Date: Fri, 27 Nov 2009 23:36:09 -0800
Reply-To: xlr82sas <xlr82sas@AOL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: xlr82sas <xlr82sas@AOL.COM>
Organization: http://groups.google.com
Subject: Re: SAS Programming Opportunites and the Normal Distribution
Content-Type: text/plain; charset=ISO-8859-1
On Nov 27, 10:09 am, "Kenneth M. Lin" <kenneth_m_...@sbcglobal.net>
wrote:
> Your argument clearly indicate that you do not understand statistics. To
> compare two normal distributions, you need to obtain their means and
> variances. Yes, India and China have a lot more people than U.S. However,
> what are of percentage of population that goes (or can afford to) to
> college? How many SAS developers are there in India? I have worked with
> Indian developers and in some cases we couldn't communicate at all because
> their English was so poor. Sure, their hourly rates are low but I spent so
> much time explaining simple tasks that at the end we didn't save any money
> and the project didn't finish on time.
>
> "xlr82sas" <xlr82...@aol.com> wrote in message
>
> news:fe3cafcb-d625-40d1-816d-496943c69859@2g2000prl.googlegroups.com...
>
>
>
> > Hi SAS_lers,
>
> > Recently I have had to write letters of references for several
> > unemployed US SAS programmers.
> > These are programmers I have worked with in the past.
>
> > One was hired for a 2 month gig with a bunch of other programmers with
> > the object to
> > keep the best programmers and fire the 'least productive'.
> > There is nothing wrong with this persons SAS programming skills.
>
> > He did not make the cut.
>
> > Recently I wrote another reference for him for a computer science
> > master degree program.
>
> > When I look at his competition, I think about the normal distribution.
> > India has over
> > 1 billion people and China has over 1.2 billion, I think there are
> > over 150,000 Indians and
> > Chinese studing Engineering and computer science in US universities
> > and many more in their
> > respective countries. The normal distribution dictates that their are
> > a lot more gifted Indian
> > and Chinese SAS programmers than US SAS programmers. I have been a
> > member of a falling minority
> > of US programmers. I work with Chinese and Indian contractors day in
> > and day out and the
> > top Chinese and Indian SAS programmers are
> > considerably better than the best US programmers. Again you cannot
> > fight the normal distribution
> > just like you cannot stop the sunrise. This is not to degrade US SAS
> > programmers, it is a fact
> > of life.
>
> > and Chinese SAS programmers than US SAS programmers. I have been a
> > member of a falling minority- Hide quoted text -
>
> - Show quoted text -
Hi SAS-Lers,
I need the assumption that there are more Chinese and Indian
programmers than US progarmmers in my industry. I know this is fact
because I have been in this industy fo 20 years at seven different
companies. No need to read on if you do not believe in this
assumption. I have done many interviews so I know the ethnicity of
SAS programmers.
Lets examine what happend to my friend and why he did not make the
cut. It was a contract job. I talked to the recruiter and he told me
before the assignment that they were providing a farily large number
of contractors and that only some of them would be kept on after the
first two months.
Suppose the recruiter provided a random sample of 100 contractors, 80
of were Chinese or Indian and 20 were US programmers. Lets assume that
only 20 would be kept on after two months. To be conservative suppose
the two populations have the same mean skill level and standard
deviations, disregard my tail theory for now. It does not matter what
the mean and standard deviation is, there will be only a 20% chance my
friend will be in the top 20 or 4 US programmers. 16 of the top 20
programmers will non US workers.
Now lets suppose thta the Chinese and Indian programmers are less
skilled with a mean skill level of 1 and a standard deviation of 1 and
that US programmers have a mean of 2 and standard deviation of 1.
Twice as skilled. I calculate the US programmers have about 45%
chance to be in the top 20% or 9 US programmers. I did a little
simulation and here are the top 10 programmers in a simulation of 1000
samples of 100 programmers with equal skill level.
Here are the Satistics for the top 11.
999990 NON-US 4.26340
999991 US 4.27822
999992 NON-US 4.28149
999993 NON-US 4.36504
999994 NON-US 4.36913
999995 NON-US 4.52429
999996 NON-US 4.61791
999997 NON-US 4.62985
999998 NON-US 4.67240
999999 NON-US 4.68363
1000000 NON-US 4.69827
