Date: Thu, 16 Aug 2012 17:17:11 -0700
From: Norm Matloff <matloff@cs.ucdavis.edu>
To: Norm Matloff <matloff@laura.cs.ucdavis.edu>
Subject: foreign student quality, university monetary incentives, etc.

To: H-1B/L-1/offshoring e-newsletter

In discussing the CSU East Bay admissions scandal this week, I've
emphasized the role of the foreign students.  In my first posting, for
instance, I noted that fully 90% of the master's students in computer
science at CSUEB are international students.  

In my posting today, I noted the San Jose State University President
Qayoumi's claim of brilliant, innovative foreign students transforming
Silicon Valley (I'm paraphrasing)--and then I pointed out that the
president's claim was contradicted by CSU professor Nieto's statement
that the grad students in her program were weak.

(My posting from this morning is archived at
http://heather.cs.ucdavis.edu/Archive/CSUEastBay1.txt)

Of course, Qayoumi's assertion came straight from the industry
lobbyists' "educational packets" ("industry" including various other
vested interests).  University presidents are happy to use these;
"ignorance is bliss," especially for university presidents, and
especially so for presidents of Silicon Valley institutions.  As I've
mentioned before, one can learn a lot through a stroll of the
engineering section of the Stanford campus.  Start with the CS/EE
building, named William Gates II Hall (guess who), and then cross the
street to Hewlett Hall and Packard Hall (HP, if you're having an off
day).  Then move down the block to the new Jen-Hsun Huang Center, named
after the cofounder of Nvidia, and the list goes on.  I have nothing
against any of these firms--Nvidia is a favorite of mine--but it's clear
that university presidents know which side of their bread is buttered.

So do CS department chairs.  Ed Lazowska, former chair of the University
of Washington (Washington as in Redmond) CS Dept., has been quite
outspoken in favor of the H-1B program, and coincidentally or not, he
does hold the Bill & Melinda Gates Chair in Computer Science &
Engineering there at UW, and sits on the boards of various tech firms.
His department has been treated well by industry too, and did I mention
Mary Gates Drive on the eastern edge of campus?  (Good for Bill for
honoring his late mother, but still...)

Some years ago when I appeared on the PBS Newshour, my "debating
opponent" was Randy Katz, then chair of CS at UC Berkeley.  In response
to my statement that H-1B displaces older (age 35+) American workers,
Katz said older people just can't learn new things, and cited himself:
"I've always wanted to learn Swedish, but I know it would be impossible
for me at my age [about 50]."  He didn't seem to see any contradiction
in the fact that he himself teaches new technologies all the time, and
the possibility does suggest itself that his "can't learn Swedish"
remark was somehow related to the largesse his department receives from
industry.

Not only might these academic officials be motivated by the goodies (you
know, say buildings) they get from industry, but they benefit hugely
from the U.S.' expansive foreign student program in other direct ways.
Recall my frequent citing of the 1989 NSF PRA position paper that
advocated bringing in a lot of foreign students to keep PhD salaries
low.  This refers not only to salaries in industry but also graduate
student stipends.  In short, universities and the NSF get more research
bang for their buck this way.

Which makes Professor Nieto downright courageous in remarking that her
department's grad students are weak.  Usually I hear such things, both
at my campus and others, only in whispers.

So, how accurate are those whispers?  As I mentioned this morning, I've
been writing about this quality issue in general for a while, and more
recently have concentrated on this very subpopulation--the former
foreign students now working in CS in the U.S.  In my recent research,
I've found among other things that

   * the former foreign students who have green cards or have become
     citizens (thus, in both cases, not exploitable) are earning 
     significantly less than their U.S. native peers,

   * the former foreign students obtain significantly fewer patents per
     capita than comparable U.S. natives, and

   * the former foreign students are significantly less likely to be
     working in R&D than are the Americans.
   
These last two points fly in the face of the industry lobbyists (and
university officials') claim that the foreign students are especially
innovative, and the first point doesn't point to brilliance either.

I also did an analysis for EE, finding that the foreign students were
the same as--i.e. not better than--the Americans on the first two points
above, and again that the foreign EEs are less likely than their
American peers to be working in R&D.

Now, here is news, some analysis I did just today:

There is an old NSF study, cited in David North's book, that found that
the foreign PhD students in engineering were concentrated in the weaker
U.S. schools; the lower the ranking of the university, the higher the
proportion of international students.

I'd been wanting to do some analysis like this for current data, for CS,
but didn't have the data on percentages of foreign students at the
various U.S. CS departments.  But last night I suddenly remembered that
the 2010 National Research Council ranking study has such data, so I was
in business.

The NRC study rated the CS programs too, but its methods were quite
controversial, being based on imputational rather than reputational
data.  So I chose to use the US News and World Report ratings of PhD CS
programs, which are based solely on surveys of professors in the field.
(The NRC ratings give similar results, though.) The USNWR ratings are on
a scale of 1-5, but they only publish those down to 2.0.  Here are my
preliminary findings:

1.  I asked the question, How good (according to the USNWR) are the
schools at which the foreign students earn their PhDs, compared to the
Americans?

The result was a mean rating of 3.71 for the American students, versus
3.44 for the foreign students.  Not a gigantic difference, but the
American mean is substantially higher, and the result is certainly
counter to the "genius foreign students" image purveyed by the industry
lobbyists and their allies.

2.  I then asked, How do the strongest and weakest CS programs compare
in terms of percentages of foreign students?

The result was that among the top 10 CS programs, the average percentage
of foreign students was 45.7%, while among the botton 10 it was 66.5%.
This is consistent with the old NSF finding that the weaker the school,
the higher the percentage of international students--and again counter
to the image put forth by industry lobbyists, the American Immigration
Law Association, Pres. Qayoum and especially Silicon Valley
congressperson Zoe Lofgren.  The congressperson, one of the most
strident promoters of H-1B in Congress, once promoted her H-1B bill with
the remark, "You can't have too many geniuses."

As I said before, the fact that a student is from a weaker school does
not necessarily mean she herself is weak.  And vice versa:  The more
highly-ranked schools rake in tons of research grant funding, and thus
"need the bodies" to spend it on; this means that their admissions
standards are not quite as high as their rankings might suggest.

Nevertheless, no one would disagree with the point that the average
quality of students at the lower-ranked schools IS lower.  Thus the above
statistics demonstrate clearly that the average foreign student is of
mediocre quality, not what the vested interests would have you believe.

Note too the fact that even at the top CS schools, about 45% of the PhD
students are foreign.  This is exactly what the 1989 NSF paper forecast
(and did not object to)--discouraged by stagnant wages, the domestic
students would avoid doctoral study.

It would be great to have corresponding data at the master's degree
level, where the difference between the Americans and foreign students
would be even more pronounced, I believe.

Norm