Just a quick post here. I wanted to note some of the observations that I have made about international students, which indirectly relates to the H-1B visa issues.
First, I would note that I am not necessarily opposed to outsourcing. I don't have a strong opinion either way right now, but I generally do believe that there is a lot of "busy work" in engineering and computer science which probably is worth outsourcing.
I am not so tolerant of the H-1B visa farce however, partly because of the political excuse that the tech industry "can't find enough workers" is so demonstrability false. However, I wanted to focus on the international students which I have meet, worked with, and even graded. There seems to be quite a lot of confusion and downright misinformation about this issue.
Some have even gone as far as suggesting that anyone who graduates with an engineering or computer science degree at a US university should automatically receive a green card. Here is my experience, which may vary from other universities and departments. But I think its a fairly good first-order approximation.
In my observation the overwhelming majority of international students in the technical fields come from one of two places, in about equal numbers, either India, or East Asia.
Those from India are almost exclusively from southern India, though there are also some from Pakistan and Bangladesh. The Indian students generally speak English natively, so they generally are good in communicating. Generally they fall into two groups, those which are very urbanized/Westernized, and those which are rural, even somewhat primitive. The East Asian students are generally much more urbanized/Western, and somewhat more individualistic. English often seems to be somewhat more of a problem, though some do quite well.
Now what I primarily wanted to focus on was academic performance. There are a lot of people which think that international students are somehow exceptional genius-level students, especially when one considers the fact that we are selecting the best and brightest out of approximately 2.5 BILLION people just in India and East Asia alone.
My experience, however, is that our international student population is much more mediocre than most realize. I am not suggesting that the students are unintelligent; in most fields of study they would likely be considered quite bright. Also, there are a very few, primarily from Japan and S. Korea it seems, which are genuinely exceptional. However, again and again my experience has been that international students are very much around average.
In practice, there seems to be some which do well, come to class on time prepared, and generally perform at or above average. What sticks out however is that there is a significant number belonging to a second category which do not seem to be hacking it as well. If this is a question of ability or simply the culture shock of suddenly being out on one's own without supervision, I do not know. However this group is quite distinctive. Typically they come to class late (especially if it is an early class), disorganized, try to turn in projects late, copy one another's HW or projects, ask/beg other students questions about the HW/projects or lectures which they do not understand, make profuse excuses, etc. Of this second group, I would have to say, in my experience, that it is much more Indian students than East Asia. Again, in my experience, those students from the Middle East, though smaller in number, tend to fall into this category overwhelmingly.
Now two points: first, I do not have access to any statistics which show international student performance versus native US students, either locally or nation wide. Secondly, it is a fair question of how much other factors, like culture shock, being personally organized, time management skills, etc, are not the more dominant factors in any performance differences between international and native students.
Also, it is worth pointing out that grades are only one, and possibly a fairly poor, measure of long term success in the technological fields. There are a lot of skills, such as interpersonal, language, communication, time management, nonlinear/creative thinking, etc, which are very hard to quantify.
One observation however, which I believe is exceptionally strong and distinct, is that US students seem to be much more creative, out-of-the-box type thinkers than the international students. The international students seem to be much more book-learned, and more comfortable with hierarchical (social/management) structures and conformity.
Ok, now for the fun part. Lets assume that intelligence/talent is approximated by a normal distribution, as shown in Figure 1. In other words, most people are grouped tightly around the average. A few people fall to either extreme tail, those on the left are severely lacking intelligence/talent, and those on the left are exceptional.

Figure 1: Normal distribution.
However, even though the chances of being on the far right part of the tail (like very high IQ's) is quite small, that does not mean that the difference with someone nearer the average is especially large. For example, the difference in actual ability between someone with an IQ in the top 0.1% (1:1000), and the top 1% (1:100) may only be a few percent despite the much lower probability, or rareness of the first individual.
Figure 2 shows how that both low and high intelligence/talent seem to "saturate" at some point, assuming one leaves out extreme probabilities.

Figure 2: Actual intelligence/skill
Now what does this all mean? Well I believe that what we can draw from this is that, if we as a nation want to bring the brightest, most exceptional students to the US to study (and often eventually to work in the ever growing H-1B program), we need to appreciate that the truly brightest, most exceptional international students are quite few in number. If we select those which are in the far right tail, I don't think that it is necessarily a bad thing, nor will it present any strain on our educational system or force out other qualified US students.
Right now however, despite the hysterical cries of academia and industry that we are loosing too many prospective students to other countries, or that they are leaving the US after graduation, we are no where near selecting only the far right tail of the bell curve. Wither you believe that the size of the international student population is motivated by the need for cheap RA's, keeping small departments alive, or for the out-of-state tuition received, it does not seem to have much to do with inviting genius-level minds to the US.
We need to do much more to get more native US students to go into engineering and computer science; however there should be no doubt that we have a long, long way to go before we tap out our base of students which are intellectually capable of competing with their global peers.
UPDATE:
This article is somewhat related to this post.