eventually you got your table. And if you changed the analysis you updated the table, sent it back to word processing  etc. With the 'data _null_" you could generate a "publication ready" table.

That noted, SPSS is thriving. IBM paid about $1.2 Billion to buy SPSS.
 http://www-03.ibm.com/press/us/en/pressrelease/27936.wss
I was recentlly in Fiji, and at the medical center in Suva Fiji they used SPSS. They commented that SAS was too expensive.









-------------------------------------------
Chris Barker, Ph.D.
President - San Francisco Bay Area Chapter of the American Statistical Association
www,barkerstats.com
-------------------------------------------






Original Message:
Sent: 08-22-2011 18:10
From: Winson Taam
Subject: Software Used for Consulting

It is sad that a regulatory agency "prefers" one commercial software over others.
I am glad that FDA does not mandate it.

-------------------------------------------
Winson Taam
The Boeing Company
-------------------------------------------

Some thoughts.

Many of the responses have been from the drug approval end to Biostat.

-- Cost of purchase is not the only component of total cost of ownership.


It has been my experience that the staff time for doing an identical task in SAS is about 15% more than the staff time to do that task in SPSS.  A lot depends on the application area.  The staff time advantage of SPSS is largely due to the easy of doing all of the data preparation and cleaning tasks that need to be done before any actual analysis can be done.  This was in the audit and evaluation world where the tasks were not production oriented. In this environment most of the effort is is cleaning and prep which can often take

In the early days of SAS it had an advantage in setting up runs on card-oriented systems that used JCL, since it provided canned sets of JCL.  However, on systems that charged for cpu and memory use such as NIH 3090, I had a run done in SAS cost a little better than twice what an run to do an identical task did in SPSS.

-- A lot depends on the goal of the consultation.

In situations where the goal is building capacity, the time to learn SPSS is a lot shorter than the time to learn SPSS.  I would attribute this to human factors such as consistency and English-like syntax, especially in the staff intensive parts of data preparation.

If one goal is to provide user files in a few different package formats, if you create the first system file in SPSS with all of the metadata and documents internal to it, it is then simple to simply savee it in one of the other common paclage formats that do not retain as much metadata.

-- A lot depends on the mix of work.  It may be useful use one package for all of the time-consuming parts and  to work with someone who has some particular piece of software for  special purposes.

-------------------------------------------
Arthur Kendall
Social Research Consultants
-------------------------------------------






Original Message:
Sent: 08-22-2011 15:04
From: Margaret Linan
Subject: Software Used for Consulting

Hi Everyone,

What about Stata? I believe the individual license is around $3,000.
I haven't used SAS yet, but I think Stata can do a few things that SAS can not.

-------------------------------------------
Margaret Linan
-------------------------------------------